Before requesting firmware, printk a message saying what we're requesting. This
makes it easier to see what's going on, and provides an explanation for the
huge silent delay that one would otherwise get after accidentally building
ipw2200 as a non-module.
Robert P. J. Day [Tue, 29 Apr 2008 07:59:34 +0000 (00:59 -0700)]
MAINTAINERS: clarify status of MN10300 mailing list as moderated
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca> Acked-by: David Howells <dhowells@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Andy Whitcroft [Tue, 29 Apr 2008 07:59:33 +0000 (00:59 -0700)]
update checkpatch.pl to version 0.18
This version brings a few fixes for the extern checks, and a couple of
new checks.
Of note:
- false is now recognised as a 0 assignment in static/external
assignments,
- printf format strings including %L are reported,
- a number of fixes for the extern in .c file detector which had
temporarily lost its ability to detect variables; undetected due to
the loss of its test.
Andy Whitcroft (8):
Version: 0.18
false should trip 0 assignment checks
tests: reinstate missing tests
tests: allow specification of the file extension for a test
fix extern checks for variables
check for and report %Lu, %Ld, and %Li
ensure we only start a statement on lines with some content
extern spacing
Signed-off-by: Andy Whitcroft <apw@shadowen.org> Cc: Ingo Molnar <mingo@elte.hu> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Andy Whitcroft [Tue, 29 Apr 2008 07:59:32 +0000 (00:59 -0700)]
update checkpatch.pl to version 0.17
This version brings improvements to external declaration detection, fixes to
quote tracking, fixes to unary tracking, some clarification of wording, and
the usual slew of fixes for false positives.
Of note:
- much better unary tracking across preprocessor directives
- UTF8 checks highlight the character at fault
- widening of mutex detection
Andy Whitcroft (17):
Version: 0.17
values: __attribute__ carries through the previous type
quotes: should only follow "positive" lines
clarify the indent tabs over spaces wording
loosen NR_CPUS check for array range initialisers
detect external function declarations without an extern prefix
function declaration arguments should be with the identifier
DEFINE_MUTEX should report in line with struct mutex
NR_CPUS is valid in preprocessor statements
comment detection should not start on the @@ line
types: add support for #undef
tighten mutex/completion reports to usage
allow export of function pointers
values: preprocessor #define is out of line maintain values
values: #define does not always have parentheses
unary '*' may be const
utf8 checks should report location of the invalid character
Wolfram Sang (1):
make checkpatch.pl really skip <asm/irq.h>
Signed-off-by: Andy Whitcroft <apw@shadowen.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Julia Lawall [Tue, 29 Apr 2008 07:59:30 +0000 (00:59 -0700)]
drivers/misc: elide a non-zero test on a result that is never 0
The function thermal_cooling_device_register always returns either a valid
pointer or a value made with ERR_PTR, so a test for non-zero on the result
will always succeed.
The problem was found using the following semantic match.
(http://www.emn.fr/x-info/coccinelle/)
//<smpl>
@a@
expression E, E1;
statement S,S1;
position p;
@@
E = thermal_cooling_device_register(...)
... when != E = E1
if@p (E) S else S1
@n@
position a.p;
expression E,E1;
statement S,S1;
@@
E = NULL
... when != E = E1
if@p (E) S else S1
@depends on !n@
expression E;
statement S,S1;
position a.p;
@@
* if@p (E)
S else S1
//</smpl>
Signed-off-by: Julia Lawall <julia@diku.dk> Cc: Thomas Sujith <sujith.thomas@intel.com> Cc: Len Brown <len.brown@intel.com> Cc: Zhang Rui <rui.zhang@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Robert P. J. Day [Tue, 29 Apr 2008 07:59:29 +0000 (00:59 -0700)]
lists: add "const" qualifier to first arg of list_splice() operations
Since neither the list_splice() nor __list_splice() routines modify their
first argument, might as well declare them "const".
[akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Robert P. J. Day [Tue, 29 Apr 2008 07:59:28 +0000 (00:59 -0700)]
kbuild: move files that don't check __KERNEL__
Move files that don't check __KERNEL__ from unifdef-y to header-y.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca> Cc: Sam Ravnborg <sam@ravnborg.org> Cc: David Woodhouse <dwmw2@infradead.org> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Robert P. J. Day [Tue, 29 Apr 2008 07:59:28 +0000 (00:59 -0700)]
kbuild: remove duplicate, conflicting entry for oom.h
oom.h is already tagged for unifdef'ing, so its entry as a simple exportable
header should be deleted.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca> Cc: Sam Ravnborg <sam@ravnborg.org> Cc: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Robert P. J. Day [Tue, 29 Apr 2008 07:59:25 +0000 (00:59 -0700)]
Remove superfluous include of string.h from percpu.h
There's nothing in percpu.h that requires an explicit inclusion of
string.h.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Robert P. J. Day [Tue, 29 Apr 2008 07:59:25 +0000 (00:59 -0700)]
kernel: explicitly include required header files under kernel/
Following an experimental deletion of the unnecessary directive
#include <linux/slab.h>
from the header file <linux/percpu.h>, these files under kernel/ were exposed
as needing to include one of <linux/slab.h> or <linux/gfp.h>, so explicit
includes were added where necessary.
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Register the ":text:E::txt::/root/cat.txt:' rule in binfmt_misc (by root) and
try launching the cat.txt file (by anyone) :) The result is - the endless
recursion in the load_misc_binary -> open_exec -> load_misc_binary chain and
stack overflow.
There's a similar problem with binfmt_script, and there's a sh_bang memner on
linux_binprm structure to handle this, but simply raising this in binfmt_misc
may break some setups when the interpreter of some misc binaries is a script.
So the proposal is to turn sh_bang into a bit, add a new one (the misc_bang)
and raise it in load_misc_binary. After this, even if we set up the misc ->
script -> misc loop for binfmts one of them will step on its own bang and
exit.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
SysRQ-P is not always useful on SMP systems, since it usually ends up showing
the backtrace of a CPU that is doing just fine, instead of the backtrace of
the CPU that is having problems.
This patch adds SysRQ show-all-cpus(L), which shows the backtrace of every
active CPU in the system. It skips idle CPUs because some SMP systems are
just too large and we already know what the backtrace of the idle task looks
like.
[akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Rik van Riel <riel@redhat.com>
Randy Dunlap <randy.dunlap@oracle.com> Cc: <lwoodman@redhat.com> Cc: Ingo Molnar <mingo@elte.hu> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
I noticed that 2.6.24.2 calculates bprm->argv_len at do_execve(). But it
doesn't update bprm->argv_len after "remove_arg_zero() +
copy_strings_kernel()" at load_script() etc.
audit_bprm() is called from search_binary_handler() and
search_binary_handler() is called from load_script() etc. Thus, I think the
condition check
if (bprm->argv_len > (audit_argv_kb << 10))
return -E2BIG;
in audit_bprm() might return wrong result when strlen(removed_arg) !=
strlen(spliced_args). Why not update bprm->argv_len at load_script() etc. ?
By the way, 2.6.25-rc3 seems to not doing the condition check. Is the field
bprm->argv_len no longer needed?
Openhaptics uses pointers in _IOC() macros, implement compat for them. Also
add _IOC alternatives which are not 32/64 bit dependent (structures
passed through aren't yet) -- libphantom will use them.
Almost all implementations of pci_iomap() in the kernel, including the generic
lib/iomap.c one, copies the content of a struct resource into unsigned long's
which will break on 32 bits platforms with 64 bits resources.
This fixes all definitions of pci_iomap() to use resource_size_t. I also
"fixed" the 64bits arch for consistency.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: <linux-arch@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
print_fn_descriptor_symbol() prints the address if we don't have a symbol, so
no need to print both.
Also, combine printing return value with elapsed time. Changes this:
Calling initcall 0xc05b7a70: pci_mmcfg_late_insert_resources+0x0/0x50()
initcall 0xc05b7a70: pci_mmcfg_late_insert_resources+0x0/0x50() returned 1.
initcall 0xc05b7a70 ran for 0 msecs: pci_mmcfg_late_insert_resources+0x0/0x50()
initcall at 0xc05b7a70: pci_mmcfg_late_insert_resources+0x0/0x50(): returned with error code 1
to this:
calling pci_mmcfg_late_insert_resources+0x0/0x50()
initcall pci_mmcfg_late_insert_resources+0x0/0x50() returned 1 after 0 msecs
initcall pci_mmcfg_late_insert_resources+0x0/0x50() returned with error code 1
Harvey Harrison [Tue, 29 Apr 2008 07:59:03 +0000 (00:59 -0700)]
autofs4: fix sparse warning in root.c
fs/autofs4/root.c:536:23: warning: symbol 'ino' shadows an earlier one
fs/autofs4/root.c:510:22: originally declared here
There is no need to redeclare, we are at the end of the loop and in
the next iteration of the loop, ino will be reset.
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Acked-by: Ian Kent <raven@themaw.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Adrian Bunk [Tue, 29 Apr 2008 07:58:58 +0000 (00:58 -0700)]
let LOG_BUF_SHIFT default to 17
16 kB is often no longer enough for a normal boot of an UP system.
And even less when people e.g. use suspend.
17 seems to be a more reasonable default for current kernels on current
hardware (it's just the default, anyone who is memory limited can still lower
it).
Signed-off-by: Adrian Bunk <bunk@kernel.org> Acked-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Adrian Bunk [Tue, 29 Apr 2008 07:58:51 +0000 (00:58 -0700)]
fs/ramfs/ extern cleanup
- internal.h shouldn't duplicate the extern declaration for
ramfs_file_operations already in include/linux/ramfs.h
- file-mmu.c needs two #include's for seeing the extern declarations
of it's global struct's
Signed-off-by: Adrian Bunk <bunk@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Sam Ravnborg [Tue, 29 Apr 2008 07:58:51 +0000 (00:58 -0700)]
cpu: fix section mismatch warning in reference to register_cpu_notifier
Fix following warnings:
WARNING: vmlinux.o(.text+0xc60): Section mismatch in reference from the function kvm_init() to the function .cpuinit.text:register_cpu_notifier()
WARNING: vmlinux.o(.text+0x33869a): Section mismatch in reference from the function xfs_icsb_init_counters() to the function .cpuinit.text:register_cpu_notifier()
WARNING: vmlinux.o(.text+0x5556a1): Section mismatch in reference from the function acpi_processor_install_hotplug_notify() to the function .cpuinit.text:register_cpu_notifier()
WARNING: vmlinux.o(.text+0xfe6b28): Section mismatch in reference from the function cpufreq_register_driver() to the function .cpuinit.text:register_cpu_notifier()
register_cpu_notifier() are only really defined when HOTPLUG_CPU is enabled.
So references to the function are OK.
Annotate it with __ref so we do not get warnings from callers and do not get
warnings for the functions/data used by register_cpu_notifier().
Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Cc: Gautham R Shenoy <ego@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Sam Ravnborg [Tue, 29 Apr 2008 07:58:50 +0000 (00:58 -0700)]
cpu: fix section mismatch warnings in *cpu_down
Fix following warnings:
WARNING: vmlinux.o(.text+0x75c8d): Section mismatch in reference from the function take_cpu_down() to the variable .cpuinit.data:cpu_chain
WARNING: vmlinux.o(.text+0x75d2a): Section mismatch in reference from the function _cpu_down() to the variable .cpuinit.data:cpu_chain
WARNING: vmlinux.o(.text+0x75d4d): Section mismatch in reference from the function _cpu_down() to the variable .cpuinit.data:cpu_chain
WARNING: vmlinux.o(.text+0x75de4): Section mismatch in reference from the function _cpu_down() to the variable .cpuinit.data:cpu_chain
WARNING: vmlinux.o(.text+0x75e33): Section mismatch in reference from the function _cpu_down() to the variable .cpuinit.data:cpu_chain
cpu_down is only used from code surrounded by HOTPLUG_CPU so any references to
__cpuinit is OK.
Add a few __ref to tech modpost to ignore the references.
This is just papering over the fact that the cpu hotplug code is fragile with
respect to use of HOTPLUG_CPU and in many cases rely on __cpuinit to get rid
of code when HOTPLUG_CPU is not enabled. For now this is the least invasive
change.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Cc: Gautham R Shenoy <ego@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Sam Ravnborg [Tue, 29 Apr 2008 07:58:48 +0000 (00:58 -0700)]
cpu: fix section mismatch warning in unregister_cpu_notifier
Fix following warning:
WARNING: vmlinux.o(.text+0x75f4e): Section mismatch in reference from the function unregister_cpu_notifier() to the variable .cpuinit.data:cpu_chain
We know that unregister_cpu_notifier is using HOTPLUG_CPU
stuff - so ignore these references.
Annotating unregister_cpu_notifier had been another option
but this caused far more warnings since not all callers were
annotated __cpuinit.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Cc: Gautham R Shenoy <ego@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Sam Ravnborg [Tue, 29 Apr 2008 07:58:47 +0000 (00:58 -0700)]
cpu: fix section mismatch warnings in hotcpu_register
Fix following warnings:
WARNING: vmlinux.o(.data+0x5020): Section mismatch in reference from the variable cpu_vsyscall_notifier_nb.12876 to the function .cpuinit.text:cpu_vsyscall_notifier()
WARNING: vmlinux.o(.data+0x9ce0): Section mismatch in reference from the variable profile_cpu_callback_nb.17654 to the function .devinit.text:profile_cpu_callback()
WARNING: vmlinux.o(.data+0xd380): Section mismatch in reference from the variable workqueue_cpu_callback_nb.15004 to the function .devinit.text:workqueue_cpu_callback()
WARNING: vmlinux.o(.data+0x11d00): Section mismatch in reference from the variable relay_hotcpu_callback_nb.19626 to the function .cpuinit.text:relay_hotcpu_callback()
WARNING: vmlinux.o(.data+0x12970): Section mismatch in reference from the variable cpu_callback_nb.24694 to the function .devinit.text:cpu_callback()
WARNING: vmlinux.o(.data+0x3fee0): Section mismatch in reference from the variable percpu_counter_hotcpu_callback_nb.10903 to the function .cpuinit.text:percpu_counter_hotcpu_callback()
WARNING: vmlinux.o(.data+0x74ce0): Section mismatch in reference from the variable topology_cpu_callback_nb.12506 to the function .cpuinit.text:topology_cpu_callback()
Functions used as argument are by definition only used in HOTPLUG_CPU
situations so thay are annotated __cpuinit. Annotate the static variable used
by hotcpu_register with __cpuinitdata to match this definition.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Cc: Gautham R Shenoy <ego@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add the RUSAGE_THREAD option for the getrusage system call. This is
essentially Roland's patch from http://lkml.org/lkml/2008/1/18/589, but the
line about RUSAGE_LWP line has been removed, as suggested by Ulrich and
Christoph.
Signed-off-by: Roland McGrath <roland@redhat.com> Signed-off-by: Sripathi Kodi <sripathik@in.ibm.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Michael Kerrisk <mtk.manpages@googlemail.com> Cc: Ulrich Drepper <drepper@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Harvey Harrison [Tue, 29 Apr 2008 07:58:41 +0000 (00:58 -0700)]
adfs: work around bogus sparse warning
fs/adfs/dir_f.c:126:4: warning: do-while statement is not a compound statement
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Cc: Russell King <rmk@arm.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Nur Hussein [Tue, 29 Apr 2008 07:58:39 +0000 (00:58 -0700)]
Taint kernel after WARN_ON(condition)
The kernel is sent to tainted within the warn_on_slowpath() function, and
whenever a warning occurs the new taint flag 'W' is set. This is useful to
know if a warning occurred before a BUG by preserving the warning as a flag
in the taint state.
This does not work on architectures where WARN_ON has its own definition.
These archs are:
1. s390
2. superh
3. avr32
4. parisc
The maintainers of these architectures have been added in the Cc: list
in this email to alert them to the situation.
The documentation in oops-tracing.txt has been updated to include the
new flag.
Signed-off-by: Nur Hussein <nurhussein@gmail.com> Cc: Arjan van de Ven <arjan@infradead.org> Cc: "Randy.Dunlap" <rdunlap@xenotime.net> Cc: Ingo Molnar <mingo@elte.hu> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Kyle McMartin <kyle@mcmartin.ca> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Haavard Skinnemoen <hskinnemoen@atmel.com> Cc: Paul Mundt <lethal@linux-sh.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
They're defined later on in the same file with bodies and nothing in
between needs them.
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> Reviewed-by: Pekka Enberg <penberg@cs.helsinki.fi> Acked-by: Jan Harkes <jaharkes@cs.cmu.edu> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Eric Dumazet [Tue, 29 Apr 2008 07:58:35 +0000 (00:58 -0700)]
Avoid divides in BITS_TO_LONGS
BITS_PER_LONG is a signed value (32 or 64)
DIV_ROUND_UP(nr, BITS_PER_LONG) performs signed arithmetic if "nr" is signed too.
Converting BITS_TO_LONGS(nr) to DIV_ROUND_UP(nr, BITS_PER_BYTE *
sizeof(long)) makes sure compiler can perform a right shift, even if "nr"
is a signed value, instead of an expensive integer divide.
Applying this patch saves 141 bytes on x86 when CONFIG_CC_OPTIMIZE_FOR_SIZE=y
and speedup bitmap operations.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Epoll calls rb_set_parent(n, n) to initialize the rb-tree node, but
rb_set_parent() accesses node's pointer in its code. This creates a
warning in kmemcheck (reported by Vegard Nossum) about an uninitialized
memory access. The warning is harmless since the following rb-tree node
insert is going to overwrite the node data. In any case I think it's
better to not have that happening at all, and fix it by simplifying the
code to get rid of a few lines that became superfluous after the previous
epoll changes.
Arjan van de Ven [Tue, 29 Apr 2008 07:58:34 +0000 (00:58 -0700)]
make /dev/kmem a config option
Make /dev/kmem a config option; /dev/kmem is VERY rarely used, and when
used, it's generally for no good (rootkits tend to be the most common
users). With this config option, users have the choice to disable
/dev/kmem, saving some size as well.
A patch to disable /dev/kmem has been in the Fedora and RHEL kernels for
4+ years now without any known problems or legit users of /dev/kmem.
[akpm@linux-foundation.org: make CONFIG_DEVKMEM default to y] Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Harvey Harrison [Tue, 29 Apr 2008 07:58:29 +0000 (00:58 -0700)]
power: replace remaining __FUNCTION__ occurrences
__FUNCTION__ is gcc-specific, use __func__
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Cc: Greg KH <greg@kroah.com> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Len Brown <lenb@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add __GFP_REPEAT to hugepage allocations. Do so to not necessitate userspace
putting pressure on the VM by repeated echo's into /proc/sys/vm/nr_hugepages
to grow the pool. With the previous patch to allow for large-order
__GFP_REPEAT attempts to loop for a bit (as opposed to indefinitely), this
increases the likelihood of getting hugepages when the system experiences (or
recently experienced) load.
Mel tested the patchset on an x86_32 laptop. With the patches, it was easier
to use the proc interface to grow the hugepage pool. The following is the
output of a script that grows the pool as much as possible running on
2.6.25-rc9.
Allocating hugepages test
-------------------------
Disabling OOM Killer for current test process
Starting page count: 0
Attempt 1: 57 pages Progress made with 57 pages
Attempt 2: 73 pages Progress made with 16 pages
Attempt 3: 74 pages Progress made with 1 pages
Attempt 4: 75 pages Progress made with 1 pages
Attempt 5: 77 pages Progress made with 2 pages
77 pages was the most it allocated but it took 5 attempts from userspace
to get it. With the 3 patches in this series applied,
Allocating hugepages test
-------------------------
Disabling OOM Killer for current test process
Starting page count: 0
Attempt 1: 75 pages Progress made with 75 pages
Attempt 2: 76 pages Progress made with 1 pages
Attempt 3: 79 pages Progress made with 3 pages
And 79 pages was the most it got. Your patches were able to allocate the
bulk of possible pages on the first attempt.
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com> Cc: Andy Whitcroft <apw@shadowen.org> Tested-by: Mel Gorman <mel@csn.ul.ie> Cc: Dave Hansen <haveblue@us.ibm.com> Cc: Christoph Lameter <clameter@sgi.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
page allocator: smarter retry of costly-order allocations
Because of page order checks in __alloc_pages(), hugepage (and similarly
large order) allocations will not retry unless explicitly marked
__GFP_REPEAT. However, the current retry logic is nearly an infinite
loop (or until reclaim does no progress whatsoever). For these costly
allocations, that seems like overkill and could potentially never
terminate. Mel observed that allowing current __GFP_REPEAT semantics for
hugepage allocations essentially killed the system. I believe this is
because we may continue to reclaim small orders of pages all over, but
never have enough to satisfy the hugepage allocation request. This is
clearly only a problem for large order allocations, of which hugepages
are the most obvious (to me).
Modify try_to_free_pages() to indicate how many pages were reclaimed.
Use that information in __alloc_pages() to eventually fail a large
__GFP_REPEAT allocation when we've reclaimed an order of pages equal to
or greater than the allocation's order. This relies on lumpy reclaim
functioning as advertised. Due to fragmentation, lumpy reclaim may not
be able to free up the order needed in one invocation, so multiple
iterations may be requred. In other words, the more fragmented memory
is, the more retry attempts __GFP_REPEAT will make (particularly for
higher order allocations).
This changes the semantics of __GFP_REPEAT subtly, but *only* for
allocations > PAGE_ALLOC_COSTLY_ORDER. With this patch, for those size
allocations, we will try up to some point (at least 1<<order reclaimed
pages), rather than forever (which is the case for allocations <=
PAGE_ALLOC_COSTLY_ORDER).
This change improves the /proc/sys/vm/nr_hugepages interface with a
follow-on patch that makes pool allocations use __GFP_REPEAT. Rather
than administrators repeatedly echo'ing a particular value into the
sysctl, and forcing reclaim into action manually, this change allows for
the sysctl to attempt a reasonable effort itself. Similarly, dynamic
pool growth should be more successful under load, as lumpy reclaim can
try to free up pages, rather than failing right away.
Choosing to reclaim only up to the order of the requested allocation
strikes a balance between not failing hugepage allocations and returning
to the caller when it's unlikely to every succeed. Because of lumpy
reclaim, if we have freed the order requested, hopefully it has been in
big chunks and those chunks will allow our allocation to succeed. If
that isn't the case after freeing up the current order, I don't think it
is likely to succeed in the future, although it is possible given a
particular fragmentation pattern.
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com> Cc: Andy Whitcroft <apw@shadowen.org> Tested-by: Mel Gorman <mel@csn.ul.ie> Cc: Dave Hansen <haveblue@us.ibm.com> Cc: Christoph Lameter <clameter@sgi.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The definition and use of __GFP_REPEAT, __GFP_NOFAIL and __GFP_NORETRY in the
core VM have somewhat differing comments as to their actual semantics.
Annoyingly, the flags definition has inline and header comments, which might
be interpreted as not being equivalent. Just add references to the header
comments in the inline ones so they don't go out of sync in the future. In
their use in __alloc_pages() clarify that the current implementation treats
low-order allocations and __GFP_REPEAT allocations as distinct cases.
To clarify, the flags' semantics are:
__GFP_NORETRY means try no harder than one run through __alloc_pages
__GFP_REPEAT means __GFP_NOFAIL
__GFP_NOFAIL means repeat forever
order <= PAGE_ALLOC_COSTLY_ORDER means __GFP_NOFAIL
usemap must be initialized only when pfn is within zone. If not, it corrupts
memory.
And this patch also reduces the number of calls to set_pageblock_migratetype()
from
(pfn & (pageblock_nr_pages -1)
to
!(pfn & (pageblock_nr_pages-1)
it should be called once per pageblock.
randconfig testing in x86.git found the following upstream build bug:
drivers/built-in.o: In function `acer_led_exit':
acer-wmi.c:(.text+0xdc76e): undefined reference to `led_classdev_unregister'
drivers/built-in.o: In function `acer_platform_probe':
acer-wmi.c:(.devinit.text+0x63e6): undefined reference to `led_classdev_register'
which was due to acer-wmi.o only depending on CONFIG_LEDS_CLASS, while
also using a symbol offered by CONFIG_NEW_LEDS. Also fix a similar bug
in CONFIG_ASUS_LAPTOP.
Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Len Brown <len.brown@intel.com>
thinkpad_acpi: fix possible NULL pointer dereference if kstrdup failed
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com> Acked-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com>
Li Zefan [Fri, 18 Apr 2008 20:27:29 +0000 (13:27 -0700)]
ACPI: check a return value correctly in acpi_power_get_context()
We should check *resource != NULL rather than resource != NULL, which will be
always true.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com> Acked-by: Zhao Yakui <yakui.zhao@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Len Brown <len.brown@intel.com>
Eric Cooper [Thu, 13 Mar 2008 11:55:46 +0000 (12:55 +0100)]
eeepc-laptop: add base driver
This patch is based on Eric Cooper's work to clean the original asus_acpi
given by Asus. It's a platform driver (/sys/devices/platform/eeepc/)
wich support:
- hotkeys - wlan on/off - camera on/off - cardr on/off
Signed-off-by: Corentin Chary <corentincj@iksaif.net> Signed-off-by: Len Brown <len.brown@intel.com>
Add missing select for BACKLIGHT_LCD_SUPPORT, as select doesn't select the
dependencies of a symbol for us.
Also, "select INPUT" in Kconfig. We are not an Input device, nor are we
anywhere close to the input subsystem in the Kconfig tree, so using
"depends on INPUT" is not user-friendly at all.
Signed-off-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br> Signed-off-by: Len Brown <len.brown@intel.com>
Switch all task workers to a private thinkpad-acpi workqueue.
This way, we don't risk causing trouble for other tasks scheduled to the
default work queue, as our workers end up needing to access the ACPI EC,
run ACPI AML code, trigger SMI traps... and none of those are exactly known
to be fast, simple operations.
Signed-off-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br> Signed-off-by: Len Brown <len.brown@intel.com>
ACPI: thinkpad-acpi: use uppercase for "LED" on user documentation
Change all occourences of the "led" word to full uppercase in user
documentation.
Signed-off-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br> Acked-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Len Brown <len.brown@intel.com>
ACPI: thinkpad-acpi: claim tpacpi as an official short handle (v1.1)
Unfortunately, a lot of stuff in the kernel has size limitations, so
"thinkpad-acpi" ends up eating up too much real estate. We were using
"tpacpi" in symbols already, but this shorthand was not visible to
userland.
Document that the driver will use tpacpi as a short hand where necessary,
and use it to name the kernel thread for NVRAM polling (now named
"ktpacpi_nvramd").
Also, register a module alias with the shorthand. One can refer to the
module using the shorthand name.
Signed-off-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br> Signed-off-by: Len Brown <len.brown@intel.com>
ACPI: thinkpad-acpi: fix brightness dimming control bug
ibm-acpi and thinkpad-acpi did not know about bit 5 of the EC backlight
level control register (EC 0x31), so it was always forced to zero on
any writes.
This would disable the BIOS option to *not* use a dimmer backlight level
scale while on battery, and who knows what else (there are two other
control bits of unknown function).
Bit 5 controls the "reduce backlight levels when on battery" optional
functionality (active low). Bits 6 and 7 are better left alone as well,
instead of being forced to zero.
Signed-off-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br> Signed-off-by: Len Brown <len.brown@intel.com>
If userspace applications mess with the CMOS NVRAM, or something causes
both the ACPI firmware and thinkpad-acpi to try to change the brightness at
the same time, it is possible to have the CMOS and EC registers for the
current brightness go out of sync.
Should that happen, thinkpad-acpi could be really obnoxious when using a
brightness_mode of 3 (both EC and CMOS). Instead of complaining a massive
number of times, make sure to complain only once until EC and CMOS are back
in sync.
Signed-off-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br> Cc: Joerg Platte <lists@naasa.net> Signed-off-by: Len Brown <len.brown@intel.com>
During initialization, thinkpad-acpi outputs some messages to make sure
releavant box identification information is easily available in-line with
the rest of the driver messages.
Enhance those messages to output the alfanumeric model number as well.
Signed-off-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br> Signed-off-by: Len Brown <len.brown@intel.com>
ACPI: thinkpad-acpi: warn once about weird hotkey masks
thinkpad-acpi knows for a while now how to best program the hotkeys by
default, and always enable them by default. Unfortunately, this
information has not filtered down everywhere it needs to, yet. Notably,
old ibm-acpi documentation and most "thinkpad setup guides" will have wrong
information on this area.
Warn the local admin once whenever any of the following patterns are met:
1. Attempts to set hotkey mask to 0xffff (artifact from docs and config
for the old ibm-acpi driver and behaviour). This mask makes no
real-world sense;
2. Attempts to set hotkey mask to 0xffffffff, which means the user is
trying to just have "everything work" without even reading the
documentation, or that we need to get a bug report, because there
is a new thinkpad out there with new exciting hot keys :-)
3. Attempts to set hotkey mask to 0xffffff, which is almost never the
correct way to set up volume and brightness event reporting (and with
the current state-of-the-art, it is known to never be right way to do
it).
The driver will perform any and all requested operations, though,
regardless of any warnings. I hope these warnings can be removed one or
two years from now.
Signed-off-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br> Signed-off-by: Len Brown <len.brown@intel.com>
Lenovo ThinkPads with generic ACPI backlight level control can be easily
set to react to keyboard brightness key presses in a more predictable way
than what they do when in "DOS / bootloader" mode after Linux brings
up the ACPI interface.
The switch to the ACPI backlight mode in the firmware is designed to be
safe to use only as an one way trapdoor. One is not to force the firmware
to switch back to "DOS/bootloader" mode except by rebooting. The mode
switch itself is performed by calling any of the ACPI _BCL methods at least
once.
When in ACPI mode, the backlight firmware just issues (standard) events for
the brightness up/down hot key presses along with the non-standard HKEY
events which thinkpad-acpi traps, and doesn't touch the hardware.
thinkpad-acpi will:
1. Place the ThinkPad firmware in ACPI backlight control mode
if one is available
2. Suppress HKEY backlight change notifications by default
to avoid double-reporting when ACPI video is loaded when
the ThinkPad is in ACPI backlight control mode
3. Urge the user to load the ACPI video driver
The user is free to use either the ACPI video driver to get the brightness
key events, or to override the thinkpad-acpi default hotkey mask to get
them from thinkpad-acpi as well (this will result in duplicate events if
ACPI video is loaded, so let's hope distros won't screw this up).
Provided userspace is sane, all should work (and *keep* working), which is
more that can be said about the non-ACPI mode of the new Lenovo ThinkPad
BIOSes when coupled to current userspace and X.org drivers.
Full guidelines for backlight hot key reporting and use of the
thinkpad-acpi backlight interface have been added to the documentation.
Signed-off-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Cc: Thomas Renninger <trenn@suse.de> Signed-off-by: Len Brown <len.brown@intel.com>
[POWERPC] mpc5200: defconfigs for CM5200, Lite5200B, Motion-PRO and TQM5200
Board-specific defconfigs based on current mpc5200_defconfig, archival
lite5200_defconfig, and [cm5200|motionpro|tqm5200]_defconfig from the
linux-2.6-denx tree. Kernels build using these defconfigs were verified
to boot with root filesystem mounted over NFS on Motion-PRO, TQM5200
and Lite5200B boards. CM5200 target was not tested due to hardware
unavailability.
Signed-off-by: Bartlomiej Sieka <tur@semihalf.com> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
[POWERPC] mpc5200: Fix FEC error handling on FIFO errors
The error handling for the mpc5200 fec interrupt is broken. The intended
behaviour is like this:
* If one of FEC_IEVENT_RFIFO_ERROR and FEC_IEVENT_XFIFO_ERROR happens,
the datasheet says (MPC5200B User's Guide R1.2, p. 14-13): "When this
occurs, software must ensure both the FIFO Controller and BestComm are
soft-reset".
* On any other error (non-TFINT) interrupt, just issue a debug message.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
[POWERPC] mpc5200: add Phytec pcm030 board support
Add board support for the Phytec pcm030 mpc5200b based board. It
does not need any platform specific fixups and as such is handled
as a mpc5200 simple platform.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
projects / linux-2.6-omap-h63xx.git / log