I originally thought this was an bug only in the SN code, but I think I
also see a hole in the generic IA64 tlb code. (Separate patch was sent
for the SN problem).
It looks like there is a bug in the TLB flushing code. During context switch,
kernel threads (kswapd, for example) inherit the mm of the task that was
previously running on the cpu. Normally, this is ok because the previous context
is still loaded into the RR registers. However, if the owner of the mm
migrates to another cpu, changes it's context number, and references a
page before kswapd issues a tlb_purge for that same page, the purge will be
done with a stale context number (& RR registers).
Signed-off-by: Tony Luck <tony.luck@intel.com>
Altix (shub2) pushes the BTE clean-up into SAL.
This patch correctly interfaces with the now implemented SAL call.
It also fixes a bug when delaying clean-up to allow busy BTEs to
complete (or error out).
Signed-off-by: Russ Anderson <rja@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
On return from INIT handler we must convert the address of the
minstate area from a kernel virtual uncached address (0xC...)
to physical uncached (0x8...). A typo (or thinko?) in the code
converted to physical cached.
Signed-off-by: Tony Luck <tony.luck@intel.com>
Cleanup a few items after moving xpc.h from arch/ia64/sn/kernel to
include/asm-ia64/sn.
Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Move xpc.h from arch/ia64/sn/kernel to include/asm-ia64/sn without change.
Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Move xpc_system_reboot() to be closer to the file it calls for readability
reasons (which are indeed subjective).
Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Allow for the loss of heartbeat while in kdebug to be ignored by remote
partitions.
Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Cleanup the XPC disengage related messages that are printed to the log.
Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
This patch fixes a problem in XPC disengage processing whereby it was not
seeing the request to disengage from a remote partition, so the disengage
wasn't happening. The disengagement is suppose to transpire during the time
a XPC channel is disconnecting, and should be completed before the channel
is declared to be disconnected.
Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
When this new syscall was added to ia64 in commit
39743889aa
fsys.S was forgotten. Add a ".data8 0" there to keep
it in step. [Reported by Stephane Eranian]
Signed-off-by: Tony Luck <tony.luck@intel.com>
on ia64 thread_info is at the constant offset from task_struct and stack
is embedded into the same beast. Set __HAVE_THREAD_FUNCTIONS, made
task_thread_info() just add a constant.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
)
From: Ingo Molnar <mingo@elte.hu>
This is the latest version of the scheduler cache-hot-auto-tune patch.
The first problem was that detection time scaled with O(N^2), which is
unacceptable on larger SMP and NUMA systems. To solve this:
- I've added a 'domain distance' function, which is used to cache
measurement results. Each distance is only measured once. This means
that e.g. on NUMA distances of 0, 1 and 2 might be measured, on HT
distances 0 and 1, and on SMP distance 0 is measured. The code walks
the domain tree to determine the distance, so it automatically follows
whatever hierarchy an architecture sets up. This cuts down on the boot
time significantly and removes the O(N^2) limit. The only assumption
is that migration costs can be expressed as a function of domain
distance - this covers the overwhelming majority of existing systems,
and is a good guess even for more assymetric systems.
[ People hacking systems that have assymetries that break this
assumption (e.g. different CPU speeds) should experiment a bit with
the cpu_distance() function. Adding a ->migration_distance factor to
the domain structure would be one possible solution - but lets first
see the problem systems, if they exist at all. Lets not overdesign. ]
Another problem was that only a single cache-size was used for measuring
the cost of migration, and most architectures didnt set that variable
up. Furthermore, a single cache-size does not fit NUMA hierarchies with
L3 caches and does not fit HT setups, where different CPUs will often
have different 'effective cache sizes'. To solve this problem:
- Instead of relying on a single cache-size provided by the platform and
sticking to it, the code now auto-detects the 'effective migration
cost' between two measured CPUs, via iterating through a wide range of
cachesizes. The code searches for the maximum migration cost, which
occurs when the working set of the test-workload falls just below the
'effective cache size'. I.e. real-life optimized search is done for
the maximum migration cost, between two real CPUs.
This, amongst other things, has the positive effect hat if e.g. two
CPUs share a L2/L3 cache, a different (and accurate) migration cost
will be found than between two CPUs on the same system that dont share
any caches.
(The reliable measurement of migration costs is tricky - see the source
for details.)
Furthermore i've added various boot-time options to override/tune
migration behavior.
Firstly, there's a blanket override for autodetection:
migration_cost=1000,2000,3000
will override the depth 0/1/2 values with 1msec/2msec/3msec values.
Secondly, there's a global factor that can be used to increase (or
decrease) the autodetected values:
migration_factor=120
will increase the autodetected values by 20%. This option is useful to
tune things in a workload-dependent way - e.g. if a workload is
cache-insensitive then CPU utilization can be maximized by specifying
migration_factor=0.
I've tested the autodetection code quite extensively on x86, on 3
P3/Xeon/2MB, and the autodetected values look pretty good:
Dual Celeron (128K L2 cache):
---------------------
migration cost matrix (max_cache_size: 131072, cpu: 467 MHz):
---------------------
[00] [01]
[00]: - 1.7(1)
[01]: 1.7(1) -
---------------------
cacheflush times [2]: 0.0 (0) 1.7 (1784008)
---------------------
Here the slow memory subsystem dominates system performance, and even
though caches are small, the migration cost is 1.7 msecs.
Dual HT P4 (512K L2 cache):
---------------------
migration cost matrix (max_cache_size: 524288, cpu: 2379 MHz):
---------------------
[00] [01] [02] [03]
[00]: - 0.4(1) 0.0(0) 0.4(1)
[01]: 0.4(1) - 0.4(1) 0.0(0)
[02]: 0.0(0) 0.4(1) - 0.4(1)
[03]: 0.4(1) 0.0(0) 0.4(1) -
---------------------
cacheflush times [2]: 0.0 (33900) 0.4 (448514)
---------------------
Here it can be seen that there is no migration cost between two HT
siblings (CPU#0/2 and CPU#1/3 are separate physical CPUs). A fast memory
system makes inter-physical-CPU migration pretty cheap: 0.4 msecs.
8-way P3/Xeon [2MB L2 cache]:
---------------------
migration cost matrix (max_cache_size: 2097152, cpu: 700 MHz):
---------------------
[00] [01] [02] [03] [04] [05] [06] [07]
[00]: - 19.2(1) 19.2(1) 19.2(1) 19.2(1) 19.2(1) 19.2(1) 19.2(1)
[01]: 19.2(1) - 19.2(1) 19.2(1) 19.2(1) 19.2(1) 19.2(1) 19.2(1)
[02]: 19.2(1) 19.2(1) - 19.2(1) 19.2(1) 19.2(1) 19.2(1) 19.2(1)
[03]: 19.2(1) 19.2(1) 19.2(1) - 19.2(1) 19.2(1) 19.2(1) 19.2(1)
[04]: 19.2(1) 19.2(1) 19.2(1) 19.2(1) - 19.2(1) 19.2(1) 19.2(1)
[05]: 19.2(1) 19.2(1) 19.2(1) 19.2(1) 19.2(1) - 19.2(1) 19.2(1)
[06]: 19.2(1) 19.2(1) 19.2(1) 19.2(1) 19.2(1) 19.2(1) - 19.2(1)
[07]: 19.2(1) 19.2(1) 19.2(1) 19.2(1) 19.2(1) 19.2(1) 19.2(1) -
---------------------
cacheflush times [2]: 0.0 (0) 19.2 (19281756)
---------------------
This one has huge caches and a relatively slow memory subsystem - so the
migration cost is 19 msecs.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Cc: <wilder@us.ibm.com>
Signed-off-by: John Hawkes <hawkes@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add per-arch sched_cacheflush() which is a write-back cacheflush used by
the migration-cost calibration code at bootup time.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
arch: Use <linux/capability.h> where capable() is used.
Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
There is a window where a probe gets removed right after the probe is hit
on some different cpu. In this case probe handlers can't find a matching
probe instance related to break address. In this case we need to read the
original instruction at break address to see if that is not a break/int3
instruction and recover safely.
Previous code had a bug where we were not checking for the above race in
case of reentrant probes and the below patch fixes this race.
Tested on IA64, Powerpc, x86_64.
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Currently arch_remove_kprobes() is only implemented/required for x86_64 and
powerpc. All other architecture like IA64, i386 and sparc64 implementes a
dummy function which is being called from arch independent kprobes.c file.
This patch removes the dummy functions and replaces it with
#define arch_remove_kprobe(p, s) do { } while(0)
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Now that all these entries in the arch ioctl32.c files are gone [1], we can
build fs/compat_ioctl.c as a normal object and kill tons of cruft. We need a
special do_ioctl32_pointer handler for s390 so the compat_ptr call is done.
This is not needed but harmless on all other architectures. Also remove some
superflous includes in fs/compat_ioctl.c
Tested on ppc64.
[1] parisc still had it's PPP handler left, which is not fully correct
for ppp and besides that ppp uses the generic SIOCPRIV ioctl so it'd
kick in for all netdevice users. We can introduce a proper handler
in one of the next patch series by adding a compat_ioctl method to
struct net_device but for now let's just kill it - parisc doesn't
compile in mainline anyway and I don't want this to block this
patchset.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Matthew Wilcox <willy@debian.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The comment in compat.c is wrong, every architecture provides a
get_compat_sigevent() for the IPC compat code already.
This basically moves the x86_64 version to common code and removes all the
others.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Paul Mackerras <paulus@samba.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: "David S. Miller" <davem@davemloft.net>
Acked-by: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add a hook so architectures can validate /dev/mem mmap requests.
This is analogous to validation we already perform in the read/write
paths.
The identity mapping scheme used on ia64 requires that each 16MB or
64MB granule be accessed with exactly one attribute (write-back or
uncacheable). This avoids "attribute aliasing", which can cause a
machine check.
Sample problem scenario:
- Machine supports VGA, so it has uncacheable (UC) MMIO at 640K-768K
- efi_memmap_init() discards any write-back (WB) memory in the first granule
- Application (e.g., "hwinfo") mmaps /dev/mem, offset 0
- hwinfo receives UC mapping (the default, since memmap says "no WB here")
- Machine check abort (on chipsets that don't support UC access to WB
memory, e.g., sx1000)
In the scenario above, the only choices are
- Use WB for hwinfo mmap. Can't do this because it causes attribute
aliasing with the UC mapping for the VGA MMIO space.
- Use UC for hwinfo mmap. Can't do this because the chipset may not
support UC for that region.
- Disallow the hwinfo mmap with -EINVAL. That's what this patch does.
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Remove various things which were checking for gcc-1.x and gcc-2.x compilers.
From: Adrian Bunk <bunk@stusta.de>
Some documentation updates and removes some code paths for gcc < 3.2.
Acked-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The ptrace_get_task_struct() helper that I added as part of the ptrace
consolidation is useful in variety of places that currently opencode it.
Switch them to the common helpers.
Add a ptrace_traceme() helper that needs to be explicitly called, and simplify
the ptrace_get_task_struct() interface. We don't need the request argument
now, and we return the task_struct directly, using ERR_PTR() for error
returns. It's a bit more code in the callers, but we have two sane routines
that do one thing well now.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
sys_migrate_pages implementation using swap based page migration
This is the original API proposed by Ray Bryant in his posts during the first
half of 2005 on linux-mm@kvack.org and linux-kernel@vger.kernel.org.
The intent of sys_migrate is to migrate memory of a process. A process may
have migrated to another node. Memory was allocated optimally for the prior
context. sys_migrate_pages allows to shift the memory to the new node.
sys_migrate_pages is also useful if the processes available memory nodes have
changed through cpuset operations to manually move the processes memory. Paul
Jackson is working on an automated mechanism that will allow an automatic
migration if the cpuset of a process is changed. However, a user may decide
to manually control the migration.
This implementation is put into the policy layer since it uses concepts and
functions that are also needed for mbind and friends. The patch also provides
a do_migrate_pages function that may be useful for cpusets to automatically
move memory. sys_migrate_pages does not modify policies in contrast to Ray's
implementation.
The current code here is based on the swap based page migration capability and
thus is not able to preserve the physical layout relative to it containing
nodeset (which may be a cpuset). When direct page migration becomes available
then the implementation needs to be changed to do a isomorphic move of pages
between different nodesets. The current implementation simply evicts all
pages in source nodeset that are not in the target nodeset.
Patch supports ia64, i386 and x86_64.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This was causing some ordering problems. Remove the up-front evaluation
and just revaluate the compiler version each time we need it.
(The up-front evaluation was problematic because some architectures modify
the value of $(CC)).
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
arch/ia64/kernel/setup.c: In function `show_cpuinfo':
arch/ia64/kernel/setup.c:576: warning: long unsigned int format, different type arg (arg 12)
arch/ia64/kernel/setup.c:576: warning: long unsigned int format, different type arg (arg 13)
Introduced by 95235ca2c2
Signed-off-by: Tony Luck <tony.luck@intel.com>
The first of these changes s/hotplug/uevent/ was needed to
compile sn2_defconfig (ia64/sn). The other three files
changed are blind changes of all remaining bus_type.hotplug
references I could find to bus_type.uevent.
This patch attempts to finish similar changes made in the
gregkh-driver-kill-hotplug-word-from-driver-core Nov 22 patch.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
The function ia64_pci_legacy_write() returns 0 for everything
except errors. This return value gets sent back to the user from
pci_write_legacy_io(), making it look like every write fails. The trivial
patch below copies the behavior of the SGI sn machvec and does what
would be expected from something implementing a write() function.
Signed-off-by: Alex Williamson <alex.williamson@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
sparc64, i386 and x86_64 have support for a special data section dedicated
to rarely updated data that is frequently read. The section was created to
avoid false sharing of those rarely read data with frequently written kernel
data.
This patch creates such a data section for ia64 and will group rarely written
data into this section.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Change the NR_CPUS default for ia64/sn up to 1024.
Signed-off-by: John Hawkes <hawkes@sgi.com>
Signed-off-by: John Hesterberg <jh@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
I see why the problem exists only on SN. SN uses a different hardware
mechanism to purge TLB entries across nodes.
It looks like there is a bug in the SN TLB flushing code. During context switch,
kernel threads inherit the mm of the task that was previously running on the
cpu. This confuses the code in sn2_global_tlb_purge().
The result is a missed TLB purge for the task that owns the "borrowed" mm.
(I hit the problem running heavy stress where kswapd was purging code pages of
a user task that woke kswapd. The user task took a SIGILL fault trying to
execute code in the page that had been ripped out from underneath it).
Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Use raw_smp_processor_id() instead of get_cpu() as we don't need the
extra features of get_cpu().
Signed-off-by: Jes Sorensen <jes@trained-monkey.org>
Signed-off-by: Tony Luck <tony.luck@intel.com>
The udelay() inline for ia64 uses the ITC. If CONFIG_PREEMPT is enabled
and the platform has unsynchronized ITCs and the calling task migrates
to another CPU while doing the udelay loop, then the effective delay may
be too short or very, very long.
This patch disables preemption around 100 usec chunks of the overall
desired udelay time. This minimizes preemption-holdoffs.
udelay() is now too big to be inline, move it out of line and export it.
Signed-off-by: John Hawkes <hawkes@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
I realized ZONE_DMA32 has a trivial bug at Kconfig for ia64. In
include/linux/gfp.h on 2.6.15-rc5-mm1, CONFIG is define like followings.
#ifdef CONFIG_DMA_IS_DMA32
#define __GFP_DMA32 ((__force gfp_t)0x01) /* ZONE_DMA is ZONE_DMA32
*/
:
:
So, CONFIG_"ZONE"_DMA_IS_DMA32 is clearly wrong.
Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
When multiple probes are registered at the same address and if due to some
recursion (probe getting triggered within a probe handler), we skip calling
pre_handlers and just increment nmissed field.
The below patch make sure it walks the list for multiple probes case.
Without the below patch we get incorrect results of nmissed count for
multiple probe case.
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
What is the value shown in "cpu MHz" of /proc/cpuinfo when CPUs are capable of
changing frequency?
Today the answer is: It depends.
On i386:
SMP kernel - It is always the boot frequency
UP kernel - Scales with the frequency change and shows that was last set.
On x86_64:
There is one single variable cpu_khz that gets written by all the CPUs. So,
the frequency set by last CPU will be seen on /proc/cpuinfo of all the
CPUs in the system. What you see also depends on whether you have constant_tsc
capable CPU or not.
On ia64:
It is always boot time frequency of a particular CPU that gets displayed.
The patch below changes this to:
Show the last known frequency of the particular CPU, when cpufreq is present. If
cpu doesnot support changing of frequency through cpufreq, then boot frequency
will be shown. The patch affects i386, x86_64 and ia64 architectures.
Signed-off-by: Venkatesh Pallipadi<venkatesh.pallipadi@intel.com>
Signed-off-by: Dave Jones <davej@redhat.com>
The patch that added support for a new platform chipset (shub2) broke
PTC deadlock recovery on older versions of the chipset. (PTCs are the
SN platform-specific method for doing a global TLB purge). This
patch fixes deadlock recovery so that it works on both the old & new
chipsets.
Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
We have a customer application which trips a bug. The problem arises
when a driver attempts to call do_munmap on an area which is mapped, but
because current->thread.task_size has been set to 0xC0000000, the call
to do_munmap fails thinking it is an unmap beyond the user's address
space.
The comment in fs/binfmt_elf.c in load_elf_library() before the call
to SET_PERSONALITY() indicates that task_size must not be changed for
the running application until flush_thread, but is for ia64 executing
ia32 binaries.
This patch moves the setting of task_size from SET_PERSONALITY() to
flush_thread() as indicated. The customer application no longer is able
to trip the bug.
Signed-off-by: Robin Holt <holt@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
The per-node data structures are allocated with strided offsets that are a
function of the node number. This prevents excessive cache-aliasing from
occurring.
On systems with a large number of nodes, the strided offset becomes
too large. This patch restricts the maximum offset to 32MB. This is far larger
than the size of any current L3 cache.
Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Altix only patch to add fixup code that sets up
pci_controller->window. This code is a temporary
fix until ACPI support on Altix is added.
Also, corrects the usage of pci_dev->sysdata,
which had previously been used to reference
platform specific device info, to now point to
a pci_controller struct.
Signed-off-by: John Keller <jpk@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Return -EINTR instead of -ERESTARTSYS when signals are delivered during
a blocked read of /proc/sal/*/event. This allows salinfo_decode to
detect signals when it is blocked on a read of those files.
Signed-off-by: Keith Owens <kaos@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
This updates the sn2_defconfig file for the Altix 330 hardware, enables
the AGP graphics for the SGI Prism, and removes prompts for the remainder
of the new features. Greg Edwards reviewed the changes.
Signed-off-by: Robin Holt <holt@sgi.com>
Signed-off-by: Greg Edwards <edwardsg@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
