Some allocations are restricted to a limited set of nodes (due to memory
policies or cpuset constraints). If the page allocator is not able to find
enough memory then that does not mean that overall system memory is low.
In particular going postal and more or less randomly shooting at processes
is not likely going to help the situation but may just lead to suicide (the
whole system coming down).
It is better to signal to the process that no memory exists given the
constraints that the process (or the configuration of the process) has
placed on the allocation behavior. The process may be killed but then the
sysadmin or developer can investigate the situation. The solution is
similar to what we do when running out of hugepages.
This patch adds a check before we kill processes. At that point
performance considerations do not matter much so we just scan the zonelist
and reconstruct a list of nodes. If the list of nodes does not contain all
online nodes then this is a constrained allocation and we should kill the
current process.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
In the badness() calculation, there's currently this piece of code:
/*
* Processes which fork a lot of child processes are likely
* a good choice. We add the vmsize of the children if they
* have an own mm. This prevents forking servers to flood the
* machine with an endless amount of children
*/
list_for_each(tsk, &p->children) {
struct task_struct *chld;
chld = list_entry(tsk, struct task_struct, sibling);
if (chld->mm = p->mm && chld->mm)
points += chld->mm->total_vm;
}
The intention is clear: If some server (apache) keeps spawning new children
and we run OOM, we want to kill the father rather than picking a child.
This -- to some degree -- also helps a bit with getting fork bombs under
control, though I'd consider this a desirable side-effect rather than a
feature.
There's one problem with this: No matter how many or few children there are,
if just one of them misbehaves, and all others (including the father) do
everything right, we still always kill the whole family. This hits in real
life; whether it's javascript in konqueror resulting in kdeinit (and thus the
whole KDE session) being hit or just a classical server that spawns children.
Sidenote: The killer does kill all direct children as well, not only the
selected father, see oom_kill_process().
The idea in attached patch is that we do want to account the memory
consumption of the (direct) children to the father -- however not fully.
This maintains the property that fathers with too many children will still
very likely be picked, whereas a single misbehaving child has the chance to
be picked by the OOM killer.
In the patch I account only half (rounded up) of the children's vm_size to
the parent. This means that if one child eats more mem than the rest of
the family, it will be picked, otherwise it's still the father and thus the
whole family that gets selected.
This is heuristics -- we could debate whether accounting for a fourth would
be better than for half of it. Or -- if people would consider it worth the
trouble -- make it a sysctl. For now I sticked to accounting for half,
which should IMHO be a significant improvement.
The patch does one more thing: As users tend to be irritated by the choice
of killed processes (mainly because the children are killed first, despite
some of them having a very low OOM score), I added some more output: The
selected (father) process will be reported first and it's oom_score printed
to syslog.
Description:
Only account for half of children's vm size in oom score calculation
This should still give the parent enough point in case of fork bombs. If
any child however has more than 50% of the vm size of all children
together, it'll get a higher score and be elected.
This patch also makes the kernel display the oom_score.
Signed-off-by: Kurt Garloff <garloff@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
acpi_rs_get_list_length() needs to account for all the vendor-defined data
bytes. Failing to include these causes buffers to be sized too small,
which causes slab corruption when we later convert AML to resources and run
off the end of the buffer.
This causes slab corruption on machines that use ACPI vendor-defined
resources. All HP ia64 machines do, and I'm told that some NEC machines
may as well.
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Cc: "Brown, Len" <len.brown@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Make sure maxnodes is safe size before calculating nlongs in
get_nodes().
Signed-off-by: Chris Wright <chrisw@sous-sol.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
I got all of these backwards. We want to return
min(input timeout, new timeout)
to userspace to prevent increasing the time-remaining value.
Thanks to Ernst Herzberg <earny@net4u.de> for reporting and diagnosing.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
One of the parameters to the __pud_free_tlb() macro for powerpc is
incorrect (see patch) . We get away with it by accident, because the one
place the macro is called, the second parameter is a variable named "pud".
Signed-off-by: David Gibson <dwg@au1.ibm.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Don't print KERN_INFO in the middle of a printk line.
printk(KERN_INFO "OEM ID: %s ",str);
is just above this. This is already fixed up in i386 copy.
Signed-off-by: Martin J. Bligh <mbligh@google.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The windfarm_pm112 module relies on smu_sat_get_sdb_partition which is in
windfarm_smu_sat.c but is not exported to modules, so despite Kconfig
having the option to build the pm112 as modules, this can never be loaded.
This patch fixes that by exporting smu_sat_get_sdb_partition with
EXPORT_SYMBOL_GPL
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
There's a rather theoretical case of the BUG triggering in
fuse_reset_request():
- iget() fails because of OOM after a successful CREATE_OPEN request
- during IO on the resulting RELEASE request the connection is aborted
Fix and add warning to fuse_reset_request().
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Restore the compatibility with the older code and make it possible to
suspend if the kernel command line doesn't contain the "resume=" argument
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Just rename the compat system call to keep the name consistent with all the
other *64 compat system calls.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The last changes that introduced the additional_cpus command line parameter
also introduced a regression regarding smp initialization speed. In
smp_setup_cpu_possible_map() cpu_present_map is set to the same value as
cpu_possible_map. Especially that means that bits in the present map will be
set for cpus that are not present. This will cause a slow down in the initial
cpu_up() loop in smp_init() since trying to take cpus online that aren't
present takes a while.
Fix this by setting only bits for present cpus in cpu_present_map and set
cpu_present_map to cpu_possible_map in smp_cpus_done().
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Introduce possible_cpus command line option. Hard sets the number of bits set
in cpu_possible_map. Unlike the additional_cpus parameter this one guarantees
that num_possible_cpus() will stay constant even if the system gets rebooted
and a different number of cpus are present at startup.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Introduce additional_cpus command line option. By default no additional cpu
can be attached to the system anymore. Only the cpus present at IPL time can
be switched on/off. If it is desired that additional cpus can be attached to
the system the maximum number of additional cpus needs to be specified with
this option.
This change is necessary in order to limit the waste of per_cpu data
structures.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Set preempt_count of idle_thread to zero before switching off cpu. Otherwise
the preempt_count will be wrong if the cpu is switched on again since the
thread will be reused.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
If __ccw_device_disband_start() fails to initiate disbanding, it should finish
with ccw_device_disband_done() (which leaves the device in offline state)
instead of ccw_device_verify_done() (which leaves the device in online state).
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Heiko Carstens <heiko.carstens@de.ibm.com> wrote:
The boot sequence on s390 sometimes takes ages and we spend a very long
time (up to one or two minutes) in calibrate_migration_costs. The time
spent there differs from boot to boot. Also the calculated costs differ
a lot. I've seen differences by up to a factor of 15 (yes, factor not
percent). Also I doubt that making these measurements make much sense on
a completely virtualized architecture where you cannot tell how much cpu
time you will get anyway.
So introduce the CONFIG_DEFAULT_MIGRATION_COST method for an architecture
to set the scheduler migration costs. This turns off automatic detection
of migration costs. Makes sense on virtual platforms, where migration
costs are hard to measure accurately.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jean-Luc Leger <reiga@dspnet.fr.eu.org> found this obvious typo.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Fix IO-port leakage from request_region in case of error during TPM
initialization, adds more pnp-verification and fixes a WTX-bug.
Signed-off-by: Marcel Selhorst <selhorst@crypto.rub.de>
Acked-by: Kylene Jo Hall <kjhall@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Fix a deadlock possible in the ext2 file system implementation. This
deadlock occurs when a file is removed from an ext2 file system which was
mounted with the "sync" mount option.
The problem is that ext2_xattr_delete_inode() was invoking the routine,
sync_dirty_buffer(), using a buffer head which was previously locked via
lock_buffer(). The first thing that sync_dirty_buffer() does is to lock
the buffer head that it was passed. It does this via lock_buffer(). Oops.
The solution is to unlock the buffer head in ext2_xattr_delete_inode()
before invoking sync_dirty_buffer(). This makes the code in
ext2_xattr_delete_inode() obey the same locking rules as all other callers
of sync_dirty_buffer() in the ext2 file system implementation.
Signed-off-by: Peter Staubach <staubach@redhat.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
- Fix the array index value in ata_rwcmd_protocol() for the added FUA commands.
- Filter out ATAPI packet command error messages in ata_pio_error()
Signed-off-by: Albert Lee <albertcc@tw.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
* libata does not care about error interrupts, so handle them locally
* the interrupts that are ignored only appear to happen at init time
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
Change the find_next_best_node algorithm to correctly skip
over holes in the node online mask. Previously it would not handle
missing nodes correctly and cause crashes at boot.
[Written by Linus, tested by AK]
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
bond_release returns EINVAL without releasing the bond lock if the
slave device is not being bonded by the bond. The following patch
ensures that the lock is released in this case.
Signed-off-by: Stephen J. Bevan <stephen@dino.dnsalias.com>
Acked-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
[patch 2/2] s390: some qeth driver fixes
From: Frank Pavlic <fpavlic@de.ibm.com>
- fixed kernel panic when using EDDP support in Layer 2 mode
- NULL pointer exception in qeth_set_offline fixed.
- setting EDDP in Layer 2 mode did not set NETIF_F_(SG/TSO)
flags when device became online.
- use sscanf for parsing and converting IPv4 addresses
from string to __u8 values.
- qeth_string_to_ipaddr6 fixed. in case of double colon
the converted IPv6 address out from the string was not correct
in previous implementation.
Signed-off-by: Frank Pavlic <fpavlic@de.ibm.com>
diffstat:
qeth.h | 112 +++++++++++++++++++++++++-----------------------------------
qeth_eddp.c | 11 ++++-
qeth_main.c | 17 +++------
3 files changed, 63 insertions(+), 77 deletions(-)
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
[patch 1/2] s390: lcs performance enhancements
From: Klaus Wacker <kdwacker@de.ibm.com>
- When flood pinging (with large packet size) an LCS device,
about 90 % of all packets are dropped by driver.
- increased number of lcs IO buffers to 32.
- use netif_stop_queue/netif_wake_queue in lcs_start_xmit routine
- don't lock the whole xmit routine but just the piece of code where
tx_buffer is touched.
Signed-off-by: Frank Pavlic <fpavlic@de.ibm.com>
diffstat:
lcs.c | 31 +++++++++++++++++--------------
lcs.h | 2 +-
2 files changed, 18 insertions(+), 15 deletions(-)
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
drivers/net/tokenring/smctr.c: In function `smctr_load_firmware':
drivers/net/tokenring/smctr.c:2981: warning: assignment discards qualifiers from pointer target type
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
Users report problems w/ auto-negotiation disabled and the link set
to 100/Half or 10/Half. Problems range from poor performance to no
link at all.
The current sky2 code does not set things properly on link up if
autonegotiation is disabled. Plus it does not contemplate a 10Mbit
setting at all. This patch corrects that.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
This is a clone of John Linville's fixed for speed setting on sky2 driver.
The skge driver has the same code (and bug). It would not allow manually forcing
100 and 10 mbit.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
Take the experimental dependency of skge driver, it is as stable as the
others.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
The sk98lin driver was changed a while ago to remove support for the
D-Link 530T card because that hardware has no working VPD data. The help
text for Kconfig was not updated.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
Some bitfields were incorrectly initialised in wavelan_cs,
causing some compiler warning. Also killed a error message that should
not be there...
Signed-off-by: Jean Tourrilhes <jt@hpl.hp.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
Do not mask TIF_SINGLESTEP bit in _TIF_WORK_MASK. Masking this stopped
do_notify_resume() from being called when it should have been.
Signed-off-by: Chuck Ebbert <76306.1226@compuserve.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
A sysfs function call uses the wrong parameter, and thus breaks a build on
SGI O2.
CC drivers/video/gbefb.o
drivers/video/gbefb.c: In function ‘gbefb_remove’:
drivers/video/gbefb.c:1246: error: ‘dev’ undeclared (first use in this function)
drivers/video/gbefb.c:1246: error: (Each undeclared identifier is reported only once
drivers/video/gbefb.c:1246: error: for each function it appears in.)
make[2]: *** [drivers/video/gbefb.o] Error 1
Signed-off-by: Joshua Kinard <kumba@gentoo.org>
Signed-off-by: Antonino Daplas <adaplas@pol.net>
Signed-off-by: Martin Michlmayr <tbm@cyrius.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This provides an interface for arch code to find out how many
nanoseconds are going to be added on to xtime by the next call to
do_timer. The value returned is a fixed-point number in 52.12 format
in nanoseconds. The reason for this format is that it gives the
full precision that the timekeeping code is using internally.
The motivation for this is to fix a problem that has arisen on 32-bit
powerpc in that the value returned by do_gettimeofday drifts apart
from xtime if NTP is being used. PowerPC is now using a lockless
do_gettimeofday based on reading the timebase register and performing
some simple arithmetic. (This method of getting the time is also
exported to userspace via the VDSO.) However, the factor and offset
it uses were calculated based on the nominal tick length and weren't
being adjusted when NTP varied the tick length.
Note that 64-bit powerpc has had the lockless do_gettimeofday for a
long time now. It also had an extremely hairy routine that got called
from the 32-bit compat routine for adjtimex, which adjusted the
factor and offset according to what it thought the timekeeping code
was going to do. Not only was this only called if a 32-bit task did
adjtimex (i.e. not if a 64-bit task did adjtimex), it was also
duplicating computations from kernel/timer.c and it wasn't clear that
it was (still) correct.
The simple solution is to ask the timekeeping code how long the
current jiffy will be on each timer interrupt, after calling
do_timer. If this jiffy will be a different length from the last one,
we then need to compute new values for the factor and offset used in
the lockless do_gettimeofday. In this way we can keep xtime and
do_gettimeofday in sync, even when NTP is varying the tick length.
Note that when adjtimex varies the tick length, it almost always
introduces the variation from the next tick on. The only case I could
see where adjtimex would vary the length of the current tick is when
an old-style adjtime adjustment is being cancelled. (It's not clear
to me why the adjustment has to be cancelled immediately rather than
from the next tick on.) Thus I don't see any real need for a hook in
adjtimex; the rare case of an old-style adjustment being cancelled can
be fixed up at the next tick.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Acked-by: john stultz <johnstul@us.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The memory allocator doesn't like empty zones (which have an
uninitialized freelist), so a x86-64 system with a node fully
in GFP_DMA32 only would crash on mbind.
Fix that up by putting all possible zones as fallback into the zonelist
and skipping the empty ones.
In fact the code always enough allocated space for all zones,
but only used it for the highest. This change just uses all the
memory that was allocated before.
This should work fine for now, but whoever implements node hot removal
needs to fix this somewhere else too (or make sure zone datastructures
by itself never go away, only their memory)
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Previously the numa hash code would be confused by holes in the node space
and stop early. This is the first part of the fix for the non boot issue
with empty nodes on Opterons.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Code was refusing good SRATs because about 12K got lost somewhere.
Allow less than 1MB of difference before rejecting it.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
But do it after everything else to risk less from recursive
crashes.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
