platform/kernel/linux-exynos.git
19 years ago[PATCH] mm: tlb_is_full_mm was obscure
Hugh Dickins [Sun, 30 Oct 2005 01:16:02 +0000 (18:16 -0700)]
[PATCH] mm: tlb_is_full_mm was obscure

tlb_is_full_mm?  What does that mean?  The TLB is full?  No, it means that the
mm's last user has gone and the whole mm is being torn down.  And it's an
inline function because sparc64 uses a different (slightly better)
"tlb_frozen" name for the flag others call "fullmm".

And now the ptep_get_and_clear_full macro used in zap_pte_range refers
directly to tlb->fullmm, which would be wrong for sparc64.  Rather than
correct that, I'd prefer to scrap tlb_is_full_mm altogether, and change
sparc64 to just use the same poor name as everyone else - is that okay?

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: tlb_gather_mmu get_cpu_var
Hugh Dickins [Sun, 30 Oct 2005 01:16:01 +0000 (18:16 -0700)]
[PATCH] mm: tlb_gather_mmu get_cpu_var

tlb_gather_mmu dates from before kernel preemption was allowed, and uses
smp_processor_id or __get_cpu_var to find its per-cpu mmu_gather.  That works
because it's currently only called after getting page_table_lock, which is not
dropped until after the matching tlb_finish_mmu.  But don't rely on that, it
will soon change: now disable preemption internally by proper get_cpu_var in
tlb_gather_mmu, put_cpu_var in tlb_finish_mmu.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: move_page_tables by extents
Hugh Dickins [Sun, 30 Oct 2005 01:16:00 +0000 (18:16 -0700)]
[PATCH] mm: move_page_tables by extents

Speeding up mremap's moving of ptes has never been a priority, but the locking
will get more complicated shortly, and is already too baroque.

Scrap the current one-by-one moving, do an extent at a time: curtailed by end
of src and dst pmds (have to use PMD_SIZE: the way pmd_addr_end gets elided
doesn't match this usage), and by latency considerations.

One nice property of the old method is lost: it never allocated a page table
unless absolutely necessary, so you could free empty page tables by mremapping
to and fro.  Whereas this way, it allocates a dst table wherever there was a
src table.  I keep diving in to reinstate the old behaviour, then come out
preferring not to clutter how it now is.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: page fault handlers tidyup
Hugh Dickins [Sun, 30 Oct 2005 01:15:59 +0000 (18:15 -0700)]
[PATCH] mm: page fault handlers tidyup

Impose a little more consistency on the page fault handlers do_wp_page,
do_swap_page, do_anonymous_page, do_no_page, do_file_page: why not pass their
arguments in the same order, called the same names?

break_cow is all very well, but what it did was inlined elsewhere: easier to
compare if it's brought back into do_wp_page.

do_file_page's fallback to do_no_page dates from a time when we were testing
pte_file by using it wherever possible: currently it's peculiar to nonlinear
vmas, so just check that.  BUG_ON if not?  Better not, it's probably page
table corruption, so just show the pte: hmm, there's a pte_ERROR macro, let's
use that for do_wp_page's invalid pfn too.

Hah!  Someone in the ppc64 world noticed pte_ERROR was unused so removed it:
restored (and say "pud" not "pmd" in its pud_ERROR).

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: exit_mmap need not reset
Hugh Dickins [Sun, 30 Oct 2005 01:15:58 +0000 (18:15 -0700)]
[PATCH] mm: exit_mmap need not reset

exit_mmap resets various mm_struct fields, but the mm is well on its way out,
and none of those fields matter by this point.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: unlink_file_vma, remove_vma
Hugh Dickins [Sun, 30 Oct 2005 01:15:57 +0000 (18:15 -0700)]
[PATCH] mm: unlink_file_vma, remove_vma

Divide remove_vm_struct into two parts: first anon_vma_unlink plus
unlink_file_vma, to unlink the vma from the list and tree by which rmap or
vmtruncate might find it; then remove_vma to close, fput and free.

The intention here is to do the anon_vma_unlink and unlink_file_vma earlier,
in free_pgtables before freeing any page tables: so we can be sure that any
page tables traversed by rmap and vmtruncate are stable (and other, ordinary
cases are stabilized by holding mmap_sem).

This will be crucial to traversing pgd,pud,pmd without page_table_lock.  But
testing the split-out patch showed that lifting the page_table_lock is
symbiotically necessary to make this change - the lock ordering is wrong to
move those unlinks into free_pgtables while it's under ptlock.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: remove_vma_list consolidation
Hugh Dickins [Sun, 30 Oct 2005 01:15:56 +0000 (18:15 -0700)]
[PATCH] mm: remove_vma_list consolidation

unmap_vma doesn't amount to much, let's put it inside unmap_vma_list.  Except
it doesn't unmap anything, unmap_region just did the unmapping: rename it to
remove_vma_list.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: vm_stat_account unshackled
Hugh Dickins [Sun, 30 Oct 2005 01:15:56 +0000 (18:15 -0700)]
[PATCH] mm: vm_stat_account unshackled

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>
19 years ago[PATCH] mm: anon is already wrprotected
Hugh Dickins [Sun, 30 Oct 2005 01:15:55 +0000 (18:15 -0700)]
[PATCH] mm: anon is already wrprotected

do_anonymous_page's pte_wrprotect causes some confusion: in such a case,
vm_page_prot must already be forcing COW, so must omit write permission, and
so the pte_wrprotect is redundant.  Replace it by a comment to that effect,
and reword the comment on unuse_pte which also caused confusion.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: zap_pte_range dont dirty anon
Hugh Dickins [Sun, 30 Oct 2005 01:15:54 +0000 (18:15 -0700)]
[PATCH] mm: zap_pte_range dont dirty anon

zap_pte_range already avoids wasting time to mark_page_accessed on anon pages:
it can also skip anon set_page_dirty - the page only needs to be marked dirty
if shared with another mm, but that will say pte_dirty too.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: msync_pte_range progress
Hugh Dickins [Sun, 30 Oct 2005 01:15:53 +0000 (18:15 -0700)]
[PATCH] mm: msync_pte_range progress

Use latency breaking in msync_pte_range like that in copy_pte_range, instead
of the ugly CONFIG_PREEMPT filemap_msync alternatives.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: copy_pte_range progress fix
Hugh Dickins [Sun, 30 Oct 2005 01:15:53 +0000 (18:15 -0700)]
[PATCH] mm: copy_pte_range progress fix

My latency breaking in copy_pte_range didn't work as intended: instead of
checking at regularish intervals, after the first interval it checked every
time around the loop, too impatient to be preempted.  Fix that.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] slab: add additional debugging to detect slabs from the wrong node
Christoph Lameter [Sun, 30 Oct 2005 01:15:52 +0000 (18:15 -0700)]
[PATCH] slab: add additional debugging to detect slabs from the wrong node

This patch adds some stack dumps if the slab logic is processing slab
blocks from the wrong node.  This is necessary in order to detect
situations as encountered by Petr.

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] shrink_list(): skip anon pages if not may_swap
Lee Schermerhorn [Sun, 30 Oct 2005 01:15:51 +0000 (18:15 -0700)]
[PATCH] shrink_list(): skip anon pages if not may_swap

Martin Hicks' page cache reclaim patch added the 'may_swap' flag to the
scan_control struct; and modified shrink_list() not to add anon pages to
the swap cache if may_swap is not asserted.

Ref:  http://marc.theaimsgroup.com/?l=linux-mm&m=111461480725322&w=4

However, further down, if the page is mapped, shrink_list() calls
try_to_unmap() which will call try_to_unmap_one() via try_to_unmap_anon ().
 try_to_unmap_one() will BUG_ON() an anon page that is NOT in the swap
cache.  Martin says he never encountered this path in his testing, but
agrees that it might happen.

This patch modifies shrink_list() to skip anon pages that are not already
in the swap cache when !may_swap, rather than just not adding them to the
cache.

Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm/msync.c cleanup
OGAWA Hirofumi [Sun, 30 Oct 2005 01:15:50 +0000 (18:15 -0700)]
[PATCH] mm/msync.c cleanup

This is not problem actually, but sync_page_range() is using for exported
function to filesystems.

The msync_xxx is more readable at least to me.

Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Acked-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Remove near all BUGs in mm/mempolicy.c
Andi Kleen [Sun, 30 Oct 2005 01:15:49 +0000 (18:15 -0700)]
[PATCH] Remove near all BUGs in mm/mempolicy.c

Most of them can never be triggered and were only for development.

Signed-off-by: "Andi Kleen" <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Convert mempolicies to nodemask_t
Andi Kleen [Sun, 30 Oct 2005 01:15:48 +0000 (18:15 -0700)]
[PATCH] Convert mempolicies to nodemask_t

The NUMA policy code predated nodemask_t so it used open coded bitmaps.
Convert everything to nodemask_t.  Big patch, but shouldn't have any actual
behaviour changes (except I removed one unnecessary check against
node_online_map and one unnecessary BUG_ON)

Signed-off-by: "Andi Kleen" <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: set per-cpu-pages lower threshold to zero
Seth, Rohit [Sun, 30 Oct 2005 01:15:48 +0000 (18:15 -0700)]
[PATCH] mm: set per-cpu-pages lower threshold to zero

Set the low water mark for hot pages in pcp to zero.

(akpm: for the life of me I cannot remember why we created pcp->low.  Neither
can Martin and the changelog is silent.  Maybe it was just a brainfart, but I
have this feeling that there was a reason.  If not, we should remove the
fields completely.  We'll see.)

Signed-off-by: Rohit Seth <rohit.seth@intel.com>
Cc: <linux-mm@kvack.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: page_alloc: increase size of per-cpu-pages
Seth, Rohit [Sun, 30 Oct 2005 01:15:47 +0000 (18:15 -0700)]
[PATCH] mm: page_alloc: increase size of per-cpu-pages

Increase the page allocator's per-cpu magazines from 1/4MB to 1/2MB.

Over 100+ runs for a workload, the difference in mean is about 2%.  The best
results for both are almost same.  Though the max variation in results with
1/2MB is only 2.2%, whereas with 1/4MB it is 12%.

Signed-off-by: Rohit Seth <rohit.seth@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] swaptoken tuning
Rik Van Riel [Sun, 30 Oct 2005 01:15:46 +0000 (18:15 -0700)]
[PATCH] swaptoken tuning

It turns out that the original swap token implementation, by Song Jiang, only
enforced the swap token while the task holding the token is handling a page
fault.  This patch approximates that, without adding an additional flag to the
mm_struct, by checking whether the mm->mmap_sem is held for reading, like the
page fault code does.

This patch has the effect of automatically, and gradually, disabling the
enforcement of the swap token when there is little or no paging going on, and
"turning up" the intensity of the swap token code the more the task holding
the token is thrashing.

Thanks to Song Jiang for pointing out this aspect of the token based thrashing
control concept.

The new code shows a slight degradation over the old swap token code, but
still a big win over running without the swap token.

2.6.12+ swap token disabled

$ for i in `seq 10` ; do /usr/bin/time ./qsbench -n 30000000 -p 3 ; done
101.74user 23.13system 8:26.91elapsed 24%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (38597major+430315minor)pagefaults 0swaps
101.98user 24.91system 8:03.06elapsed 26%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (33939major+430457minor)pagefaults 0swaps
101.93user 22.12system 7:34.90elapsed 27%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (33166major+421267minor)pagefaults 0swaps
101.82user 22.38system 8:31.40elapsed 24%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (39338major+433262minor)pagefaults 0swaps

2.6.12+ swap token enabled, timeout 300 seconds

$ for i in `seq 4` ; do /usr/bin/time ./qsbench -n 30000000 -p 3 ; done
102.58user 16.08system 3:41.44elapsed 53%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (19707major+285786minor)pagefaults 0swaps
102.07user 19.56system 4:00.64elapsed 50%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (19012major+299259minor)pagefaults 0swaps
102.64user 18.25system 4:07.31elapsed 48%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (21990major+304831minor)pagefaults 0swaps
101.39user 19.41system 5:15.81elapsed 38%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (24850major+323321minor)pagefaults 0swaps

2.6.12+ with new swap token code, timeout 300 seconds

$ for i in `seq 4` ; do /usr/bin/time ./qsbench -n 30000000 -p 3 ; done
101.87user 24.66system 5:53.20elapsed 35%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (26848major+363497minor)pagefaults 0swaps
102.83user 19.95system 4:17.25elapsed 47%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (19946major+305722minor)pagefaults 0swaps
102.09user 19.46system 5:12.57elapsed 38%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (25461major+334994minor)pagefaults 0swaps
101.67user 20.61system 4:52.97elapsed 41%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (22190major+329508minor)pagefaults 0swaps

Signed-off-by: Rik Van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] add sem_is_read/write_locked()
Rik Van Riel [Sun, 30 Oct 2005 01:15:44 +0000 (18:15 -0700)]
[PATCH] add sem_is_read/write_locked()

Add sem_is_read/write_locked functions to the read/write semaphores, along the
same lines of the *_is_locked spinlock functions.  The swap token tuning patch
uses sem_is_read_locked; sem_is_write_locked is added for completeness.

Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] fix alpha breakage
Ivan Kokshaysky [Sun, 30 Oct 2005 01:15:43 +0000 (18:15 -0700)]
[PATCH] fix alpha breakage

barrier.h uses barrier() in non-SMP case.  And doesn't include compiler.h.

Cc: Al Viro <viro@ftp.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] TIMERS: add missing compensation for HZ == 250
YOSHIFUJI Hideaki [Sun, 30 Oct 2005 01:15:42 +0000 (18:15 -0700)]
[PATCH] TIMERS: add missing compensation for HZ == 250

Add missing compensation for (HZ == 250) != (1 << SHIFT_HZ) in
second_overflow().

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] vmalloc_node
Christoph Lameter [Sun, 30 Oct 2005 01:15:41 +0000 (18:15 -0700)]
[PATCH] vmalloc_node

This patch adds

vmalloc_node(size, node) -> Allocate necessary memory on the specified node

and

get_vm_area_node(size, flags, node)

and the other functions that it depends on.

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Sat, 29 Oct 2005 21:02:16 +0000 (14:02 -0700)]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

19 years ago[ARM] 3061/1: cleanup the XIP link address mess
Nicolas Pitre [Sat, 29 Oct 2005 20:44:56 +0000 (21:44 +0100)]
[ARM] 3061/1: cleanup the XIP link address mess

Patch from Nicolas Pitre

Since vmlinux.lds.S is preprocessed, we can use the defines already
present in asm/memory.h (allowed by patch #3060) for the XIP kernel link
address instead of relying on a duplicated Makefile hardcoded value, and
also get rid of its dependency on awk to handle it at the same time.

While at it let's clean XIP stuff even further and make things clearer
in head.S with a nice code reduction.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
19 years ago[ARM] 3060/1: allow constants found in asm/memory.h to be used in asm code
Nicolas Pitre [Sat, 29 Oct 2005 20:44:55 +0000 (21:44 +0100)]
[ARM] 3060/1: allow constants found in asm/memory.h to be used in asm code

Patch from Nicolas Pitre

This patch allows for assorted type of cleanups by letting assembly code
use the same set of defines for constant values and avoid duplicated
definitions that might not always be in sync, or that might simply be
confusing due to the different names for the same thing.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
19 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Sat, 29 Oct 2005 19:19:15 +0000 (12:19 -0700)]
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

19 years agoUpdate MIPS defconfig files.
Ralf Baechle [Sat, 29 Oct 2005 18:32:54 +0000 (19:32 +0100)]
Update MIPS defconfig files.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years ago prom_free_prom_memory() returns unsigned long
Arthur Othieno [Fri, 28 Oct 2005 04:42:56 +0000 (00:42 -0400)]
prom_free_prom_memory() returns unsigned long

    Some boards declare prom_free_prom_memory as a void function but the
    caller free_initmem() expects a return value.

    Fix those up and return 0 instead, just like everyone else does.

Signed-off-by: Arthur Othieno <a.othieno@bluewin.ch>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoGet rid of SINGLE_ONLY_FPU. Linux does not support half FPU other than
Ralf Baechle [Sun, 23 Oct 2005 14:05:47 +0000 (15:05 +0100)]
Get rid of SINGLE_ONLY_FPU.  Linux does not support half FPU other than
by emulation of a full FPU.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoFix all the get_user / put_user related sparse warnings.
Ralf Baechle [Sun, 23 Oct 2005 12:58:21 +0000 (13:58 +0100)]
Fix all the get_user / put_user related sparse warnings.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoDelete unused ieee754_cname[] and declaration.
Ralf Baechle [Sun, 23 Oct 2005 12:48:12 +0000 (13:48 +0100)]
Delete unused ieee754_cname[] and declaration.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoInclude for prototypes.
Ralf Baechle [Sun, 23 Oct 2005 12:46:25 +0000 (13:46 +0100)]
Include for prototypes.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoProtect against multiple inclusion.
Ralf Baechle [Sun, 23 Oct 2005 12:44:31 +0000 (13:44 +0100)]
Protect against multiple inclusion.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoRemove useless casts of kmalloc return values.
Ralf Baechle [Fri, 21 Oct 2005 21:26:07 +0000 (22:26 +0100)]
Remove useless casts of kmalloc return values.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoHack to resolve longstanding prefetch issue
Ralf Baechle [Thu, 20 Oct 2005 21:55:26 +0000 (22:55 +0100)]
Hack to resolve longstanding prefetch issue

Prefetching may be fatal on some systems if we're prefetching beyond the
end of memory on some systems.  It's also a seriously bad idea on non
dma-coherent systems.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoMore foolproofing of the CPU configuration.
Ralf Baechle [Thu, 20 Oct 2005 21:33:09 +0000 (22:33 +0100)]
More foolproofing of the CPU configuration.

Limit the number of cpu type options in the cpu menu to just those
types that are actually available for the select platform.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agopci-expmem-hack
Andrew Isaacson [Thu, 20 Oct 2005 06:59:46 +0000 (23:59 -0700)]
pci-expmem-hack

CFE 1.2.5 and earlier fails to turn on the ExpMemEn bit in the
PCIFeatureControl register, which means that DMA does not work
beyond physical address 01_0000_0000, ergo to DRAM beyond 1GB.

With ExpMemEn turned on, 01_0000_0000-0f_ffff_ffff is mapped,
so DMA works for up to 61 GB of DRAM.

Will be fixed in CFE 1.2.6 (yet to be released).

Signed-Off-By: Andy Isaacson <adi@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoBCM1480 HT support
Andrew Isaacson [Thu, 20 Oct 2005 06:59:11 +0000 (23:59 -0700)]
BCM1480 HT support

PCI support code for PLX 7250 PCI-X tunnel on BCM91480B BigSur board.

Signed-Off-By: Andy Isaacson <adi@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoSupport for the BCM1480 on-chip PCI-X bridge.
Andrew Isaacson [Thu, 20 Oct 2005 06:58:49 +0000 (23:58 -0700)]
Support for the BCM1480 on-chip PCI-X bridge.

Signed-Off-By: Andy Isaacson <adi@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoSB1 cache exception handling.
Andrew Isaacson [Thu, 20 Oct 2005 06:57:40 +0000 (23:57 -0700)]
SB1 cache exception handling.

Expand SB1 cache error handling by adding SB1_CEX_ALWAYS_FATAL and
SB1_CEX_STALL, allowing configurable behavior on cache errors.

Signed-Off-By: Andy Isaacson <adi@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoSupport for BigSur board.
Andrew Isaacson [Thu, 20 Oct 2005 06:57:11 +0000 (23:57 -0700)]
Support for BigSur board.

Signed-Off-By: Andy Isaacson <adi@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoAdd support for BCM1480 family of chips.
Andrew Isaacson [Thu, 20 Oct 2005 06:56:38 +0000 (23:56 -0700)]
Add support for BCM1480 family of chips.

 - Kconfig and Makefile changes
 - arch/mips/sibyte/bcm1480/
 - changes to sibyte common code to support 1480

Signed-Off-By: Andy Isaacson <adi@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoAdd support for SB1A CPU.
Andrew Isaacson [Thu, 20 Oct 2005 06:56:20 +0000 (23:56 -0700)]
Add support for SB1A CPU.

Signed-Off-By: Andy Isaacson <adi@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoSibyte header cleanup
Andrew Isaacson [Thu, 20 Oct 2005 06:55:57 +0000 (23:55 -0700)]
Sibyte header cleanup

Update sibyte headers to match Broadcom internal copies:
 - comment cleanup and updates
 - fix LittleSur part number to match the board silkscreen

Signed-Off-By: Andy Isaacson <adi@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoBCM1480 headers
Andrew Isaacson [Thu, 20 Oct 2005 06:55:11 +0000 (23:55 -0700)]
BCM1480 headers

Add header files for BCM1480/1280/1455/1255 family of chips, and
update sb1250 headers which are shared by BCM1480 family.

Signed-Off-By: Andy Isaacson <adi@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
diff --git a/include/asm-mips/sibyte/bcm1480_int.h b/include/asm-mips/sibyte/bcm1480_int.h
new file mode 100644

19 years agoSibyte fixes
Andrew Isaacson [Thu, 20 Oct 2005 06:54:43 +0000 (23:54 -0700)]
Sibyte fixes

Fix typo in cpu_probe_sibyte.

Signed-Off-By: Andy Isaacson <adi@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoMake UL what should be UL.
Ralf Baechle [Wed, 19 Oct 2005 13:45:09 +0000 (14:45 +0100)]
Make UL what should be UL.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoFix zero length sys_cacheflush
Atsushi Nemoto [Wed, 19 Oct 2005 10:57:14 +0000 (19:57 +0900)]
Fix zero length sys_cacheflush

Cacheflush(0, 0, 0) was crashing the system.  This is because
flush_icache_range(start, end) tries to flushing whole address space
(0 - ~0UL) if both start and end are zero.

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoGet 64-bit right in the kgdb stub.
Ralf Baechle [Tue, 18 Oct 2005 12:25:29 +0000 (13:25 +0100)]
Get 64-bit right in the kgdb stub.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoSys_lookup_dcookie arguments occupy 4 argument slots.
Ralf Baechle [Tue, 18 Oct 2005 11:48:31 +0000 (12:48 +0100)]
Sys_lookup_dcookie arguments occupy 4 argument slots.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoFPU emulator garbage collection.
Ralf Baechle [Tue, 18 Oct 2005 09:26:46 +0000 (10:26 +0100)]
FPU emulator garbage collection.

First argument of fpu_emulator_cop1Handler() was unused.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoDon't print file name and line in die and die_if_kernel.
Ralf Baechle [Thu, 13 Oct 2005 16:07:54 +0000 (17:07 +0100)]
Don't print file name and line in die and die_if_kernel.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoRename page argument of flush_cache_page to something more descriptive.
Ralf Baechle [Tue, 11 Oct 2005 23:02:34 +0000 (00:02 +0100)]
Rename page argument of flush_cache_page to something more descriptive.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoDefine EOWNERDEAD and ENOTRECOVERABLE.
Ralf Baechle [Sun, 9 Oct 2005 17:56:01 +0000 (18:56 +0100)]
Define EOWNERDEAD and ENOTRECOVERABLE.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoSliceup Kconfig; it's grown too large.
Ralf Baechle [Sat, 29 Oct 2005 18:32:41 +0000 (19:32 +0100)]
Sliceup Kconfig; it's grown too large.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoMore configcheck fixes.
Ralf Baechle [Sat, 29 Oct 2005 18:32:40 +0000 (19:32 +0100)]
More configcheck fixes.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years ago2.6.14-rc1 updates for MIPS compat types.
Ralf Baechle [Sat, 29 Oct 2005 18:32:40 +0000 (19:32 +0100)]
2.6.14-rc1 updates for MIPS compat types.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoComplete the fcntl.h cleanup.
Ralf Baechle [Sat, 29 Oct 2005 18:32:40 +0000 (19:32 +0100)]
Complete the fcntl.h cleanup.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoCleanup Sibyte Kconfig a bit further.
Ralf Baechle [Sat, 29 Oct 2005 18:32:39 +0000 (19:32 +0100)]
Cleanup Sibyte Kconfig a bit further.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoDate: Fri Jan 14 03:03:23 2005 +0000
Ralf Baechle [Fri, 14 Jan 2005 03:03:23 +0000 (03:03 +0000)]
Date:   Fri Jan 14 03:03:23 2005 +0000

Locking cleanups.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoFix weirdness in <asm/bug.h>
Ralf Baechle [Sat, 29 Oct 2005 18:32:38 +0000 (19:32 +0100)]
Fix weirdness in <asm/bug.h>

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoFix wrong comment.
Ralf Baechle [Sat, 29 Oct 2005 18:32:38 +0000 (19:32 +0100)]
Fix wrong comment.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoFixup a few lose ends in explicit support for MIPS R1/R2.
Ralf Baechle [Fri, 7 Oct 2005 15:58:15 +0000 (16:58 +0100)]
Fixup a few lose ends in explicit support for MIPS R1/R2.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoDocument the meaning of the CPU_MIPS32, CPU_MIPS64, CPU_MIPSR1 and
Ralf Baechle [Fri, 7 Oct 2005 11:06:12 +0000 (12:06 +0100)]
Document the meaning of the CPU_MIPS32, CPU_MIPS64, CPU_MIPSR1 and
CPU_MIPSR2.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoProtect manipulation of c0_status against preemption and multithreading.
Ralf Baechle [Thu, 6 Oct 2005 16:39:32 +0000 (17:39 +0100)]
Protect manipulation of c0_status against preemption and multithreading.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoDetect 4KSD and treat it like 4KSc.
Ralf Baechle [Tue, 4 Oct 2005 14:01:26 +0000 (15:01 +0100)]
Detect 4KSD and treat it like 4KSc.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoWe're no longer hosted on oss for ages ...
Ralf Baechle [Tue, 4 Oct 2005 12:30:10 +0000 (13:30 +0100)]
We're no longer hosted on oss for ages ...

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoConvert the remaining SPIN_LOCK_UNLOCKED instances to DEFINE_SPINLOCK.
Ralf Baechle [Mon, 3 Oct 2005 12:41:19 +0000 (13:41 +0100)]
Convert the remaining SPIN_LOCK_UNLOCKED instances to DEFINE_SPINLOCK.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoDefine and initialize kdb_lock using DEFINE_SPINLOCK.
Ralf Baechle [Mon, 3 Oct 2005 12:40:26 +0000 (13:40 +0100)]
Define and initialize kdb_lock using DEFINE_SPINLOCK.
Convert kgdb_cpulock into a raw_spinlock_t.

SPIN_LOCK_UNLOCKED is deprecated and it's replacement DEFINE_SPINLOCK is
not suitable for arrays of spinlocks.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoMake kgdb_wait static.
Ralf Baechle [Mon, 3 Oct 2005 12:30:57 +0000 (13:30 +0100)]
Make kgdb_wait static.

Nothing outside gdb-stub.c uses kgdb_wait, so change it's definition to
static.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoDon't copy SB1 cache error handler to uncached memory.
Ralf Baechle [Sat, 1 Oct 2005 19:22:39 +0000 (20:22 +0100)]
Don't copy SB1 cache error handler to uncached memory.

This may have made sense on a paranoid day with pass 1 BCM1250 processors
that were throwing cache error exception left and right for no good
reason.  On modern silicion that hardly makes sense and the code had
gotten just an obscurity ...

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoProvide 64-bit address space definitions for the Sibyte SB1 CPU core.
Ralf Baechle [Sat, 1 Oct 2005 16:34:35 +0000 (17:34 +0100)]
Provide 64-bit address space definitions for the Sibyte SB1 CPU core.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoNo need to explicitly call __read_64bit_c0_split; __read_64bit_c0_register
Ralf Baechle [Sat, 1 Oct 2005 12:14:58 +0000 (13:14 +0100)]
No need to explicitly call __read_64bit_c0_split; __read_64bit_c0_register
will do that itself iff needed.  Fix format string.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoFix stale comment in c-sb1.c.
Andrew Isaacson [Wed, 22 Jun 2005 23:02:03 +0000 (16:02 -0700)]
Fix stale comment in c-sb1.c.

Signed-Off-By: Andrew Isaacson <adi@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoCleanup the mess in cpu_cache_init.
Ralf Baechle [Sat, 1 Oct 2005 12:06:32 +0000 (13:06 +0100)]
Cleanup the mess in cpu_cache_init.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoUse cpumask_t rather than hand-rolled bitmask code in sb1250_set_affinity.
Andrew Isaacson [Wed, 22 Jun 2005 23:01:09 +0000 (16:01 -0700)]
Use cpumask_t rather than hand-rolled bitmask code in sb1250_set_affinity.

Signed-Off-By: Andrew Isaacson <adi@broadcom.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoUse R4000 TLB routines for SB1 also.
Ralf Baechle [Sat, 1 Oct 2005 10:14:17 +0000 (11:14 +0100)]
Use R4000 TLB routines for SB1 also.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoFix build error caused by missmatching duplicate declaration.
Ralf Baechle [Sat, 1 Oct 2005 09:17:54 +0000 (10:17 +0100)]
Fix build error caused by missmatching duplicate declaration.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoDon't call memset to clean irq_desc; these data fields have already
Ralf Baechle [Fri, 30 Sep 2005 23:03:42 +0000 (00:03 +0100)]
Don't call memset to clean irq_desc; these data fields have already
previously been initialized statically in kernel/irq/handle.c.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoFix ARCH_KMALLOC_MINALIGN values on MIPS
Ralf Baechle [Fri, 30 Sep 2005 23:00:00 +0000 (00:00 +0100)]
Fix ARCH_KMALLOC_MINALIGN values on MIPS

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoSupport for MIPSsim, the cycle accurate MIPS simulator.
Ralf Baechle [Fri, 30 Sep 2005 12:59:37 +0000 (13:59 +0100)]
Support for MIPSsim, the cycle accurate MIPS simulator.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoSwitch Sibyte profiling driver to ->compat_ioctl
Ralf Baechle [Fri, 30 Sep 2005 00:51:21 +0000 (01:51 +0100)]
Switch Sibyte profiling driver to ->compat_ioctl

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoRevise MIPS 64-bit ptrace interface
Daniel Jacobowitz [Wed, 28 Sep 2005 22:11:15 +0000 (18:11 -0400)]
Revise MIPS 64-bit ptrace interface

Change the N32 debugging ABI to something more sane, and add support
for o32 and n32 debuggers to trace n64 programs.

Signed-off-by: Daniel Jacobowitz <dan@codesourcery.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoSync c-tx39.c with c-r4k.c.
Atsushi Nemoto [Wed, 28 Sep 2005 11:24:58 +0000 (20:24 +0900)]
Sync c-tx39.c with c-r4k.c.

tx39_flush_cache_range() does nothing if !cpu_has_dc_aliases.  It should
flush d-cache and invalidate i-cache since the TX39(H2) has separate I/D
cache.

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoGlue again after removal of BUILD_BUG().
Ralf Baechle [Fri, 23 Sep 2005 20:02:38 +0000 (20:02 +0000)]
Glue again after removal of BUILD_BUG().

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoLet it compile again after i2c algo id removal. Sort headers.
Ladislav Michl [Fri, 23 Sep 2005 10:52:27 +0000 (10:52 +0000)]
Let it compile again after i2c algo id removal. Sort headers.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoSMP on Malta needs to define ARCH_HAS_IRQ_PER_CPU since 2.6.14-rc1.
Ralf Baechle [Thu, 22 Sep 2005 14:55:21 +0000 (14:55 +0000)]
SMP on Malta needs to define ARCH_HAS_IRQ_PER_CPU since 2.6.14-rc1.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
diff --git a/include/asm-mips/mach-mips/irq.h b/include/asm-mips/mach-mips/irq.h
new file mode 100644

19 years agosmc91x platform support; requires patch to smc91x.h which was sent
Pete Popov [Wed, 21 Sep 2005 21:39:44 +0000 (21:39 +0000)]
smc91x platform support; requires patch to smc91x.h which was sent
upstream.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoAu1[12]00 mmc driver. Only tested on the Au1200 at this point though
Pete Popov [Wed, 21 Sep 2005 06:18:27 +0000 (06:18 +0000)]
Au1[12]00 mmc driver. Only tested on the Au1200 at this point though
it should work on the Au1100 as well. Updated defconfig to include driver.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoAdd SOCK_DCCP definition for MIPS also.
Ralf Baechle [Tue, 20 Sep 2005 13:43:51 +0000 (13:43 +0000)]
Add SOCK_DCCP definition for MIPS also.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoThe type of sum in csum_tcpudp_nofold is "unsigned int", so when we assign
Ralf Baechle [Tue, 20 Sep 2005 10:56:26 +0000 (10:56 +0000)]
The type of sum in csum_tcpudp_nofold is "unsigned int", so when we assign
to it in an asm() block, and we're running on a system with 64-bit
registers, it is vitally important that we sign extend it correctly before
returning to C.  Otherwise the stray high bits will be preserved into
csum_fold, and on the SB-1 processor, 32-bit arithmetic on a non
sign-extended register will yield surprising results.

This caused incorrect checksums in some UDP packets for NFS root.  The
problem was mild when using a 10.0.1.x IP address, but severe when
using 192.168.1.x.

Signed-off-by: Daniel Jacobowitz <dan@codesourcery.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoKernel gpio/2 routines that will be used by some drivers.
Pete Popov [Sun, 18 Sep 2005 11:18:10 +0000 (11:18 +0000)]
Kernel gpio/2 routines that will be used by some drivers.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoFix a cache snooping issue.
Pete Popov [Sun, 18 Sep 2005 01:10:46 +0000 (01:10 +0000)]
Fix a cache snooping issue.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoUpdated pcmcia driver with pb1200 and db1200 support.
Pete Popov [Sat, 17 Sep 2005 00:38:10 +0000 (00:38 +0000)]
Updated pcmcia driver with pb1200 and db1200 support.
Updated db1200_defconfig so pcmcia is enabled by default.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years ago- moved platform structure to platform.c
Pete Popov [Thu, 15 Sep 2005 23:42:27 +0000 (23:42 +0000)]
- moved platform structure to platform.c
- fixed an iounmap warning
- export fixup_xx, needed by the module

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoFix excessive signal latencies.
Ralf Baechle [Thu, 15 Sep 2005 16:43:50 +0000 (16:43 +0000)]
Fix excessive signal latencies.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoReorder & reformat a bit.
Ralf Baechle [Thu, 15 Sep 2005 08:52:34 +0000 (08:52 +0000)]
Reorder & reformat a bit.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
19 years agoUse ARCH_MAY_HAVE_PC_FDC where needed.
Ralf Baechle [Thu, 15 Sep 2005 08:52:34 +0000 (08:52 +0000)]
Use ARCH_MAY_HAVE_PC_FDC where needed.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>