We do the exact same printk about a dozen lines above
with no intermediate printk's.
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Here is a small patch for i386 which adds a cpufeature flag and
detection code for Intel's Branch Trace Store (BTS) feature. This
feature can be found on Intel P4 and Core 2 processors among others.
It can also be used by perfmon.
changelog:
- add CPU_FEATURE_BTS
- add Branch Trace Store detection
signed-off-by: stephane eranian <eranian@hpl.hp.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Make the needlessly global alloc_gdt() static.
(against) pda-percpu-init
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: Andi Kleen <ak@muc.de>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Until not so long ago, there were system log messages pointing to
inconsistent MTRR setup of the video frame buffer caused by the way vesafb
and X worked. While vesafb was fixed meanwhile, I believe fixing it there
only hides a shortcoming in the MTRR code itself, in that that code is not
symmetric with respect to the ordering of attempts to set up two (or more)
regions where one contains the other. In the current shape, it permits
only setting up sub-regions of pre-exisiting ones. The patch below makes
this symmetric.
While working on that I noticed a few more inconsistencies in that code,
namely
- use of 'unsigned int' for sizes in many, but not all places (the patch
is converting this to use 'unsigned long' everywhere, which specifically
might be necessary for x86-64 once a processor supporting more than 44
physical address bits would become available)
- the code to correct inconsistent settings during secondary processor
startup tried (if necessary) to correct, among other things, the value
in IA32_MTRR_DEF_TYPE, however the newly computed value would never get
used (i.e. stored in the respective MSR)
- the generic range validation code checked that the end of the
to-be-added range would be above 1MB; the value checked should have been
the start of the range
- when contained regions are detected, previously this was allowed only
when the old region was uncacheable; this can be symmetric (i.e. the new
region can also be uncacheable) and even further as per Intel's
documentation write-trough and write-back for either region is also
compatible with the respective opposite in the other
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Allow selected bug checks to be skipped by paravirt kernels. The two most
important are the F00F workaround (which is either done by the hypervisor,
or not required), and the 'hlt' instruction check, which can break under
some hypervisors.
Signed-off-by: Zachary Amsden <zach@vmware.com>
Signed-off-by: Chris Wright <chrisw@sous-sol.org>
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Both lhype and Xen want to call the core of the x86 cpu detect code before
calling start_kernel.
(extracted from larger patch)
AK: folded in start_kernel header patch
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Use the pcurrent field in the PDA to implement the "current" macro. This ends
up compiling down to a single instruction to get the current task.
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: Chuck Ebbert <76306.1226@compuserve.com>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Jan Beulich <jbeulich@novell.com>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Use the cpu_number in the PDA to implement raw_smp_processor_id. This is a
little simpler than using thread_info, though the cpu field in thread_info
cannot be removed since it is used for things other than getting the current
CPU in common code.
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: Chuck Ebbert <76306.1226@compuserve.com>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Jan Beulich <jbeulich@novell.com>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
This patch is the meat of the PDA change. This patch makes several related
changes:
1: Most significantly, %gs is now used in the kernel. This means that on
entry, the old value of %gs is saved away, and it is reloaded with
__KERNEL_PDA.
2: entry.S constructs the stack in the shape of struct pt_regs, and this
is passed around the kernel so that the process's saved register
state can be accessed.
Unfortunately struct pt_regs doesn't currently have space for %gs
(or %fs). This patch extends pt_regs to add space for gs (no space
is allocated for %fs, since it won't be used, and it would just
complicate the code in entry.S to work around the space).
3: Because %gs is now saved on the stack like %ds, %es and the integer
registers, there are a number of places where it no longer needs to
be handled specially; namely context switch, and saving/restoring the
register state in a signal context.
4: And since kernel threads run in kernel space and call normal kernel
code, they need to be created with their %gs == __KERNEL_PDA.
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: Chuck Ebbert <76306.1226@compuserve.com>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Jan Beulich <jbeulich@novell.com>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
When a CPU is brought up, a PDA and GDT are allocated for it. The GDT's
__KERNEL_PDA entry is pointed to the allocated PDA memory, so that all
references using this segment descriptor will refer to the PDA.
This patch rearranges CPU initialization a bit, so that the GDT/PDA are set up
as early as possible in cpu_init(). Also for secondary CPUs, GDT+PDA are
preallocated and initialized so all the secondary CPU needs to do is set up
the ldt and load %gs. This will be important once smp_processor_id() and
current use the PDA.
In all cases, the PDA is set up in head.S, before a CPU starts running C code,
so the PDA is always available.
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: Chuck Ebbert <76306.1226@compuserve.com>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Jan Beulich <jbeulich@novell.com>
Cc: Andi Kleen <ak@suse.de>
Cc: James Bottomley <James.Bottomley@SteelEye.com>
Cc: Matt Tolentino <matthew.e.tolentino@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Clean up the espfix code:
- Introduced PER_CPU() macro to be used from asm
- Introduced GET_DESC_BASE() macro to be used from asm
- Rewrote the fixup code in asm, as calling a C code with the altered %ss
appeared to be unsafe
- No longer altering the stack from a .fixup section
- 16bit per-cpu stack is no longer used, instead the stack segment base
is patched the way so that the high word of the kernel and user %esp
are the same.
- Added the limit-patching for the espfix segment. (Chuck Ebbert)
[jeremy@goop.org: use the x86 scaling addressing mode rather than shifting]
Signed-off-by: Stas Sergeev <stsp@aknet.ru>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: Zachary Amsden <zach@vmware.com>
Acked-by: Chuck Ebbert <76306.1226@compuserve.com>
Acked-by: Jan Beulich <jbeulich@novell.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Here is a patch (used by perfmon2) to detect the presence of the Precise Event
Based Sampling (PEBS) feature for i386. The patch also adds the cpu_has_pebs
macro.
- adds X86_FEATURE_PEBS
- adds cpu_has_pebs to test for X86_FEATURE_PEBS
Signed-off-by: stephane eranian <eranian@hpl.hp.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Fix checks that failed to realize that values are 4-kB-unit-sized (note the
format strings in this same diff context which *do* realize the unit size,
via appended "000"!). Also fix an incorrect below-1MB area check (as
gathered from Jan Beulich's unapplied patch at
http://www.ussg.iu.edu/hypermail/linux/kernel/0411.1/1378.html ) Update
mtrr_add_page() docu to make 4-kB-sized calculation more obvious.
Given several further items mentioned in Jan's patch mail, all in all MTRR
code seems surprisingly buggy, for a surprisingly long period of time (many
years). Further work/investigation would be useful.
TBD Note that my patch is pretty much UNTESTED, since I can only verify that it
TBD successfully boots my machine, but I cannot test against actual buggy
TBD hardware which would require these (formerly broken) checks. Long -mm
TBD simmering would make sense, especially since these now-working checks might
TBD turn out to have adverse effects on unaffected hardware.
Signed-off-by: Andreas Mohr <andi@lisas.de>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: Jan Beulich <jbeulich@novell.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
arch/x86_64/kernel/cpufreq/../../../i386/kernel/cpu/cpufreq/speedstep-lib.c:131: error: 'MSR_FSB_FREQ' undeclared (first use in this function)
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Dave Jones <davej@redhat.com>
On some systems such as the IBM x3650 there are bits set in the
upper half of the control values provided by the _PSS object.
These bits are only relevant for cpufreq drivers that use IO ports
which are not currently supported by the speedstep-centrino driver.
The current MSR oriented code assumes that upper bits are not set
and thus fails to work correctly when they are. e.g. the control
and status value equality check fails even though the ACPI spec
allows the inequality.
Signed-off-by: Gary Hade <garyh@us.ibm.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Several more Intel CPUs are now capable using the p4-clockmod cpufreq
driver. As it is of limited use most of the time, print a big bold warning
if a better cpufreq driver might be available.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Dave Jones <davej@redhat.com>
acpi-cpufreq needs the same patch as the previous speedstep-centrino change.
Additionally, the centrino driver can have its ifdef moved out a little
further to eliminate some more code/variables.
Signed-off-by: Dave Jones <davej@redhat.com>
arch/i386/kernel/cpu/cpufreq/speedstep-centrino.c:396: warning: 'sw_any_bug_dmi_table' defined but not used
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Dave Jones <davej@redhat.com>
acpi-cpufreq.c, speedstep-centrino.c: warning: 'sw_any_bug_dmi_table' defined but not used
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Len Brown <len.brown@intel.com>
If someone inserts speedstep-smi on a mobile P4, it prevents other cpufreq
modules from loading until it is unloaded.
Signed-off-by: Hiroshi Miura <miura@da-cha.org>
Signed-off-by: Dave Jones <davej@redhat.com>
ioremap must be balanced by an iounmap and failing to do so can result
in a memory leak.
Tested (compilation only):
- using allmodconfig
- making sure the files are compiling without any warning/error due to
new changes
Signed-off-by: Amol Lad <amol@verismonetworks.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Enable ondemand governor and acpi-cpufreq to use IA32_APERF and IA32_MPERF MSR
to get active frequency feedback for the last sampling interval. This will
make ondemand take right frequency decisions when hardware coordination of
frequency is going on.
Without APERF/MPERF, ondemand can take wrong decision at times due
to underlying hardware coordination or TM2.
Example:
* CPU 0 and CPU 1 are hardware cooridnated.
* CPU 1 running at highest frequency.
* CPU 0 was running at highest freq. Now ondemand reduces it to
some intermediate frequency based on utilization.
* Due to underlying hardware coordination with other CPU 1, CPU 0 continues to
run at highest frequency (as long as other CPU is at highest).
* When ondemand samples CPU 0 again next time, without actual frequency
feedback from APERF/MPERF, it will think that previous frequency change
was successful and can go to wrong target frequency. This is because it
thinks that utilization it has got this sampling interval is when running at
intermediate frequency, rather than actual highest frequency.
More information about IA32_APERF IA32_MPERF MSR:
Refer to IA-32 Intel® Architecture Software Developer's Manual at
http://developer.intel.com
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Recent speedstep-centrino unification onto acpi-cpufreq patchset broke
cpuinfo_cur_freq interface in /sys/../cpuinfo/, when MSR was used for
transitions. Attached patch fixes that breakage.
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Only change the frequency if the state previously set is different
from what we are trying to set. We don't really have to get the current
frequency at this point.
Signed-off-by: Denis Sadykov <denis.m.sadykov@intel.com>
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Mark ACPI hooks in speedstep-centrino as deprecated. Change the order in which
speedstep-centrino and acpi-cpufreq (when both are in kernel) will be
added. First driver to be tried is now acpi-cpufreq, followed by
speedstep-centrino.
Add a note in feature-removal-schedule to mark this deprecation.
Signed-off-by: Denis Sadykov <denis.m.sadykov@intel.com>
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Add in the support for Intel Enhanced Speedstep - MSR based transitions.
With this change, the ACPI based support in speedstep-centrino can be
deprecated and duplicate code in that driver can be marked for removal.
Much easier to maintain and support this way. This also reduces the
user misconfigurations and questions on which driver is to be used
under which CPUs to support Enhanced Speedstep.
Signed-off-by: Denis Sadykov <denis.m.sadykov@intel.com>
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Some clean up and redsign of the driver. Mainly making it easier to add
support for multiple sub-mechanisms of changing frequency. Currently this
driver supports only ACPI SYSTEM_IO address space. With the changes
below it is easier to add support for other address spaces like Intel
Enhanced Speedstep which uses MSR (ACPI FIXED_FEATURE_HARDWARE) to do the
transitions.
Signed-off-by: Denis Sadykov <denis.m.sadykov@intel.com>
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Dave Jones <davej@redhat.com>
This patchset has refresh/rebase of a bunch of patches/bugfixes related to
acpi-cpufreq that were sent earlier on this list.
patch 1/8
Patch that fixes a bug in swcoordination code in acpi-cpufreq
patch 2/8 through patch 7/8
Grand unification of ACPI based speedstep-centrino and acpi-cpufreq drivers.
ACPI allows P-state transitions in multiple ways. Like using IO ports or using
processor native method (MSR). Without this patch, IO port based P-state
transitions are handled in acpi-cpufreq driver and MSR based transitions on
Intel CPUs are handled in speedstep-centrino driver. Even though most of the
code in these two drivers should be similar, except for final changing/checking
of frequency (one driver does it using IO port and other does it through
MSR), we have duplicated code in these two drivers. There are also issues
around BIOSes supporting both MSR and IO port and which driver should be
loaded first in standard installations.
The patchset combines functionality of these two driver into acpi-cpufreq
driver. ACPI based functionality in speedstep-centrino is marked deprecated
and will be removed in future. speedstep-centrino will continue to work
on systems that depend on older non-ACPI table based P-state chanes.
* 2/8 - Patch that reorganizes the code in acpi-cpufreq, cleaning it up
a little and making it easier to add MSR support later.
* 3/8 - Pull in the MSR based transition support into acpi-cpufreq.
* 4/8 - Mark speedstep-centrino deprecated. Change the order in Makefile to
load acpi-cpufreq first and speedstep-centrino later, in cases where both
are configured in.
* 5/8 - lindent acpi-cpufreq.c
* 6/8 - Minor change to eliminate the check of current frequency on
notifications. We can use last set frequency instead.
* 7/8 - Make cpufreq->get of acpi_cpufreq work correctly again.
There will be a patch in future that removes ACPI based support in
speedstep-centrino in coming months.
patch 8/8
Add support for IA32_APERF and IA32_MPERF MSR and get the actual frequency
from these MSRs and use it to determine the next frequency target in ondemand
governor
This patch:
There is a bug in software coordination patch in acpi-cpufreq, due to which
frequency will only be set on first CPU of any coordinated group.
Bug identified by Denis, was not recognised earlier as there are no platforms
yet that use software coordination with acpi-cpufreq driver.
Signed-off-by: Denis Sadykov <denis.m.sadykov@intel.com>
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Get rid of warning in the thermal throttling code about not checking
sysfs return values.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This changes a couple of if() BUG(); constructs to
BUG_ON(); so it can be safely optimized away.
Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Jesper Juhl reported that testing the software math-emulation by forcing
"no387" doesn't work on modern CPU's.
The reason was two-fold:
- you also need to pass in "nofxsr" to make sure that we not only don't
touch the old i387 legacy hardware, it also needs to disable the
modern XMM/FXSR sequences
- "nofxsr" didn't actually clear the capability bits immediately,
leaving the early boot sequence still using FXSR until we got to
the identify_cpu() stage.
This fixes the "nofxsr" flag to take effect immediately on the boot CPU.
Debugging by Randy Dunlap
Acked-by: Randy Dunlap <rdunlap@xenotime.net>
Cc: Jesper Juhl <jesper.juhl@gmail.com>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Please ignore previous message.
This patch is adding support for CPU connected to CLE266
chipset. For older CPU this is only way. For "Powersaver"
processor this way will be used if ACPI C3 isn't supported.
I have tested it. It seems to work exacly like ACPI.
But it is less safe. On CLE266 chipset port 0x22 is
blocking processor access to PCI bus too.
Signed-off-by: Rafa³ Bilski <rafalbilski@interia.pl>
Signed-off-by: Dave Jones <davej@redhat.com>
Make the sections proper and get rid of section mismatch warnings.
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Dave Jones <davej@redhat.com>
The functions prepare_set and post_set in kernel/cpu/mtrr/generic.c wrap
the spinlock set_atomicity_lock: prepare_set returns with the lock held,
and post_set releases the lock without acquiring it. Add lock annotations
to these two functions so that sparse can check callers for lock pairing,
and so that sparse will not complain about these functions since they
intentionally use locks in this manner.
Signed-off-by: Josh Triplett <josh@freedesktop.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add supplemental SSE3 instructions flag, and Direct Cache Access flag.
As described in "Intel Processor idenfication and the CPUID instruction
AP485 Sept 2006"
AK: also added for x86-64
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Andi Kleen <ak@suse.de>
The counter is exported to /sys that keeps track of the
number of thermal events, such that the user knows how bad the
thermal problem might be (since the logging to syslog and mcelog
is rate limited).
AK: Fixed cpu hotplug locking
Signed-off-by: Dmitriy Zavin <dmitriyz@google.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Refactor the event processing (syslog messaging and rate limiting)
into separate file therm_throt.c. This allows consistent reporting
of CPU thermal throttle events.
After ACK'ing the interrupt, if the event is current, the user
(p4.c/mce_intel.c) calls therm_throt_process to log (and rate limit)
the event. If that function returns 1, the user has the option to log
things further (such as to mce_log in x86_64).
AK: minor cleanup
Signed-off-by: Dmitriy Zavin <dmitriyz@google.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Burman Yan