OpenCloudOS-Kernel/arch/powerpc/mm/slice.c

814 lines
22 KiB
C
Raw Normal View History

[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/*
* address space "slices" (meta-segments) support
*
* Copyright (C) 2007 Benjamin Herrenschmidt, IBM Corporation.
*
* Based on hugetlb implementation
*
* Copyright (C) 2003 David Gibson, IBM Corporation.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
#undef DEBUG
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/err.h>
#include <linux/spinlock.h>
#include <linux/export.h>
#include <linux/hugetlb.h>
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
#include <asm/mman.h>
#include <asm/mmu.h>
#include <asm/copro.h>
#include <asm/hugetlb.h>
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
static DEFINE_SPINLOCK(slice_convert_lock);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
#ifdef DEBUG
int _slice_debug = 1;
static void slice_print_mask(const char *label, const struct slice_mask *mask)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
{
if (!_slice_debug)
return;
pr_devel("%s low_slice: %*pbl\n", label,
(int)SLICE_NUM_LOW, &mask->low_slices);
pr_devel("%s high_slice: %*pbl\n", label,
(int)SLICE_NUM_HIGH, mask->high_slices);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
#define slice_dbg(fmt...) do { if (_slice_debug) pr_devel(fmt); } while (0)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
#else
static void slice_print_mask(const char *label, const struct slice_mask *mask) {}
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
#define slice_dbg(fmt...)
#endif
static void slice_range_to_mask(unsigned long start, unsigned long len,
struct slice_mask *ret)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
{
unsigned long end = start + len - 1;
ret->low_slices = 0;
if (SLICE_NUM_HIGH)
bitmap_zero(ret->high_slices, SLICE_NUM_HIGH);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
if (start < SLICE_LOW_TOP) {
unsigned long mend = min(end,
(unsigned long)(SLICE_LOW_TOP - 1));
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
ret->low_slices = (1u << (GET_LOW_SLICE_INDEX(mend) + 1))
- (1u << GET_LOW_SLICE_INDEX(start));
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
if ((start + len) > SLICE_LOW_TOP) {
unsigned long start_index = GET_HIGH_SLICE_INDEX(start);
unsigned long align_end = ALIGN(end, (1UL << SLICE_HIGH_SHIFT));
unsigned long count = GET_HIGH_SLICE_INDEX(align_end) - start_index;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
bitmap_set(ret->high_slices, start_index, count);
}
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
static int slice_area_is_free(struct mm_struct *mm, unsigned long addr,
unsigned long len)
{
struct vm_area_struct *vma;
if ((mm->context.slb_addr_limit - len) < addr)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
return 0;
vma = find_vma(mm, addr);
mm: larger stack guard gap, between vmas Stack guard page is a useful feature to reduce a risk of stack smashing into a different mapping. We have been using a single page gap which is sufficient to prevent having stack adjacent to a different mapping. But this seems to be insufficient in the light of the stack usage in userspace. E.g. glibc uses as large as 64kB alloca() in many commonly used functions. Others use constructs liks gid_t buffer[NGROUPS_MAX] which is 256kB or stack strings with MAX_ARG_STRLEN. This will become especially dangerous for suid binaries and the default no limit for the stack size limit because those applications can be tricked to consume a large portion of the stack and a single glibc call could jump over the guard page. These attacks are not theoretical, unfortunatelly. Make those attacks less probable by increasing the stack guard gap to 1MB (on systems with 4k pages; but make it depend on the page size because systems with larger base pages might cap stack allocations in the PAGE_SIZE units) which should cover larger alloca() and VLA stack allocations. It is obviously not a full fix because the problem is somehow inherent, but it should reduce attack space a lot. One could argue that the gap size should be configurable from userspace, but that can be done later when somebody finds that the new 1MB is wrong for some special case applications. For now, add a kernel command line option (stack_guard_gap) to specify the stack gap size (in page units). Implementation wise, first delete all the old code for stack guard page: because although we could get away with accounting one extra page in a stack vma, accounting a larger gap can break userspace - case in point, a program run with "ulimit -S -v 20000" failed when the 1MB gap was counted for RLIMIT_AS; similar problems could come with RLIMIT_MLOCK and strict non-overcommit mode. Instead of keeping gap inside the stack vma, maintain the stack guard gap as a gap between vmas: using vm_start_gap() in place of vm_start (or vm_end_gap() in place of vm_end if VM_GROWSUP) in just those few places which need to respect the gap - mainly arch_get_unmapped_area(), and and the vma tree's subtree_gap support for that. Original-patch-by: Oleg Nesterov <oleg@redhat.com> Original-patch-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Michal Hocko <mhocko@suse.com> Tested-by: Helge Deller <deller@gmx.de> # parisc Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-06-19 19:03:24 +08:00
return (!vma || (addr + len) <= vm_start_gap(vma));
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
static int slice_low_has_vma(struct mm_struct *mm, unsigned long slice)
{
return !slice_area_is_free(mm, slice << SLICE_LOW_SHIFT,
1ul << SLICE_LOW_SHIFT);
}
static int slice_high_has_vma(struct mm_struct *mm, unsigned long slice)
{
unsigned long start = slice << SLICE_HIGH_SHIFT;
unsigned long end = start + (1ul << SLICE_HIGH_SHIFT);
#ifdef CONFIG_PPC64
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/* Hack, so that each addresses is controlled by exactly one
* of the high or low area bitmaps, the first high area starts
* at 4GB, not 0 */
if (start == 0)
start = SLICE_LOW_TOP;
#endif
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
return !slice_area_is_free(mm, start, end - start);
}
static void slice_mask_for_free(struct mm_struct *mm, struct slice_mask *ret,
unsigned long high_limit)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
{
unsigned long i;
ret->low_slices = 0;
if (SLICE_NUM_HIGH)
bitmap_zero(ret->high_slices, SLICE_NUM_HIGH);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
for (i = 0; i < SLICE_NUM_LOW; i++)
if (!slice_low_has_vma(mm, i))
ret->low_slices |= 1u << i;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
if (high_limit <= SLICE_LOW_TOP)
return;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
for (i = 0; i < GET_HIGH_SLICE_INDEX(high_limit); i++)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
if (!slice_high_has_vma(mm, i))
__set_bit(i, ret->high_slices);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
#ifdef CONFIG_PPC_BOOK3S_64
static struct slice_mask *slice_mask_for_size(struct mm_struct *mm, int psize)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
{
#ifdef CONFIG_PPC_64K_PAGES
if (psize == MMU_PAGE_64K)
return &mm->context.mask_64k;
#endif
if (psize == MMU_PAGE_4K)
return &mm->context.mask_4k;
#ifdef CONFIG_HUGETLB_PAGE
if (psize == MMU_PAGE_16M)
return &mm->context.mask_16m;
if (psize == MMU_PAGE_16G)
return &mm->context.mask_16g;
#endif
BUG();
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
#elif defined(CONFIG_PPC_8xx)
static struct slice_mask *slice_mask_for_size(struct mm_struct *mm, int psize)
{
if (psize == mmu_virtual_psize)
return &mm->context.mask_base_psize;
#ifdef CONFIG_HUGETLB_PAGE
if (psize == MMU_PAGE_512K)
return &mm->context.mask_512k;
if (psize == MMU_PAGE_8M)
return &mm->context.mask_8m;
#endif
BUG();
}
#else
#error "Must define the slice masks for page sizes supported by the platform"
#endif
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
static bool slice_check_range_fits(struct mm_struct *mm,
const struct slice_mask *available,
unsigned long start, unsigned long len)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
{
unsigned long end = start + len - 1;
u64 low_slices = 0;
if (start < SLICE_LOW_TOP) {
unsigned long mend = min(end,
(unsigned long)(SLICE_LOW_TOP - 1));
low_slices = (1u << (GET_LOW_SLICE_INDEX(mend) + 1))
- (1u << GET_LOW_SLICE_INDEX(start));
}
if ((low_slices & available->low_slices) != low_slices)
return false;
if (SLICE_NUM_HIGH && ((start + len) > SLICE_LOW_TOP)) {
unsigned long start_index = GET_HIGH_SLICE_INDEX(start);
unsigned long align_end = ALIGN(end, (1UL << SLICE_HIGH_SHIFT));
unsigned long count = GET_HIGH_SLICE_INDEX(align_end) - start_index;
unsigned long i;
for (i = start_index; i < start_index + count; i++) {
if (!test_bit(i, available->high_slices))
return false;
}
}
return true;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
static void slice_flush_segments(void *parm)
{
#ifdef CONFIG_PPC64
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
struct mm_struct *mm = parm;
unsigned long flags;
if (mm != current->active_mm)
return;
copy_mm_to_paca(current->active_mm);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
local_irq_save(flags);
slb_flush_and_rebolt();
local_irq_restore(flags);
#endif
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
static void slice_convert(struct mm_struct *mm,
const struct slice_mask *mask, int psize)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
{
int index, mask_index;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/* Write the new slice psize bits */
unsigned char *hpsizes, *lpsizes;
struct slice_mask *psize_mask, *old_mask;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
unsigned long i, flags;
int old_psize;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
slice_dbg("slice_convert(mm=%p, psize=%d)\n", mm, psize);
slice_print_mask(" mask", mask);
psize_mask = slice_mask_for_size(mm, psize);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/* We need to use a spinlock here to protect against
* concurrent 64k -> 4k demotion ...
*/
spin_lock_irqsave(&slice_convert_lock, flags);
lpsizes = mm->context.low_slices_psize;
for (i = 0; i < SLICE_NUM_LOW; i++) {
if (!(mask->low_slices & (1u << i)))
continue;
mask_index = i & 0x1;
index = i >> 1;
/* Update the slice_mask */
old_psize = (lpsizes[index] >> (mask_index * 4)) & 0xf;
old_mask = slice_mask_for_size(mm, old_psize);
old_mask->low_slices &= ~(1u << i);
psize_mask->low_slices |= 1u << i;
/* Update the sizes array */
lpsizes[index] = (lpsizes[index] & ~(0xf << (mask_index * 4))) |
(((unsigned long)psize) << (mask_index * 4));
}
hpsizes = mm->context.high_slices_psize;
for (i = 0; i < GET_HIGH_SLICE_INDEX(mm->context.slb_addr_limit); i++) {
if (!test_bit(i, mask->high_slices))
continue;
mask_index = i & 0x1;
index = i >> 1;
/* Update the slice_mask */
old_psize = (hpsizes[index] >> (mask_index * 4)) & 0xf;
old_mask = slice_mask_for_size(mm, old_psize);
__clear_bit(i, old_mask->high_slices);
__set_bit(i, psize_mask->high_slices);
/* Update the sizes array */
hpsizes[index] = (hpsizes[index] & ~(0xf << (mask_index * 4))) |
(((unsigned long)psize) << (mask_index * 4));
}
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
slice_dbg(" lsps=%lx, hsps=%lx\n",
(unsigned long)mm->context.low_slices_psize,
(unsigned long)mm->context.high_slices_psize);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
spin_unlock_irqrestore(&slice_convert_lock, flags);
copro_flush_all_slbs(mm);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
/*
* Compute which slice addr is part of;
* set *boundary_addr to the start or end boundary of that slice
* (depending on 'end' parameter);
* return boolean indicating if the slice is marked as available in the
* 'available' slice_mark.
*/
static bool slice_scan_available(unsigned long addr,
const struct slice_mask *available,
int end, unsigned long *boundary_addr)
{
unsigned long slice;
if (addr < SLICE_LOW_TOP) {
slice = GET_LOW_SLICE_INDEX(addr);
*boundary_addr = (slice + end) << SLICE_LOW_SHIFT;
return !!(available->low_slices & (1u << slice));
} else {
slice = GET_HIGH_SLICE_INDEX(addr);
*boundary_addr = (slice + end) ?
((slice + end) << SLICE_HIGH_SHIFT) : SLICE_LOW_TOP;
return !!test_bit(slice, available->high_slices);
}
}
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
static unsigned long slice_find_area_bottomup(struct mm_struct *mm,
unsigned long len,
const struct slice_mask *available,
powerpc/mm: Enable mappings above 128TB Not all user space application is ready to handle wide addresses. It's known that at least some JIT compilers use higher bits in pointers to encode their information. It collides with valid pointers with 512TB addresses and leads to crashes. To mitigate this, we are not going to allocate virtual address space above 128TB by default. But userspace can ask for allocation from full address space by specifying hint address (with or without MAP_FIXED) above 128TB. If hint address set above 128TB, but MAP_FIXED is not specified, we try to look for unmapped area by specified address. If it's already occupied, we look for unmapped area in *full* address space, rather than from 128TB window. This approach helps to easily make application's memory allocator aware about large address space without manually tracking allocated virtual address space. This is going to be a per mmap decision. ie, we can have some mmaps with larger addresses and other that do not. A sample memory layout looks like: 10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB 10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB 10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB 10029630000-10029660000 rw-p 00000000 00:00 0 [heap] 7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0 7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83690000-7fff836a0000 rw-p 00000000 00:00 0 7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso] 7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack] 1000000000000-1000000010000 rw-p 00000000 00:00 0 1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0 Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-03-30 19:05:21 +08:00
int psize, unsigned long high_limit)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
{
int pshift = max_t(int, mmu_psize_defs[psize].shift, PAGE_SHIFT);
unsigned long addr, found, next_end;
struct vm_unmapped_area_info info;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
info.flags = 0;
info.length = len;
info.align_mask = PAGE_MASK & ((1ul << pshift) - 1);
info.align_offset = 0;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
addr = TASK_UNMAPPED_BASE;
powerpc/mm: Enable mappings above 128TB Not all user space application is ready to handle wide addresses. It's known that at least some JIT compilers use higher bits in pointers to encode their information. It collides with valid pointers with 512TB addresses and leads to crashes. To mitigate this, we are not going to allocate virtual address space above 128TB by default. But userspace can ask for allocation from full address space by specifying hint address (with or without MAP_FIXED) above 128TB. If hint address set above 128TB, but MAP_FIXED is not specified, we try to look for unmapped area by specified address. If it's already occupied, we look for unmapped area in *full* address space, rather than from 128TB window. This approach helps to easily make application's memory allocator aware about large address space without manually tracking allocated virtual address space. This is going to be a per mmap decision. ie, we can have some mmaps with larger addresses and other that do not. A sample memory layout looks like: 10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB 10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB 10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB 10029630000-10029660000 rw-p 00000000 00:00 0 [heap] 7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0 7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83690000-7fff836a0000 rw-p 00000000 00:00 0 7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso] 7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack] 1000000000000-1000000010000 rw-p 00000000 00:00 0 1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0 Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-03-30 19:05:21 +08:00
/*
* Check till the allow max value for this mmap request
*/
while (addr < high_limit) {
info.low_limit = addr;
if (!slice_scan_available(addr, available, 1, &addr))
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
continue;
next_slice:
/*
* At this point [info.low_limit; addr) covers
* available slices only and ends at a slice boundary.
* Check if we need to reduce the range, or if we can
* extend it to cover the next available slice.
*/
if (addr >= high_limit)
addr = high_limit;
else if (slice_scan_available(addr, available, 1, &next_end)) {
addr = next_end;
goto next_slice;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
info.high_limit = addr;
found = vm_unmapped_area(&info);
if (!(found & ~PAGE_MASK))
return found;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
return -ENOMEM;
}
static unsigned long slice_find_area_topdown(struct mm_struct *mm,
unsigned long len,
const struct slice_mask *available,
powerpc/mm: Enable mappings above 128TB Not all user space application is ready to handle wide addresses. It's known that at least some JIT compilers use higher bits in pointers to encode their information. It collides with valid pointers with 512TB addresses and leads to crashes. To mitigate this, we are not going to allocate virtual address space above 128TB by default. But userspace can ask for allocation from full address space by specifying hint address (with or without MAP_FIXED) above 128TB. If hint address set above 128TB, but MAP_FIXED is not specified, we try to look for unmapped area by specified address. If it's already occupied, we look for unmapped area in *full* address space, rather than from 128TB window. This approach helps to easily make application's memory allocator aware about large address space without manually tracking allocated virtual address space. This is going to be a per mmap decision. ie, we can have some mmaps with larger addresses and other that do not. A sample memory layout looks like: 10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB 10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB 10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB 10029630000-10029660000 rw-p 00000000 00:00 0 [heap] 7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0 7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83690000-7fff836a0000 rw-p 00000000 00:00 0 7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso] 7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack] 1000000000000-1000000010000 rw-p 00000000 00:00 0 1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0 Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-03-30 19:05:21 +08:00
int psize, unsigned long high_limit)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
{
int pshift = max_t(int, mmu_psize_defs[psize].shift, PAGE_SHIFT);
unsigned long addr, found, prev;
struct vm_unmapped_area_info info;
info.flags = VM_UNMAPPED_AREA_TOPDOWN;
info.length = len;
info.align_mask = PAGE_MASK & ((1ul << pshift) - 1);
info.align_offset = 0;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
addr = mm->mmap_base;
powerpc/mm: Enable mappings above 128TB Not all user space application is ready to handle wide addresses. It's known that at least some JIT compilers use higher bits in pointers to encode their information. It collides with valid pointers with 512TB addresses and leads to crashes. To mitigate this, we are not going to allocate virtual address space above 128TB by default. But userspace can ask for allocation from full address space by specifying hint address (with or without MAP_FIXED) above 128TB. If hint address set above 128TB, but MAP_FIXED is not specified, we try to look for unmapped area by specified address. If it's already occupied, we look for unmapped area in *full* address space, rather than from 128TB window. This approach helps to easily make application's memory allocator aware about large address space without manually tracking allocated virtual address space. This is going to be a per mmap decision. ie, we can have some mmaps with larger addresses and other that do not. A sample memory layout looks like: 10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB 10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB 10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB 10029630000-10029660000 rw-p 00000000 00:00 0 [heap] 7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0 7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83690000-7fff836a0000 rw-p 00000000 00:00 0 7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso] 7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack] 1000000000000-1000000010000 rw-p 00000000 00:00 0 1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0 Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-03-30 19:05:21 +08:00
/*
* If we are trying to allocate above DEFAULT_MAP_WINDOW
* Add the different to the mmap_base.
* Only for that request for which high_limit is above
* DEFAULT_MAP_WINDOW we should apply this.
*/
if (high_limit > DEFAULT_MAP_WINDOW)
addr += mm->context.slb_addr_limit - DEFAULT_MAP_WINDOW;
powerpc/mm: Enable mappings above 128TB Not all user space application is ready to handle wide addresses. It's known that at least some JIT compilers use higher bits in pointers to encode their information. It collides with valid pointers with 512TB addresses and leads to crashes. To mitigate this, we are not going to allocate virtual address space above 128TB by default. But userspace can ask for allocation from full address space by specifying hint address (with or without MAP_FIXED) above 128TB. If hint address set above 128TB, but MAP_FIXED is not specified, we try to look for unmapped area by specified address. If it's already occupied, we look for unmapped area in *full* address space, rather than from 128TB window. This approach helps to easily make application's memory allocator aware about large address space without manually tracking allocated virtual address space. This is going to be a per mmap decision. ie, we can have some mmaps with larger addresses and other that do not. A sample memory layout looks like: 10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB 10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB 10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB 10029630000-10029660000 rw-p 00000000 00:00 0 [heap] 7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0 7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83690000-7fff836a0000 rw-p 00000000 00:00 0 7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso] 7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack] 1000000000000-1000000010000 rw-p 00000000 00:00 0 1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0 Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-03-30 19:05:21 +08:00
while (addr > PAGE_SIZE) {
info.high_limit = addr;
if (!slice_scan_available(addr - 1, available, 0, &addr))
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
continue;
prev_slice:
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/*
* At this point [addr; info.high_limit) covers
* available slices only and starts at a slice boundary.
* Check if we need to reduce the range, or if we can
* extend it to cover the previous available slice.
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
*/
if (addr < PAGE_SIZE)
addr = PAGE_SIZE;
else if (slice_scan_available(addr - 1, available, 0, &prev)) {
addr = prev;
goto prev_slice;
}
info.low_limit = addr;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
found = vm_unmapped_area(&info);
if (!(found & ~PAGE_MASK))
return found;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
/*
* A failed mmap() very likely causes application failure,
* so fall back to the bottom-up function here. This scenario
* can happen with large stack limits and large mmap()
* allocations.
*/
powerpc/mm: Enable mappings above 128TB Not all user space application is ready to handle wide addresses. It's known that at least some JIT compilers use higher bits in pointers to encode their information. It collides with valid pointers with 512TB addresses and leads to crashes. To mitigate this, we are not going to allocate virtual address space above 128TB by default. But userspace can ask for allocation from full address space by specifying hint address (with or without MAP_FIXED) above 128TB. If hint address set above 128TB, but MAP_FIXED is not specified, we try to look for unmapped area by specified address. If it's already occupied, we look for unmapped area in *full* address space, rather than from 128TB window. This approach helps to easily make application's memory allocator aware about large address space without manually tracking allocated virtual address space. This is going to be a per mmap decision. ie, we can have some mmaps with larger addresses and other that do not. A sample memory layout looks like: 10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB 10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB 10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB 10029630000-10029660000 rw-p 00000000 00:00 0 [heap] 7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0 7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83690000-7fff836a0000 rw-p 00000000 00:00 0 7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso] 7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack] 1000000000000-1000000010000 rw-p 00000000 00:00 0 1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0 Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-03-30 19:05:21 +08:00
return slice_find_area_bottomup(mm, len, available, psize, high_limit);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
static unsigned long slice_find_area(struct mm_struct *mm, unsigned long len,
const struct slice_mask *mask, int psize,
powerpc/mm: Enable mappings above 128TB Not all user space application is ready to handle wide addresses. It's known that at least some JIT compilers use higher bits in pointers to encode their information. It collides with valid pointers with 512TB addresses and leads to crashes. To mitigate this, we are not going to allocate virtual address space above 128TB by default. But userspace can ask for allocation from full address space by specifying hint address (with or without MAP_FIXED) above 128TB. If hint address set above 128TB, but MAP_FIXED is not specified, we try to look for unmapped area by specified address. If it's already occupied, we look for unmapped area in *full* address space, rather than from 128TB window. This approach helps to easily make application's memory allocator aware about large address space without manually tracking allocated virtual address space. This is going to be a per mmap decision. ie, we can have some mmaps with larger addresses and other that do not. A sample memory layout looks like: 10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB 10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB 10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB 10029630000-10029660000 rw-p 00000000 00:00 0 [heap] 7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0 7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83690000-7fff836a0000 rw-p 00000000 00:00 0 7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso] 7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack] 1000000000000-1000000010000 rw-p 00000000 00:00 0 1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0 Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-03-30 19:05:21 +08:00
int topdown, unsigned long high_limit)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
{
if (topdown)
powerpc/mm: Enable mappings above 128TB Not all user space application is ready to handle wide addresses. It's known that at least some JIT compilers use higher bits in pointers to encode their information. It collides with valid pointers with 512TB addresses and leads to crashes. To mitigate this, we are not going to allocate virtual address space above 128TB by default. But userspace can ask for allocation from full address space by specifying hint address (with or without MAP_FIXED) above 128TB. If hint address set above 128TB, but MAP_FIXED is not specified, we try to look for unmapped area by specified address. If it's already occupied, we look for unmapped area in *full* address space, rather than from 128TB window. This approach helps to easily make application's memory allocator aware about large address space without manually tracking allocated virtual address space. This is going to be a per mmap decision. ie, we can have some mmaps with larger addresses and other that do not. A sample memory layout looks like: 10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB 10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB 10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB 10029630000-10029660000 rw-p 00000000 00:00 0 [heap] 7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0 7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83690000-7fff836a0000 rw-p 00000000 00:00 0 7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso] 7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack] 1000000000000-1000000010000 rw-p 00000000 00:00 0 1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0 Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-03-30 19:05:21 +08:00
return slice_find_area_topdown(mm, len, mask, psize, high_limit);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
else
powerpc/mm: Enable mappings above 128TB Not all user space application is ready to handle wide addresses. It's known that at least some JIT compilers use higher bits in pointers to encode their information. It collides with valid pointers with 512TB addresses and leads to crashes. To mitigate this, we are not going to allocate virtual address space above 128TB by default. But userspace can ask for allocation from full address space by specifying hint address (with or without MAP_FIXED) above 128TB. If hint address set above 128TB, but MAP_FIXED is not specified, we try to look for unmapped area by specified address. If it's already occupied, we look for unmapped area in *full* address space, rather than from 128TB window. This approach helps to easily make application's memory allocator aware about large address space without manually tracking allocated virtual address space. This is going to be a per mmap decision. ie, we can have some mmaps with larger addresses and other that do not. A sample memory layout looks like: 10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB 10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB 10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB 10029630000-10029660000 rw-p 00000000 00:00 0 [heap] 7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0 7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83690000-7fff836a0000 rw-p 00000000 00:00 0 7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso] 7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack] 1000000000000-1000000010000 rw-p 00000000 00:00 0 1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0 Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-03-30 19:05:21 +08:00
return slice_find_area_bottomup(mm, len, mask, psize, high_limit);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
static inline void slice_copy_mask(struct slice_mask *dst,
const struct slice_mask *src)
{
dst->low_slices = src->low_slices;
if (!SLICE_NUM_HIGH)
return;
bitmap_copy(dst->high_slices, src->high_slices, SLICE_NUM_HIGH);
}
static inline void slice_or_mask(struct slice_mask *dst,
const struct slice_mask *src1,
const struct slice_mask *src2)
{
dst->low_slices = src1->low_slices | src2->low_slices;
if (!SLICE_NUM_HIGH)
return;
bitmap_or(dst->high_slices, src1->high_slices, src2->high_slices, SLICE_NUM_HIGH);
}
static inline void slice_andnot_mask(struct slice_mask *dst,
const struct slice_mask *src1,
const struct slice_mask *src2)
{
dst->low_slices = src1->low_slices & ~src2->low_slices;
if (!SLICE_NUM_HIGH)
return;
bitmap_andnot(dst->high_slices, src1->high_slices, src2->high_slices, SLICE_NUM_HIGH);
}
#ifdef CONFIG_PPC_64K_PAGES
#define MMU_PAGE_BASE MMU_PAGE_64K
#else
#define MMU_PAGE_BASE MMU_PAGE_4K
#endif
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
unsigned long slice_get_unmapped_area(unsigned long addr, unsigned long len,
unsigned long flags, unsigned int psize,
int topdown)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
{
struct slice_mask good_mask;
struct slice_mask potential_mask;
const struct slice_mask *maskp;
const struct slice_mask *compat_maskp = NULL;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
int fixed = (flags & MAP_FIXED);
int pshift = max_t(int, mmu_psize_defs[psize].shift, PAGE_SHIFT);
unsigned long page_size = 1UL << pshift;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
struct mm_struct *mm = current->mm;
unsigned long newaddr;
powerpc/mm: Enable mappings above 128TB Not all user space application is ready to handle wide addresses. It's known that at least some JIT compilers use higher bits in pointers to encode their information. It collides with valid pointers with 512TB addresses and leads to crashes. To mitigate this, we are not going to allocate virtual address space above 128TB by default. But userspace can ask for allocation from full address space by specifying hint address (with or without MAP_FIXED) above 128TB. If hint address set above 128TB, but MAP_FIXED is not specified, we try to look for unmapped area by specified address. If it's already occupied, we look for unmapped area in *full* address space, rather than from 128TB window. This approach helps to easily make application's memory allocator aware about large address space without manually tracking allocated virtual address space. This is going to be a per mmap decision. ie, we can have some mmaps with larger addresses and other that do not. A sample memory layout looks like: 10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB 10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB 10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB 10029630000-10029660000 rw-p 00000000 00:00 0 [heap] 7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0 7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83690000-7fff836a0000 rw-p 00000000 00:00 0 7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso] 7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack] 1000000000000-1000000010000 rw-p 00000000 00:00 0 1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0 Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-03-30 19:05:21 +08:00
unsigned long high_limit;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
high_limit = DEFAULT_MAP_WINDOW;
if (addr >= high_limit || (fixed && (addr + len > high_limit)))
high_limit = TASK_SIZE;
if (len > high_limit)
return -ENOMEM;
if (len & (page_size - 1))
return -EINVAL;
if (fixed) {
if (addr & (page_size - 1))
return -EINVAL;
if (addr > high_limit - len)
return -ENOMEM;
}
if (high_limit > mm->context.slb_addr_limit) {
/*
* Increasing the slb_addr_limit does not require
* slice mask cache to be recalculated because it should
* be already initialised beyond the old address limit.
*/
mm->context.slb_addr_limit = high_limit;
powerpc/mm: Enable mappings above 128TB Not all user space application is ready to handle wide addresses. It's known that at least some JIT compilers use higher bits in pointers to encode their information. It collides with valid pointers with 512TB addresses and leads to crashes. To mitigate this, we are not going to allocate virtual address space above 128TB by default. But userspace can ask for allocation from full address space by specifying hint address (with or without MAP_FIXED) above 128TB. If hint address set above 128TB, but MAP_FIXED is not specified, we try to look for unmapped area by specified address. If it's already occupied, we look for unmapped area in *full* address space, rather than from 128TB window. This approach helps to easily make application's memory allocator aware about large address space without manually tracking allocated virtual address space. This is going to be a per mmap decision. ie, we can have some mmaps with larger addresses and other that do not. A sample memory layout looks like: 10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB 10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB 10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB 10029630000-10029660000 rw-p 00000000 00:00 0 [heap] 7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0 7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83690000-7fff836a0000 rw-p 00000000 00:00 0 7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso] 7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack] 1000000000000-1000000010000 rw-p 00000000 00:00 0 1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0 Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-03-30 19:05:21 +08:00
on_each_cpu(slice_flush_segments, mm, 1);
}
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/* Sanity checks */
BUG_ON(mm->task_size == 0);
BUG_ON(mm->context.slb_addr_limit == 0);
VM_BUG_ON(radix_enabled());
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
slice_dbg("slice_get_unmapped_area(mm=%p, psize=%d...\n", mm, psize);
slice_dbg(" addr=%lx, len=%lx, flags=%lx, topdown=%d\n",
addr, len, flags, topdown);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/* If hint, make sure it matches our alignment restrictions */
if (!fixed && addr) {
addr = _ALIGN_UP(addr, page_size);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
slice_dbg(" aligned addr=%lx\n", addr);
/* Ignore hint if it's too large or overlaps a VMA */
if (addr > high_limit - len ||
!slice_area_is_free(mm, addr, len))
addr = 0;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
/* First make up a "good" mask of slices that have the right size
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
* already
*/
maskp = slice_mask_for_size(mm, psize);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/*
* Here "good" means slices that are already the right page size,
* "compat" means slices that have a compatible page size (i.e.
* 4k in a 64k pagesize kernel), and "free" means slices without
* any VMAs.
*
* If MAP_FIXED:
* check if fits in good | compat => OK
* check if fits in good | compat | free => convert free
* else bad
* If have hint:
* check if hint fits in good => OK
* check if hint fits in good | free => convert free
* Otherwise:
* search in good, found => OK
* search in good | free, found => convert free
* search in good | compat | free, found => convert free.
*/
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/*
* If we support combo pages, we can allow 64k pages in 4k slices
* The mask copies could be avoided in most cases here if we had
* a pointer to good mask for the next code to use.
*/
if (IS_ENABLED(CONFIG_PPC_64K_PAGES) && psize == MMU_PAGE_64K) {
compat_maskp = slice_mask_for_size(mm, MMU_PAGE_4K);
if (fixed)
slice_or_mask(&good_mask, maskp, compat_maskp);
else
slice_copy_mask(&good_mask, maskp);
} else {
slice_copy_mask(&good_mask, maskp);
}
slice_print_mask(" good_mask", &good_mask);
if (compat_maskp)
slice_print_mask(" compat_mask", compat_maskp);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/* First check hint if it's valid or if we have MAP_FIXED */
if (addr != 0 || fixed) {
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/* Check if we fit in the good mask. If we do, we just return,
* nothing else to do
*/
if (slice_check_range_fits(mm, &good_mask, addr, len)) {
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
slice_dbg(" fits good !\n");
newaddr = addr;
goto return_addr;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
} else {
/* Now let's see if we can find something in the existing
* slices for that size
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
*/
newaddr = slice_find_area(mm, len, &good_mask,
powerpc/mm: Enable mappings above 128TB Not all user space application is ready to handle wide addresses. It's known that at least some JIT compilers use higher bits in pointers to encode their information. It collides with valid pointers with 512TB addresses and leads to crashes. To mitigate this, we are not going to allocate virtual address space above 128TB by default. But userspace can ask for allocation from full address space by specifying hint address (with or without MAP_FIXED) above 128TB. If hint address set above 128TB, but MAP_FIXED is not specified, we try to look for unmapped area by specified address. If it's already occupied, we look for unmapped area in *full* address space, rather than from 128TB window. This approach helps to easily make application's memory allocator aware about large address space without manually tracking allocated virtual address space. This is going to be a per mmap decision. ie, we can have some mmaps with larger addresses and other that do not. A sample memory layout looks like: 10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB 10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB 10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB 10029630000-10029660000 rw-p 00000000 00:00 0 [heap] 7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0 7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so 7fff83690000-7fff836a0000 rw-p 00000000 00:00 0 7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso] 7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so 7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack] 1000000000000-1000000010000 rw-p 00000000 00:00 0 1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0 Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-03-30 19:05:21 +08:00
psize, topdown, high_limit);
if (newaddr != -ENOMEM) {
/* Found within the good mask, we don't have to setup,
* we thus return directly
*/
slice_dbg(" found area at 0x%lx\n", newaddr);
goto return_addr;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
}
/*
* We don't fit in the good mask, check what other slices are
* empty and thus can be converted
*/
slice_mask_for_free(mm, &potential_mask, high_limit);
slice_or_mask(&potential_mask, &potential_mask, &good_mask);
slice_print_mask(" potential", &potential_mask);
if (addr != 0 || fixed) {
if (slice_check_range_fits(mm, &potential_mask, addr, len)) {
slice_dbg(" fits potential !\n");
newaddr = addr;
goto convert;
}
}
/* If we have MAP_FIXED and failed the above steps, then error out */
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
if (fixed)
return -EBUSY;
slice_dbg(" search...\n");
/* If we had a hint that didn't work out, see if we can fit
* anywhere in the good area.
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
*/
if (addr) {
newaddr = slice_find_area(mm, len, &good_mask,
psize, topdown, high_limit);
if (newaddr != -ENOMEM) {
slice_dbg(" found area at 0x%lx\n", newaddr);
goto return_addr;
}
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
/* Now let's see if we can find something in the existing slices
* for that size plus free slices
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
*/
newaddr = slice_find_area(mm, len, &potential_mask,
psize, topdown, high_limit);
#ifdef CONFIG_PPC_64K_PAGES
if (newaddr == -ENOMEM && psize == MMU_PAGE_64K) {
/* retry the search with 4k-page slices included */
slice_or_mask(&potential_mask, &potential_mask, compat_maskp);
newaddr = slice_find_area(mm, len, &potential_mask,
psize, topdown, high_limit);
}
#endif
if (newaddr == -ENOMEM)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
return -ENOMEM;
slice_range_to_mask(newaddr, len, &potential_mask);
slice_dbg(" found potential area at 0x%lx\n", newaddr);
slice_print_mask(" mask", &potential_mask);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
convert:
/*
* Try to allocate the context before we do slice convert
* so that we handle the context allocation failure gracefully.
*/
if (need_extra_context(mm, newaddr)) {
if (alloc_extended_context(mm, newaddr) < 0)
return -ENOMEM;
}
slice_andnot_mask(&potential_mask, &potential_mask, &good_mask);
if (compat_maskp && !fixed)
slice_andnot_mask(&potential_mask, &potential_mask, compat_maskp);
if (potential_mask.low_slices ||
(SLICE_NUM_HIGH &&
!bitmap_empty(potential_mask.high_slices, SLICE_NUM_HIGH))) {
slice_convert(mm, &potential_mask, psize);
if (psize > MMU_PAGE_BASE)
on_each_cpu(slice_flush_segments, mm, 1);
}
return newaddr;
return_addr:
if (need_extra_context(mm, newaddr)) {
if (alloc_extended_context(mm, newaddr) < 0)
return -ENOMEM;
}
return newaddr;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
EXPORT_SYMBOL_GPL(slice_get_unmapped_area);
unsigned long arch_get_unmapped_area(struct file *filp,
unsigned long addr,
unsigned long len,
unsigned long pgoff,
unsigned long flags)
{
return slice_get_unmapped_area(addr, len, flags,
current->mm->context.user_psize, 0);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
unsigned long arch_get_unmapped_area_topdown(struct file *filp,
const unsigned long addr0,
const unsigned long len,
const unsigned long pgoff,
const unsigned long flags)
{
return slice_get_unmapped_area(addr0, len, flags,
current->mm->context.user_psize, 1);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
unsigned int get_slice_psize(struct mm_struct *mm, unsigned long addr)
{
unsigned char *psizes;
int index, mask_index;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
VM_BUG_ON(radix_enabled());
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
if (addr < SLICE_LOW_TOP) {
psizes = mm->context.low_slices_psize;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
index = GET_LOW_SLICE_INDEX(addr);
} else {
psizes = mm->context.high_slices_psize;
index = GET_HIGH_SLICE_INDEX(addr);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
mask_index = index & 0x1;
return (psizes[index >> 1] >> (mask_index * 4)) & 0xf;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
EXPORT_SYMBOL_GPL(get_slice_psize);
void slice_init_new_context_exec(struct mm_struct *mm)
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
{
unsigned char *hpsizes, *lpsizes;
struct slice_mask *mask;
unsigned int psize = mmu_virtual_psize;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
slice_dbg("slice_init_new_context_exec(mm=%p)\n", mm);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/*
* In the case of exec, use the default limit. In the
* case of fork it is just inherited from the mm being
* duplicated.
*/
#ifdef CONFIG_PPC64
mm->context.slb_addr_limit = DEFAULT_MAP_WINDOW_USER64;
#else
mm->context.slb_addr_limit = DEFAULT_MAP_WINDOW;
#endif
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
mm->context.user_psize = psize;
/*
* Set all slice psizes to the default.
*/
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
lpsizes = mm->context.low_slices_psize;
memset(lpsizes, (psize << 4) | psize, SLICE_NUM_LOW >> 1);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
hpsizes = mm->context.high_slices_psize;
memset(hpsizes, (psize << 4) | psize, SLICE_NUM_HIGH >> 1);
/*
* Slice mask cache starts zeroed, fill the default size cache.
*/
mask = slice_mask_for_size(mm, psize);
mask->low_slices = ~0UL;
if (SLICE_NUM_HIGH)
bitmap_fill(mask->high_slices, SLICE_NUM_HIGH);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
void slice_set_range_psize(struct mm_struct *mm, unsigned long start,
unsigned long len, unsigned int psize)
{
struct slice_mask mask;
VM_BUG_ON(radix_enabled());
slice_range_to_mask(start, len, &mask);
slice_convert(mm, &mask, psize);
}
#ifdef CONFIG_HUGETLB_PAGE
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
/*
* is_hugepage_only_range() is used by generic code to verify whether
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
* a normal mmap mapping (non hugetlbfs) is valid on a given area.
*
* until the generic code provides a more generic hook and/or starts
* calling arch get_unmapped_area for MAP_FIXED (which our implementation
* here knows how to deal with), we hijack it to keep standard mappings
* away from us.
*
* because of that generic code limitation, MAP_FIXED mapping cannot
* "convert" back a slice with no VMAs to the standard page size, only
* get_unmapped_area() can. It would be possible to fix it here but I
* prefer working on fixing the generic code instead.
*
* WARNING: This will not work if hugetlbfs isn't enabled since the
* generic code will redefine that function as 0 in that. This is ok
* for now as we only use slices with hugetlbfs enabled. This should
* be fixed as the generic code gets fixed.
*/
int slice_is_hugepage_only_range(struct mm_struct *mm, unsigned long addr,
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
unsigned long len)
{
const struct slice_mask *maskp;
unsigned int psize = mm->context.user_psize;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
VM_BUG_ON(radix_enabled());
maskp = slice_mask_for_size(mm, psize);
#ifdef CONFIG_PPC_64K_PAGES
/* We need to account for 4k slices too */
if (psize == MMU_PAGE_64K) {
const struct slice_mask *compat_maskp;
struct slice_mask available;
compat_maskp = slice_mask_for_size(mm, MMU_PAGE_4K);
slice_or_mask(&available, maskp, compat_maskp);
return !slice_check_range_fits(mm, &available, addr, len);
}
#endif
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
return !slice_check_range_fits(mm, maskp, addr, len);
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 14:27:27 +08:00
}
#endif