Doug Ledford [Thu, 3 May 2012 05:44:52 +0000 (15:44 +1000)]
ipc/mqueue: correct mq_attr_ok test
While working on the other parts of the mqueue stuff, I noticed that the
calculation for overflow in mq_attr_ok didn't actually match reality (this
is especially true since my last patch which changed how we account memory
slightly).
In particular, we used to test for overflow using:
msgs * msgsize + msgs * sizeof(struct msg_msg *)
That was never really correct because each message we allocate via
load_msg() is actually a struct msg_msg followed by the data for the
message (and if struct msg_msg + data exceeds PAGE_SIZE we end up
allocating struct msg_msgseg structs too, but accounting for them would
get really tedious, so let's ignore those...they're only a pointer in size
anyway). This patch updates the calculation to be more accurate in
regards to maximum possible memory consumption by the mqueue.
Signed-off-by: Doug Ledford <dledford@redhat.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Manfred Spraul <manfred@colorfullife.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Doug Ledford [Thu, 3 May 2012 05:44:51 +0000 (15:44 +1000)]
ipc/mqueue: improve performance of send/recv
The existing implementation of the POSIX message queue send and recv
functions is, well, abysmal. Even worse than abysmal. I submitted a
patch to increase the maximum POSIX message queue limit to 65536 due to
customer needs, however, upon looking over the send/recv implementation, I
realized that my customer needs help with that too even if they don't know
it. The basic problem is that, given the fairly typical use case scenario
for a large queue of queueing lots of messages all at the same priority (I
verified with my customer that this is indeed what their app does), the
msg_insert routine is basically a frikkin' bubble sort. I mean, whoa,
that's *so* middle school.
OK, OK, to not slam the original author too much, I'm sure they didn't
envision a queue depth of 50,000+ messages. No one would think that
moving elements in an array, one at a time, and dereferencing each pointer
in that array to check priority of the message being pointed too, again
one at a time, for 50,000+ times would be good. So let's assume that, as
is typical, the users have found a way to break our code simply by using
it in a way we didn't envision. Fair enough.
"So, just how broken is it?", you ask. I wondered the same thing, so I
wrote an app to let me know. It's my next patch. It gave me some
interesting results. Here's what it tested:
Interference with other apps - In continuous mode, the app just sits there
and hits a message queue forever, while you go do something productive on
another terminal using other CPUs. You then measure how long it takes you
to do that something productive. Then you restart the app in fake
continuous mode, and it sits in a tight loop on a CPU while you repeat
your tests. The whole point of this is to keep one CPU tied up (so it
can't be used in your other work) but in one case tied up hitting the
mqueue code so we can see the effect of walking that 65,528 element array
one pointer at a time on the global CPU cache. If it's bad, then it will
slow down your app on the other CPUs just by polluting cache mercilessly.
In the fake case, it will be in a tight loop, but not polluting cache.
Testing the mqueue subsystem directly - Here we just run a number of tests
to see how the mqueue subsystem performs under different conditions. A
couple conditions are known to be worst case for the old system, and some
routines, so this tests all of them.
So, on to the results already:
Subsystem/Test Old New
Time to compile linux
kernel (make -j12 on a
6 core CPU)
Running mqueue test user 49m10.744s user 45m26.294s
sys 5m51.924s sys 4m59.894s
total 55m02.668s total 50m26.188s
Running fake test user 45m32.686s user 45m18.552s
sys 5m12.465s sys 4m56.468s
total 50m45.151s total 50m15.020s
% slowdown from mqueue
cache thrashing ~8% ~.5%
Avg time to send/recv (in nanoseconds per message)
when queue empty 305/288 349/318
when queue full (65528 messages)
constant priority 526589/823 362/314
increasing priority 403105/916 495/445
decreasing priority 73420/594 482/409
random priority 280147/920 546/436
Time to fill/drain queue (65528 messages, in seconds)
constant priority 17.37/.12 .13/.12
increasing priority 4.14/.14 .21/.18
decreasing priority 12.93/.13 .21/.18
random priority 8.88/.16 .22/.17
So, I think the results speak for themselves. It's possible this
implementation could be improved by cacheing at least one priority level
in the node tree (that would bring the queue empty performance more in
line with the old implementation), but this works and is *so* much better
than what we had, especially for the common case of a single priority in
use, that further refinements can be in follow on patches.
Signed-off-by: Doug Ledford <dledford@redhat.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Manfred Spraul <manfred@colorfullife.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Doug Ledford [Thu, 3 May 2012 05:44:51 +0000 (15:44 +1000)]
selftests: add mq_open_tests
Add a directory to house POSIX message queue subsystem specific tests.
Add first test which checks the operation of mq_open() under various
corner conditions.
Signed-off-by: Doug Ledford <dledford@redhat.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Doug Ledford <dledford@redhat.com> Cc: Joe Korty <joe.korty@ccur.com> Cc: Amerigo Wang <amwang@redhat.com> Cc: Serge E. Hallyn <serue@us.ibm.com> Cc: Jiri Slaby <jslaby@suse.cz> Cc: Manfred Spraul <manfred@colorfullife.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
KOSAKI Motohiro [Thu, 3 May 2012 05:44:50 +0000 (15:44 +1000)]
mqueue: separate mqueue default value from maximum value
commit b231cca438 ("message queues: increase range limits") changed mqueue
default value when attr parameter is specified NULL from hard coded value
to fs.mqueue.{msg,msgsize}_max sysctl value.
This made large side effect. When user need to use two mqueue
applications 1) using !NULL attr parameter and it require big message size
and 2) using NULL attr parameter and only need small size message, app (1)
require to raise fs.mqueue.msgsize_max and app (2) consume large memory
size even though it doesn't need.
Doug Ledford propsed to switch back it to static hard coded value.
However it also has a compatibility problem. Some applications might
started depend on the default value is tunable.
The solution is to separate default value from maximum value.
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Doug Ledford <dledford@redhat.com> Acked-by: Doug Ledford <dledford@redhat.com> Acked-by: Joe Korty <joe.korty@ccur.com> Cc: Amerigo Wang <amwang@redhat.com> Acked-by: Serge E. Hallyn <serue@us.ibm.com> Cc: Jiri Slaby <jslaby@suse.cz> Cc: Manfred Spraul <manfred@colorfullife.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
KOSAKI Motohiro [Thu, 3 May 2012 05:44:50 +0000 (15:44 +1000)]
mqueue: don't use kmalloc with KMALLOC_MAX_SIZE
KMALLOC_MAX_SIZE is not a good threshold. It is extremely high and
problematic. Unfortunately, some silly drivers depend on this and we
can't change it. But any new code needn't use such extreme ugly high
order allocations. It brings us awful fragmentation issues and system
slowdown.
Signed-off-by: KOSAKI Motohiro <mkosaki@jp.fujitsu.com> Acked-by: Doug Ledford <dledford@redhat.com> Acked-by: Joe Korty <joe.korty@ccur.com> Cc: Amerigo Wang <amwang@redhat.com> Cc: Serge E. Hallyn <serue@us.ibm.com> Cc: Jiri Slaby <jslaby@suse.cz> Cc: Joe Korty <joe.korty@ccur.com> Cc: Manfred Spraul <manfred@colorfullife.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
KOSAKI Motohiro [Thu, 3 May 2012 05:44:49 +0000 (15:44 +1000)]
mqueue: revert bump up DFLT_*MAX
Mqueue limitation is slightly naieve parameter likes other ipcs because
unprivileged user can consume kernel memory by using ipcs.
Thus, too aggressive raise bring us security issue. Example, current
setting allow evil unprivileged user use 256GB (= 256 * 1024 * 1024*1024)
and it's enough large to system will belome unresponsive. Don't do that.
Instead, every admin should adjust the knobs for their own systems.
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Acked-by: Doug Ledford <dledford@redhat.com> Acked-by: Joe Korty <joe.korty@ccur.com> Cc: Amerigo Wang <amwang@redhat.com> Acked-by: Serge E. Hallyn <serue@us.ibm.com> Cc: Jiri Slaby <jslaby@suse.cz> Cc: Manfred Spraul <manfred@colorfullife.com> Cc: Dave Hansen <haveblue@us.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Doug Ledford [Thu, 3 May 2012 05:44:49 +0000 (15:44 +1000)]
ipc/mqueue: update maximums for the mqueue subsystem
Commit b231cca4381e ("message queues: increase range limits") changed the
maximum size of a message in a message queue from INT_MAX to 8192*128.
Unfortunately, we had customers that relied on a size much larger than
8192*128 on their production systems. After reviewing POSIX, we found
that it is silent on the maximum message size. We did find a couple other
areas in which it was not silent. Fix up the mqueue maximums so that the
customer's system can continue to work, and document both the POSIX and
real world requirements in ipc_namespace.h so that we don't have this
issue crop back up.
Also, commit 9cf18e1dd74cd0 ("ipc: HARD_MSGMAX should be higher not lower
on 64bit") fiddled with HARD_MSGMAX without realizing that the number was
intentionally in place to limit the msg queue depth to one that was small
enough to kmalloc an array of pointers (hence why we divided 128k by
sizeof(long)). If we wish to meet POSIX requirements, we have no choice
but to change our allocation to a vmalloc instead (at least for the large
queue size case). With that, it's possible to increase our allowed
maximum to the POSIX requirements (or more if we choose).
Signed-off-by: Doug Ledford <dledford@redhat.com> Cc: Serge E. Hallyn <serue@us.ibm.com> Cc: Amerigo Wang <amwang@redhat.com> Cc: Joe Korty <joe.korty@ccur.com> Cc: Jiri Slaby <jslaby@suse.cz> Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Manfred Spraul <manfred@colorfullife.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Doug Ledford [Thu, 3 May 2012 05:44:48 +0000 (15:44 +1000)]
ipc/mqueue: enforce hard limits
In two places we don't enforce the hard limits for CAP_SYS_RESOURCE apps.
In preparation for making more reasonable hard limits, start enforcing
them even on CAP_SYS_RESOURCE.
Signed-off-by: Doug Ledford <dledford@redhat.com> Cc: Serge E. Hallyn <serue@us.ibm.com> Cc: Amerigo Wang <amwang@redhat.com> Cc: Joe Korty <joe.korty@ccur.com> Cc: Jiri Slaby <jslaby@suse.cz> Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Manfred Spraul <manfred@colorfullife.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Doug Ledford [Thu, 3 May 2012 05:44:48 +0000 (15:44 +1000)]
ipc/mqueue: switch back to using non-max values on create
Commit b231cca4381 ("message queues: increase range limits") changed how
we create a queue that does not include an attr struct passed to open so
that it creates the queue with whatever the maximum values are. However,
if the admin has set the maximums to allow flexibility in creating a queue
(aka, both a large size and large queue are allowed, but combined they
create a queue too large for the RLIMIT_MSGQUEUE of the user), then
attempts to create a queue without an attr struct will fail. Switch back
to using acceptable defaults regardless of what the maximums are.
Note: so far, we only know of a few applications that rely on this
behavior (specifically, set the maximums in /proc, then run the
application which calls mq_open() without passing in an attr struct, and
the application expects the newly created message queue to have the
maximum sizes that were set in /proc used on the mq_open() call, and all
of those applications that we know of are actually part of regression test
suites that were coded to do something like this:
for size in 4096 65536 $((1024 * 1024)) $((16 * 1024 * 1024)); do
echo $size > /proc/sys/fs/mqueue/msgsize_max
mq_open || echo "Error opening mq with size $size"
done
These test suites that depend on any behavior like this are broken. The
concept that programs should rely upon the system wide maximum in order to
get their desired results instead of simply using a attr struct to specify
what they want is fundamentally unfriendly programming practice for any
multi-tasking OS.
Fixing this will break those few apps that we know of (and those app
authors recognize the brokenness of their code and the need to fix it).
However, the following patch "mqueue: separate mqueue default value"
allows a workaround in the form of new knobs for the default msg queue
creation parameters for any software out there that we don't already know
about that might rely on this behavior at the moment.
Signed-off-by: Doug Ledford <dledford@redhat.com> Cc: Serge E. Hallyn <serue@us.ibm.com> Cc: Amerigo Wang <amwang@redhat.com> Cc: Joe Korty <joe.korty@ccur.com> Cc: Jiri Slaby <jslaby@suse.cz> Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Manfred Spraul <manfred@colorfullife.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Doug Ledford [Thu, 3 May 2012 05:44:48 +0000 (15:44 +1000)]
ipc/mqueue: cleanup definition names and locations
Since commit b231cca4381 ("message queues: increase range limits") on Oct
18, 2008, calls to mq_open() that did not pass in an attribute struct and
expected to get default values for the size of the queue and the max
message size now get the system wide maximums instead of hardwired
defaults like they used to get.
This was uncovered when one of the earlier patches in this patch set
increased the default system wide maximums at the same time it increased
the hard ceiling on the system wide maximums (a customer specifically
needed the hard ceiling brought back up, the new ceiling that commit b231cca4381 introduced was too low for their production systems). By
increasing the default maximums and not realising they were tied to any
attempt to create a message queue without an attribute struct, I had
inadvertently made it such that all message queue creation attempts
without an attribute struct were failing because the new default maximums
would create a queue that exceeded the default rlimit for message queue
bytes.
As a result, the system wide defaults were brought back down to their
previous levels, and the system wide ceilings on the maximums were raised
to meet the customer's needs. However, the fact that the no attribute
struct behavior of mq_open() could be broken by changing the system wide
maximums for message queues was seen as fundamentally broken itself. So
we hardwired the no attribute case back like it used to be. But, then we
realized that on the very off chance that some piece of software in the
wild depended on that behavior, we could work around that issue by adding
two new knobs to /proc that allowed setting the defaults for message
queues created without an attr struct separately from the system wide
maximums.
What is not an option IMO is to leave the current behavior in place. No
piece of software should ever rely on setting the system wide maximums in
order to get a desired message queue. Such a reliance would be so
fundamentally multitasking OS unfriendly as to not really be tolerable.
Fortunately, we don't know of any software in the wild that uses this
except for a regression test program that caught the issue in the first
place. If there is though, we have made accommodations with the two new
/proc knobs (and that's all the accommodations such fundamentally broken
software can be allowed)..
This patch:
The various defines for minimums and maximums of the sysctl controllable
mqueue values are scattered amongst different files and named
inconsistently. Move them all into ipc_namespace.h and make them have
consistent names. Additionally, make the number of queues per namespace
also have a minimum and maximum and use the same sysctl function as the
other two settable variables.
Signed-off-by: Doug Ledford <dledford@redhat.com> Acked-by: Serge E. Hallyn <serue@us.ibm.com> Cc: Amerigo Wang <amwang@redhat.com> Cc: Joe Korty <joe.korty@ccur.com> Cc: Jiri Slaby <jslaby@suse.cz> Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Manfred Spraul <manfred@colorfullife.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Manfred Spraul [Thu, 3 May 2012 05:44:47 +0000 (15:44 +1000)]
ipc/sem.c: alternatives to preempt_disable()
ipc/sem.c uses a custom wakeup scheme that relies on preempt_disable().
On -RT, this causes increased latencies and debug warnings.
The patch adds two additional schemes:
- one built around a completion - could be better for -RT kernels
- one built around a spinlock - unfortunately it's broken
- and the current one
My preferred solution would be the spinlock implementation: RT would use
premptible spinlocks, mainline normal spinlocks. Thus both get the
optimal implementation without any special code in ipc/sem.c.
Unfortunately, I don't see how it could be fixed.
Signed-off-by: Manfred Spraul <manfred@colorfullife.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Mike Galbraith <efault@gmx.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Anton Vorontsov [Thu, 3 May 2012 05:44:47 +0000 (15:44 +1000)]
um: properly check all process' threads for a live mm
kill_off_processes() might miss a valid process, this is because checking
for process->mm is not enough. Process' main thread may exit or detach
its mm via use_mm(), but other threads may still have a valid mm.
To catch this we use find_lock_task_mm(), which walks up all threads and
returns an appropriate task (with task lock held).
Suggested-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Cc: Richard Weinberger <richard@nod.at> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Anton Vorontsov [Thu, 3 May 2012 05:44:46 +0000 (15:44 +1000)]
blackfin: fix possible deadlock in decode_address()
Oleg Nesterov found an interesting deadlock possibility:
> sysrq_showregs_othercpus() does smp_call_function(showacpu)
> and showacpu() show_stack()->decode_address(). Now suppose that IPI
> interrupts the task holding read_lock(tasklist).
To fix this, blackfin should not grab the write_ variant of the
tasklist lock, read_ one is enough.
Suggested-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Cc: Mike Frysinger <vapier@gentoo.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Anton Vorontsov [Thu, 3 May 2012 05:44:45 +0000 (15:44 +1000)]
blackfin: a couple of task->mm handling fixes
The patch fixes two problems:
1. Working with task->mm w/o getting mm or grabing the task lock is
dangerous as ->mm might disappear (exit_mm() assigns NULL under
task_lock(), so tasklist lock is not enough).
We can't use get_task_mm()/mmput() pair as mmput() might sleep,
so we have to take the task lock while handle its mm.
2. Checking for process->mm is not enough because process' main
thread may exit or detach its mm via use_mm(), but other threads
may still have a valid mm.
To catch this we use find_lock_task_mm(), which walks up all
threads and returns an appropriate task (with task lock held).
Suggested-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Cc: Mike Frysinger <vapier@gentoo.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Anton Vorontsov [Thu, 3 May 2012 05:44:45 +0000 (15:44 +1000)]
sh: use clear_tasks_mm_cpumask()
Checking for process->mm is not enough because process' main thread may
exit or detach its mm via use_mm(), but other threads may still have a
valid mm.
To fix this we would need to use find_lock_task_mm(), which would walk up
all threads and returns an appropriate task (with task lock held).
clear_tasks_mm_cpumask() has the issue fixed, so let's use it.
Suggested-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Cc: Paul Mundt <lethal@linux-sh.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Anton Vorontsov [Thu, 3 May 2012 05:44:45 +0000 (15:44 +1000)]
powerpc: use clear_tasks_mm_cpumask()
Current CPU hotplug code has some task->mm handling issues:
1. Working with task->mm w/o getting mm or grabing the task lock is
dangerous as ->mm might disappear (exit_mm() assigns NULL under
task_lock(), so tasklist lock is not enough).
We can't use get_task_mm()/mmput() pair as mmput() might sleep,
so we must take the task lock while handle its mm.
2. Checking for process->mm is not enough because process' main
thread may exit or detach its mm via use_mm(), but other threads
may still have a valid mm.
To fix this we would need to use find_lock_task_mm(), which would
walk up all threads and returns an appropriate task (with task
lock held).
clear_tasks_mm_cpumask() has all the issues fixed, so let's use it.
Suggested-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Anton Vorontsov [Thu, 3 May 2012 05:44:44 +0000 (15:44 +1000)]
arm: use clear_tasks_mm_cpumask()
Checking for process->mm is not enough because process' main thread may
exit or detach its mm via use_mm(), but other threads may still have a
valid mm.
To fix this we would need to use find_lock_task_mm(), which would walk up
all threads and returns an appropriate task (with task lock held).
clear_tasks_mm_cpumask() has this issue fixed, so let's use it.
Suggested-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Cc: Russell King <rmk@arm.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Anton Vorontsov [Thu, 3 May 2012 05:44:44 +0000 (15:44 +1000)]
cpu: introduce clear_tasks_mm_cpumask() helper
Many architectures clear tasks' mm_cpumask like this:
read_lock(&tasklist_lock);
for_each_process(p) {
if (p->mm)
cpumask_clear_cpu(cpu, mm_cpumask(p->mm));
}
read_unlock(&tasklist_lock);
Depending on the context, the code above may have several problems,
such as:
1. Working with task->mm w/o getting mm or grabing the task lock is
dangerous as ->mm might disappear (exit_mm() assigns NULL under
task_lock(), so tasklist lock is not enough).
2. Checking for process->mm is not enough because process' main
thread may exit or detach its mm via use_mm(), but other threads
may still have a valid mm.
This patch implements a small helper function that does things
correctly, i.e.:
1. We take the task's lock while whe handle its mm (we can't use
get_task_mm()/mmput() pair as mmput() might sleep);
2. To catch exited main thread case, we use find_lock_task_mm(),
which walks up all threads and returns an appropriate task
(with task lock held).
Also, Per Peter Zijlstra's idea, now we don't grab tasklist_lock in
the new helper, instead we take the rcu read lock. We can do this
because the function is called after the cpu is taken down and marked
offline, so no new tasks will get this cpu set in their mm mask.
Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Cc: Richard Weinberger <richard@nod.at> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Russell King <rmk@arm.linux.org.uk> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Mike Frysinger <vapier@gentoo.org> Cc: Paul Mundt <lethal@linux-sh.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
fork: call complete_vfork_done() after clearing child_tid and flushing rss-counters
Child should wake up the parent from vfork() only after finishing all
operations with shared mm. There is no sense in using
CLONE_CHILD_CLEARTID together with CLONE_VFORK, but it looks more accurate
now.
Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Markus Trippelsdorf <markus@trippelsdorf.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Add the line "HWPoinson: <size> kB" into /proc/pid/smaps if
CONFIG_MEMORY_FAILURE=y and some HWPoison pages were found. This may be
useful for searching applications which use a broken memory.
Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Andi Kleen <andi@firstfloor.org> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Currently, nonlinear mappings can not be distinguished from ordinary
mappings. This patch adds into /proc/pid/smaps line "Nonlinear: <size>
kB", where size is amount of nonlinear ptes in vma, this line appears only
if VM_NONLINEAR is set. This information may be useful not only for
checkpoint/restore project.
Requested by Pavel Emelyanov.
Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Andi Kleen <andi@firstfloor.org> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
This is an implementation of Andrew's proposal to extend the pagemap file
bits to report what is missing about tasks' working set.
The problem with the working set detection is multilateral. In the criu
(checkpoint/restore) project we dump the tasks' memory into image files
and to do it properly we need to detect which pages inside mappings are
really in use. The mincore syscall I though could help with this did not.
First, it doesn't report swapped pages, thus we cannot find out which
parts of anonymous mappings to dump. Next, it does report pages from page
cache as present even if they are not mapped, and it doesn't make that has
not been cow-ed.
Note, that issue with swap pages is critical -- we must dump swap pages to
image file. But the issues with file pages are optimization -- we can
take all file pages to image, this would be correct, but if we know that a
page is not mapped or not cow-ed, we can remove them from dump file. The
dump would still be self-consistent, though significantly smaller in size
(up to 10 times smaller on real apps).
Andrew noticed, that the proc pagemap file solved 2 of 3 above issues --
it reports whether a page is present or swapped and it doesn't report not
mapped page cache pages. But, it doesn't distinguish cow-ed file pages
from not cow-ed.
I would like to make the last unused bit in this file to report whether the
page mapped into respective pte is PageAnon or not.
[comment stolen from Pavel Emelyanov's v1 patch]
Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: Matt Mackall <mpm@selenic.com> Cc: Hugh Dickins <hughd@google.com> Cc: Rik van Riel <riel@redhat.com> Acked-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tim Bird [Thu, 3 May 2012 05:44:38 +0000 (15:44 +1000)]
stack usage: add pid to warning printk in check_stack_usage
In embedded systems, sometimes the same program (busybox) is the cause of
multiple warnings. Outputting the pid with the program name in the
warning printk helps distinguish which instances of a program are using
the stack most.
This is a small patch, but useful.
Signed-off-by: Tim Bird <tim.bird@am.sony.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Oleg Nesterov [Thu, 3 May 2012 05:44:38 +0000 (15:44 +1000)]
cred: remove task_is_dead() from __task_cred() validation
commit 8f92054e ("CRED: Fix __task_cred()'s lockdep check and banner
comment"):
add the following validation condition:
task->exit_state >= 0
to permit the access if the target task is dead and therefore
unable to change its own credentials.
OK, but afaics currently this can only help wait_task_zombie() which calls
__task_cred() without rcu lock.
Remove this validation and change wait_task_zombie() to use task_uid()
instead. This means we do rcu_read_lock() only to shut up the lockdep,
but we already do the same in, say, wait_task_stopped().
task_is_dead() should die, task->exit_state != 0 means that this task has
passed exit_notify(), only do_wait-like code paths should use this.
Unfortunately, we can't kill task_is_dead() right now, it has already
acquired buggy users in drivers/staging. The fix already exists.
Signed-off-by: Oleg Nesterov <oleg@redhat.com> Reviewed-by: "Eric W. Biederman" <ebiederm@xmission.com> Acked-by: David Howells <dhowells@redhat.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: James Morris <jmorris@namei.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tetsuo Handa [Thu, 3 May 2012 05:44:37 +0000 (15:44 +1000)]
kmod: avoid deadlock from recursive kmod call
The system deadlocks (at least since 2.6.10) when
call_usermodehelper(UMH_WAIT_EXEC) request triggered
call_usermodehelper(UMH_WAIT_PROC) request.
This is because "khelper thread is waiting for the worker thread at
wait_for_completion() in do_fork() since the worker thread was created
with CLONE_VFORK flag" and "the worker thread cannot call complete()
because do_execve() is blocked at UMH_WAIT_PROC request" and "the khelper
thread cannot start processing UMH_WAIT_PROC request because the khelper
thread is waiting for the worker thread at wait_for_completion() in
do_fork()".
In order to avoid deadlock, do not try to call wait_for_completion() in
call_usermodehelper_exec() if the worker thread was created by khelper
thread with CLONE_VFORK flag.
The easiest example to observe this deadlock is to use a corrupted
/sbin/hotplug binary (like shown below).
call_usermodehelper("/tmp/dummy", UMH_WAIT_EXEC) is called from
kobject_uevent_env() in lib/kobject_uevent.c upon loading/unloading a
module. do_execve("/tmp/dummy") triggers a call to
request_module("binfmt-0000") from search_binary_handler() which in turn
calls call_usermodehelper(UMH_WAIT_PROC).
There are various hooks called during do_execve() operation (e.g.
security_bprm_check(), audit_bprm(), "struct
linux_binfmt"->load_binary()). If one of such hooks triggers
UMH_WAIT_EXEC, this deadlock will happen even if /sbin/hotplug is not
corrupted.
[akpm@linux-foundation.org: add comment to kmod_thread_locker] Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Boaz Harrosh [Thu, 3 May 2012 05:44:37 +0000 (15:44 +1000)]
kmod: move call_usermodehelper_fns() to .c file and unexport all it's helpers
If we move call_usermodehelper_fns() to kmod.c file and EXPORT_SYMBOL it
we can avoid exporting all it's helper functions:
call_usermodehelper_setup
call_usermodehelper_setfns
call_usermodehelper_exec
And make all of them static to kmod.c
Since the optimizer will see all these as a single call site it will
inline them inside call_usermodehelper_fns(). So we loose the call to
_fns but gain 3 calls to the helpers. (Not that it matters)
Boaz Harrosh [Thu, 3 May 2012 05:44:36 +0000 (15:44 +1000)]
kmod: convert two call sites to call_usermodehelper_fns()
Both kernel/sys.c && security/keys/request_key.c where inlining the exact
same code as call_usermodehelper_fns(); So simply convert these sites to
directly use call_usermodehelper_fns().
Nikolaus Voss [Thu, 3 May 2012 05:44:35 +0000 (15:44 +1000)]
drivers/rtc/rtc-m41t93.c: don't let get_time() reset M41T93_FLAG_OF
If the rtc reports the time might be invalid due to oscillator failure,
M41T93_FLAG_OF flag must not be reset by get_time() as the read operation
doesn't make the time valid.
Without this patch, only the first get_time() reported an invalid time,
the second get_time() reported a valid time althought the reported time is
probably wrong due to oscillator failure.
Instead of resetting in get_time(), with this patch M41T93_FLAG_OF is
reset in set_time() when a valid time is to be written.
Signed-off-by: Nikolaus Voss <n.voss@weinmann.de> Cc: Alessandro Zummo <a.zummo@towertech.it> Cc: Grant Likely <grant.likely@secretlab.ca> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Wolfram Sang [Thu, 3 May 2012 05:44:35 +0000 (15:44 +1000)]
rtc: ds1307: add trickle charger support
Some DS13XX devices have "trickle chargers". Its configuration register
is at different locations, the setup is the same, though. Since the
configuration is board specific, introduce a platform_data to this driver.
Tested with a DS1339 on a custom board.
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de> Cc: Alessandro Zummo <alessandro.zummo@towertech.it> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Alexander Stein [Thu, 3 May 2012 05:44:32 +0000 (15:44 +1000)]
rtc: add ioctl to get/clear battery low voltage status
Currently there is no generic way to get the RTC battery status within an
application. So add an ioctl to read the status bit. The idea is that
the bit is set once a low voltage is detected. It stays there until it is
reset using the RTC_VL_CLR ioctl.
Signed-off-by: Alexander Stein <alexander.stein@systec-electronic.com> Cc: Alessandro Zummo <a.zummo@towertech.it> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Viresh Kumar [Thu, 3 May 2012 05:44:32 +0000 (15:44 +1000)]
rtc/spear: add Device Tree probing capability
SPEAr platforms now support DT and so must convert all drivers support DT.
This patch adds DT probing support for rtc and updates its documentation
too.
Signed-off-by: Viresh Kumar <viresh.kumar@st.com> Cc: Stefan Roese <sr@denx.de> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Alessandro Zummo <a.zummo@towertech.it> Cc: Rajeev Kumar <rajeev-dlh.kumar@st.com> Cc: Rob Herring <robherring2@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
init: disable sparse checking of the mount.o source files
The init/mount.o source files produce a number of sparse warnings of the
type:
warning: incorrect type in argument 1 (different address spaces)
expected char [noderef] <asn:1>*dev_name
got char *name
This is due to the syscalls expecting some of the arguments to be user
pointers but they are being passed as kernel pointers. This is harmless
but adds a lot of noise to a sparse build.
To limit the noise just disable the sparse checking in the relevant source
files, but still display a warning so that the user knows this has been
done.
Since the sparse checking has been disabled we can also remove the __user
__force casts that are scattered thru the source.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Stephen Boyd [Thu, 3 May 2012 05:44:30 +0000 (15:44 +1000)]
spinlock_debug: print kallsyms name for lock
When a spinlock warning is printed we usually get
BUG: spinlock bad magic on CPU#0, modprobe/111
lock: 0xdff09f38, .magic: 00000000, .owner: /0, .owner_cpu: 0
but it's nicer to print the symbol for the lock if we have it so that we
can avoid 'grep dff09f38 /proc/kallsyms' to find out which lock it was.
Use kallsyms to print the symbol name so we get something a bit easier to
read
BUG: spinlock bad magic on CPU#0, modprobe/112
lock: test_lock, .magic: 00000000, .owner: <none>/-1, .owner_cpu: 0
If the lock is not in kallsyms %ps will fall back to printing the address
directly.
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org> Cc: Ingo Molnar <mingo@elte.hu> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Stephen Boyd [Thu, 3 May 2012 05:44:30 +0000 (15:44 +1000)]
vsprintf: fix %ps on non symbols when using kallsyms
Using %ps in a printk format will sometimes fail silently and print the
empty string if the address passed in does not match a symbol that
kallsyms knows about. But using %pS will fall back to printing the full
address if kallsyms can't find the symbol. Make %ps act the same as %pS
by falling back to printing the address.
While we're here also make %ps print the module that a symbol comes from
so that it matches what %pS already does. Take this simple function for
example (in a module):
Andrew Morton [Thu, 3 May 2012 05:44:29 +0000 (15:44 +1000)]
lib/bitmap.c: fix documentation for scnprintf() functions
The code comments for bscnl_emit() and bitmap_scnlistprintf() are
describing snprintf() return semantics, but these functions use
scnprintf() return semantics. Fix that, and document the
bitmap_scnprintf() return value as well.
Cc: Ryota Ozaki <ozaki.ryota@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
lib/test-kstrtox.c: mark const init data with __initconst instead of __initdata
As long as there is no other non-const variable marked __initdata in the
same compilation unit it doesn't hurt. If there were one however
compilation would fail with
error: $variablename causes a section type conflict
because a section containing const variables is marked read only and so
cannot contain non-const variables.
Chris Metcalf [Thu, 3 May 2012 05:44:28 +0000 (15:44 +1000)]
list_debug: WARN for adding something already in the list
We were bitten by this at one point and added an additional sanity test
for DEBUG_LIST. You can't validly add a list_head to a list where either
prev or next is the same as the thing you're adding.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com> Cc: Andi Kleen <andi@firstfloor.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Shuah Khan [Thu, 3 May 2012 05:44:28 +0000 (15:44 +1000)]
leds: add new transient trigger for one shot timer activation
The leds timer trigger does not currently have an interface to activate a
one shot timer. The current support allows for setting two timers, one
for specifying how long a state to be on, and the second for how long the
state to be off. The delay_on value specifies the time period an LED
should stay in on state, followed by a delay_off value that specifies how
long the LED should stay in off state. The on and off cycle repeats until
the trigger gets deactivated. There is no provision for one time
activation to implement features that require an on or off state to be
held just once and then stay in the original state forever.
Without one shot timer interface, user space can still use timer trigger
to set a timer to hold a state, however when user space application
crashes or goes away without deactivating the timer, the hardware will be
left in that state permanently.
As a specific example of this use-case, let's look at vibrate feature on
phones. Vibrate function on phones is implemented using PWM pins on SoC
or PMIC. There is a need to activate one shot timer to control the
vibrate feature, to prevent user space crashes leaving the phone in
vibrate mode permanently causing the battery to drain.
This trigger exports three properties, activate, state, and duration When
transient trigger is activated these properties are set to default values.
- duration allows setting timer value in msecs. The initial value is 0.
- activate allows activating and deactivating the timer specified by
duration as needed. The initial and default value is 0. This will allow
duration to be set after trigger activation.
- state allows user to specify a transient state to be held for the specified
duration.
Signed-off-by: Shuah Khan <shuahkhan@gmail.com> Cc: Jonas Bonn <jonas@southpole.se> Cc: Richard Purdie <rpurdie@rpsys.net> Cc: NeilBrown <neilb@suse.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Shuah Khan [Thu, 3 May 2012 05:44:26 +0000 (15:44 +1000)]
leds: change ledtrig-timer to use activated flag
Change existing timer trigger to use the new ->activated flag to set
activate successful status in activate routine and check it in deactivate
routine to do cleanup.
Signed-off-by: Shuah Khan <shuahkhan@gmail.com> Cc: Richard Purdie <rpurdie@rpsys.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Shuah Khan [Thu, 3 May 2012 05:44:25 +0000 (15:44 +1000)]
leds: change existing triggers to use activated flag
Change existing triggers backlight, gpio, and heartbeat to use the new
->activated flag to set activate successful status in their activate
routines and check it in their deactivate routines to do cleanup.
Signed-off-by: Shuah Khan <shuahkhan@gmail.com> Cc: Richard Purdie <rpurdie@rpsys.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Shuah Khan [Thu, 3 May 2012 05:44:25 +0000 (15:44 +1000)]
leds: add new field to led_classdev struct to save activation state
Add a new field to led_classdev to save activattion state after activate
routine is successful. This saved state is used in deactivate routine to
do cleanup such as removing device files, and free memory allocated during
activation. Currently trigger_data not being null is used for this
purpose.
Existing triggers will need changes to use this new field.
Signed-off-by: Shuah Khan <shuahkhan@gmail.com> Cc: Richard Purdie <rpurdie@rpsys.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Thomas Meyer [Thu, 3 May 2012 05:44:25 +0000 (15:44 +1000)]
leds: Use kcalloc instead of kzalloc to allocate array
The advantage of kcalloc is that will prevent integer overflows which
could result from the multiplication of number of elements and size and it
is also a bit nicer to read.
The semantic patch that makes this change is available
in https://lkml.org/lkml/2011/11/25/107
Signed-off-by: Thomas Meyer <thomas@m3y3r.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Axel Lin [Thu, 3 May 2012 05:44:24 +0000 (15:44 +1000)]
leds: lm3556: don't call kfree for the memory allocated by devm_kzalloc
The devm_* functions eliminate the need for manual resource releasing
and simplify error handling. Resources allocated by devm_* are freed
automatically on driver detach.
Thus adding kfree calls here will introduce double free bug.
Signed-off-by: Axel Lin <axel.lin@gmail.com> Cc: Geon Si Jeong <gshark.jeong@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Geon Si Jeong [Thu, 3 May 2012 05:44:23 +0000 (15:44 +1000)]
leds: add LED driver for lm3556 chip
A simple driver for the Texas Instruments LM3556 chip.
The LM3556 is a 4 MHz fixed-frequency synchronous boost converter plus
1.5A constant current driver for a high-current white LED. Datasheet:
www.national.com/ds/LM/LM3556.pdf
Tested on OMAP4430
Signed-off-by: Geon Si Jeong <gshark.jeong@gmail.com> Cc: Richard Purdie <rpurdie@rpsys.net> Cc: Daniel Jeong <daniel.jeong@ti.com> Cc: Mark Brown <broonie@opensource.wolfsonmicro.com> Cc: Greg KH <greg@kroah.com> Cc: Wolfram Sang <w.sang@pengutronix.de> Cc: Shuah Khan <shuahkhan@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Gcc 4.6.2 complains that:
drivers/leds/leds-lp5521.c: In function `lp5521_load_program':
drivers/leds/leds-lp5521.c:214:21: warning: `mode' may be used uninitialized in this function [-Wuninitialized]
drivers/leds/leds-lp5521.c: In function `lp5521_probe':
drivers/leds/leds-lp5521.c:788:5: warning: `buf' may be used uninitialized in this function [-Wuninitialized]
drivers/leds/leds-lp5521.c:740:6: warning: `ret' may be used uninitialized in this function [-Wuninitialized]
These are real problems if lp5521_read() returns an error. When that
happens we should handle it, instead of ignoring it or doing a bitwise OR
with all the other error codes and continuing.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Cc: Milo <Milo.Kim@ti.com> Cc: Richard Purdie <rpurdie@rpsys.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Inki Dae [Thu, 3 May 2012 05:44:22 +0000 (15:44 +1000)]
fbdev: add events for early fb event support
Add FB_EARLY_EVENT_BLANK and FB_R_EARLY_EVENT_BLANK event mode supports.
first, fb_notifier_call_chain() is called with FB_EARLY_EVENT_BLANK and
fb_blank() of specific fb driver is called and then
fb_notifier_call_chain() is called with FB_EVENT_BLANK again at
fb_blank(). and if fb_blank() was failed then fb_nitifier_call_chain()
would be called with FB_R_EARLY_EVENT_BLANK to revert the previous
effects.
Signed-off-by: Inki Dae <inki.dae@samsung.com> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com> Cc: Lars-Peter Clausen <lars@metafoo.de> Cc: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Richard Purdie <rpurdie@rpsys.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Inki Dae [Thu, 3 May 2012 05:44:21 +0000 (15:44 +1000)]
lcd: add callbacks for early fb event blank support
This patchset adds early fb blank feature that a callback of lcd panel
driver is called prior to specific fb driver's one. In the case of
MIPI-DSI based video mode LCD Panel, for lcd power off, the power off
commands should be transferred to lcd panel with display and mipi-dsi
controller enabled because the commands is set to lcd panel at vsync porch
period. and in opposite case, the callback of fb driver should be called
prior to lcd panel driver's one because of same issue. Also if fb_blank
mode is changed to FB_BLANK_POWERDOWN then display controller would be
off(clock disable) but lcd panel would be still on. at this time, you
could see some issue like sparkling on lcd panel because video clock to be
delivered to ldi module of lcd panel was disabled. this issue could
occurs for all lcd panels.
The callback order is as the following:
at fb_blank function of fbmem.c
-> fb_notifier_call_chain(FB_EARLY_EVENT_BLANK)
-> lcd panel driver's early_set_power()
-> info->fbops->fb_blank()
-> spcefic fb driver's fb_blank()
-> fb_notifier_call_chain(FB_EVENT_BLANK)
-> lcd panel driver's set_power()
-> fb_notifier_call_chain(FB_R_EARLY_EVENT_BLANK) if
info->fops->fb_blank() was failed.
fb_notifier_call_chain(FB_R_EARLY_EVENT_BLANK) would be called to revert
the effects of previous FB_EARLY_EVENT_BLANK call. and note that if
early_set_power() of lcd_ops is NULL then early fb blank callback would be
ignored.
This patch:
Add early_set_power and r_early_set_power callbacks. early_set_power
callback is called prior to fb_blank() of fbmem.c and r_early_set_power
callback is called if fb_blank() was failed to revert the effects of the
early_set_power call of lcd panel driver.
Signed-off-by: Inki Dae <inki.dae@samsung.com> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com> Cc: Lars-Peter Clausen <lars@metafoo.de> Cc: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> Cc: Richard Purdie <rpurdie@rpsys.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
blacklight: remove redundant spi driver bus initialization
In ancient times it was necessary to manually initialize the bus field of
an spi_driver to spi_bus_type. These days this is done in
spi_driver_register() so we can drop the manual assignment.
The patch was generated using the following coccinelle semantic patch:
// <smpl>
@@
identifier _driver;
@@
struct spi_driver _driver = {
.driver = {
- .bus = &spi_bus_type,
},
};
// </smpl>
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de> Cc: Richard Purdie <rpurdie@rpsys.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Denys Vlasenko [Thu, 3 May 2012 05:44:19 +0000 (15:44 +1000)]
vsprintf-further-optimize-decimal-conversion-v2
Here is a new version. I also plugged a hole in num_to_str() -
it was assuming it's safe to call put_dec() for num=0.
(We never tripped over it before because the single caller
of num_to_str() takes care of that case).
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com> Cc: Michal Nazarewicz <mina86@mina86.com> Cc: Douglas W Jones <jones@cs.uiowa.edu> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Denys Vlasenko [Thu, 3 May 2012 05:44:19 +0000 (15:44 +1000)]
vsprintf: further optimize decimal conversion
Previous code was using optimizations which were developed to work well
even on narrow-word CPUs (by today's standards). But Linux runs only on
32-bit and wider CPUs. We can use that.
First: using 32x32->64 multiply and trivial 32-bit shift, we can correctly
divide by 10 much larger numbers, and thus we can print groups of 9 digits
instead of groups of 5 digits.
Next: there are two algorithms to print larger numbers. One is generic:
divide by 1000000000 and repeatedly print groups of (up to) 9 digits.
It's conceptually simple, but requires an (unsigned long long) / 1000000000 division.
Second algorithm splits 64-bit unsigned long long into 16-bit chunks,
manipulates them cleverly and generates groups of 4 decimal digits. It so
happens that it does NOT require long long division.
If long is > 32 bits, division of 64-bit values is relatively easy, and we
will use the first algorithm. If long long is > 64 bits (strange
architecture with VERY large long long), second algorithm can't be used,
and we again use the first one.
Else (if long is 32 bits and long long is 64 bits) we use second one.
And third: there is a simple optimization which takes fast path not only
for zero as was done before, but for all one-digit numbers.
In all tested cases new code is faster than old one, in many cases by 30%,
in few cases by more than 50% (for example, on x86-32, conversion of 12345678). Code growth is ~0 in 32-bit case and ~130 bytes in 64-bit
case.
This patch is based upon an original from Michal Nazarewicz.
Signed-off-by: Michal Nazarewicz <mina86@mina86.com> Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com> Cc: Douglas W Jones <jones@cs.uiowa.edu> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Oleg Nesterov [Thu, 3 May 2012 05:44:17 +0000 (15:44 +1000)]
keys: kill the dummy key_replace_session_keyring()
After the previouse change key_replace_session_keyring() becomes a nop.
Remove the dummy definition in key.h and update the callers in
arch/*/kernel/signal.c.
Signed-off-by: Oleg Nesterov <oleg@redhat.com> Acked-by: David Howells <dhowells@redhat.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Richard Kuo <rkuo@codeaurora.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Alexander Gordeev <agordeev@redhat.com> Cc: Chris Zankel <chris@zankel.net> Cc: David Smith <dsmith@redhat.com> Cc: "Frank Ch. Eigler" <fche@redhat.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Larry Woodman <lwoodman@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Tejun Heo <tj@kernel.org> Cc: Ingo Molnar <mingo@elte.hu> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Oleg Nesterov [Thu, 3 May 2012 05:44:16 +0000 (15:44 +1000)]
keys: change keyctl_session_to_parent() to use task_work_add()
Change keyctl_session_to_parent() to use task_work_add() and move
key_replace_session_keyring() logic into task_work->func().
Note that we do task_work_cancel() before task_work_add() to ensure that
only one work can be pending at any time. This is important, we must not
allow user-space to abuse the parent's ->task_works list.
The callback, replace_session_keyring(), checks PF_EXITING. I guess this
is not really needed but looks better.
As a side effect, this fixes the (unlikely) race. The callers of
key_replace_session_keyring() and keyctl_session_to_parent() lack the
necessary barriers, the parent can miss the request.
Now we can remove task_struct->replacement_session_keyring and related
code.
Signed-off-by: Oleg Nesterov <oleg@redhat.com> Acked-by: David Howells <dhowells@redhat.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Richard Kuo <rkuo@codeaurora.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Alexander Gordeev <agordeev@redhat.com> Cc: Chris Zankel <chris@zankel.net> Cc: David Smith <dsmith@redhat.com> Cc: "Frank Ch. Eigler" <fche@redhat.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Larry Woodman <lwoodman@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Tejun Heo <tj@kernel.org> Cc: Ingo Molnar <mingo@elte.hu> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Oleg Nesterov [Thu, 3 May 2012 05:44:15 +0000 (15:44 +1000)]
task_work_add: generic process-context callbacks
Provide a simple mechanism that allows running code in the (nonatomic)
context of the arbitrary task.
The caller does task_work_add(task, task_work) and this task executes
task_work->func() either from do_notify_resume() or from do_exit(). The
callback can rely on PF_EXITING to detect the latter case.
"struct task_work" can be embedded in another struct, still it has "void
*data" to handle the most common/simple case.
This allows us to kill the ->replacement_session_keyring hack, and
potentially this can have more users.
Performance-wise, this adds 2 "unlikely(!hlist_empty())" checks into
tracehook_notify_resume() and do_exit(). But at the same time we can
remove the "replacement_session_keyring != NULL" checks from
arch/*/signal.c and exit_creds().
Note: task_work_add/task_work_run abuses ->pi_lock. This is only because
this lock is already used by lookup_pi_state() to synchronize with
do_exit() setting PF_EXITING. Fortunately the scope of this lock in
task_work.c is really tiny, and the code is unlikely anyway.
Signed-off-by: Oleg Nesterov <oleg@redhat.com> Acked-by: David Howells <dhowells@redhat.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Richard Kuo <rkuo@codeaurora.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Alexander Gordeev <agordeev@redhat.com> Cc: Chris Zankel <chris@zankel.net> Cc: David Smith <dsmith@redhat.com> Cc: "Frank Ch. Eigler" <fche@redhat.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Larry Woodman <lwoodman@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Tejun Heo <tj@kernel.org> Cc: Ingo Molnar <mingo@elte.hu> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
projects / karo-tx-linux.git / log