Instead of pinning per-cpu TLB into a DTR, use DTC. This will free up
one TLB entry for application, or even kernel if access pattern to
per-cpu data area has high temporal locality.
Since per-cpu is mapped at the top of region 7 address, we just need to
add special case in alt_dtlb_miss. The physical address of per-cpu data
is already conveniently stored in IA64_KR(PER_CPU_DATA). Latency for
alt_dtlb_miss is not affected as we can hide all the latency. It was
measured that alt_dtlb_miss handler has 23 cycles latency before and
after the patch.
The performance effect is massive for applications that put lots of tlb
pressure on CPU. Workload environment like database online transaction
processing or application uses tera-byte of memory would benefit the most.
Measurement with industry standard database benchmark shown an upward
of 1.6% gain. While smaller workloads like cpu, java also showing small
improvement.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
This patch fixes
- marking I-cache clean of pages DMAed to now only done for IA64
- broken multiple inclusion in include/asm-x86_64/swiotlb.h
- missing call to mark_clean in swiotlb_sync_sg()
- a (perhaps only theoretical) issue in swiotlb_dma_supported() when
io_tlb_end is exactly at the end of memory
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Eliminate arch specific memory_present call ia64 NUMA by utilizing
sparse_memory_present_with_active_regions.
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Bob Picco <bob.picco@hp.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Tony Luck <tony.luck@intel.com>
On the ia64 architecture only this patch upgrades show_mem() for sparse
memory to be the same as it was for discontig memory. It has been shown to
work on NUMA and flatmem architectures.
Signed-off-by: George Beshers <gbeshers@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Tony Luck <tony.luck@intel.com>
While pursuing and unrelated issue with 64Mb granules I noticed a problem
related to inconsistent use of add_active_range. There doesn't appear any
reason to me why FLATMEM versus DISCONTIG_MEM should register memory to
add_active_range with different code. So I've changed the code into a
common implementation.
The other subtle issue fixed by this patch was calling add_active_range in
count_node_pages before granule aligning is performed. We were lucky with
16MB granules but not so with 64MB granules. count_node_pages has reserved
regions filtered out and as a consequence linked kernel text and data
aren't covered by calls to count_node_pages. So linked kernel regions
wasn't reported to add_active_regions. This resulted in free_initmem
causing numerous bad_page reports. This won't occur with this patch
because now all known memory regions are reported by
register_active_ranges.
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Bob Picco <bob.picco@hp.com>
Acked-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Fix a typo in the saved_max_pfn description in contig.c
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Set saved_max_pfn when discontig memory is in use.
This sets up saved_max_pfn when disctontig memory is in use.
This mirrors the code for contig memory.
This patch does not entirely solve the problem of making vmcore work,
however it does appear to be neccessary. Please consider applying.
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Fix an oops experienced on the Cell architecture when init-time functions,
early_*(), are called at runtime. It alters the call paths to make sure
that the callers explicitly say whether the call is being made on behalf of
a hotplug even, or happening at boot-time.
It has been compile tested on ppc64, ia64, s390, i386 and x86_64.
Acked-by: Arnd Bergmann <arndb@de.ibm.com>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This warning only shows up with CONFIG_VIRTUAL_MEM_MAP=y and
CONFIG_FLATMEM=y.
There is only one caller left for register_active_ranges() from the
contig.c code ... so it doesn't need to pick up the node number, the
node number is always zero.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Actually, on reflection I think that there is a good case for
keeping the options separate. I am thinking particularly of people
who want a very small crashdump kernel and thus don't want to compile
in kexec.
The patch below should fix things up so that all valid combinations of
KEXEC, CRASH_DUMP and VMCORE compile cleanly - VMCORE depends on
CRASH_DUMP which is why I said valid combinations. In a nutshell
it just untangles unrelated code and switches around a few defines.
Please note that it creats a new file, arch/ia64/kernel/crash_dump.c
This is in keeping with the i386 implementation.
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Tony Luck <tony.luck@intel.com>
SLAB_KERNEL is an alias of GFP_KERNEL.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Following up with the work on shared page table done by Dave McCracken. This
set of patch target shared page table for hugetlb memory only.
The shared page table is particular useful in the situation of large number of
independent processes sharing large shared memory segments. In the normal
page case, the amount of memory saved from process' page table is quite
significant. For hugetlb, the saving on page table memory is not the primary
objective (as hugetlb itself already cuts down page table overhead
significantly), instead, the purpose of using shared page table on hugetlb is
to allow faster TLB refill and smaller cache pollution upon TLB miss.
With PT sharing, pte entries are shared among hundreds of processes, the cache
consumption used by all the page table is smaller and in return, application
gets much higher cache hit ratio. One other effect is that cache hit ratio
with hardware page walker hitting on pte in cache will be higher and this
helps to reduce tlb miss latency. These two effects contribute to higher
application performance.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Acked-by: Hugh Dickins <hugh@veritas.com>
Cc: Dave McCracken <dmccr@us.ibm.com>
Cc: William Lee Irwin III <wli@holomorphy.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Adam Litke <agl@us.ibm.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
(David:)
If hugetlbfs_file_mmap() returns a failure to do_mmap_pgoff() - for example,
because the given file offset is not hugepage aligned - then do_mmap_pgoff
will go to the unmap_and_free_vma backout path.
But at this stage the vma hasn't been marked as hugepage, and the backout path
will call unmap_region() on it. That will eventually call down to the
non-hugepage version of unmap_page_range(). On ppc64, at least, that will
cause serious problems if there are any existing hugepage pagetable entries in
the vicinity - for example if there are any other hugepage mappings under the
same PUD. unmap_page_range() will trigger a bad_pud() on the hugepage pud
entries. I suspect this will also cause bad problems on ia64, though I don't
have a machine to test it on.
(Hugh:)
prepare_hugepage_range() should check file offset alignment when it checks
virtual address and length, to stop MAP_FIXED with a bad huge offset from
unmapping before it fails further down. PowerPC should apply the same
prepare_hugepage_range alignment checks as ia64 and all the others do.
Then none of the alignment checks in hugetlbfs_file_mmap are required (nor
is the check for too small a mapping); but even so, move up setting of
VM_HUGETLB and add a comment to warn of what David Gibson discovered - if
hugetlbfs_file_mmap fails before setting it, do_mmap_pgoff's unmap_region
when unwinding from error will go the non-huge way, which may cause bad
behaviour on architectures (powerpc and ia64) which segregate their huge
mappings into a separate region of the address space.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Acked-by: Adam Litke <agl@us.ibm.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Arch-independent zone-sizing is using indices instead of symbolic names to
offset within an array related to zones (max_zone_pfns). The unintended
impact is that ZONE_DMA and ZONE_NORMAL is initialised on powerpc instead
of ZONE_DMA and ZONE_HIGHMEM when CONFIG_HIGHMEM is set. As a result, the
the machine fails to boot but will boot with CONFIG_HIGHMEM turned off.
The following patch properly initialises the max_zone_pfns[] array and uses
symbolic names instead of indices in each architecture using
arch-independent zone-sizing. Two users have successfully booted their
powerpcs with it (one an ibook G4). It has also been boot tested on x86,
x86_64, ppc64 and ia64. Please merge for 2.6.19-rc2.
Credit to Benjamin Herrenschmidt for identifying the bug and rolling the
first fix. Additional credit to Johannes Berg and Andreas Schwab for
reporting the problem and testing on powerpc.
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
In cases where the acpi memory-add event does not containe the pxm (node)
infomation allow the driver to look up node info based on the address. The
acpi_get_node call returns -1 if it can't decode the pxm info, this causes
add_memory to panic. acpi_get_node would have to decode the resource from the
handle (a lenghty proposition). This seems to be the cleanist point to
interject the hook.
[kamezawa.hiroyu@jp.fujitsu.com: build fixes]
[y-goto@jp.fujitsu.com: build fixes]
Signed-off-by: Keith Mannthey <kmannth@us.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This is an updated version of Eric Biederman's is_init() patch.
(http://lkml.org/lkml/2006/2/6/280). It applies cleanly to 2.6.18-rc3 and
replaces a few more instances of ->pid == 1 with is_init().
Further, is_init() checks pid and thus removes dependency on Eric's other
patches for now.
Eric's original description:
There are a lot of places in the kernel where we test for init
because we give it special properties. Most significantly init
must not die. This results in code all over the kernel test
->pid == 1.
Introduce is_init to capture this case.
With multiple pid spaces for all of the cases affected we are
looking for only the first process on the system, not some other
process that has pid == 1.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Sukadev Bhattiprolu <sukadev@us.ibm.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Serge Hallyn <serue@us.ibm.com>
Cc: Cedric Le Goater <clg@fr.ibm.com>
Cc: <lxc-devel@lists.sourceforge.net>
Acked-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Make PROT_WRITE imply PROT_READ for a number of architectures which don't
support write only in hardware.
While looking at this, I noticed that some architectures which do not
support write only mappings already take the exact same approach. For
example, in arch/alpha/mm/fault.c:
"
if (cause < 0) {
if (!(vma->vm_flags & VM_EXEC))
goto bad_area;
} else if (!cause) {
/* Allow reads even for write-only mappings */
if (!(vma->vm_flags & (VM_READ | VM_WRITE)))
goto bad_area;
} else {
if (!(vma->vm_flags & VM_WRITE))
goto bad_area;
}
"
Thus, this patch brings other architectures which do not support write only
mappings in-line and consistent with the rest. I've verified the patch on
ia64, x86_64 and x86.
Additional discussion:
Several architectures, including x86, can not support write-only mappings.
The pte for x86 reserves a single bit for protection and its two states are
read only or read/write. Thus, write only is not supported in h/w.
Currently, if i 'mmap' a page write-only, the first read attempt on that page
creates a page fault and will SEGV. That check is enforced in
arch/blah/mm/fault.c. However, if i first write that page it will fault in
and the pte will be set to read/write. Thus, any subsequent reads to the page
will succeed. It is this inconsistency in behavior that this patch is
attempting to address. Furthermore, if the page is swapped out, and then
brought back the first read will also cause a SEGV. Thus, any arbitrary read
on a page can potentially result in a SEGV.
According to the SuSv3 spec, "if the application requests only PROT_WRITE, the
implementation may also allow read access." Also as mentioned, some
archtectures, such as alpha, shown above already take the approach that i am
suggesting.
The counter-argument to this raised by Arjan, is that the kernel is enforcing
the write only mapping the best it can given the h/w limitations. This is
true, however Alan Cox, and myself would argue that the inconsitency in
behavior, that is applications can sometimes work/sometimes fails is highly
undesireable. If you read through the thread, i think people, came to an
agreement on the last patch i posted, as nobody has objected to it...
Signed-off-by: Jason Baron <jbaron@redhat.com>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Roman Zippel <zippel@linux-m68k.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Andi Kleen <ak@muc.de>
Acked-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Cc: Kazumoto Kojima <kkojima@rr.iij4u.or.jp>
Cc: Ian Molton <spyro@f2s.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
[IA64] minor reformatting to vmlinux.lds.S
[IA64] CMC/CPE: Reverse the order of fetching log and checking poll threshold
[IA64] PAL calls need physical mode, stacked
[IA64] ar.fpsr not set on MCA/INIT kernel entry
[IA64] printing support for MCA/INIT
[IA64] trim output of show_mem()
[IA64] show_mem() printk levels
[IA64] Make gp value point to Region 5 in mca handler
Revert "[IA64] Unwire set/get_robust_list"
[IA64] Implement futex primitives
[IA64-SGI] Do not request DMA memory for BTE
[IA64] Move perfmon tables from thread_struct to pfm_context
[IA64] Add interface so modules can discover whether multithreading is on.
[IA64] kprobes: fixup the pagefault exception caused by probehandlers
[IA64] kprobe opcode 16 bytes alignment on IA64
[IA64] esi-support
[IA64] Add "model name" to /proc/cpuinfo
Cut the number of lines of memory info output per node from five
to one line.
Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Use the default sysrq printk level for printing show_mem() output both
for disconfig and contig versions. This is consistent with the printk
level used on other architectures (well ia32 at least).
Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
contig.c (FLATMEM) requires the same optimization as in discontig.c for show_mem
when VIRTUAL_MEM_MAP is in use. Otherwise FLATMEM has softlockup timeouts.
This was boot tested for memory configuration: SPARSEMEM,
DISCONTIG+VIRTUAL_MEM_MAP, FLATMEM, FLATMEM+VIRTUAL_MEM_MAP and
FLATMEM+VIRTUAL_MEM_MAP with largest memory gap less than LARGE_GAP by
using boot parameter "mem=".
This was boot tested and "echo m >/proc/sysrq-trigger" output evaluated for
: FLATMEM, FLATMEM+VIRTUAL_MEM_MAP, DISCONTIGMEM+VIRTUAL_MEM_MAP and
SPARSEMEM.
Signed-off-by: Bob Picco <bob.picco@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Assure that vmem_map's high endpoint is MAX_ORDER aligned. Not doing so violates
the buddy allocator algorithm. Also anyone using mem=XXX on boot line and
not aligned to MAX_ORDER requires this patch in order to satisfy buddy
allocator. vmem_map always starts at pfn 0. The potentially large MAX_ORDER
on ia64 (due to hugetlbfs) requires that the end of vmem_map be aligned
to MAX_ORDER_NR_PAGES.
This was boot tested for: FLATMEM, FLATMEM+VIRTUAL_MEM_MAP,
DISCONTIGMEM+VIRTUAL_MEM_MAP and SPARSEMEM.
Signed-off-by: Bob Picco <bob.picco@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Fix some sparse warnings on ia64. Large constants that should be long
instead of int. Use NULL instead of 0. Add some missing __iomem
casts. Replace a non-C99 structure assignment.
Signed-off-by: Keith Owens <kaos@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
I found a bug in memory hot-add code for ia64.
IA64's code has copies of pgdat's array on each node to reduce memory
access over crossing node. This array is used by NODE_DATA() macro. When
new node is hot-added, this pgdat's array should be updated and copied on
new node too.
However, I used for_each_online_node() in scatter_node_data() to copy
it. This meant its array is not copied on new node.
Because initialization of structures for new node was halfway,
so online_node_map couldn't be set at this time.
To copy arrays on new node, I changed it to check value of pgdat_list[]
which is source array of copies. I tested this patch with my Memory Hotadd
emulation on Tiger4. This patch is for 2.6.17-git20.
Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This is a patch to allocate pgdat and per node data area for ia64. The size
for them can be calculated by compute_pernodesize().
Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This is to refresh node_data[] array for ia64. As I mentioned previous
patches, ia64 has copies of information of pgdat address array on each node as
per node data.
At v2 of node_add, this function used stop_machine_run() to update them. (I
wished that they were copied safety as much as possible.) But, in this patch,
this arrays are just copied simply, and set node_online_map bit after
completion of pgdat initialization.
So, kernel must touch NODE_DATA() macro after checking node_online_map().
(Current code has already done it.) This is more simple way for just
hot-add.....
Note : It will be problem when hot-remove will occur,
because, even if online_map bit is set, kernel may
touch NODE_DATA() due to race condition. :-(
Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This is a preparatory patch to make common code for updating of NODE_DATA() of
ia64 between boottime and hotplug.
Current code remembers pgdat address in mem_data which is used at just boot
time. But its information can be used at hotplug time by moving to global
value. The next patch uses this array.
Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Change the name of old add_memory() to arch_add_memory. And use node id to
get pgdat for the node at NODE_DATA().
Note: Powerpc's old add_memory() is defined as __devinit. However,
add_memory() is usually called only after bootup.
I suppose it may be redundant. But, I'm not well known about powerpc.
So, I keep it. (But, __meminit is better at least.)
Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: "Brown, Len" <len.brown@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Overloading of page fault notification with the notify_die() has performance
issues(since the only interested components for page fault is kprobes and/or
kdb) and hence this patch introduces the new notifier call chain exclusively
for page fault notifications their by avoiding notifying unnecessary
components in the do_page_fault() code path.
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Len Brown <len.brown@intel.com>
This closes a couple holes in our attribute aliasing avoidance scheme:
- The current kernel fails mmaps of some /dev/mem MMIO regions because
they don't appear in the EFI memory map. This keeps X from working
on the Intel Tiger box.
- The current kernel allows UC mmap of the 0-1MB region of
/sys/.../legacy_mem even when the chipset doesn't support UC
access. This causes an MCA when starting X on HP rx7620 and rx8620
boxes in the default configuration.
There's more detail in the Documentation/ia64/aliasing.txt file this
adds, but the general idea is that if a region might be covered by
a granule-sized kernel identity mapping, any access via /dev/mem or
mmap must use the same attribute as the identity mapping.
Otherwise, we fall back to using an attribute that is supported
according to the EFI memory map, or to using UC if the EFI memory
map doesn't mention the region.
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
This patch modifies ia64's show_mem() to walk the vmem_map page tables and
rapidly skip forward across regions where the page tables are missing.
This prevents the pfn_valid() check from causing numerous unnecessary
page faults.
Without this patch on a 512 node 512 cpu system where every node has four
memory holes, the show_mem() call takes 1 hour 18 minutes. With this
patch, it takes less than 3 seconds.
Signed-off-by: Robin Holt <holt@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Take a hint from an x86_64 optimization by Arjan van de Ven and use it
for ia64. See a9ba9a3b38
Prefetch the mmap_sem, which is critical for the performance of the page fault
handler.
Note: mm may be NULL but I guess that is safe.
See 458f935527
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
[IA64] ioremap() should prefer WB over UC
[IA64] Add __mca_table to the DISCARD list in gate.lds
[IA64] Move __mca_table out of the __init section
[IA64] simplify some condition checks in iosapic_check_gsi_range
[IA64] correct some messages and fixes some minor things
[IA64-SGI] fix for-loop in sn_hwperf_geoid_to_cnode()
[IA64-SGI] sn_hwperf use of num_online_cpus()
[IA64] optimize flush_tlb_range on large numa box
[IA64] lazy_mmu_prot_update needs to be aware of huge pages
efi_memmap_init() collects full granules of WB memory, without
regard for whether they also support UC. So in order for ioremap()
to work for main memory, it must prefer WB mappings when possible.
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
It was reported from a field customer that global spin lock ptcg_lock
is giving a lot of grief on munmap performance running on a large numa
machine. What appears to be a problem coming from flush_tlb_range(),
which currently unconditionally calls platform_global_tlb_purge().
For some of the numa machines in existence today, this function is
mapped into ia64_global_tlb_purge(), which holds ptcg_lock spin lock
while executing ptc.ga instruction.
Here is a patch that attempt to avoid global tlb purge whenever
possible. It will use local tlb purge as much as possible. Though the
conditions to use local tlb purge is pretty restrictive. One of the
side effect of having flush tlb range instruction on ia64 is that
kernel don't get a chance to clear out cpu_vm_mask. On ia64, this mask
is sticky and it will accumulate if process bounces around. Thus
diminishing the possible use of ptc.l. Thoughts?
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Acked-by: Jack Steiner <steiner@sgi.com>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Function lazy_mmu_prot_update is also used on huge pages when it is called
by set_huge_ptep_writable, but it isn't aware of huge pages.
Signed-off-by: Zhang Yanmin <yanmin.zhang@intel.com>
Acked-by: Ken Chen <kenneth.w.chen@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Because pgdat_list was linked to pgdat_list in *reverse* order, (By default)
some of arch has to sort it by themselves.
for_each_pgdat has gone..for_each_online_pgdat() uses node_online_map, which
doesn't need to be sorted.
This patch removes codes for sorting pgdat.
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Check the EFI memory map so we can use the correct memory attributes for
ioremap(). Previously, we always used uncacheable access, which blows up on
some machines for regular system memory.
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Cc: Matt Domsch <Matt_Domsch@dell.com>
Cc: "Tolentino, Matthew E" <matthew.e.tolentino@intel.com>
Cc: "Brown, Len" <len.brown@intel.com>
Cc: Andi Kleen <ak@muc.de>
Acked-by: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add init declaration to bunch of patch functions and gate
page setup function.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Add init declaration to variables/functions used for memory
initialization. I don't think they would clash with memory
hotplug. If they do, please yell.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Add init declaration to cpu initialization functions.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
fix is_hugepage_only_range() definition to be "overlaps"
instead of "within architectural restricted hugetlb address
range". Simplify the ia64 specific code that used to use
is_hugepage_only_range() to just check which region the
address is in.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Quite a long time back, prepare_hugepage_range() replaced
is_aligned_hugepage_range() as the callback from mm/mmap.c to arch code to
verify if an address range is suitable for a hugepage mapping.
is_aligned_hugepage_range() stuck around, but only to implement
prepare_hugepage_range() on archs which didn't implement their own.
Most archs (everything except ia64 and powerpc) used the same
implementation of is_aligned_hugepage_range(). On powerpc, which
implements its own prepare_hugepage_range(), the custom version was never
used.
In addition, "is_aligned_hugepage_range()" was a bad name, because it
suggests it returns true iff the given range is a good hugepage range,
whereas in fact it returns 0-or-error (so the sense is reversed).
This patch cleans up by abolishing is_aligned_hugepage_range(). Instead
prepare_hugepage_range() is defined directly. Most archs use the default
version, which simply checks the given region is aligned to the size of a
hugepage. ia64 and powerpc define custom versions. The ia64 one simply
checks that the range is in the correct address space region in addition to
being suitably aligned. The powerpc version (just as previously) checks
for suitable addresses, and if necessary performs low-level MMU frobbing to
set up new areas for use by hugepages.
No libhugetlbfs testsuite regressions on ppc64 (POWER5 LPAR).
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Zhang Yanmin <yanmin.zhang@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: William Lee Irwin III <wli@holomorphy.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
set_page_count usage outside mm/ is limited to setting the refcount to 1.
Remove set_page_count from outside mm/, and replace those users with
init_page_count() and set_page_refcounted().
This allows more debug checking, and tighter control on how code is allowed
to play around with page->_count.
Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
First step to memory hotplug for ia64 (add only,
all new memory is added to node 0, does not use
ZONE_EASY_RECLAIM yet).
Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
I originally thought this was an bug only in the SN code, but I think I
also see a hole in the generic IA64 tlb code. (Separate patch was sent
for the SN problem).
It looks like there is a bug in the TLB flushing code. During context switch,
kernel threads (kswapd, for example) inherit the mm of the task that was
previously running on the cpu. Normally, this is ok because the previous context
is still loaded into the RR registers. However, if the owner of the mm
migrates to another cpu, changes it's context number, and references a
page before kswapd issues a tlb_purge for that same page, the purge will be
done with a stale context number (& RR registers).
Signed-off-by: Tony Luck <tony.luck@intel.com>
here is the BSP removal support for IA64. Its pretty much the same thing that
was released a while back, but has your feedback incorporated.
- Removed CONFIG_BSP_REMOVE_WORKAROUND and associated cmdline param
- Fixed compile issue with sn2/zx1 due to a undefined fix_b0_for_bsp
- some formatting nits (whitespace etc)
This has been tested on tiger and long back by alex on hp systems as well.
Signed-off-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
The per-node data structures are allocated with strided offsets that are a
function of the node number. This prevents excessive cache-aliasing from
occurring.
On systems with a large number of nodes, the strided offset becomes
too large. This patch restricts the maximum offset to 32MB. This is far larger
than the size of any current L3 cache.
Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
The original memory less node allocation attempted to use NODEDATA_ALIGN for
alignment. The bootmem allocator only allows a power of two alignments. This
causes a BUG_ON for some nodes. For cpu only nodes just allocate with a
PERCPU_PAGE_SIZE alignment.
Some older firmware reports SLIT distances of 0xff and results in bestnode
not being computed. This is now treated correctly.
The failed allocation check was removed because it's redundant. The
bootmem allocator already makes this check.
This fix has been boot tested on 4 node machine which has 4 cpu only nodes
and 1 memory node. Thanks to Pete Keilty for reporting this and helping me
test it.
Signed-off-by: Bob Picco <bob.picco@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
wrap_mmu_context(), delayed_tlb_flush(), get_mmu_context() all
have an extra { } block which cause one extra indentation.
get_mmu_context() is particularly bad with 5 indentations to
the most inner "if". It finally gets on my nerve that I can't
keep the code within 80 columns. Remove the extra { } block
and while I'm at it, reformat all the comments to 80-column
friendly. No functional change at all with this patch.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Corrects the very inefficent method of finding free context_ids in
get_mmu_context(). Instead of walking the task_list of all processes,
2 bitmaps are used to efficently store and lookup state, inuse and
needs flushing. The entire rid address space is now used before calling
wrap_mmu_context and global tlb flushing.
Special thanks to Ken and Rohit for their review and modifications in
using a bit flushmap.
Signed-off-by: Peter Keilty <peter.keilty@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
pgdat->node_size_lock is basically only neeeded in one place in the normal
code: show_mem(), which is the arch-specific sysrq-m printing function.
Strictly speaking, the architectures not doing memory hotplug do no need this
locking in show_mem(). However, they are all included for completeness. This
should also make any future consolidation of all of the implementations a
little more straightforward.
This lock is also held in the sparsemem code during a memory removal, as
sections are invalidated. This is the place there pfn_valid() is made false
for a memory area that's being removed. The lock is only required when doing
pfn_valid() operations on memory which the user does not already have a
reference on the page, such as in show_mem().
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
There was one small but very significant change in the previous patch:
mprotect's flush_tlb_range fell outside the page_table_lock: as it is in 2.4,
but that doesn't prove it safe in 2.6.
On some architectures flush_tlb_range comes to the same as flush_tlb_mm, which
has always been called from outside page_table_lock in dup_mmap, and is so
proved safe. Others required a deeper audit: I could find no reliance on
page_table_lock in any; but in ia64 and parisc found some code which looks a
bit as if it might want preemption disabled. That won't do any actual harm,
so pending a decision from the maintainers, disable preemption there.
Remove comments on page_table_lock from flush_tlb_mm, flush_tlb_range and
flush_tlb_page entries in cachetlb.txt: they were rather misleading (what
generic code does is different from what usually happens), the rules are now
changing, and it's not yet clear where we'll end up (will the generic
tlb_flush_mmu happen always under lock? never under lock? or sometimes under
and sometimes not?).
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
First step in pushing down the page_table_lock. init_mm.page_table_lock has
been used throughout the architectures (usually for ioremap): not to serialize
kernel address space allocation (that's usually vmlist_lock), but because
pud_alloc,pmd_alloc,pte_alloc_kernel expect caller holds it.
Reverse that: don't lock or unlock init_mm.page_table_lock in any of the
architectures; instead rely on pud_alloc,pmd_alloc,pte_alloc_kernel to take
and drop it when allocating a new one, to check lest a racing task already
did. Similarly no page_table_lock in vmalloc's map_vm_area.
Some temporary ugliness in __pud_alloc and __pmd_alloc: since they also handle
user mms, which are converted only by a later patch, for now they have to lock
differently according to whether or not it's init_mm.
If sources get muddled, there's a danger that an arch source taking
init_mm.page_table_lock will be mixed with common source also taking it (or
neither take it). So break the rules and make another change, which should
break the build for such a mismatch: remove the redundant mm arg from
pte_alloc_kernel (ppc64 scrapped its distinct ioremap_mm in 2.6.13).
Exceptions: arm26 used pte_alloc_kernel on user mm, now pte_alloc_map; ia64
used pte_alloc_map on init_mm, now pte_alloc_kernel; parisc had bad args to
pmd_alloc and pte_alloc_kernel in unused USE_HPPA_IOREMAP code; ppc64
map_io_page forgot to unlock on failure; ppc mmu_mapin_ram and ppc64 im_free
took page_table_lock for no good reason.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
ia64 has expand_backing_store function for growing its Register Backing Store
vma upwards. But more complete code for this purpose is found in the
CONFIG_STACK_GROWSUP part of mm/mmap.c. Uglify its #ifdefs further to provide
expand_upwards for ia64 as well as expand_stack for parisc.
The Register Backing Store vma should be marked VM_ACCOUNT. Implement the
intention of growing it only a page at a time, instead of passing an address
outside of the vma to handle_mm_fault, with unknown consequences.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The original vm_stat_account has fallen into disuse, with only one user, and
only one user of vm_stat_unaccount. It's easier to keep track if we convert
them all to __vm_stat_account, then free it from its __shackles.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
flush_tlb_all() can be a scaling issue on large SGI Altix systems
since it uses the global call_lock and always executes on all cpus.
When a process enters flush_tlb_range() to purge TLBs for another
process, it is possible to avoid flush_tlb_all() and instead allow
sn2_global_tlb_purge() to purge TLBs only where necessary.
This patch modifies flush_tlb_range() so that this case can be handled
by platform TLB purge functions and updates ia64_global_tlb_purge()
accordingly. sn2_global_tlb_purge() now calculates the region register
value from the mm argument introduced with this patch.
Signed-off-by: Dean Roe <roe@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
In arch/ia64 change the explicit use of for-loops and NR_CPUS into the
general for_each_cpu() or for_each_online_cpu() constructs, as
appropriate. This widens the scope of potential future optimizations
of the general constructs, as well as takes advantage of the existing
optimizations of first_cpu() and next_cpu().
Signed-off-by: John Hawkes <hawkes@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
This patch is the minimal set of changes required by ia64 to use SPARSEMEM.
Signed-off-by: Bob Picco <bob.picco@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
For FLATMEM contig_page_data has been made transparent to the arch code.
This patch conforms to that change.
Signed-off-by: Bob Picco <bob.picco@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
The patch modifies the Kconfig file to introduce the new memory model
options and other related SPARSEMEM changes. There is also a minor change
in the Makefile.
Signed-off-by: Bob Picco <bob.picco@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
This patch contains the ia64 architecture specific changes to prevent the
possible race conditions.
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
When copying data from user-space to kernel-space by __copy_user(),
a page_not_present fault sometimes occurs at vmalloced kernel address
because of VHPT pre-fetching.
Ignore the page_not_present fault in ia64_do_page_fault() before
jumping into exception handlers.
Signed-off-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com>
Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
I've solved the problem I was having with the simulator and not
booting Debian.
The problem is that the number of bits for the virtual linear array
short-format VHPT (Virtually mapped linear page table, VMLPT for
short) is being tested incorrectly.
There are two problems:
1. The PAL call that should tell the kernel the size of the
virtual address space isn't implemented for the simulator, so
the kernel uses the default 50. This is addressed separately
in dc90e95f31
2. In arch/ia64/mm/init.c there's code to calcualte the size
of the VMLPT based on the number of implemented virtual address
bits and the page size. It checks to see if the VMLPT base
address overlaps the top of the mapped region, but this check
doesn't allow for the address space hole, and in fact will
never trigger.
Here's an alternative test and panic, that I think is more accurate.
Signed-off-by: Peter Chubb <peterc@gelato.unsw.edu.au>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Currently, region numbers are defined in several files, with several
names. For example, we have REGION_KERNEL in asm/page.h and
RGN_KERNEL in pgtable.h
We also have address definitions that should depend on the
RGN_XXX macros, but are currently just long constants.
The following patch reorganises all the definitions so that they have
the same form (RGN_XXX), are in one place, and that addresses that
depend on RGN_XXX are derived from them.
(This is a necessary but not sufficient patch to allow UML-like
operation on IA64).
Thanks to David Mosberger for catching the change I missed in mmu_context.h.
Signed-off-by: Peter Chubb <peterc@gelato.unsw.edu.au>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Jesse Barnes provided the original version of this patch months ago, but
other changes kept conflicting with it, so it got deferred. Greg Edwards
dug it out of obscurity just over a week ago, and almost immediately
another conflicting patch appeared (Bob Picco's memory-less nodes).
I've resolved the conflicts and got it running again. CONFIG_SGI_TIOCX
is set to "y" in defconfig, which causes a Tiger to not boot (oops in
tiocx_init). But that can be resolved later ... get this in now before it
gets stale again.
Signed-off-by: Tony Luck <tony.luck@intel.com>
I reworked how nodes with only CPUs are treated. The patch below seems
simpler to me and has eliminated the complicated routine
reassign_cpu_only_nodes. There isn't any longer the requirement
to modify ACPI NUMA information which was in large part the
complexity introduced in reassign_cpu_only_nodes.
This patch will produce a different number of nodes. For example,
reassign_cpu_only_nodes would reduce two CPUonly nodes and one memory node
configuration to one memory+CPUs node configuration. This patch
doesn't change the number of nodes which means the user will see three. Two
nodes without memory and one node with all the memory.
While doing this patch, I noticed that early_nr_phys_cpus_node isn't serving
any useful purpose. It is called once in find_pernode_space but the value
isn't used to computer pernode space.
Signed-off-by: bob.picco <bob.picco@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
As many of you know that kprobes exist in the main line kernel for various
architecture including i386, x86_64, ppc64 and sparc64. Attached patches
following this mail are a port of Kprobes and Jprobes for IA64.
I have tesed this patches for kprobes and Jprobes and this seems to work fine.
I have tested this patch by inserting kprobes on various slots and various
templates including various types of branch instructions.
I have also tested this patch using the tool
http://marc.theaimsgroup.com/?l=linux-kernel&m=111657358022586&w=2 and the
kprobes for IA64 works great.
Here is list of TODO things and pathes for the same will appear soon.
1) Support kprobes on "mov r1=ip" type of instruction
2) Support Kprobes and Jprobes to exist on the same address
3) Support Return probes
3) Architecture independent cleanup of kprobes
This patch adds the kdebug die notification mechanism needed by Kprobes.
For break instruction on Branch type slot, imm21 is ignored and value
zero is placed in IIM register, hence we need to handle kprobes
for switch case zero.
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Signed-off-by: Rusty Lynch <Rusty.lynch@intel.com>
From: Rusty Lynch <rusty.lynch@intel.com>
At the point in traps.c where we recieve a break with a zero value, we can
not say if the break was a result of a kprobe or some other debug facility.
This simple patch changes the informational string to a more correct "break
0" value, and applies to the 2.6.12-rc2-mm2 tree with all the kprobes
patches that were just recently included for the next mm cut.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch effectively eliminates direct use of pgdat->node_mem_map outside
of the DISCONTIG code. On a flat memory system, these fields aren't
currently used, neither are they on a sparsemem system.
There was also a node_mem_map(nid) macro on many architectures. Its use
along with the use of ->node_mem_map itself was not consistent. It has
been removed in favor of two new, more explicit, arch-independent macros:
pgdat_page_nr(pgdat, pagenr)
nid_page_nr(nid, pagenr)
I called them "pgdat" and "nid" because we overload the term "node" to mean
"NUMA node", "DISCONTIG node" or "pg_data_t" in very confusing ways. I
believe the newer names are much clearer.
These macros can be overridden in the sparsemem case with a theoretically
slower operation using node_start_pfn and pfn_to_page(), instead. We could
make this the only behavior if people want, but I don't want to change too
much at once. One thing at a time.
This patch removes more code than it adds.
Compile tested on alpha, alpha discontig, arm, arm-discontig, i386, i386
generic, NUMAQ, Summit, ppc64, ppc64 discontig, and x86_64. Full list
here: http://sr71.net/patches/2.6.12/2.6.12-rc1-mhp2/configs/
Boot tested on NUMAQ, x86 SMP and ppc64 power4/5 LPARs.
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Martin J. Bligh <mbligh@aracnet.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
A lot of the code in arch/*/mm/hugetlbpage.c is quite similar. This patch
attempts to consolidate a lot of the code across the arch's, putting the
combined version in mm/hugetlb.c. There are a couple of uglyish hacks in
order to covert all the hugepage archs, but the result is a very large
reduction in the total amount of code. It also means things like hugepage
lazy allocation could be implemented in one place, instead of six.
Tested, at least a little, on ppc64, i386 and x86_64.
Notes:
- this patch changes the meaning of set_huge_pte() to be more
analagous to set_pte()
- does SH4 need s special huge_ptep_get_and_clear()??
Acked-by: William Lee Irwin <wli@holomorphy.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Thanks to Mark for tracking down this one. Users of __copy_from_user_inatomic()
will be sad if we don't handle lfetch faults for the "no_context" case.
Signed-off-by: Tony Luck <tony.luck@intel.com>
This patch introduces using the quicklists for pgd, pmd, and pte levels
by combining the alloc and free functions into a common set of routines.
This greatly simplifies the reading of this header file.
This patch is simple but necessary for large numa configurations.
It simply ensures that only pages from the local node are added to a
cpus quicklist. This prevents the trapping of pages on a remote nodes
quicklist by starting a process, touching a large number of pages to
fill pmd and pte entries, migrating to another node, and then unmapping
or exiting. With those conditions, the pages get trapped and if the
machine has more than 100 nodes of the same size, the calculation of
the pgtable high water mark will be larger than any single node so page
table cache flushing will never occur.
I ran lmbench lat_proc fork and lat_proc exec on a zx1 with and without
this patch and did not notice any change.
On an sn2 machine, there was a slight improvement which is possibly
due to pages from other nodes trapped on the test node before starting
the run. I did not investigate further.
This patch shrinks the quicklist based upon free memory on the node
instead of the high/low water marks. I have written it to enable
preemption periodically and recalculate the amount to shrink every time
we have freed enough pages that the quicklist size should have grown.
I rescan the nodes zones each pass because other processess may be
draining node memory at the same time as we are adding.
Signed-off-by: Robin Holt <holt@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
ia64 and ppc64 had hugetlb_free_pgtables functions which were no longer being
called, and it wasn't obvious what to do about them.
The ppc64 case turns out to be easy: the associated tables are noted elsewhere
and freed later, safe to either skip its hugetlb areas or go through the
motions of freeing nothing. Since ia64 does need a special case, restore to
ppc64 the special case of skipping them.
The ia64 hugetlb case has been broken since pgd_addr_end went in, though it
probably appeared to work okay if you just had one such area; in fact it's
been broken much longer if you consider a long munmap spanning from another
region into the hugetlb region.
In the ia64 hugetlb region, more virtual address bits are available than in
the other regions, yet the page tables are structured the same way: the page
at the bottom is larger. Here we need to scale down each addr before passing
it to the standard free_pgd_range. Was about to write a hugely_scaled_down
macro, but found htlbpage_to_page already exists for just this purpose. Fixed
off-by-one in ia64 is_hugepage_only_range.
Uninline free_pgd_range to make it available to ia64. Make sure the
vma-gathering loop in free_pgtables cannot join a hugepage_only_range to any
other (safe to join huges? probably but don't bother).
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Recent woes with some arches needing their own pgd_addr_end macro; and 4-level
clear_page_range regression since 2.6.10's clear_page_tables; and its
long-standing well-known inefficiency in searching throughout the higher-level
page tables for those few entries to clear and free: all can be blamed on
ignoring the list of vmas when we free page tables.
Replace exit_mmap's clear_page_range of the total user address space by
free_pgtables operating on the mm's vma list; unmap_region use it in the same
way, giving floor and ceiling beyond which it may not free tables. This
brings lmbench fork/exec/sh numbers back to 2.6.10 (unless preempt is enabled,
in which case latency fixes spoil unmap_vmas throughput).
Beware: the do_mmap_pgoff driver failure case must now use unmap_region
instead of zap_page_range, since a page table might have been allocated, and
can only be freed while it is touched by some vma.
Move free_pgtables from mmap.c to memory.c, where its lower levels are adapted
from the clear_page_range levels. (Most of free_pgtables' old code was
actually for a non-existent case, prev not properly set up, dating from before
hch gave us split_vma.) Pass mmu_gather** in the public interfaces, since we
might want to add latency lockdrops later; but no attempt to do so yet, going
by vma should itself reduce latency.
But what if is_hugepage_only_range? Those ia64 and ppc64 cases need careful
examination: put that off until a later patch of the series.
What of x86_64's 32bit vdso page __map_syscall32 maps outside any vma?
And the range to sparc64's flush_tlb_pgtables? It's less clear to me now that
we need to do more than is done here - every PMD_SIZE ever occupied will be
flushed, do we really have to flush every PGDIR_SIZE ever partially occupied?
A shame to complicate it unnecessarily.
Special thanks to David Miller for time spent repairing my ceilings.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
Chen, Kenneth W