do a proper idle-wakeup event on HLT as well - some CPUs stop the TSC
in HLT too, not just when going through the ACPI methods.
(the ACPI idle code already does this.)
[ update the 64-bit side too, as noticed by Jiri Slaby. ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
scale the sched_clock() cyc_2_nsec scaling factor according to
CPU frequency changes.
[ mingo@elte.hu: simplified it and fixed it for SMP. ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
cf http://lkml.org/lkml/2007/10/3/41
To summarize: on Linux, SA_ONSTACK decides whether you are already on the
signal stack based on the value of the SP at the time of a signal. If
you are not already inside the range, you are not "on the signal stack"
and so the new signal handler frame starts over at the base of the signal
stack.
sigaltstack (and sigstack before it) was invented in BSD. There, the
SA_ONSTACK behavior has always been different. It uses a kernel state
flag to decide, rather than the SP value. When you first take an
SA_ONSTACK signal and switch to the alternate signal stack, it sets the
SS_ONSTACK flag in the thread's sigaltstack state in the kernel.
Thereafter you are "on the signal stack" and don't switch SP before
pushing a handler frame no matter what the SP value is. Only when you
sigreturn from the original handler context do you clear the SS_ONSTACK
flag so that a new handler frame will start over at the base of the
alternate signal stack.
The undesireable effect of the Linux behavior is that an overflow of the
alternate signal stack can not only go undetected, but lead to a ring
buffer effect of clobbering the original handler frame at the base of the
signal stack for each successive signal that comes just after the
overflow. This is what Shi Weihua's test case demonstrates. Normally
this does not come up because of the signal mask, but the test case uses
SA_NODEFER for its SIGSEGV handler.
The other subtle part of the existing Linux semantics is that a simple
longjmp out of a signal handler serves to take you off the signal stack
in a safe and reliable fashion without having used sigreturn (nor having
just returned from the handler normally, which means the same). After
the longjmp (or even informal stack switching not via any proper libc or
kernel interface), the alternate signal stack stands ready to be used
again.
A paranoid program would allocate a PROT_NONE red zone around its
alternate signal stack. Then a small overflow would trigger a SIGSEGV in
handler setup, and be fatal (core dump) whether or not SIGSEGV is
blocked. As with thread stack red zones, that cannot catch all overflows
(or underflows). e.g., a local array as large as page size allocated in
a function called from a handler, but not actually touched before more
calls push more stack, could cause an overflow that silently pushes into
some unrelated allocated pages.
The BSD behavior does not do anything in particular about overflow. But
it does at least avoid the wraparound or "ring buffer effect", so you'll
just get a straightforward all-out overflow down your address space past
the low end of the alternate signal stack. I don't know what the BSD
behavior is for longjmp out of an SA_ONSTACK handler.
The POSIX wording relating to sigaltstack is pretty minimal. I don't
think it speaks to this issue one way or another. (The program that
overflows its stack is clearly in undefined behavior territory of one
sort or another anyhow.)
Given the longjmp issue and the potential for highly subtle complications
in existing programs relying on this in arcane ways deep in their code, I
am very dubious about changing the behavior to the BSD style persistent
flag. I think Shi Weihua's patches have a similar effect by tracking the
SP used in the last handler setup.
I think it would be sensible for the signal handler setup code to detect
when it would itself be causing a stack overflow. Maybe something like
the following patch (untested). This issue exists in the same way on all
machines, so ideally they would all do a similar check.
When it's the handler function itself or its callees that cause the
overflow, rather than the signal handler frame setup alone crossing the
boundary, this still won't help. But I don't see any way to distinguish
that from the valid longjmp case.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
add the DMI strings provided by Islam Amer <pharon@gmail.com>, for
the Compaq Presario V6000 (Quanta/30B7).
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
make io_delay=0xed the default. This frees up port 0x80 which is
a debug port on some machines and locks up certain laptops.
Testing only for now. Try the io_delay=0x80 boot option if this does not
work for you.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
various changes to the in_p/out_p delay details:
- add the io_delay=none method
- make each method selectable from the kernel config
- simplify the delay code a bit by getting rid of an indirect function call
- add the /proc/sys/kernel/io_delay_type sysctl
- change 'io_delay=standard|alternate' to io_delay=0x80 and io_delay=0xed
- make the io delay config not depend on CONFIG_DEBUG_KERNEL
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: "David P. Reed" <dpreed@reed.com>
x86: provide a DMI based port 0x80 I/O delay override.
Certain (HP) laptops experience trouble from our port 0x80 I/O delay
writes. This patch provides for a DMI based switch to the "alternate
diagnostic port" 0xed (as used by some BIOSes as well) for these.
David P. Reed confirmed that port 0xed works for him and provides a
proper delay. The symptoms of _not_ working are a hanging machine,
with "hwclock" use being a direct trigger.
Earlier versions of this attempted to simply use udelay(2), with the
2 being a value tested to be a nicely conservative upper-bound with
help from many on the linux-kernel mailinglist but that approach has
two problems.
First, pre-loops_per_jiffy calibration (which is post PIT init while
some implementations of the PIT are actually one of the historically
problematic devices that need the delay) udelay() isn't particularly
well-defined. We could initialise loops_per_jiffy conservatively (and
based on CPU family so as to not unduly delay old machines) which
would sort of work, but...
Second, delaying isn't the only effect that a write to port 0x80 has.
It's also a PCI posting barrier which some devices may be explicitly
or implicitly relying on. Alan Cox did a survey and found evidence
that additionally some drivers may be racy on SMP without the bus
locking outb.
Switching to an inb() makes the timing too unpredictable and as such,
this DMI based switch should be the safest approach for now. Any more
invasive changes should get more rigid testing first. It's moreover
only very few machines with the problem and a DMI based hack seems
to fit that situation.
This also introduces a command-line parameter "io_delay" to override
the DMI based choice again:
io_delay=<standard|alternate>
where "standard" means using the standard port 0x80 and "alternate"
port 0xed.
This retains the udelay method as a config (CONFIG_UDELAY_IO_DELAY) and
command-line ("io_delay=udelay") choice for testing purposes as well.
This does not change the io_delay() in the boot code which is using
the same port 0x80 I/O delay but those do not appear to be a problem
as David P. Reed reported the problem was already gone after using the
udelay version. He moreover reported that booting with "acpi=off" also
fixed things and seeing as how ACPI isn't touched until after this DMI
based I/O port switch I believe it's safe to leave the ones in the boot
code be.
The DMI strings from David's HP Pavilion dv9000z are in there already
and we need to get/verify the DMI info from other machines with the
problem, notably the HP Pavilion dv6000z.
This patch is partly based on earlier patches from Pavel Machek and
David P. Reed.
Signed-off-by: Rene Herman <rene.herman@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Document the fact that __save_processor_state() has to save all CPU
registers referred to by the kernel in case a different kernel is
used to load and restore a hibernation image containing it.
Sigend-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Michael Opdenacker reported:
For backward compatibility with earlier (< 2.6.24) kernels,
arch/i386/boot/bzImage or arch/x86_64/boot/bzImage symbolic links to
arch/x86/boot/bzImage are created when you build an x86 kernel. The
arch/i386 or arch/x86_64 directories are then created for this only
purpose.
Issue: these generated directories and symbolic links are *not cleaned
up* when you run "make mrproper" (and thus "make distclean"). This
disturbs the production of patches, because the source tree is left with
generated files and directories.
Sam has an alternative fix:
The directory is killed during make clean as opposed to make mrproper.
Reported-by: Michael Opdenacker <michael-lists@free-electrons.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Current idle time in kstat is based on jiffies and is coarse grained.
tick_sched.idle_sleeptime is making some attempt to keep track of idle time
in a fine grained manner. But, it is not handling the time spent in
interrupts fully.
Make tick_sched.idle_sleeptime accurate with respect to time spent on
handling interrupts and also add tick_sched.idle_lastupdate, which keeps
track of last time when idle_sleeptime was updated.
This statistics will be crucial for cpufreq-ondemand governor, which can
shed some conservative gaurd band that is uses today while setting the
frequency. The ondemand changes that uses the exact idle time is coming
soon.
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
I recently noticed on one of my boxes that when synched with an NTP
server, the drift value reported for the system was ~283ppm. While in
some cases, clock hardware can be that bad, it struck me as unusual as
the system was using the acpi_pm clocksource, which is one of the more
trustworthy and accurate clocksources on x86 hardware.
I brought up another system and let it sync to the same NTP server, and
I noticed a similar 280some ppm drift.
In looking at the code, I found that the acpi_pm's constant frequency
was being computed correctly at boot-up, however once the system was up,
even without the ntp daemon running, the clocksource's frequency was
being modified by the clocksource_adjust() function.
Digging deeper, I realized that in the code that keeps track of how much
the clocksource is skewing from the ntp desired time, we were using
different lengths to establish how long an time interval was.
The clocksource was being setup with the following interval:
NTP_INTERVAL_LENGTH = NSEC_PER_SEC/NTP_INTERVAL_FREQ
While the ntp code was using the tick_length_base value:
tick_length_base ~= (tick_usec * NSEC_PER_USEC * USER_HZ)
/NTP_INTERVAL_FREQ
The subtle difference is:
(tick_usec * NSEC_PER_USEC * USER_HZ) != NSEC_PER_SEC
This difference in calculation was causing the clocksource correction
code to apply a correction factor to the clocksource so the two
intervals were the same, however this results in the actual frequency of
the clocksource to be made incorrect. I believe this difference would
affect all clocksources, although to differing degrees depending on the
clocksource resolution.
The issue was introduced when my HZ free ntp patch landed in 2.6.21-rc1,
so my apologies for the mistake, and for not noticing it until now.
The following patch, corrects the clocksource's initialization code so
it uses the same interval length as the code in ntp.c. After applying
this patch, the drift value for the same system went from ~283ppm to
only 2.635ppm.
I believe this patch to be good, however it does affect all arches and
I've only tested on x86, so some caution is advised. I do think it would
be a likely candidate for a stable 2.6.24.x release.
Any thoughts or feedback would be appreciated.
Signed-off-by: John Stultz <johnstul@us.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Looks like IRQ 31 is assigned to timer 3, even without the patch!
I wonder who wrote the number 31. But the manual says that it is
zero by default.
I think we should check whether the timer has been allocated an IRQ before
proceeding to assign one to it. Here is a patch that does this.
Signed-off-by: Balaji Rao <balajirrao@gmail.com>
Tested-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
The userspace API for the HPET (see Documentation/hpet.txt) did not work. The
HPET_IE_ON ioctl was failing as there was no IRQ assigned to the timer
device. This patch fixes it by allocating IRQs to timer blocks in the HPET.
arch/x86/kernel/hpet.c | 13 +++++--------
drivers/char/hpet.c | 45 ++++++++++++++++++++++++++++++++++++++-------
include/linux/hpet.h | 2 +-
3 files changed, 44 insertions(+), 16 deletions(-)
Signed-off-by: Balaji Rao <balajirrao@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
detect zero event-device multiplicators - they then cause
division-by-zero crashes if a clockevent has been initialized
incorrectly.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
The following scenario might leave PIT as a disfunctional clock source:
PIT is registered as clocksource
PM_TIMER is registered as clocksource and enables highres/dyntick mode
PIT is switched to oneshot mode
-> now the readout of PIT is bogus, but the user might select PIT
via the sysfs override, which would break the box as the time
readout is unusable.
Unregister the PIT clocksource when the PIT clock event device is switched
into shutdown / oneshot mode.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
On x86 the PIT might become an unusable clocksource. Add an unregister
function to provide a possibilty to remove the PIT from the list of
available clock sources.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
PIT clocksource is registered unconditionally even when HPET is enabled
or when PIT is replaced by the local APIC timer. In both cases PIT can
not be used as it is stopped and the readout would be stale.
Prevent registering PIT in those cases.
patch depends on:
x86: offer is_hpet_enabled() on !CONFIG_HPET_TIMER too
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
This way it checks if the clocks are synchronized between CPUs too.
This might be able to detect slowly drifting TSCs which only
go wrong over longer time.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
clocksource_watchdog can use a deferrable timer - reduces wakeups from
idle per second.
Signed-off-by: Parag Warudkar <parag.warudkar@gmail.com>
Cc: john stultz <johnstul@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
- getnstimeofday() was just a wrapper around __get_realtime_clock_ts()
- Replace calls to __get_realtime_clock_ts() by calls to getnstimeofday()
- Fix bogus reference to get_realtime_clock_ts(), which never existed
Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com>
Cc: john stultz <johnstul@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
I was confused by FSEC = 10^15 NSEC statement, plus small whitespace
fixes. When there's copyright, there should be GPL.
Signed-off-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Small cleanups to tick-related code. Wrong preempt count is followed
by BUG(), so it is hardly KERN_WARNING.
Signed-off-by: Pavel Machek <pavel@suse.cz>
Cc: john stultz <johnstul@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Clean up hungarian notation from timer code.
Signed-off-by: Pavel Machek <pavel@suse.cz>
Cc: john stultz <johnstul@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
the logic in this function is just crazy. It's recursive, but we
can circumvent the creation for the kobject and whole creation of the
threshold_block if some conditions are met. That's why we see the
allocate_threshold_blocks so many times in the callstack, yet only a few
kobjects created.
Then we blow up in kobject_uevent_env() on the first debug printk.
Which means that we are just passing in garbage.
Man, this is one time that comments in code would have been very nice to
have, and why forward goto's into major code blocks are just evil...
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
* git://git.linux-nfs.org/pub/linux/nfs-2.6: (118 commits)
NFSv4: Iterate through all nfs_clients when the server recalls a delegation
NFSv4: Deal more correctly with duplicate delegations
NFS: Fix a potential race between umount and nfs_access_cache_shrinker()
NFS: Add an asynchronous delegreturn operation for use in nfs_clear_inode
nfs: convert NFS_*(inode) helpers to static inline
nfs: obliterate NFS_FLAGS macro
NFS: Address memory leaks in the NFS client mount option parser
nfs4: allow nfsv4 acls on non-regular-files
NFS: Optimise away the sigmask code in aio/dio reads and writes
SUNRPC: Don't bother changing the sigmask for asynchronous RPC calls
SUNRPC: rpcb_getport_sync() passes incorrect address size to rpc_create()
SUNRPC: Clean up block comment preceding rpcb_getport_sync()
SUNRPC: Use appropriate argument types in rpcb client
SUNRPC: rpcb_getport_sync() should use built-in hostname generator
SUNRPC: Clean up functions that free address_strings array
NFS: NFS version number is unsigned
NLM: Fix a bogus 'return' in nlmclnt_rpc_release
NLM: Introduce an arguments structure for nlmclnt_init()
NLM/NFS: Use cached nlm_host when calling nlmclnt_proc()
NFS: Invoke nlmclnt_init during NFS mount processing
...
If the two requests belong to the same io context, we will attempt
to lock the same lock twice. But swapping contexts is pointless in
that case, so just check for rioc == nioc before doing the double
lock and copy.
Tested-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
The same delegation may have been handed out to more than one nfs_client.
Ensure that if a recall occurs, we return all instances.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
If a (broken?) server hands out two different delegations for the same
file, then we should return one of them.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Otherwise, there is a potential deadlock if the last dput() from an NFSv4
close() or other asynchronous operation leads to nfs_clear_inode calling
the synchronous delegreturn.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
David Howells noticed that repeating the same mount option twice during an
NFS mount request can result in orphaned memory in certain cases.
Only the client_address and mount_server.hostname strings are initialized
in the mount parsing loop, so those appear to be the only two pointers that
might be written over by repeating a mount option. The strings in the
nfs_server section of the nfs_parsed_mount_data structure are set only once
after the options are parsed, thus these are not susceptible to being
overwritten.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
The rfc doesn't give any reason it shouldn't be possible to set an
attribute on a non-regular file. And if the server supports it, then it
shouldn't be up to us to prevent it.
Thanks to Erez for the report and Trond for further analysis.
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Tested-by: Erez Zadok <ezk@cs.sunysb.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
There are no interruptible waits for asynchronous RPC tasks, so we don't
need to wrap calls to rpc_run_task() with an
rpc_clnt_sigmask/rpc_clnt_unsigmask pair.
Instead we can wrap the wait_for_completion_interruptible() in
nfs_direct_wait(). This means that we completely optimise away sigmask
setting for the case of non-blocking aio/dio.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
The variable "sin" is a pointer, so sizeof(sin) is the size of a pointer,
not the size of thing that sin points to.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Clean up: Follow recommendations of Chapter 5 of Documentation/CodingStyle
and use "u32" instead of "__u32" for types in definitions that are not
shared with user space.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
rpc_create() can already fill in the hostname with a string representation
of the server's IP address, so remove redundant logic in in
rpcb_getport_sync() that does that.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Clean up: document the rule (kfree) and the exceptions
(RPC_DISPLAY_PROTO and RPC_DISPLAY_NETID) when freeing the objects in
a transport's address_strings array.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
RPC protocol version numbers are unsigned.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Clean up: pass 5 arguments to nlmclnt_init() in a structure similar to the
new nfs_client_initdata structure.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Now that each NFS mount point caches its own nlm_host structure, it can be
passed to nlmclnt_proc() for each lock request. By pinning an nlm_host for
each mount point, we trade the overhead of looking up or creating a fresh
nlm_host struct during every NLM procedure call for a little extra memory.
We also restrict the nlmclnt_proc symbol to limit the use of this call to
in-tree modules.
Note that nlm_lookup_host() (just removed from the client's per-request
NLM processing) could also trigger an nlm_host garbage collection. Now
client-side nlm_host garbage collection occurs only during NFS mount
processing. Since the NFS client now holds a reference on these nlm_host
structures, they wouldn't have been affected by garbage collection
anyway.
Given that nlm_lookup_host() reorders the global nlm_host chain after
every successful lookup, and that a garbage collection could be triggered
during the call, we've removed a significant amount of per-NLM-request
CPU processing overhead.
Sidebar: there are only a few remaining references to the internals of
NFS inodes in the client-side NLM code. The only references I found are
related to extracting or comparing the inode's file handle via NFS_FH().
One is in nlmclnt_grant(); the other is in nlmclnt_setlockargs().
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
