Fixes a problem where an extraneous error message was emitted during
initialization if there is a GPE block larger than 255 bits. Any
GPE block larger than 120 GPEs could generate the error.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
With the NULL entry sanity check implemented, the XSDT validation is
useless because:
1. If XSDT contains NULL entries, it can be bypassed by the new sanity
check mechanism;
2. If RSDP contains a bad XSDT address, invoking XSDT validation will still
lead to a kernel crash.
This patch deletes the old XSDT validation solution and thus enables the
new NULL entry sanity check solution.
Note that if there are reports reporting regressions caused by the enabling
of the new feature and disabling of the old feature, this commit should be
bisected and reverted. Lv Zheng.
References: https://bugzilla.kernel.org/show_bug.cgi?id=73911
References: https://bugs.archlinux.org/task/39811
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Reported-and-tested-by: Bruce Chiarelli <mano155@gmail.com>
Reported-and-tested-by: Spyros Stathopoulos <spystath@gmail.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
It is reported that there are buggy BIOSes in the world: AMI uses an XSDT
compiler for early BIOSes, this compiler will generate XSDT with a NULL
entry. The affected BIOS versions are "AMI BIOS F2-F4".
Original solution on Linux is to use an alternative heathy root table
instead of the ill one. This commit is:
Commit: 671cc68dc6
Subject: ACPICA: Back port and refine validation of the XSDT root table.
This is an example of such XSDT dumped from B85-HD3 (AMI F3 BIOS):
[000h 0000 4] Signature : "XSDT" [Extended System Description Table]
[004h 0004 4] Table Length : 00000074
[008h 0008 1] Revision : 01
[009h 0009 1] Checksum : 18
[00Ah 0010 6] Oem ID : "ALASKA"
[010h 0016 8] Oem Table ID : "A M I"
[018h 0024 4] Oem Revision : 01072009
[01Ch 0028 4] Asl Compiler ID : "AMI "
[020h 0032 4] Asl Compiler Revision : 00010013
[024h 0036 8] ACPI Table Address 0 : 00000000BA5F8180
[02Ch 0044 8] ACPI Table Address 1 : 00000000BA5F8290
[034h 0052 8] ACPI Table Address 2 : 00000000BA5F8308
[03Ch 0060 8] ACPI Table Address 3 : 00000000BA5F8848
[044h 0068 8] ACPI Table Address 4 : 00000000BA5F9320
[04Ch 0076 8] ACPI Table Address 5 : 00000000BA5F9360
[054h 0084 8] ACPI Table Address 6 : 00000000BA5F9398
[05Ch 0092 8] ACPI Table Address 7 : 00000000BA5F9708
[064h d100 8] ACPI Table Address 8 : 00000000BA5FC9A8
[06Ch 0108 8] ACPI Table Address 9 : 0000000000000000
But according to the bug report, the XSDT in fact is not broken. In the
above XSDT, ACPI Table Address 1-8 contains the same value as RSDT. The
differences can only be seen on the following 2 entries:
1. The first entry points to a FADT whose Revision is 5 while the first
entry in RSDT points to a FADT whose Revision is 2.
The FADT dumped from the address indicated by the first entry of XSDT:
FACP @ 0x00000000BA5F8180
0000: 46 41 43 50 0C 01 00 00<05>4B 41 4C 41 53 4B 41 FACP.....KALASKA
...
The FADT dumped from the address indicated by the first entry of RSDT:
FACP @ 0x00000000BA5ED0F0
0000: 46 41 43 50 84 00 00 00<02>A7 41 4C 41 53 4B 41 FACP......ALASKA
...
2. The last entry is a NULL terminator.
According to the test result, the Revision 5 FADT is accessible. Thus the
original solution turns out to be a work around that is preventing the
higher revision tables to be used for such platforms (they are all x86-64
platforms, and should use XSDT and higher revision FADT).
This patch offers a new solution, where a sanity check is performed before
installing a table address from XSDT. If the entry is NULL, it is simply
discarded.
Note that, this patch doesn't remove the original solution, so for Linux
kernel, this commit is actually a no-op, but it allows acpidump to be
working on such platforms. By doing so, we allow another easy revertable
commit to enable this feature so that when that commit is reverted, the
useful sanity check will not be affected. Lv Zheng.
References: https://bugzilla.kernel.org/show_bug.cgi?id=73911
References: https://bugs.archlinux.org/task/39811
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Reported-and-tested-by: Bruce Chiarelli <mano155@gmail.com>
Reported-and-tested-by: Spyros Stathopoulos <spystath@gmail.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch adds "-x" and "-x -x" options to disable XSDT for acpidump.
The single "-x" can be used to stop using XSDT, RSDT will be forced to find
static tables, note that XSDT will still be dumped. The double "-x" can
stop dumping XSDT, which is useful when the XSDT address reported by RSDP
is pointing to an invalid address.
It is reported there are platforms having broken XSDT shipped, acpidump
will stop working while accessing such XSDT. This patch adds new option so
that users can force acpidump to dump tables listed in the RSDT. Lv Zheng.
Buglink: https://bugzilla.kernel.org/show_bug.cgi?id=73911
Buglink: https://bugs.archlinux.org/task/39811
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Reported-and-tested-by: Bruce Chiarelli <mano155@gmail.com>
Reported-and-tested-by: Spyros Stathopoulos <spystath@gmail.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This is the linuxize result of the following commit:
Subject: ACPICA: Improve handling of exception code blocks.
Split exception codes into three distinct blocks; for the main
ASL compiler, Table compiler, and the preprocessor. This allows
easy addition of new codes into each block without disturbing
the others. Adds one new file, aslmessages.c
The iASL changes are not in this patch as iASL currently is not
shipped in the kernel.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch is the linuxize result of the following commit:
Subject: ACPICA: Add check for _PRP/_HID dependency, with error message.
_PRP requires that a _HID appears in the same scope.
The iASL changes are not in this patch as iASL currently is not
shipped in the kernel.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch enforces a rule to always use ACPI_VALIDATE_RSDP_SIG for RSDP
signatures passed from table header or ACPI_SIG_RSDP so that truncated
string comparison can be avoided. This could help to fix the issue that
"RSD " matches but "RSD PTR " doesn't match. Lv Zheng.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Linux wants to include all header files but leave empty inline
stub variables for a feature that is not configured during build.
This patch configures ACPICA external globals/macros/functions out and
defines them into no-op when CONFIG_ACPI is not enabled. Lv Zheng.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
OSPMs like Linux trend to include all header files but leave empty inline
stub functions for a feature that is not configured during build.
This patch adds wrappers mechanism to be used around ACPICA external
interfaces to facilitate OSPM with such configurability.
This patch doesn't include code for Linux to use this new mechanism, thus
no functional change. Lv Zheng.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch re-orders the interface prototypes defined in acpixf.h, moving
those having not back ported to ACPICA into a seperate section to reduce
the source code differences between Linux and ACPICA.
This can help to reduce the cost of linuxizing the follow up commits.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch extends ACPI_HW_DEPENDENT_x mechanism to all debugging output
related functions so that the OSPMs can have full control to configure
them into stub functions.
This patch doesn't include code for Linux to use this new mechanism, thus
no functional change. Lv Zheng.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch extends ACPI_HW_DEPENDENT_x mechanism to all error message
related functions so that the OSPMs can have full control to configure them
into stub functions.
This patch doesn't include code for Linux to use this new mechanism, thus
no functional change. Lv Zheng.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
OSPMs like Linux trend to include all header files but leave empty stub
macros for a feature that is not configured during build.
For macros defined without other symbols referencesd it is safe to leave
them without protections.
By investigation, there are only the following internal/external
symbols referenced by the ACPICA macros:
1. C library symbols, including string, ctype, stdarg APIs. Since such
symbols are always accessbile in the kernel source tree, it is safe to
leave macros referencing them without protected for Linux.
2. ACPICA OSL symbols, such symbols are designed to be used only by ACPICA
internal APIs. And there are macros directly referencing mutex and
memory allocation OSL symbols. We need to examine the external usages
of such macros.
For macros referencing the mutex OSL symbols, fortunately, there is no
external user directly invoking such macros.
========================================================================
!! IMPORTANT !!
========================================================================
For macros referencing memory allocation OSL symbols -
1. 'free' - ACPI_FREE
2. 'alloc' - ACPI_ALLOCATE, ACPI_ALLOCATE_ZEROED, ACPI_ALLOCATE_BUFFER,
ACPI_ALLOCATE_LOCAL_BUFFER
there are external users directly invoking 'alloc' macros. And the more
complicated situation is the reversals of such macros are not ACPI_FREE
but acpi_os_free (or kfree) in Linux. Though we can define such macros
into no-op, we in fact cannot define their reversals into no-op.
This patch adds mechanism to protect ACPICA memory allocation APIs for
Linux so that acpi_os_free (or kfree) invoked in Linux can have a zero
address returned by 'alloc' macros to free. In this
way, acpi_os_free (or kfree) can be converted into no-op.
========================================================================
3. ACPI_OFFSET and other macros that would access structure members, we
need to check if such structure members are not accessible under a
specific configuration. Fortunately, currently Linux doesn't use such
structure members when CONFIG_ACPI is disabled.
This patch thus only adds mechanism useful for implementing stubs for
ACPICA provided macros - the configurability of memory allocation APIs.
This patch doesn't include code for Linux to use this new mechanism, thus
no functional changes. Lv Zheng.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
OSPMs like Linux trend to include all header files but leave empty stub
macros for a feature that is not configured during build.
This patch cleans up global variables that are defined in utglobal.c using
ACPI_INIT_GLOBAL mechanism. In Linux, such global variables are used by
the subsystems external to ACPICA.
This patch also cleans up global variables that are defined in utglobal.c
using ACPI_GLOBAL mechanism. In Linux, such global variables are not used
or should not be used by the subsystems external to ACPICA.
External global variables can be redefined by OSPMs using
ACPI_INIT_GLOBAL/ACPI_GLOBAL macros. Thus the ACPI_GLOBAL/ACPI_INIT_GLOBAL
mechanisms can be used by OSPM to implement stubs for such external
globals.
This patch doesn't include code for Linux to use this new mechanism, thus
no functional changes. Lv Zheng.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Change all instances of "sub-table" to "subtable" for
consistency.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
More of a style cleanup. If hw_build_pci_list is to return a non-zero
status, it now deletes any partial ID list that has been constructed.
If it returns AE_OK, the caller is responsible for list deletion.
David Box.
Signed-off-by: David E. Box <david.e.box@linux.intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch currently only affects acpihelp and iASL which are not shipped
in the Linux kernel.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Adds header, disassembler, table compiler, and template support
for the Low Power Idle Table (LPIT).
Note that the disassembler and table compiler are not shipped in
the kernel.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
1) Add standard trace mechanism.
2) Add ACPI_EXPORT_SYMBOL macro.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Move all of the public globals to acpixf.h for the convenience
of users. Also:
Adds #ifndef/#endif conditions arround ACPI_GLOBAL and
ACPI_INIT_GLOBAL definition so that OSPMs might be able to:
1. Redefine ACPI_GLOBAL/ACPI_INIT_GLOBAL into no-op, and
2. Redefine external global variables into immediates to implement stubs
for them.
Lv Zheng.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch fixes an issue that the while loop is not needed as fread()
should return exact the bytes of expected.
The patch is tested by runing diff against the output of "-c" mode and
the normal mode, and only finds the following differences:
1. table addresses: the "-c" mode will always fill 0x0000000000000000 for
the address.
2. RSDP/RSDT/XSDT: there is no generation of such tables for "-c" mode.
So the test result shows the fix is valid. Lv Zheng.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch deploys ACPI_DEBUGGER_EXEC usage to utglobal.c to reduce "ifdef"
of ACPI_DEBUGGER. No functional changes. Lv Zheng.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This patch deletes global variable declarations that are no longer used by
ACPICA. No functional changes. Lv Zheng.
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Enable CPUidle so it's easier for maintainers to notice
if some future code changes cause regressions.
Cc: Kevin Hilman <khilman@linaro.org>
Cc: Nishanth Menon <nm@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Tero Kristo <t-kristo@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
On N900 there are nice LEDs that show the state of the
sys_clkreq and sys_off_mode pins.
These LEDs go low when the system enters deeper idle
states. The left LED shows the state of the sys_clkreq
pin, and goes off during retention idle. The right LED
shows the state of sys_off_mode pin and both go off
during off idle.
As N900 is a battery operated device, these LEDs should
be off most of the time. So let's enable them by default
so we can make sure the system is mostly idle.
This allows the maintainers to also immediately test
patches for PM regressions by looking at the LEDs,
which certainly makes my life easier.
The LED can naturally be disabled during runtime with:
# echo none > /sys/class/leds/debug::sleep/trigger
Note that we don't currently have support for omap3
errata 1.158 that remuxes GPIO pins to INPUT_PULLUP |
MUX_MODE7 for the duration of idle. This means that the
GPIO pins set high will go down during off idle. In this
case it does not matter as the sys_off_mode goes down
too, but there's still a slim chance of false off idle
LED signals. If in doubt, false LED signals can be
verified by the sys_off_mode or vdd_core values.
Also note that to allow the UARTs to autoidle, the
following needs to be run on N900 to enable off idle:
#!/bin/sh
uarts=$(find /sys/class/tty/ttyO*/device/power/ -type d)
for uart in $uarts; do
echo 3000 > $uart/autosuspend_delay_ms
done
uarts=$(find /sys/class/tty/ttyO*/power/ -type d)
for uart in $uarts; do
echo enabled > $uart/wakeup
echo auto > $uart/control
done
echo 1 > /sys/kernel/debug/pm_debug/enable_off_mode
For retention idle, change the above to set 0 to
enable_off_mode.
Also note that without the twl4030 PM scripts the actual
voltage scaling won't happen for off idle so we only get
voltage scaling over I2C4 for retention idle. I'll do
some device tree patches for those also a bit later on.
Cc: Kevin Hilman <khilman@linaro.org>
Cc: Nishanth Menon <nm@ti.com>
Cc: Pali Rohár <pali.rohar@gmail.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: Sebastian Reichel <sre@kernel.org>
Cc: Tero Kristo <t-kristo@ti.com>
Acked-by: Aaro Koskinen <aaro.koskinen@iki.fi>
[tony@atomide.com: also make sure the LEDs get built to see PM regressions]
Signed-off-by: Tony Lindgren <tony@atomide.com>
We are currently disabling the init of voltage scaling
for device tree. With the voltage scaling problems fixed
for omap3 in general, there's no need to disable the voltage
scaling init for device tree based booting.
Cc: Kevin Hilman <khilman@linaro.org>
Cc: Nishanth Menon <nm@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Tero Kristo <t-kristo@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Almost certainly any sane board has the twl4030 has the I2C4
pins connected as those are needed for voltage control during
idle. If the I2C4 lines are not properly muxed, any voltage
scaling over I2C4 will fail.
Let's mux those pins by default, the boards that are not using
them can still configure things separately.
Cc: Kevin Hilman <khilman@linaro.org>
Cc: Nishanth Menon <nm@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Tero Kristo <t-kristo@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Patch 01f8fa4f01 "genirq: Allow forcing cpu affinity of interrupts" added
an irq_force_affinity() function, and 30ccf03b4a "clocksource: Exynos_mct:
Use irq_force_affinity() in cpu bringup" subsequently uses it. However, the
driver can be used with CONFIG_SMP disabled, but the function declaration
is only available for CONFIG_SMP, leading to this build error:
drivers/clocksource/exynos_mct.c:431:3: error: implicit declaration of function 'irq_force_affinity' [-Werror=implicit-function-declaration]
irq_force_affinity(mct_irqs[MCT_L0_IRQ + cpu], cpumask_of(cpu));
This patch introduces a dummy helper function for the non-SMP case
that always returns success, to get rid of the build error.
Since the patches causing the problem are marked for stable backports,
this one should be as well.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Acked-by: Kukjin Kim <kgene.kim@samsung.com>
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/5619084.0zmrrIUZLV@wuerfel
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Currently we're attempting to use a static value for the
voltctrl register that only works for controlling the PMIC
over I2C4. For using sys_off_mode signaling, we need to update
update clksetup, voltsetup1, voltsetup2 and voltctrl registers
dynamically depending on the idle state.
So let's fix this by configuring things for I2C4 controlled idle
and sys_off_mode pin controlled idle, and then write the
configured register values depending on the idle state. This
is similar what N900 kernel is doing too.
Cc: Kevin Hilman <khilman@linaro.org>
Cc: Nishanth Menon <nm@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Tero Kristo <t-kristo@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Commit c589eb3869 (ARM: OMAP3: VC: calculate ramp times)
started using regulator slew rates for calculating the idle
mode start-up times. This works fine for I2C4 controlled
regulator scaling as the regulators are never completely
turned off.
For sys_off_mode pin controlled PMIC scripts, the slew rate
based calculations won't work at all as the regulators are
completely turned off and the start-up time is much longer.
This means currently omap3_set_off_timings currently has
zero chance of working on any real hardare. The current code
is broken in at least the following ways:
1. It attempts to use the default ULONG_MAX value for the
oscillator start-up value as we're currently never
initializing the start-up value.
2. It relies on a magic number potentially set by the
bootloader for volsetup2 register.
3. If no magic value is passed, it attempts to calculate
voltsetup2 register based on the regulator slew rate.
This won't work as there is roughly at least five
times the delay needed for turning on vdd1 and vdd2
regulators.
4. It does duplicate register write to OMAP3_PRM_VOLTOFFSET
5. It duplicates the code for omap_usec_to_32k unnecessarily
6. It initialized global registers twice, once for each channel
Let's just remove the broken code and call omap3_set_i2c_timings
directly, we're better off with this function doing nothing until
it's fixed. And otherwise further fixes to omap3_set_off_timings
will be unreadable.
And let's get rid of omap3_set_clksetup as that's not needed
for off-idle controlled by I2C4 as in that case the oscillator
is never shut down.
Cc: Kevin Hilman <khilman@linaro.org>
Cc: Nishanth Menon <nm@ti.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Tero Kristo <t-kristo@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
CPUFreq specific helper functions for OPP (Operating Performance Points)
now use generic OPP functions that allow CPUFreq to be be moved back
into CPUFreq framework. This allows for independent modifications
or future enhancements as needed isolated to just CPUFreq framework
alone.
Here, we just move relevant code and documentation to make this part of
CPUFreq infrastructure.
Cc: Kevin Hilman <khilman@deeprootsystems.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
CPUFREQ custom functions for OPP (Operating Performance Points)
currently exist inside the OPP library. These custom functions currently
depend on internal data structures to pick up OPP information to create
the cpufreq table. For example, the cpufreq table is created precisely
in the same order of how OPP entries are stored inside the list implementation.
This kind of tight interdependency is purely artificial since the same
functionality can be achieved using the generic OPP functions
meant to do the same. This interdependency also limits the independent
modification of cpufreq and OPP library.
So use the generic dev_pm_opp_find_freq_ceil function that achieves the
table organization as we currently use.
As a result of this, we dont need to use the internal device_opp
structure anymore, and we hence we can switch over to rcu lock instead
of the mutex holding the internal list lock.
This breaking of dependency on internal data structure imposes no change
to usage of these.
NOTE: This change is a precursor to moving this cpufreq specific logic
out of the generic library into cpufreq.
Cc: Kevin Hilman <khilman@deeprootsystems.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Some cpufreq drivers were redundantly invoking the _begin() and _end()
APIs around frequency transitions, and this double invocation (one from
the cpufreq core and the other from the cpufreq driver) used to result
in a self-deadlock, leading to system hangs during boot. (The _begin()
API makes contending callers wait until the previous invocation is
complete. Hence, the cpufreq driver would end up waiting on itself!).
Now all such drivers have been fixed, but debugging this issue was not
very straight-forward (even lockdep didn't catch this). So let us add a
debug infrastructure to the cpufreq core to catch such issues more easily
in the future.
We add a new field called 'transition_task' to the policy structure, to keep
track of the task which is performing the frequency transition. Using this
field, we make note of this task during _begin() and print a warning if we
find a case where the same task is calling _begin() again, before completing
the previous frequency transition using the corresponding _end().
We have left out ASYNC_NOTIFICATION drivers from this debug infrastructure
for 2 reasons:
1. At the moment, we have no way to avoid a particular scenario where this
debug infrastructure can emit false-positive warnings for such drivers.
The scenario is depicted below:
Task A Task B
/* 1st freq transition */
Invoke _begin() {
...
...
}
Change the frequency
/* 2nd freq transition */
Invoke _begin() {
... //waiting for B to
... //finish _end() for
... //the 1st transition
... | Got interrupt for successful
... | change of frequency (1st one).
... |
... | /* 1st freq transition */
... | Invoke _end() {
... | ...
... V }
...
...
}
This scenario is actually deadlock-free because, once Task A changes the
frequency, it is Task B's responsibility to invoke the corresponding
_end() for the 1st frequency transition. Hence it is perfectly legal for
Task A to go ahead and attempt another frequency transition in the meantime.
(Of course it won't be able to proceed until Task B finishes the 1st _end(),
but this doesn't cause a deadlock or a hang).
The debug infrastructure cannot handle this scenario and will treat it as
a deadlock and print a warning. To avoid this, we exclude such drivers
from the purview of this code.
2. Luckily, we don't _need_ this infrastructure for ASYNC_NOTIFICATION drivers
at all! The cpufreq core does not automatically invoke the _begin() and
_end() APIs during frequency transitions in such drivers. Thus, the driver
alone is responsible for invoking _begin()/_end() and hence there shouldn't
be any conflicts which lead to double invocations. So, we can skip these
drivers, since the probability that such drivers will hit this problem is
extremely low, as outlined above.
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Since commit d37e2b7644 ("intel_pstate: remove unneeded sample buffers")
we use only one sample. So, there is no need to pass the sample
pointer to intel_pstate_calc_busy. Instead, get the pointer from
cpudata. Also, remove the unused SAMPLE_COUNT macro.
While at it, reformat the first line in this function.
Signed-off-by: Stratos Karafotis <stratosk@semaphore.gr>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Acked-by: Dirk Brandewie <dirk.j.brandewie@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The command "cpupower frequency-info" can be used when using cpupower to
monitor and test processor behaviour to determine if the processor is
behaving as expected. This data can be compared to the output of
/proc/cpuinfo or the output of
/sys/devices/system/cpu/cpuX/cpufreq/scaling_available_frequencies
to determine if the cpu is in an expected state.
When doing this I noticed comparison test failures due to the way the
data is displayed in cpupower. For example,
[root@intel-s3e37-02 cpupower]# cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies
2262000 2261000 2128000 1995000 1862000 1729000 1596000 1463000 1330000
1197000 1064000
compared to
[root@intel-s3e37-02 cpupower]# cpupower frequency-info
analyzing CPU 0:
driver: acpi-cpufreq
CPUs which run at the same hardware frequency: 0
CPUs which need to have their frequency coordinated by software: 0
maximum transition latency: 10.0 us.
hardware limits: 1.06 GHz - 2.26 GHz
available frequency steps: 2.26 GHz, 2.26 GHz, 2.13 GHz, 2.00 GHz, 1.86 GHz, 1.73 GHz, 1.60 GHz, 1.46 GHz, 1.33 GHz, 1.20 GHz, 1.06 GHz
available cpufreq governors: conservative, userspace, powersave, ondemand, performance
current policy: frequency should be within 1.06 GHz and 2.26 GHz.
The governor "performance" may decide which speed to use
within this range.
current CPU frequency is 2.26 GHz (asserted by call to hardware).
boost state support:
Supported: yes
Active: yes
shows very different values for the available frequency steps. The cpupower
output rounds off values at 2 decimal points and this causes problems with
test scripts. For example, with the data above,
1.064 is 1.06
1.197 is 1.20
1.596 is 1.60
1.995 is 2.00
2.128 is 2.13
and most confusingly,
2.261 is 2.26
2.262 is 2.26
Truncating these values serves no real purpose other than making the output
pretty. Since the default has been to round off these values I am adding
a -n/--no-rounding option to the cpupower utility that will display the
data without rounding off the still significant digits.
After patch,
analyzing CPU 0:
driver: acpi-cpufreq
CPUs which run at the same hardware frequency: 0
CPUs which need to have their frequency coordinated by software: 0
maximum transition latency: 10.000 us.
hardware limits: 1.064000 GHz - 2.262000 GHz
available frequency steps: 2.262000 GHz, 2.261000 GHz, 2.128000 GHz, 1.995000 GHz, 1.862000 GHz, 1.729000 GHz, 1.596000 GHz, 1.463000 GHz, 1.330000 GHz, 1.197000 GHz, 1.064000 GHz
available cpufreq governors: conservative, userspace, powersave, ondemand, performance
current policy: frequency should be within 1.064000 GHz and 2.262000 GHz.
The governor "performance" may decide which speed to use
within this range.
current CPU frequency is 2.262000 GHz (asserted by call to hardware).
boost state support:
Supported: yes
Active: yes
Acked-by: Thomas Renninger <trenn@suse.de>
Signed-off-by: Prarit Bhargava <prarit@redhat.com>
[rjw: Subject]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The Intel 64 and IA-32 Architectures Software Developer's Manual says
that TjMax is stored in bits 23:16 of MSR_TEMPERATURE TARGET (0x1a2).
That's 8 bits, not 7, so it must be masked with 0xFF rather than 0x7F.
The manual has no mention of which values should be considered valid,
which kind of implies that they all are. Arbitrarily discarding values
outside a specific range is wrong. The upper range check had to be
fixed recently (commit 144b44b1) and the lower range check is just as
wrong. See bug #75071:
https://bugzilla.kernel.org/show_bug.cgi?id=75071
There are many Xeon processor series with TjMax of 70, 71 or 80
degrees Celsius, way below the arbitrary 85 degrees Celsius limit.
There may be other (past or future) models with even lower limits.
So drop this arbitrary check. The only value that would be clearly
invalid is 0. Everything else should be accepted.
After these changes, turbostat is aligned with what the coretemp
driver does.
Signed-off-by: Jean Delvare <jdelvare@suse.de>
Cc: Len Brown <len.brown@intel.com>
Acked-by: Guenter Roeck <linux@roeck-us.net>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Directory readahead can throw loud scary but harmless warnings
when multiblock directories are in use a specific pattern of
discontiguous blocks are found in the directory. That is, if a hole
follows a discontiguous block, it will throw a warning like:
XFS (dm-1): xfs_da_do_buf: bno 637 dir: inode 34363923462
XFS (dm-1): [00] br_startoff 637 br_startblock 1917954575 br_blockcount 1 br_state 0
XFS (dm-1): [01] br_startoff 638 br_startblock -2 br_blockcount 1 br_state 0
And dump a stack trace.
This is because the readahead offset increment loop does a double
increment of the block index - it does an increment for the loop
iteration as well as increase the loop counter by the number of
blocks in the extent. As a result, the readahead offset does not get
incremented correctly for discontiguous blocks and hence can ask for
readahead of a directory block from an offset part way through a
directory block. If that directory block is followed by a hole, it
will trigger a mapping warning like the above.
The bad readahead will be ignored, though, because the main
directory block read loop uses the correct mapping offsets rather
than the readahead offset and so will ignore the bad readahead
altogether.
Fix the warning by ensuring that the readahead offset is correctly
incremented.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Reports of a shutdown hang when fsyncing a directory have surfaced,
such as this:
[ 3663.394472] Call Trace:
[ 3663.397199] [<ffffffff815f1889>] schedule+0x29/0x70
[ 3663.402743] [<ffffffffa01feda5>] xlog_cil_force_lsn+0x185/0x1a0 [xfs]
[ 3663.416249] [<ffffffffa01fd3af>] _xfs_log_force_lsn+0x6f/0x2f0 [xfs]
[ 3663.429271] [<ffffffffa01a339d>] xfs_dir_fsync+0x7d/0xe0 [xfs]
[ 3663.435873] [<ffffffff811df8c5>] do_fsync+0x65/0xa0
[ 3663.441408] [<ffffffff811dfbc0>] SyS_fsync+0x10/0x20
[ 3663.447043] [<ffffffff815fc7d9>] system_call_fastpath+0x16/0x1b
If we trigger a shutdown in xlog_cil_push() from xlog_write(), we
will never wake waiters on the current push sequence number, so
anything waiting in xlog_cil_force_lsn() for that push sequence
number to come up will not get woken and hence stall the shutdown.
Fix this by ensuring we call wake_up_all(&cil->xc_commit_wait) in
the push abort handling, in the log shutdown code when waking all
waiters, and adding a shutdown check in the sequence completion wait
loops to ensure they abort when a wakeup due to a shutdown occurs.
Reported-by: Boris Ranto <branto@redhat.com>
Reported-by: Eric Sandeen <esandeen@redhat.com>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
truncate_setsize() removes pages from the page cache, and hence
requires page locks to be held. It is not valid to lock a page cache
page inside a transaction context as we can hold page locks when we
we reserve space for a transaction. If we do, then we expose an ABBA
deadlock between log space reservation and page locks.
That is, both the write path and writeback lock a page, then start a
transaction for block allocation, which means they can block waiting
for a log reservation with the page lock held. If we hold a log
reservation and then do something that locks a page (e.g.
truncate_setsize in xfs_setattr_size) then that page lock can block
on the page locked and waiting for a log reservation. If the
transaction that is waiting for the page lock is the only active
transaction in the system that can free log space via a commit,
then writeback will never make progress and so log space will never
free up.
This issue with xfs_setattr_size() was introduced back in 2010 by
commit fa9b227 ("xfs: new truncate sequence") which moved the page
cache truncate from outside the transaction context (what was
xfs_itruncate_data()) to inside the transaction context as a call to
truncate_setsize().
The reason truncate_setsize() was located where in this place was
that we can't shouldn't change the file size until after we are in
the transaction context and the operation will either succeed or
shut down the filesystem on failure. However, block_truncate_page()
already modifies the file contents before we enter the transaction
context, so we can't really fulfill this guarantee in any way. Hence
we may as well ensure that on success or failure, the in-memory
inode and data is truncated away and that the application cleans up
the mess appropriately.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
They do not need to be used outside fs/nfsd/nfs4state.c
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
There is almost nothing left it in, just merge it into the only file
that includes it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
There are no legitimate users outside of fs/nfsd, so move it there.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
There are no legitimate users outside of fs/nfsd, so move it there.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
The only real user of this header is fs/nfsd/nfsfh.h, so merge the
two. Various lockѕ source files used it to indirectly get other
sunrpc or nfs headers, so fix those up.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
It is large, it is used in more than one place, and it is not performance
critical. Let gcc figure out whether it should be inlined...
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
When building without CONFIG_SYSCTL, the compiler saw an unused
label. This moves the label into the #ifdef it is used under.
fs/lockd/svc.c: In function ‘init_nlm’:
fs/lockd/svc.c:626:1: warning: label ‘err_sysctl’ defined but not used [-Wunused-label]
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
New variant of iov_iter - ITER_BVEC in iter->type, backed with
bio_vec array instead of iovec one. Primitives taught to deal
with such beasts, __swap_write() switched to using that kind
of iov_iter.
Note that bio_vec is just a <page, offset, length> triple - there's
nothing block-specific about it. I've left the definition where it
was, but took it from under ifdef CONFIG_BLOCK.
Next target: ->splice_write()...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
if we'd ended up in the end of a segment, jump to the
beginning of the next one (iov_offset = 0, iov++),
rather than having the next primitive deal with that.
Ought to be folded back...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
