The TASK_SIZE macro should reflect the size of a user process virtual
address space. Previously for 64-bit kernels, this was not the case.
The immediate cause of pain was in
hugetlbfs/inode.c:hugetlb_get_unmapped_area() where 32-bit processes
trying to mmap a huge page would be served a page with an address
outside of the 32-bit address range. But there are other uses of
TASK_SIZE in the kernel as well that would like an accurate value.
The new definition is nice because it now makes TASK_SIZE and
TASK_SIZE_OF() yield the same value for any given process.
For 32-bit kernels there should be no change, although I did factor
out some code in asm/processor.h that became identical for the 32-bit and
64-bit cases.
__UA_LIMIT is now set to ~((1 << SEGBITS) - 1) for 64-bit kernels.
This should eliminate the possibility of getting a
AddressErrorException in the kernel for addresses that pass the
access_ok() test.
With the patch applied, I can still run o32, n32 and n64 processes,
and have an o32 shell fork/exec both n32 and n64 processes.
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/1701/
Since we no longer need to provide KM_type, the whole pte_*map_nested()
API is now redundant, remove it.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Chris Metcalf <cmetcalf@tilera.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: David Miller <davem@davemloft.net>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In the 64-bit kernel we use swapper_pg_dir for three different things.
1) xuseg mappings for kernel threads.
2) vmap mappings for all kernel-space accesses in xkseg.
3) vmap mappings for kernel modules in ksseg (kseg2).
Due to how the TLB refill handlers work, any mapping established in
xkseg or ksseg will also establish a xuseg mapping that should never
be used by the kernel.
In order to be able to use exceptions to trap NULL pointer
dereferences, we need to ensure that nothing is mapped at address
zero. Since vmap mappings in xkseg are reflected in xuseg, this means
we need to ensure that there are no vmap mappings established at the
start of xkseg. So we move back VMALLOC_START to avoid establishing
vmap mappings at the start of xkseg.
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1129/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
The SmartMIPS ASE specifies how Read Inhibit (RI) and eXecute Inhibit
(XI) bits in the page tables work. The upper two bits of EntryLo{0,1}
are RI and XI when the feature is enabled in the PageGrain register.
SmartMIPS only covers 32-bit systems. Cavium Octeon+ extends this to
64-bit systems by continuing to place the RI and XI bits in the top of
EntryLo even when EntryLo is 64-bits wide.
Because we need to carry the RI and XI bits in the PTE, the layout of
the PTE is changed. There is a two instruction overhead in the TLB
refill hot path to get the EntryLo bits into the proper position.
Also the TLB load exception has to probe the TLB to check if RI or XI
caused the exception.
Also of note is that the layout of the PTE bits is done at compile and
runtime rather than statically. In the 32-bit case this allows for
the same number of PFN bits as before the patch as the _PAGE_HUGE is
not supported in 32-bit kernels (we have _PAGE_NO_EXEC and
_PAGE_NO_READ instead of _PAGE_READ and _PAGE_HUGE).
The patch is tested on Cavium Octeon+, but should also work on 32-bit
systems with the Smart-MIPS ASE.
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/952/
Patchwork: http://patchwork.linux-mips.org/patch/956/
Patchwork: http://patchwork.linux-mips.org/patch/962/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
For 64-bit kernels with 64KB pages and two level page tables, there are
42 bits worth of virtual address space This is larger than the 40 bits of
virtual address space obtained with the default 4KB Page size and three
levels, so there are no draw backs for using two level tables with this
configuration.
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/761/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Linux kernel 2.6.32 and later allocate address space from the top of the
kernel virtual memory address space.
This patch implements virtual memory size detection for 64 bit MIPS CPUs
to avoid resulting crashes.
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/935/
Reviewed-by: David Daney <ddaney@caviumnetworks.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
By combining swapper_pg_dir and module_pg_dir, several if conditions
can be eliminated from the tlb exception handler. The reason they
can be combined is that, the effective virtual address of vmalloc
returned is at the bottom, and of module_alloc returned is at the
top. It also fixes the bug in vmalloc(), which happens when its
return address is not covered by the first pgd.
Signed-off-by: Wu Fei <at.wufei@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Beyond the requirements of the architecture standard Cavium also supports
8k and 32k pages.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Acked-by: David Daney <ddaney@caviumnetworks.com>
David Daney