Tejun Heo [Fri, 11 Mar 2011 09:33:31 +0000 (10:33 +0100)]
x86-64, NUMA: Don't call numa_set_distanc() for all possible node combinations during emulation
The distance transforming in numa_emulation() used to call
numa_set_distance() for all MAX_NUMNODES * MAX_NUMNODES node
combinations regardless of which are enabled. As numa_set_distance()
ignores all out-of-bound distance settings, this doesn't cause any
problem other than looping unnecessarily many times during boot.
However, as MAX_NUMNODES * MAX_NUMNODES can be pretty high, update the
code such that it iterates through only the enabled combinations.
Yinghai Lu identified the issue and provided an initial patch to
address the issue; however, the patch was incorrect in that it didn't
build emulated distance table when there's no physical distance table
and unnecessarily complex.
http://thread.gmane.org/gmane.linux.kernel/
1107986/focus=
1107988
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Yinghai Lu <yinghai@kernel.org>
Acked-by: Yinghai Lu <yinghai@kernel.org>
Ingo Molnar [Sat, 5 Mar 2011 06:32:45 +0000 (07:32 +0100)]
Merge branch 'x86-mm' of git://git./linux/kernel/git/tj/misc into x86/mm
Tejun Heo [Fri, 4 Mar 2011 15:32:02 +0000 (16:32 +0100)]
x86-64, NUMA: Don't assume phys node 0 is always online in numa_emulation()
Undetermined entries in emu_nid_to_phys[] are filled with zero
assuming that physical node 0 is always online; however, this might
not be true depending on hardware configuration. Find a physical node
which is actually online and use it instead.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: David Rientjes <rientjes@google.com>
LKML-Reference: <alpine.DEB.2.00.
1103020628210.31626@chino.kir.corp.google.com>
David Rientjes [Fri, 4 Mar 2011 14:17:21 +0000 (15:17 +0100)]
x86-64, NUMA: Clean up initmem_init()
This patch cleans initmem_init() so that it is more readable and doesn't
use an unnecessary array of function pointers to convolute the flow of
the code. It also makes it obvious that dummy_numa_init() will always
succeed (and documents that requirement) so that the existing BUG() is
never actually reached.
No functional change.
-tj: Updated comment for dummy_numa_init() slightly.
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Yinghai Lu [Fri, 4 Mar 2011 13:49:28 +0000 (14:49 +0100)]
x86-64, NUMA: Fix numa_emulation code with node0 without RAM
On one system that does not have RAM on node0.
When numa_emulation is compiled in, and
1. boot system without numa=fake...
2. or boot system with numa=fake=128 to make emulation fail
will get:
[ 0.092026] ------------[ cut here ]------------
[ 0.096005] kernel BUG at arch/x86/mm/numa_emulation.c:439!
[ 0.096005] invalid opcode: 0000 [#1] SMP
[ 0.096005] last sysfs file:
[ 0.096005] CPU 0
[ 0.096005] Modules linked in:
[ 0.096005]
[ 0.096005] Pid: 0, comm: swapper Not tainted
2.6.38-rc6-tip-yh-03869-gcb0491d-dirty #684 Sun Microsystems Sun Fire X4240/Sun Fire X4240
[ 0.096005] RIP: 0010:[<
ffffffff81cdc65b>] [<
ffffffff81cdc65b>] numa_add_cpu+0x56/0xcf
[ 0.096005] RSP: 0000:
ffffffff82437ed8 EFLAGS:
00010246
...
[ 0.096005] Call Trace:
[ 0.096005] [<
ffffffff81cd7931>] identify_cpu+0x2d7/0x2df
[ 0.096005] [<
ffffffff827e54fa>] identify_boot_cpu+0x10/0x30
[ 0.096005] [<
ffffffff827e5704>] check_bugs+0x9/0x2d
[ 0.096005] [<
ffffffff827dceda>] start_kernel+0x3d7/0x3f1
[ 0.096005] [<
ffffffff827dc2cc>] x86_64_start_reservations+0x9c/0xa0
[ 0.096005] [<
ffffffff827dc4ad>] x86_64_start_kernel+0x1dd/0x1e8
[ 0.096005] Code: 74 06 48 8d 04 90 eb 0f 48 c7 c0 30 d9 00 00 48 03 04 d5 90 0f 60 82 8b 00 83 f8 ff 74 0d 0f a3 05 8b 7e 92 00 19 d2 85 d2 75 02 <0f> 0b 48 98 be 00 01 00 00 48 c7 c7 e0 44 60 82 44 8b 2c 85 e0
[ 0.096005] RIP [<
ffffffff81cdc65b>] numa_add_cpu+0x56/0xcf
[ 0.096005] RSP <
ffffffff82437ed8>
[ 0.096026] ---[ end trace
a7919e7f17c0a725 ]---
We need to use early_cpu_to_node() directly, because numa_cpu_node()
will return node0 that is not onlined.
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Tejun Heo [Fri, 4 Mar 2011 09:26:36 +0000 (10:26 +0100)]
x86-64, NUMA: Revert NUMA affine page table allocation
This patch reverts NUMA affine page table allocation added by commit
1411e0ec31 (x86-64, numa: Put pgtable to local node memory).
The commit made an undocumented change where the kernel linear mapping
strictly follows intersection of e820 memory map and NUMA
configuration. If the physical memory configuration has holes or NUMA
nodes are not properly aligned, this leads to using unnecessarily
smaller mapping size which leads to increased TLB pressure. For
details,
http://thread.gmane.org/gmane.linux.kernel/
1104672
Patches to fix the problem have been proposed but the underlying code
needs more cleanup and the approach itself seems a bit heavy handed
and it has been determined to revert the feature for now and come back
to it in the next developement cycle.
http://thread.gmane.org/gmane.linux.kernel/
1105959
As init_memory_mapping_high() callsites have been consolidated since
the commit, reverting is done manually. Also, the RED-PEN comment in
arch/x86/mm/init.c is not restored as the problem no longer exists
with memblock based top-down early memory allocation.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Jan Beulich [Thu, 3 Mar 2011 10:55:29 +0000 (10:55 +0000)]
x86: Work around old gas bug
Add extra parentheses around a couple of definitions introduced
by "x86: Cleanup vector usage" and used in assembly macro
arguments, and remove spaces. Without that old (2.16.1) gas
would see more macro arguments than were actually specified.
Reported-and-tested-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Cc: Shaohua Li <shaohua.li@intel.com>
LKML-Reference: <
4D6F81B10200007800034B0B@vpn.id2.novell.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Tejun Heo [Wed, 2 Mar 2011 10:32:47 +0000 (11:32 +0100)]
x86-64, NUMA: Better explain numa_distance handling
Handling of out-of-bounds distances and allocation failure can use
better documentation. Add it.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Acked-by: David Rientjes <rientjes@google.com>
Yinghai Lu [Wed, 2 Mar 2011 10:22:14 +0000 (11:22 +0100)]
x86-64, NUMA: Fix distance table handling
NUMA distance table handling has the following problems.
* numa_reset_distance() uses numa_distance * sizeof(numa_distance[0])
as the table size when it should be using the square of
numa_distance.
* The same size miscalculation when allocation space for phys_dist in
numa_emulation().
* In numa_emulation(), phys_dist must be reserved; otherwise, the new
emulated distance table may overlap it.
Fix them and, while at it, take numa_distance_cnt resetting in
numa_reset_distance() out of the if block to simplify the code a bit.
David Rientjes reported incorrect handling of distance table during
emulation.
-tj: Edited out numa_alloc_distance() related changes which weren't
necessary and rewrote patch description.
-v2: Ingo was unhappy with 80-column limit induced linebreaks. Let
lines run over 80-column.
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Reported-by: David Rientjes <rientjes@google.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: David Rientjes <rientjes@google.com>
Yinghai Lu [Sat, 26 Feb 2011 12:05:43 +0000 (13:05 +0100)]
mm: Move early_node_map[] reverse scan helpers under HAVE_MEMBLOCK
Heiko found recent memblock change triggers these warnings on s390:
mm/page_alloc.c:3623:22: warning: 'last_active_region_index_in_nid' defined but not used
mm/page_alloc.c:3638:22: warning: 'previous_active_region_index_in_nid' defined but not used
Need to move those two function under HAVE_MEMBLOCK with its only
user, find_memory_core_early().
-tj: Minor updates to description.
Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
David Rientjes [Fri, 25 Feb 2011 09:06:39 +0000 (10:06 +0100)]
x86-64, NUMA: Fix size of numa_distance array
numa_distance should be sized like the SLIT, an NxN matrix where N is
the highest node id + 1. This patch fixes the calculation to avoid
overflowing the array on the subsequent iteration.
-tj: The original patch used last index to calculate size. Yinghai
pointed out it should be incremented so it is the number of
elements instead of the last index to calculate the size of the
table. Updated accordingly.
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Yinghai Lu [Thu, 24 Feb 2011 13:46:24 +0000 (14:46 +0100)]
x86: Rename e820_table_* to pgt_buf_*
e820_table_{start|end|top}, which are used to buffer page table
allocation during early boot, are now derived from memblock and don't
have much to do with e820. Change the names so that they reflect what
they're used for.
This patch doesn't introduce any behavior change.
-v2: Ingo found that earlier patch "x86: Use early pre-allocated page
table buffer top-down" caused crash on 32bit and needed to be
dropped. This patch was updated to reflect the change.
-tj: Updated commit description.
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Yinghai Lu [Thu, 24 Feb 2011 13:43:06 +0000 (14:43 +0100)]
bootmem: Move __alloc_memory_core_early() to nobootmem.c
Now that bootmem.c and nobootmem.c are separate, there's no reason to
define __alloc_memory_core_early(), which is used only by nobootmem,
inside #ifdef in page_alloc.c. Move it to nobootmem.c and make it
static.
This patch doesn't introduce any behavior change.
-tj: Updated commit description.
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Acked-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Yinghai Lu [Thu, 24 Feb 2011 13:43:06 +0000 (14:43 +0100)]
bootmem: Move contig_page_data definition to bootmem.c/nobootmem.c
Now that bootmem.c and nobootmem.c are separate, it's cleaner to
define contig_page_data in each file than in page_alloc.c with #ifdef.
Move it.
This patch doesn't introduce any behavior change.
-v2: According to Andrew, fixed the struct layout.
-tj: Updated commit description.
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Acked-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Yinghai Lu [Thu, 24 Feb 2011 13:43:05 +0000 (14:43 +0100)]
bootmem: Separate out CONFIG_NO_BOOTMEM code into nobootmem.c
mm/bootmem.c contained code paths for both bootmem and no bootmem
configurations. They implement about the same set of APIs in
different ways and as a result bootmem.c contains massive amount of
#ifdef CONFIG_NO_BOOTMEM.
Separate out CONFIG_NO_BOOTMEM code into mm/nobootmem.c. As the
common part is relatively small, duplicate them in nobootmem.c instead
of creating a common file or ifdef'ing in bootmem.c.
The followings are duplicated.
* {min|max}_low_pfn, max_pfn, saved_max_pfn
* free_bootmem_late()
* ___alloc_bootmem()
* __alloc_bootmem_low()
The followings are applicable only to nobootmem and moved verbatim.
* __free_pages_memory()
* free_all_memory_core_early()
The followings are not applicable to nobootmem and omitted in
nobootmem.c.
* reserve_bootmem_node()
* reserve_bootmem()
The rest split function bodies according to CONFIG_NO_BOOTMEM.
Makefile is updated so that only either bootmem.c or nobootmem.c is
built according to CONFIG_NO_BOOTMEM.
This patch doesn't introduce any behavior change.
-tj: Rewrote commit description.
Suggested-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Acked-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Yinghai Lu [Tue, 22 Feb 2011 10:18:49 +0000 (11:18 +0100)]
x86-64, NUMA: Seperate out numa_alloc_distance() from numa_set_distance()
Alloc code is much bigger the distance setting. Separate it out into
numa_alloc_distance() for readability.
-v2: Let alloc_numa_distance to return -ENOMEM on failing path,
requested by tj.
-tj: Description update. Minor tweaks including function name,
location and return value check.
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Tejun Heo [Tue, 22 Feb 2011 10:10:08 +0000 (11:10 +0100)]
x86-64, NUMA: Add proper function comments to global functions
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Tejun Heo [Tue, 22 Feb 2011 10:10:08 +0000 (11:10 +0100)]
x86-64, NUMA: Move NUMA emulation into numa_emulation.c
Create numa_emulation.c and move all NUMA emulation code there. The
definitions of struct numa_memblk and numa_meminfo are moved to
numa_64.h. Also, numa_remove_memblk_from(), numa_cleanup_meminfo(),
numa_reset_distance() along with numa_emulation() are made global.
- v2: Internal declarations moved to numa_internal.h as suggested by
Yinghai.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Yinghai Lu <yinghai@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Tejun Heo [Tue, 22 Feb 2011 10:10:08 +0000 (11:10 +0100)]
x86-64, NUMA: Prepare numa_emulation() for moving NUMA emulation into a separate file
Update numa_emulation() such that, it
- takes @numa_meminfo and @numa_dist_cnt instead of directly
referencing the global variables.
- copies the distance table by iterating each distance with
node_distance() instead of memcpy'ing the distance table.
- tests emu_cmdline to determine whether emulation is requested and
fills emu_nid_to_phys[] with identity mapping if emulation is not
used. This allows the caller to call numa_emulation()
unconditionally and makes return value unncessary.
- defines dummy version if CONFIG_NUMA_EMU is disabled.
This patch doesn't introduce any behavior change.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Yinghai Lu [Mon, 21 Feb 2011 09:58:13 +0000 (10:58 +0100)]
x86-64, NUMA: Do not scan two times for setup_node_bootmem()
By the time setup_node_bootmem() is called, all the memblocks are
already registered. As node_data is allocated from these memblocks,
calling it more than once doesn't make any difference. Drop the loop.
tj: Dropped comment referencing to the old behavior as suggested by
David and rephrased the description.
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Yinghai Lu [Thu, 17 Feb 2011 13:53:20 +0000 (14:53 +0100)]
x86-64, NUMA: Put dummy_numa_init() in the init section
dummy_numa_init() is used only during system boot. Put it in .init
like other NUMA init functions.
- tj: Description update.
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Yinghai Lu [Thu, 17 Feb 2011 13:46:37 +0000 (14:46 +0100)]
x86-64, NUMA: Don't call __pa() with invalid address in numa_reset_distance()
Do not call __pa(numa_distance) if it was not allocated before.
Calling with invalid address triggers VIRTUAL_BUG_ON() in
__phys_addr() if CONFIG_DEBUG_VIRTUAL.
Also reported by Ingo.
http://thread.gmane.org/gmane.linux.kernel/
1101306/focus=
1101785
- v2: Change to check existing path as tj requested.
- tj: Description update.
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Ingo Molnar <mingo@elte.hu>
Tejun Heo [Wed, 16 Feb 2011 16:11:10 +0000 (17:11 +0100)]
x86-64, NUMA: Unify emulated distance mapping
NUMA emulation needs to update node distance information. It did it
by remapping apicid to PXM mapping, even when amdtopology is being
used. There is no reason to go through such convolution. The generic
code has all the information necessary to transform the distance table
to the emulated nid space.
Implement generic distance table transformation in numa_emulation()
and drop private implementations in srat_64 and amdtopology_64. This
makes find_node_by_addr() and fake_physnodes() and related functions
unnecessary, drop them.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:10 +0000 (17:11 +0100)]
x86-64, NUMA: Unify emulated apicid -> node mapping transformation
NUMA emulation changes node mappings and thus apicid -> node mapping
needs to be updated accordingly. srat_64 and amdtopology_64 did this
separately; however, all the necessary information is the mapping from
emulated nodes to physical nodes which is available in
emu_nid_to_phys[].
Implement common __apicid_to_node[] transformation in numa_emulation()
and drop duplicate implementations.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:10 +0000 (17:11 +0100)]
x86-64, NUMA: Emulate directly from numa_meminfo
NUMA emulation built physnodes[] array which could only represent
configurations from the physical meminfo and emulated nodes using the
information. There's no reason to take this extra level of
indirection. Update emulation functions so that they operate directly
on numa_meminfo. This simplifies the code and makes emulation layout
behave better with interleaved physical nodes.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:10 +0000 (17:11 +0100)]
x86-64, NUMA: Wrap node ID during emulation
Both emulation layout functions - split_nodes[_size]_interleave() -
didn't wrap emulated nid while laying out the fake nodes and tried to
avoid interating over the specified number of nodes, which is fragile.
Now that the emulation code generates numa_meminfo, the node memblks
don't need to be consecutive and emulated node IDs can simply wrap.
This makes the code more robust and is necessary for updates to better
handle the cases where the physical nodes are interleaved.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:10 +0000 (17:11 +0100)]
x86-64, NUMA: Make emulation code build numa_meminfo and share the registration path
NUMA emulation code built nodes[] array and had its own registration
path to set up the emulated nodes. Update it such that it generates
emulated numa_meminfo and returns control to initmem_init() and shares
the same registration path with non-emulated cases.
Because {acpi|amd}_fake_nodes() expect nodes[] parameter,
fake_physnodes() now generates nodes[] from numa_meminfo. This will
go away with further updates.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:10 +0000 (17:11 +0100)]
x86-64, NUMA: Build and use direct emulated nid -> phys nid mapping
NUMA emulation copied physical NUMA configuration into physnodes[] and
used it to reverse-map emulated nodes to physical nodes, which is
unnecessarily convoluted. Build emu_nid_to_phys[] array to map
emulated nids directly to the matching physical nids and use it in
numa_add_cpu().
physnodes[] will be removed with further patches.
- v2: Build failure when CONFIG_DEBUG_PER_CPU_MAPS due to missing
local variable definition fixed. Reported by Ingo.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:10 +0000 (17:11 +0100)]
x86-64, NUMA: Trivial changes to prepare for emulation updates
* Separate out numa_add_memblk_to() from numa_add_memblk() so that
different numa_meminfo can be used.
* Rename cmdline to emu_cmdline.
* Drop @start/last_pfn from numa_emulation() and use max_pfn directly.
This patch doesn't introduce any behavior change.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:09 +0000 (17:11 +0100)]
x86-64, NUMA: Implement generic node distance handling
Node distance either used direct node comparison, ACPI PXM comparison
or ACPI SLIT table lookup. This patch implements generic node
distance handling. NUMA init methods can call numa_set_distance() to
set distance between nodes and the common __node_distance()
implementation will report the set distance.
Due to the way NUMA emulation is implemented, the generic node
distance handling is used only when emulation is not used. Later
patches will update NUMA emulation to use the generic distance
mechanism.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:09 +0000 (17:11 +0100)]
x86-64, NUMA: Kill mem_nodes_parsed
With all memory configuration information now carried in numa_meminfo,
there's no need to keep mem_nodes_parsed separate. Drop it and use
numa_nodes_parsed for CPU / memory-less nodes.
A new helper numa_nodemask_from_meminfo() is added to calculate
memnode mask on the fly which is currently used to set
node_possible_map.
This simplifies NUMA init methods a bit and removes a source of
possible inconsistencies.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:09 +0000 (17:11 +0100)]
x86-64, NUMA: Rename cpu_nodes_parsed to numa_nodes_parsed
It's no longer necessary to keep both cpu_nodes_parsed and
mem_nodes_parsed. In preparation for merge, rename cpu_nodes_parsed
to numa_nodes_parsed.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:09 +0000 (17:11 +0100)]
x86-64, NUMA: Kill numa_nodes[]
numa_nodes[] doesn't carry any information which isn't present in
numa_meminfo. Each entry is simply min/max range of all the memblks
for the node. This is not only redundant but also inaccurate when
memblks for different nodes interleave - for example,
find_node_by_addr() can return the wrong nodeid.
Kill numa_nodes[] and always use numa_meminfo instead.
* nodes_cover_memory() is renamed to numa_meminfo_cover_memory() and
now operations on numa_meminfo and returns bool.
* setup_node_bootmem() needs min/max range. Compute the range on the
fly. setup_node_bootmem() invocation is restructured to use outer
loop instead of hardcoding the double invocations.
* find_node_by_addr() now operates on numa_meminfo.
* setup_physnodes() builds physnodes[] from memblks. This will go
away when emulation code is updated to use struct numa_meminfo.
This patch also makes the following misc changes.
* Clearing of nodes_add[] clearing is converted to memset().
* numa_add_memblk() in amd_numa_init() is moved down a bit for
consistency.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:09 +0000 (17:11 +0100)]
x86-64, NUMA: Add common find_node_by_addr()
srat_64.c and amdtopology_64.c had their own versions of
find_node_by_addr() which were basically the same. Add common one in
numa_64.c and remove the duplicates.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:09 +0000 (17:11 +0100)]
x86-64, NUMA: consolidate and improve memblk sanity checks
memblk sanity check was scattered around and incomplete. Consolidate
and improve.
* Confliction detection and cutoff_node() logic are moved to
numa_cleanup_meminfo().
* numa_cleanup_meminfo() clears the unused memblks before returning.
* Check and warn about invalid input parameters in numa_add_memblk().
* Check the maximum number of memblk isn't exceeded in
numa_add_memblk().
* numa_cleanup_meminfo() is now called before numa_emulation() so that
the emulation code also uses the cleaned up version.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:09 +0000 (17:11 +0100)]
x86-64, NUMA: make numa_cleanup_meminfo() prettier
* Factor out numa_remove_memblk_from().
* Hole detection doesn't need separate start/end. Calculate start/end
once.
* Relocate comment.
* Define iterators at the top and remove unnecessary prefix
increments.
This prepares for further improvements to the function.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:09 +0000 (17:11 +0100)]
x86-64, NUMA: Separate out numa_cleanup_meminfo()
Separate out numa_cleanup_meminfo() from numa_register_memblks().
node_possible_map initialization is moved to the top of the split
numa_register_memblks().
This patch doesn't cause behavior change.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:08 +0000 (17:11 +0100)]
x86-64, NUMA: Introduce struct numa_meminfo
Arrays for memblks and nodeids and their length lived in separate
variables making things unnecessarily cumbersome. Introduce struct
numa_meminfo which contains all memory configuration info. This patch
doesn't cause any behavior change.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:08 +0000 (17:11 +0100)]
x86-64, NUMA: Remove %NULL @nodeids handling from compute_hash_shift()
numa_emulation() called compute_hash_shift() with %NULL @nodeids which
meant identity mapping between index and nodeid. Make
numa_emulation() build identity array and drop %NULL @nodeids handling
from populate_memnodemap() and thus from compute_hash_shift(). This
is to prepare for transition to using memblks instead.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:08 +0000 (17:11 +0100)]
x86-64, NUMA: Kill {acpi|amd|dummy}_scan_nodes()
They are empty now. Kill them.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:08 +0000 (17:11 +0100)]
x86-64, NUMA: Unify the rest of memblk registration
Move the remaining memblk registration logic from acpi_scan_nodes() to
numa_register_memblks() and initmem_init().
This applies nodes_cover_memory() sanity check, memory node sorting
and node_online() checking, which were only applied to acpi, to all
init methods.
As all memblk registration is moved to common code, active range
clearing is moved to initmem_init() too and removed from bad_srat().
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:08 +0000 (17:11 +0100)]
x86-64, NUMA: Unify use of memblk in all init methods
Make both amd and dummy use numa_add_memblk() to describe the detected
memory blocks. This allows initmem_init() to call
numa_register_memblk() regardless of init method in use. Drop custom
memory registration codes from amd and dummy.
After this change, memblk merge/cleanup in numa_register_memblks() is
applied to all init methods.
As this makes compute_hash_shift() and numa_register_memblks() used
only inside numa_64.c, make them static.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 16:11:07 +0000 (17:11 +0100)]
x86-64, NUMA: Factor out memblk handling into numa_{add|register}_memblk()
Factor out memblk handling from srat_64.c into two functions in
numa_64.c. This patch doesn't introduce any behavior change. The
next patch will make all init methods use these functions.
- v2: Fixed build failure on 32bit due to misplaced NR_NODE_MEMBLKS.
Reported by Ingo.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 11:13:07 +0000 (12:13 +0100)]
x86-64, NUMA: Kill {acpi|amd}_get_nodes()
With common numa_nodes[], common code in numa_64.c can access it
directly. Copy directly and kill {acpi|amd}_get_nodes().
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 11:13:07 +0000 (12:13 +0100)]
x86-64, NUMA: Use common numa_nodes[]
ACPI and amd are using separate nodes[] array. Add numa_nodes[] and
use them in all NUMA init methods. cutoff_node() cleanup is moved
from srat_64.c to numa_64.c and applied in initmem_init() regardless
of init methods.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 11:13:07 +0000 (12:13 +0100)]
x86-64, NUMA: Move apicid to numa mapping initialization from amd_scan_nodes() to amd_numa_init()
This brings amd initialization behavior closer to that of acpi.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 11:13:07 +0000 (12:13 +0100)]
x86-64, NUMA: Remove local variable found from amd_numa_init()
Use weight count on mem_nodes_parsed instead.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 11:13:07 +0000 (12:13 +0100)]
x86-64, NUMA: Use common {cpu|mem}_nodes_parsed
ACPI and amd are using separate nodes_parsed masks. Add
{cpu|mem}_nodes_parsed and use them in all NUMA init methods.
Initialization of the masks and building node_possible_map are now
handled commonly by initmem_init().
dummy_numa_init() is updated to set node 0 on both masks. While at
it, move the info messages from scan to init.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 11:13:06 +0000 (12:13 +0100)]
x86-64, NUMA: Restructure initmem_init()
Reorganize initmem_init() such that,
* Different NUMA init methods are iterated in a consistent way.
* Each iteration re-initializes all the parameters and different
method can be tried after a failure.
* Dummy init is handled the same as other methods.
Apart from how retry after failure, this patch doesn't change the
behavior. The call sequences are kept equivalent across the
conversion.
After the change, bad_srat() doesn't need to clear apic to node
mapping or worry about numa_off. Simplified accordingly.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 11:13:06 +0000 (12:13 +0100)]
x86, NUMA: Move *_numa_init() invocations into initmem_init()
There's no reason for these to live in setup_arch(). Move them inside
initmem_init().
- v2: x86-32 initmem_init() weren't updated breaking 32bit builds.
Fixed. Found by Ankita.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Ankita Garg <ankita@in.ibm.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 11:13:06 +0000 (12:13 +0100)]
x86-64, NUMA: Wrap acpi_numa_init() so that failure can be indicated by return value
Because of the way ACPI tables are parsed, the generic
acpi_numa_init() couldn't return failure when error was detected by
arch hooks. Instead, the failure state was recorded and later arch
dependent init hook - acpi_scan_nodes() - would fail.
Wrap acpi_numa_init() with x86_acpi_numa_init() so that failure can be
indicated as return value immediately. This is in preparation for
further NUMA init cleanups.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 11:13:06 +0000 (12:13 +0100)]
x86-64, NUMA: Unify {acpi|amd}_{numa_init|scan_nodes}() arguments and return values
The functions used during NUMA initialization - *_numa_init() and
*_scan_nodes() - have different arguments and return values. Unify
them such that they all take no argument and return 0 on success and
-errno on failure. This is in preparation for further NUMA init
cleanups.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 11:13:06 +0000 (12:13 +0100)]
x86, NUMA: Drop @start/last_pfn from initmem_init()
initmem_init() extensively accesses and modifies global data
structures and the parameters aren't even followed depending on which
path is being used. Drop @start/last_pfn and let it deal with
@max_pfn directly. This is in preparation for further NUMA init
cleanups.
- v2: x86-32 initmem_init() weren't updated breaking 32bit builds.
Fixed. Found by Yinghai.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 11:13:06 +0000 (12:13 +0100)]
x86-64, NUMA: Simplify hotplug node handling in acpi_numa_memory_affinity_init()
Hotplug node handling in acpi_numa_memory_affinity_init() was
unnecessarily complicated with storing the original nodes[] entry and
restoring it afterwards. Simplify it by not modifying the nodes[]
entry for hotplug nodes from the beginning.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Tejun Heo [Wed, 16 Feb 2011 11:13:06 +0000 (12:13 +0100)]
x86-64, NUMA: Make dummy node initialization path similar to non-dummy ones
Dummy node initialization in initmem_init() didn't initialize apicid
to node mapping and set cpu to node mapping directly by caling
numa_set_node(), which is different from non-dummy init paths.
Update it such that they behave similarly. Initialize apicid to node
mapping and call numa_init_array(). The actual cpu to node mapping is
handled by init_cpu_to_node() later.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Yinghai Lu <yinghai@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Shaohui Zheng <shaohui.zheng@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Ingo Molnar [Wed, 16 Feb 2011 08:45:33 +0000 (09:45 +0100)]
Merge branch 'x86/amd-nb' into x86/mm
Merge reason: consolidate it into the more generic x86/mm tree to prevent conflicts
with ongoing NUMA work.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Ingo Molnar [Wed, 16 Feb 2011 08:44:04 +0000 (09:44 +0100)]
Merge branch 'x86/numa' into x86/mm
Merge reason: consolidate it into the more generic x86/mm tree to prevent conflicts.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Ingo Molnar [Wed, 16 Feb 2011 08:42:50 +0000 (09:42 +0100)]
Merge branch 'x86/bootmem' into x86/mm
Merge reason: the topic is ready - consolidate it into the more generic x86/mm tree
and prevent conflicts.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Borislav Petkov [Mon, 14 Feb 2011 17:14:51 +0000 (18:14 +0100)]
x86, amd: Initialize variable properly
Commit
d518573de63f ("x86, amd: Normalize compute unit IDs on
multi-node processors") introduced compute unit normalization
but causes a compiler warning:
arch/x86/kernel/cpu/amd.c: In function 'amd_detect_cmp':
arch/x86/kernel/cpu/amd.c:268: warning: 'cores_per_cu' may be used uninitialized in this function
arch/x86/kernel/cpu/amd.c:268: note: 'cores_per_cu' was declared here
The compiler is right - initialize it with a proper value.
Also, fix up a comment while at it.
Reported-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
Cc: Andreas Herrmann <andreas.herrmann3@amd.com>
LKML-Reference: <
20110214171451.GB10076@kryptos.osrc.amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
David Rientjes [Mon, 7 Feb 2011 22:08:53 +0000 (14:08 -0800)]
x86, numa: Add error handling for bad cpu-to-node mappings
CONFIG_DEBUG_PER_CPU_MAPS may return NUMA_NO_NODE when an
early_cpu_to_node() mapping hasn't been initialized. In such a
case, it emits a warning and continues without an issue but
callers may try to use the return value to index into an array.
We can catch those errors and fail silently since a warning has
already been emitted. No current user of numa_add_cpu()
requires this error checking to avoid a crash, but it's better
to be proactive in case a future user happens to have a bug and
a user tries to diagnose it with CONFIG_DEBUG_PER_CPU_MAPS.
Reported-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: Tejun Heo <tj@kernel.org>
LKML-Reference: <alpine.DEB.2.00.
1102071407250.7812@chino.kir.corp.google.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Ingo Molnar [Mon, 14 Feb 2011 12:28:29 +0000 (13:28 +0100)]
Merge commit 'v2.6.38-rc4' into x86/numa
Merge reason: Merge latest fixes before applying new patch.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Kamal Mostafa [Fri, 4 Feb 2011 01:38:05 +0000 (17:38 -0800)]
x86: Emit "mem=nopentium ignored" warning when not supported
Emit warning when "mem=nopentium" is specified on any arch other
than x86_32 (the only that arch supports it).
Signed-off-by: Kamal Mostafa <kamal@canonical.com>
BugLink: http://bugs.launchpad.net/bugs/553464
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Len Brown <len.brown@intel.com>
Cc: Rafael J. Wysocki <rjw@sisk.pl>
LKML-Reference: <
1296783486-23033-2-git-send-email-kamal@canonical.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: <stable@kernel.org>
Kamal Mostafa [Fri, 4 Feb 2011 01:38:04 +0000 (17:38 -0800)]
x86: Fix panic when handling "mem={invalid}" param
Avoid removing all of memory and panicing when "mem={invalid}"
is specified, e.g. mem=blahblah, mem=0, or mem=nopentium (on
platforms other than x86_32).
Signed-off-by: Kamal Mostafa <kamal@canonical.com>
BugLink: http://bugs.launchpad.net/bugs/553464
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Len Brown <len.brown@intel.com>
Cc: Rafael J. Wysocki <rjw@sisk.pl>
Cc: <stable@kernel.org> # .3x: as far back as it applies
LKML-Reference: <
1296783486-23033-1-git-send-email-kamal@canonical.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Shaohua Li [Mon, 17 Jan 2011 02:52:10 +0000 (10:52 +0800)]
x86: Avoid tlbstate lock if not enough cpus
This one isn't related to previous patch. If online cpus are
below NUM_INVALIDATE_TLB_VECTORS, we don't need the lock. The
comments in the code declares we don't need the check, but a hot
lock still needs an atomic operation and expensive, so add the
check here.
Uses nr_cpu_ids here as suggested by Eric Dumazet.
Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Andi Kleen <andi@firstfloor.org>
LKML-Reference: <
1295232730.1949.710.camel@sli10-conroe>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Shaohua Li [Mon, 17 Jan 2011 02:52:07 +0000 (10:52 +0800)]
x86: Scale up the number of TLB invalidate vectors with NR_CPUs, up to 32
Make the maxium TLB invalidate vectors depend on NR_CPUS linearly,
with a maximum of 32 vectors.
We currently only have 8 vectors for TLB invalidation and that is clearly
inadequate. If we have a lot of CPUs, the CPUs need share the 8 vectors and
tlbstate_lock is used to protect them. flush_tlb_page() is
heavily used in page reclaim, which will cause a lot of lock
contention for tlbstate_lock.
Andi Kleen suggested increasing the vectors number to 32, which should be
good for current typical systems to reduce the tlbstate_lock contention.
My test system has 4 sockets and 64G memory, and 64 CPUs. My
workload creates 64 processes. Each process mmap reads a big
empty sparse file. The total size of the files are 2*total_mem,
so this will cause a lot of page reclaim.
Below is the result I get from perf call-graph profiling:
without the patch:
------------------
24.25% usemem [kernel] [k] _raw_spin_lock
|
--- _raw_spin_lock
|
|--42.15%-- native_flush_tlb_others
with the patch:
------------------
14.96% usemem [kernel] [k] _raw_spin_lock
|
--- _raw_spin_lock
|--13.89%-- native_flush_tlb_others
So this heavily reduces the tlbstate_lock contention.
Suggested-by: Andi Kleen <andi@firstfloor.org>
Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <
1295232727.1949.709.camel@sli10-conroe>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Shaohua Li [Mon, 17 Jan 2011 02:52:05 +0000 (10:52 +0800)]
x86: Allocate 32 tlb_invalidate_interrupt handler stubs
Add up to 32 invalidate_interrupt handlers. How many handlers are
added depends on NUM_INVALIDATE_TLB_VECTORS. So if
NUM_INVALIDATE_TLB_VECTORS is smaller than 32, we reduce code
size.
Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
LKML-Reference: <
1295232725.1949.708.camel@sli10-conroe>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Shaohua Li [Mon, 17 Jan 2011 02:52:02 +0000 (10:52 +0800)]
x86: Cleanup vector usage
Cleanup the vector usage and make them continuous if possible.
Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
LKML-Reference: <
1295232722.1949.707.camel@sli10-conroe>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Ingo Molnar [Mon, 14 Feb 2011 10:55:18 +0000 (11:55 +0100)]
Merge branch 'linus' into x86/bootmem
Conflicts:
arch/x86/mm/numa_64.c
Merge reason: fix the conflict, update to latest -rc and pick up this
dependent fix from Yinghai:
e6d2e2b2b1e1: memblock: don't adjust size in memblock_find_base()
Signed-off-by: Ingo Molnar <mingo@elte.hu>
David Miller [Mon, 14 Feb 2011 00:37:07 +0000 (16:37 -0800)]
klist: Fix object alignment on 64-bit.
Commit
c0e69a5bbc6f ("klist.c: bit 0 in pointer can't be used as flag")
intended to make sure that all klist objects were at least pointer size
aligned, but used the constant "4" which only works on 32-bit.
Use "sizeof(void *)" which is correct in all cases.
Signed-off-by: David S. Miller <davem@davemloft.net>
Acked-by: Jesper Nilsson <jesper.nilsson@axis.com>
Cc: stable <stable@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Linus Torvalds [Sun, 13 Feb 2011 15:59:48 +0000 (07:59 -0800)]
Merge branch 'spi/merge' of git://git.secretlab.ca/git/linux-2.6
* 'spi/merge' of git://git.secretlab.ca/git/linux-2.6:
devicetree-discuss is moderated for non-subscribers
MAINTAINERS: Add entry for GPIO subsystem
dt: add documentation of ARM dt boot interface
dt: Remove obsolete description of powerpc boot interface
dt: Move device tree documentation out of powerpc directory
spi/spi_sh_msiof: fix wrong address calculation, which leads to an Oops
Linus Torvalds [Sun, 13 Feb 2011 15:58:50 +0000 (07:58 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/tiwai/sound-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
ALSA: hda - add quirk for Ordissimo EVE using a realtek ALC662
ALSA: hrtimer: remove superfluous tasklet invocation
ALSA: hrtimer: handle delayed timer interrupts
ALSA: HDA: Add subwoofer quirk for Acer Aspire 8942G
ALSA: hda - Don't handle empty patch files
ALSA: hda - Fix missing CA initialization for HDMI/DP
ALSA: usbaudio - Enable the E-MU 0204 USB
ALSA: hda - switch lfe with side in mixer for 4930g
ASoC: Improve WM8994 digital power sequencing
ASoC: Create an AIF1ADCDAT signal widget to match AIF2
asoc: davinci: da830/omap-l137: correct cpu_dai_name
ASoC: fill in snd_soc_pcm_runtime.card before calling snd_soc_dai_link.init()
Linus Torvalds [Sun, 13 Feb 2011 15:50:50 +0000 (07:50 -0800)]
Revert "pci: use security_capable() when checking capablities during config space read"
This reverts commit
47970b1b2aa64464bc0a9543e86361a622ae7c03.
It turns out it breaks several distributions. Looks like the stricter
selinux checks fail due to selinux policies not being set to allow the
access - breaking X, but also lspci.
So while the change was clearly the RightThing(tm) to do in theory, in
practice we have backwards compatibility issues making it not work.
Reported-by: Dave Young <hidave.darkstar@gmail.com>
Acked-by: David Airlie <airlied@linux.ie>
Acked-by: Alex Riesen <raa.lkml@gmail.com>
Cc: Eric Paris <eparis@redhat.com>
Cc: Chris Wright <chrisw@sous-sol.org>
Cc: James Morris <jmorris@namei.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Takashi Iwai [Sun, 13 Feb 2011 09:05:30 +0000 (10:05 +0100)]
Merge branch 'fix/asoc' into for-linus
Grant Likely [Sun, 13 Feb 2011 06:53:34 +0000 (23:53 -0700)]
Merge branch 'devicetree/merge' into spi/merge
Paul Bolle [Sat, 12 Feb 2011 11:33:59 +0000 (12:33 +0100)]
devicetree-discuss is moderated for non-subscribers
Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Grant Likely [Sat, 12 Feb 2011 08:48:14 +0000 (01:48 -0700)]
MAINTAINERS: Add entry for GPIO subsystem
I'll probably regret this....
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Linus Torvalds [Sat, 12 Feb 2011 17:10:24 +0000 (09:10 -0800)]
Merge branch 'for_linus' of git://git./linux/kernel/git/tytso/ext4
* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
jbd2: call __jbd2_log_start_commit with j_state_lock write locked
ext4: serialize unaligned asynchronous DIO
ext4: make grpinfo slab cache names static
ext4: Fix data corruption with multi-block writepages support
ext4: fix up ext4 error handling
ext4: unregister features interface on module unload
ext4: fix panic on module unload when stopping lazyinit thread
Theodore Ts'o [Sat, 12 Feb 2011 13:18:24 +0000 (08:18 -0500)]
jbd2: call __jbd2_log_start_commit with j_state_lock write locked
On an SMP ARM system running ext4, I've received a report that the
first J_ASSERT in jbd2_journal_commit_transaction has been triggering:
J_ASSERT(journal->j_running_transaction != NULL);
While investigating possible causes for this problem, I noticed that
__jbd2_log_start_commit() is getting called with j_state_lock only
read-locked, in spite of the fact that it's possible for it might
j_commit_request. Fix this by grabbing the necessary information so
we can test to see if we need to start a new transaction before
dropping the read lock, and then calling jbd2_log_start_commit() which
will grab the write lock.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Eric Sandeen [Sat, 12 Feb 2011 13:17:34 +0000 (08:17 -0500)]
ext4: serialize unaligned asynchronous DIO
ext4 has a data corruption case when doing non-block-aligned
asynchronous direct IO into a sparse file, as demonstrated
by xfstest 240.
The root cause is that while ext4 preallocates space in the
hole, mappings of that space still look "new" and
dio_zero_block() will zero out the unwritten portions. When
more than one AIO thread is going, they both find this "new"
block and race to zero out their portion; this is uncoordinated
and causes data corruption.
Dave Chinner fixed this for xfs by simply serializing all
unaligned asynchronous direct IO. I've done the same here.
The difference is that we only wait on conversions, not all IO.
This is a very big hammer, and I'm not very pleased with
stuffing this into ext4_file_write(). But since ext4 is
DIO_LOCKING, we need to serialize it at this high level.
I tried to move this into ext4_ext_direct_IO, but by then
we have the i_mutex already, and we will wait on the
work queue to do conversions - which must also take the
i_mutex. So that won't work.
This was originally exposed by qemu-kvm installing to
a raw disk image with a normal sector-63 alignment. I've
tested a backport of this patch with qemu, and it does
avoid the corruption. It is also quite a lot slower
(14 min for package installs, vs. 8 min for well-aligned)
but I'll take slow correctness over fast corruption any day.
Mingming suggested that we can track outstanding
conversions, and wait on those so that non-sparse
files won't be affected, and I've implemented that here;
unaligned AIO to nonsparse files won't take a perf hit.
[tytso@mit.edu: Keep the mutex as a hashed array instead
of bloating the ext4 inode]
[tytso@mit.edu: Fix up namespace issues so that global
variables are protected with an "ext4_" prefix.]
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Eric Sandeen [Sat, 12 Feb 2011 13:12:18 +0000 (08:12 -0500)]
ext4: make grpinfo slab cache names static
In 2.6.37 I was running into oopses with repeated module
loads & unloads. I tracked this down to:
fb1813f4 ext4: use dedicated slab caches for group_info structures
(this was in addition to the features advert unload problem)
The kstrdup & subsequent kfree of the cache name was causing
a double free. In slub, at least, if I read it right it allocates
& frees the name itself, slab seems to do something different...
so in slub I think we were leaking -our- cachep->name, and double
freeing the one allocated by slub.
After getting lost in slab/slub/slob a bit, I just looked at other
sized-caches that get allocated. jbd2, biovec, sgpool all do it
more or less the way jbd2 does. Below patch follows the jbd2
method of dynamically allocating a cache at mount time from
a list of static names.
(This might also possibly fix a race creating the caches with
parallel mounts running).
[Folded in a fix from Dan Carpenter which fixed an off-by-one error in
the original patch]
Cc: stable@kernel.org
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Grant Likely [Sat, 12 Feb 2011 08:45:55 +0000 (01:45 -0700)]
MAINTAINERS: Add entry for GPIO subsystem
I'll probably regret this....
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Linus Torvalds [Sat, 12 Feb 2011 00:30:09 +0000 (16:30 -0800)]
Merge branch 'kvm-updates/2.6.38' of git://git./virt/kvm/kvm
* 'kvm-updates/2.6.38' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
KVM: SVM: Make sure KERNEL_GS_BASE is valid when loading gs_index
Linus Torvalds [Sat, 12 Feb 2011 00:30:05 +0000 (16:30 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/bp/bp
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/bp/bp:
amd64_edac: Fix DIMMs per DCTs output
Linus Torvalds [Sat, 12 Feb 2011 00:29:57 +0000 (16:29 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/teigland/dlm
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm:
dlm: use single thread workqueues
Linus Torvalds [Sat, 12 Feb 2011 00:29:50 +0000 (16:29 -0800)]
Merge git://git./linux/kernel/git/sfrench/cifs-2.6
* git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6:
cifs: don't always drop malformed replies on the floor (try #3)
cifs: clean up checks in cifs_echo_request
[CIFS] Do not send SMBEcho requests on new sockets until SMBNegotiate
Linus Torvalds [Sat, 12 Feb 2011 00:16:25 +0000 (16:16 -0800)]
Merge branch 'hwmon-for-linus' of git://git./linux/kernel/git/groeck/staging
* 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/staging:
hwmon: (emc1403) Fix I2C address range
hwmon: (lm63) Consider LM64 temperature offset
Linus Torvalds [Sat, 12 Feb 2011 00:16:03 +0000 (16:16 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jmorris/security-testing-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6:
pci: use security_capable() when checking capablities during config space read
security: add cred argument to security_capable()
tpm_tis: Use timeouts returned from TPM
Linus Torvalds [Sat, 12 Feb 2011 00:15:15 +0000 (16:15 -0800)]
Merge branch 's5p-fixes-for-linus' of git://git./linux/kernel/git/kgene/linux-samsung
* 's5p-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
ARM: SAMSUNG: Ensure struct sys_device is declared in plat/pm.h
ARM: S5PV310: Cleanup System MMU
ARM: S5PV310: Add support System MMU on SMDKV310
Linus Torvalds [Sat, 12 Feb 2011 00:13:53 +0000 (16:13 -0800)]
Merge branch 'next' of git://git.monstr.eu/linux-2.6-microblaze
* 'next' of git://git.monstr.eu/linux-2.6-microblaze:
microblaze: Fix msr instruction detection
microblaze: Fix pte_update function
microblaze: Fix asm compilation warning
microblaze: Fix IRQ flag handling for MSR=0
Julia Lawall [Thu, 10 Feb 2011 23:01:37 +0000 (15:01 -0800)]
drivers/w1/masters/omap_hdq.c: add missing clk_put
This code makes two calls to clk_get, then test both return values and
fails if either failed.
The problem is that in the first inner if, where the first call to
clk_get has failed, it don't know if the second call has failed as well.
So it don't know whether clk_get should be called on the result of the
second call. Of course, it would be possible to test that value again.
A simpler solution is just to test the result of calling clk_get
directly after each call.
The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)
// <smpl>
@r@
position p1,p2;
expression e;
statement S;
@@
e = clk_get@p1(...)
...
if@p2 (IS_ERR(e)) S
@@
expression e;
statement S;
identifier l;
position r.p1, p2 != r.p2;
@@
*e = clk_get@p1(...)
... when != clk_put(e)
*if@p2 (...)
{
... when != clk_put(e)
* return ...;
}// </smpl>
Signed-off-by: Julia Lawall <julia@diku.dk>
Cc: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
Acked-by: Tony Lindgren <tony@atomide.com>
Acked-by: Amit Kucheria <amit.kucheria@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
KAMEZAWA Hiroyuki [Thu, 10 Feb 2011 23:01:36 +0000 (15:01 -0800)]
memcg: fix leak of accounting at failure path of hugepage collapsing
mem_cgroup_uncharge_page() should be called in all failure cases after
mem_cgroup_charge_newpage() is called in huge_memory.c::collapse_huge_page()
[ 4209.076861] BUG: Bad page state in process khugepaged pfn:1e9800
[ 4209.077601] page:
ffffea0006b14000 count:0 mapcount:0 mapping: (null) index:0x2800
[ 4209.078674] page flags: 0x40000000004000(head)
[ 4209.079294] pc:
ffff880214a30000 pc->flags:
2146246697418756 pc->mem_cgroup:
ffffc9000177a000
[ 4209.082177] (/A)
[ 4209.082500] Pid: 31, comm: khugepaged Not tainted 2.6.38-rc3-mm1 #1
[ 4209.083412] Call Trace:
[ 4209.083678] [<
ffffffff810f4454>] ? bad_page+0xe4/0x140
[ 4209.084240] [<
ffffffff810f53e6>] ? free_pages_prepare+0xd6/0x120
[ 4209.084837] [<
ffffffff8155621d>] ? rwsem_down_failed_common+0xbd/0x150
[ 4209.085509] [<
ffffffff810f5462>] ? __free_pages_ok+0x32/0xe0
[ 4209.086110] [<
ffffffff810f552b>] ? free_compound_page+0x1b/0x20
[ 4209.086699] [<
ffffffff810fad6c>] ? __put_compound_page+0x1c/0x30
[ 4209.087333] [<
ffffffff810fae1d>] ? put_compound_page+0x4d/0x200
[ 4209.087935] [<
ffffffff810fb015>] ? put_page+0x45/0x50
[ 4209.097361] [<
ffffffff8113f779>] ? khugepaged+0x9e9/0x1430
[ 4209.098364] [<
ffffffff8107c870>] ? autoremove_wake_function+0x0/0x40
[ 4209.099121] [<
ffffffff8113ed90>] ? khugepaged+0x0/0x1430
[ 4209.099780] [<
ffffffff8107c236>] ? kthread+0x96/0xa0
[ 4209.100452] [<
ffffffff8100dda4>] ? kernel_thread_helper+0x4/0x10
[ 4209.101214] [<
ffffffff8107c1a0>] ? kthread+0x0/0xa0
[ 4209.101842] [<
ffffffff8100dda0>] ? kernel_thread_helper+0x0/0x10
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Reviewed-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Johannes Weiner [Thu, 10 Feb 2011 23:01:34 +0000 (15:01 -0800)]
vmscan: fix zone shrinking exit when scan work is done
Commit
3e7d34497067 ("mm: vmscan: reclaim order-0 and use compaction
instead of lumpy reclaim") introduced an indefinite loop in
shrink_zone().
It meant to break out of this loop when no pages had been reclaimed and
not a single page was even scanned. The way it would detect the latter
is by taking a snapshot of sc->nr_scanned at the beginning of the
function and comparing it against the new sc->nr_scanned after the scan
loop. But it would re-iterate without updating that snapshot, looping
forever if sc->nr_scanned changed at least once since shrink_zone() was
invoked.
This is not the sole condition that would exit that loop, but it
requires other processes to change the zone state, as the reclaimer that
is stuck obviously can not anymore.
This is only happening for higher-order allocations, where reclaim is
run back to back with compaction.
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reported-by: Michal Hocko <mhocko@suse.cz>
Tested-by: Kent Overstreet<kent.overstreet@gmail.com>
Reported-by: Kent Overstreet <kent.overstreet@gmail.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Michel Lespinasse [Thu, 10 Feb 2011 23:01:33 +0000 (15:01 -0800)]
mlock: do not munlock pages in __do_fault()
If the page is going to be written to, __do_page needs to break COW.
However, the old page (before breaking COW) was never mapped mapped into
the current pte (__do_fault is only called when the pte is not present),
so vmscan can't have marked the old page as PageMlocked due to being
mapped in __do_fault's VMA. Therefore, __do_fault() does not need to
worry about clearing PageMlocked() on the old page.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Hugh Dickins <hughd@google.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Michel Lespinasse [Thu, 10 Feb 2011 23:01:32 +0000 (15:01 -0800)]
mlock: fix race when munlocking pages in do_wp_page()
vmscan can lazily find pages that are mapped within VM_LOCKED vmas, and
set the PageMlocked bit on these pages, transfering them onto the
unevictable list. When do_wp_page() breaks COW within a VM_LOCKED vma,
it may need to clear PageMlocked on the old page and set it on the new
page instead.
This change fixes an issue where do_wp_page() was clearing PageMlocked
on the old page while the pte was still pointing to it (as well as
rmap). Therefore, we were not protected against vmscan immediately
transfering the old page back onto the unevictable list. This could
cause pages to get stranded there forever.
I propose to move the corresponding code to the end of do_wp_page(),
after the pte (and rmap) have been pointed to the new page.
Additionally, we can use munlock_vma_page() instead of
clear_page_mlock(), so that the old page stays mlocked if there are
still other VM_LOCKED vmas mapping it.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Hugh Dickins <hughd@google.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Yinghai Lu [Thu, 10 Feb 2011 23:01:30 +0000 (15:01 -0800)]
memblock: don't adjust size in memblock_find_base()
While applying patch to use memblock to find aperture for 64bit x86.
Ingo found system with 1g + force_iommu
> No AGP bridge found
> Node 0: aperture @
38000000 size 32 MB
> Aperture pointing to e820 RAM. Ignoring.
> Your BIOS doesn't leave a aperture memory hole
> Please enable the IOMMU option in the BIOS setup
> This costs you 64 MB of RAM
> Cannot allocate aperture memory hole (0,65536K)
the corresponding code:
addr = memblock_find_in_range(0, 1ULL<<32, aper_size, 512ULL<<20);
if (addr == MEMBLOCK_ERROR || addr + aper_size > 0xffffffff) {
printk(KERN_ERR
"Cannot allocate aperture memory hole (%lx,%uK)\n",
addr, aper_size>>10);
return 0;
}
memblock_x86_reserve_range(addr, addr + aper_size, "aperture64")
fails because memblock core code align the size with 512M. That could
make size way too big.
So don't align the size in that case.
actually __memblock_alloc_base, the another caller already align that
before calling that function.
BTW. x86 does not use __memblock_alloc_base...
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: David Miller <davem@davemloft.net>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Dave Airlie <airlied@linux.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Soren Hansen [Thu, 10 Feb 2011 23:01:28 +0000 (15:01 -0800)]
nbd: remove module-level ioctl mutex
Commit
2a48fc0ab242417 ("block: autoconvert trivial BKL users to private
mutex") replaced uses of the BKL in the nbd driver with mutex
operations. Since then, I've been been seeing these lock ups:
INFO: task qemu-nbd:16115 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
qemu-nbd D
0000000000000001 0 16115 16114 0x00000004
ffff88007d775d98 0000000000000082 ffff88007d775fd8 ffff88007d774000
0000000000013a80 ffff8800020347e0 ffff88007d775fd8 0000000000013a80
ffff880133730000 ffff880002034440 ffffea0004333db8 ffffffffa071c020
Call Trace:
[<
ffffffff815b9997>] __mutex_lock_slowpath+0xf7/0x180
[<
ffffffff815b93eb>] mutex_lock+0x2b/0x50
[<
ffffffffa071a21c>] nbd_ioctl+0x6c/0x1c0 [nbd]
[<
ffffffff812cb970>] blkdev_ioctl+0x230/0x730
[<
ffffffff811967a1>] block_ioctl+0x41/0x50
[<
ffffffff81175c03>] do_vfs_ioctl+0x93/0x370
[<
ffffffff81175f61>] sys_ioctl+0x81/0xa0
[<
ffffffff8100c0c2>] system_call_fastpath+0x16/0x1b
Instrumenting the nbd module's ioctl handler with some extra logging
clearly shows the NBD_DO_IT ioctl being invoked which is a long-lived
ioctl in the sense that it doesn't return until another ioctl asks the
driver to disconnect. However, that other ioctl blocks, waiting for the
module-level mutex that replaced the BKL, and then we're stuck.
This patch removes the module-level mutex altogether. It's clearly
wrong, and as far as I can see, it's entirely unnecessary, since the nbd
driver maintains per-device mutexes, and I don't see anything that would
require a module-level (or kernel-level, for that matter) mutex.
Signed-off-by: Soren Hansen <soren@linux2go.dk>
Acked-by: Serge Hallyn <serge.hallyn@canonical.com>
Acked-by: Paul Clements <paul.clements@steeleye.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: <stable@kernel.org> [2.6.37.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Alexander Strakh [Thu, 10 Feb 2011 23:01:25 +0000 (15:01 -0800)]
drivers/rtc/rtc-proc.c: add module_put on error path in rtc_proc_open()
In file drivers/rtc/rtc-proc.c seq_open() can return -ENOMEM.
86 if (!try_module_get(THIS_MODULE))
87 return -ENODEV;
88
89 return single_open(file, rtc_proc_show, rtc);
In this case before exiting (line 89) from rtc_proc_open the
module_put(THIS_MODULE) must be called.
Found by Linux Device Drivers Verification Project
Signed-off-by: Alexander Strakh <strakh@ispras.ru>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Roland Stigge [Thu, 10 Feb 2011 23:01:23 +0000 (15:01 -0800)]
drivers/gpio/pca953x.c: add a mutex to fix race condition
Add a mutex to register communication and handling. Without the mutex,
GPIOs didn't switch as expected when toggled in a fast sequence of
status changes of multiple outputs.
Signed-off-by: Roland Stigge <stigge@antcom.de>
Acked-by: Eric Miao <eric.y.miao@gmail.com>
Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: Marc Zyngier <maz@misterjones.org>
Cc: Ben Gardner <bgardner@wabtec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Tejun Heo [Thu, 10 Feb 2011 23:01:22 +0000 (15:01 -0800)]
ptrace: use safer wake up on ptrace_detach()
The wake_up_process() call in ptrace_detach() is spurious and not
interlocked with the tracee state. IOW, the tracee could be running or
sleeping in any place in the kernel by the time wake_up_process() is
called. This can lead to the tracee waking up unexpectedly which can be
dangerous.
The wake_up is spurious and should be removed but for now reduce its
toxicity by only waking up if the tracee is in TRACED or STOPPED state.
This bug can possibly be used as an attack vector. I don't think it
will take too much effort to come up with an attack which triggers oops
somewhere. Most sleeps are wrapped in condition test loops and should
be safe but we have quite a number of places where sleep and wakeup
conditions are expected to be interlocked. Although the window of
opportunity is tiny, ptrace can be used by non-privileged users and with
some loading the window can definitely be extended and exploited.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Roland McGrath <roland@redhat.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Boaz Harrosh [Thu, 10 Feb 2011 23:01:20 +0000 (15:01 -0800)]
vfs: call rcu_barrier after ->kill_sb()
In commit
fa0d7e3de6d6 ("fs: icache RCU free inodes"), we use rcu free
inode instead of freeing the inode directly. It causes a crash when we
rmmod immediately after we umount the volume[1].
So we need to call rcu_barrier after we kill_sb so that the inode is
freed before we do rmmod. The idea is inspired by Aneesh Kumar.
rcu_barrier will wait for all callbacks to end before preceding. The
original patch was done by Tao Ma, but synchronize_rcu() is not enough
here.
1. http://marc.info/?l=linux-fsdevel&m=
129680863330185&w=2
Tested-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Cc: Nick Piggin <npiggin@kernel.dk>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Chris Mason <chris.mason@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>