| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Add a virtio-based inter-processor communication bus, which enables
kernel drivers to communicate with entities, running on remote
processors, over shared memory using a simple messaging protocol.
Every pair of AMP processors share two vrings, which are used to send
and receive the messages over shared memory.
The header of every message sent on the rpmsg bus contains src and dst
addresses, which make it possible to multiplex several rpmsg channels on
the same vring.
Every rpmsg channel is a device on this bus. When a channel is added,
and an appropriate rpmsg driver is found and probed, it is also assigned
a local rpmsg address, which is then bound to the driver's callback.
When inbound messages carry the local address of a bound driver,
its callback is invoked by the bus.
This patch provides a kernel interface only; user space interfaces
will be later exposed by kernel users of this rpmsg bus.
Designed with Brian Swetland <swetland@google.com>.
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
Acked-by: Rusty Russell <rusty@rustcorp.com.au> (virtio_ids.h)
Cc: Brian Swetland <swetland@google.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: Tony Lindgren <tony@atomide.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Greg KH <greg@kroah.com>
Cc: Stephen Boyd <sboyd@codeaurora.org>
|
|
Add a remoteproc driver for OMAP4, so we can boot the dual-M3 and
and DSP subsystems.
Use the omap_device_* API to control the hardware state, and utilize
the OMAP mailbox to interrupt the remote processor when a new message
is pending (the mailbox payload is used to tell it which virtqueue was
the message placed in).
Conversely, when an inbound mailbox message arrives, tell the remoteproc
core which virtqueue is triggered.
Later we will also use the mailbox payload to signal omap-specific
events like remote crashes (which will be used to trigger remoteproc
recovery) and power management transitions. At that point we will also
extend the remoteproc core to support this.
Based on (but now quite far from) work done by Fernando Guzman Lugo
<fernando.lugo@ti.com> and Hari Kanigeri <h-kanigeri2@ti.com>.
Designed with Brian Swetland <swetland@google.com>.
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
Acked-by: Tony Lindgren <tony@atomide.com>
Cc: Brian Swetland <swetland@google.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Greg KH <greg@kroah.com>
Cc: Stephen Boyd <sboyd@codeaurora.org>
|
|
Create an rpmsg virtio device to allow message-based communication
with the remote processor (but only if supported by its firmware).
There are several advantages to provide this functionality at
the remoteproc-level:
- to support it, platforms only have to provide their own ->kick()
handler; no need to duplicate the rest of the code.
- the virtio device is created only when the remote processor is
registered and ready to go. No need to depend on initcall magic.
moreover, we only add the virtio device if the firmware really
supports it, and only after we know the supported virtio device features.
- correct device model hierarchy can be set, and that is useful
for natural power management and DMA API behavior.
- when the remote processor crashes (or removed) we only need
to remove the virtio device, and the driver core will take care of
the rest. No need to implement any out-of-bound notifiers.
- we can now easily bind the virtio device to its rproc handle, and
this way we don't need any name-based remoteproc ->get() API.
Currently we only support creating a single rpmsg virtio device per
remote processor, but later this is going to be extended to support
creating numerous virtio devices of other types too (block, net,
console...).
Designed with Brian Swetland <swetland@google.com>.
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
Cc: Brian Swetland <swetland@google.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: Tony Lindgren <tony@atomide.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Greg KH <greg@kroah.com>
Cc: Stephen Boyd <sboyd@codeaurora.org>
|
|
Expose several remote processor properties (name, state, trace buffer)
that are helpful for debugging.
This part is extracted to a separate patch just to keep the review load
down.
Designed with Brian Swetland <swetland@google.com>.
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Cc: Brian Swetland <swetland@google.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Tony Lindgren <tony@atomide.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Greg KH <greg@kroah.com>
Cc: Stephen Boyd <sboyd@codeaurora.org>
|
|
Modern SoCs typically employ a central symmetric multiprocessing (SMP)
application processor running Linux, with several other asymmetric
multiprocessing (AMP) heterogeneous processors running different instances
of operating system, whether Linux or any other flavor of real-time OS.
Booting a remote processor in an AMP configuration typically involves:
- Loading a firmware which contains the OS image
- Allocating and providing it required system resources (e.g. memory)
- Programming an IOMMU (when relevant)
- Powering on the device
This patch introduces a generic framework that allows drivers to do
that. In the future, this framework will also include runtime power
management and error recovery.
Based on (but now quite far from) work done by Fernando Guzman Lugo
<fernando.lugo@ti.com>.
ELF loader was written by Mark Grosen <mgrosen@ti.com>, based on
msm's Peripheral Image Loader (PIL) by Stephen Boyd <sboyd@codeaurora.org>.
Designed with Brian Swetland <swetland@google.com>.
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Cc: Brian Swetland <swetland@google.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Tony Lindgren <tony@atomide.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Greg KH <greg@kroah.com>
Cc: Stephen Boyd <sboyd@codeaurora.org>
|
|
|
|
In checkin
303395ac3bf3 x86: Generate system call tables and unistd_*.h from tables
the feature macros in <asm/unistd.h> were unified between 32 and 64
bits. Unfortunately 32 bits requires __ARCH_WANT_SYS_IPC and this was
inadvertently dropped.
Reported-by: Dmitry Kasatkin <dmitry.kasatkin@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Link: http://lkml.kernel.org/r/CALLzPKbeXN5gdngo8uYYU8mAow=XhrwBFBhKfG811f37BubQOg@mail.gmail.com
|
|
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
|
|
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
|
|
(struct qnx4_inode_entry *)(bh->b_data + some_offset)->di_fname
is not going to be NULL, TYVM...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
|
|
Permit key_serial() to be called with a const key pointer.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
|
|
Replace the rcu_assign_pointer() calls with rcu_assign_keypointer().
Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <jmorris@namei.org>
|
|
Fix ima_policy.c sparse "warning: dereference of noderef expression"
message, by accessing cred->uid using current_cred().
Changelog v1:
- Change __cred to just cred (based on David Howell's comment)
Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
|
|
Randy Dunlap reports that we get
arch/x86/um/shared/sysdep/ptrace.h:7:20: error: redefinition of 'regs_return_value'
arch/x86/um/shared/sysdep/ptrace.h:7:20: note: previous definition of 'regs_return_value' was here
when compiling UML for x86-64.
Stephen Rothwell root-caused it and says:
"Caused by commit d7e7528bcd45 ("Audit: push audit success and retcode
into arch ptrace.h") (another patch that was never in linux-next :-().
This file now needs protection against double inclusion."
so let's do as the man says.
Reported-by: Randy Dunlap <rdunlap@xenotime.net>
Analyzed-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Add a missing ENOMEM check.
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Mimi Zohar <zohar@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
|
|
This patch fixes an (ACPI S3) suspend regression introduced in commit
68d6e6713fcb ("tpm: Introduce function to poll for result of self test")
and occurring with an Infineon TPM and tpm_tis and tpm_infineon drivers
active.
The suspend problem occurred if the TPM was disabled and/or deactivated
and therefore the TPM_PCRRead checking the result of the (asynchronous)
self test returned an error code which then caused the tpm_tis driver to
become inactive and this then seemed to have negatively influenced the
suspend support by the tpm_infineon driver... Besides that the tpm_tis
drive may stay active even if the TPM is disabled and/or deactivated.
Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
Tested-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Rajiv Andrade <srajiv@linux.vnet.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
The type of 'make_request_fn' changed in 5a7bbad27a4 ("block: remove
support for bio remapping from ->make_request"), but the merge of the
nvme driver didn't take that into account, and as a result the driver
would compile with a warning:
drivers/block/nvme.c: In function 'nvme_alloc_ns':
drivers/block/nvme.c:1336:2: warning: passing argument 2 of 'blk_queue_make_request' from incompatible pointer type [enabled by default]
include/linux/blkdev.h:830:13: note: expected 'void (*)(struct request_queue *, struct bio *)' but argument is of type 'int (*)(struct request_queue *, struct bio *)'
It's benign, but the warning is annoying.
Reported-by: Stephen Rothwell <sfr@canb.auug.org>
Cc: Matthew Wilcox <matthew.r.wilcox@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Fix these warnings:
drivers/xen/biomerge.c:14:1: warning: data definition has no type or storage class [enabled by default]
drivers/xen/biomerge.c:14:1: warning: type defaults to 'int' in declaration of 'EXPORT_SYMBOL' [-Wimplicit-int]
drivers/xen/biomerge.c:14:1: warning: parameter names (without types) in function declaration [enabled by default]
And this build error:
ERROR: "xen_biovec_phys_mergeable" [drivers/block/nvme.ko] undefined!
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Using the correct gpio offset for setting the initial value
of gpio when setting output direction.
Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
|
|
ACPI 5.0 provides extensions to the EINJ mechanism to specify the
target for the error injection - by APICID for cpu related errors,
by address for memory related errors, and by segment/bus/device/function
for PCIe related errors. Also extensions for vendor specific error
injections.
Tested-by: Chen Gong <gong.chen@linux.intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
|
|
Function split up, should have no functional change.
Provides entry point for physically hotplugged CPUs
to initialize and activate cpuidle.
Signed-off-by: Thomas Renninger <trenn@suse.de>
CC: Deepthi Dharwar <deepthi@linux.vnet.ibm.com>
CC: Shaohua Li <shaohua.li@intel.com>
CC: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Len Brown <len.brown@intel.com>
|
|
V2: Fix typo: pr->handle -> pr, here: acpi_processor_hotadd_init(pr)
This is a very small part taken from patches which afaik
are coming from Yunhong Jiang (for a Xen not a Linus repo?).
Cleanup only: no functional change.
Advantage (beside cleanup) is that other data of the pr (acpi_processor) struct
in the acpi_processor_hotadd_init() is needed later, for example a newly
introduced flag:
pr->flags.need_hotplug_init
Signed-off-by: Thomas Renninger <trenn@suse.de>
CC: Bjorn Helgaas <bhelgaas@google.com>
CC: Jiang, Yunhong <yunhong.jiang@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
|
|
Kdump kernels leave MSI-X interrupts (as setup by the crashed kernel)
enabled. However, kdump only enables one CPU in the new environment,
thus causing tg3 to abort MSI-X setup. When the driver attempts to
enable INTA or MSI interrupt modes on a kdump kernel, interrupt
delivery fails.
This patch attempts to workaround the problem by forcing the driver
to enable a single MSI-X interrupt. In such a configuration, the
device's multivector interrupt mode must be disabled.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The logic to split up the list of datapaths into multiple Netlink messages
was simply wrong, causing the list to be terminated after the first part.
Only about the first 50 datapaths would be dumped. This fixes the
problem.
Reported-by: Paul Ingram <paul@nicira.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
In commit 4ce3c183fca (snmp: 64bit ipstats_mib for all arches), I forgot
to change the /proc/net/dev_snmp6/xxx output for IP counters.
percpu array is 64bit per counter but the folding still used the 'long'
variant, and output garbage on 32bit arches.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
JONGMAN HEO reports:
With current linus git (commit a25a2b84), I got following build error,
arch/x86/kernel/vm86_32.c: In function 'do_sys_vm86':
arch/x86/kernel/vm86_32.c:340: error: implicit declaration of function '__audit_syscall_exit'
make[3]: *** [arch/x86/kernel/vm86_32.o] Error 1
OK, I can reproduce it (32bit allmodconfig with AUDIT=y, AUDITSYSCALL=n)
It's due to commit d7e7528bcd45: "Audit: push audit success and retcode
into arch ptrace.h".
Reported-by: JONGMAN HEO <jongman.heo@samsung.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
due to broken suspend functionality.
Signed-off-by: Benjamin Larsson <benjamin@southpole.se>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
BF54x on-chip ATAPI controller allows maximum 0x1fffe bytes to be transfered
in one ATAPI transfer. So, set the max sg_tablesize to 4.
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
Configure the FSL SATA controller to the preferred, enterprise mode.
Signed-off-by: Yutaka Ando <r46913@freescale.com>
Signed-off-by: Jerry Huang <Chang-Ming.Huang@freescale.com>
CC: Jeff Garzik <jgarzik@pobox.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
This saves devices suspend/resume time.
Tested system suspend/resume with SATA IDE/AHCI mode 3 times.
Below is the time took for devices suspend/resume.
SATA mode vanilla-kernel patched-kernel
--------- --------------------- ---------------------
IDE suspend: 0.744 suspend: 0.432
(0.716, 0.768, 0.748) (0.440, 0.428, 0.428)
resume: 5.084 resume: 2.209
(5.100, 5.064, 5.088) (2.168, 2.232, 2.228)
AHCI: suspend: 0.725 suspend: 0.449
(0.740, 0.708, 0.728) (0.456, 0.448, 0.444)
resume: 2.556 resume: 1.896
(2.604, 2.492, 2.572) (1.932, 1.872, 1.884)
Signed-off-by: Lin Ming <ming.m.lin@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
|
|
The as3645a driver calls kzalloc(). Include slab.h.
Reported-by: Fredrik Lingvall <fredrik.lingvall@gmail.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
The intent here was to test if the FE_HAS_LOCK was set. The current
test is equivalent to "if (status) { ..."
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
Several control names used inconsistent capitalization or were inconsistent
in other ways. I also corrected a spelling mistake and fixed four strings
that were too long (>31 characters). Harmless, but the string is cut off when
it is returned with QUERYCTRL.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
Add an override of read_status to intercept lock status. This allows
us to switch LEDs appropriately on and off with signal un/locked.
The second phase is to override sleep to properly turn off both.
This is a hackish way to achieve that.
Thanks to Mike Krufky for his help.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Acked-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Patrick Boettcher <pboettcher@kernellabs.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
This means cut & paste from the former f. attach. But while at it write
to the right GPIO to turn on the right LED. Also turn the other two
off jsut for sure.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Acked-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Patrick Boettcher <pboettcher@kernellabs.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
The start is common for both stk7070pd and novatd specific routine.
This is just a preparation for the next patch.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Acked-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Patrick Boettcher <pboettcher@kernellabs.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
To properly support the three LEDs which are on the stick, we need
a special handling in the ->frontend_attach function. Thus let's have
a separate ->frontend_attach instead of ifs in the common one.
The hadnling itself will be added in further patches.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Acked-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Patrick Boettcher <pboettcher@kernellabs.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
The dib0700 needs a binary firmware file. This patch added the MODULE_FIRMWARE-macro.
Signed-off-by: Christoph Anton Mitterer <calestyo@scientia.net>
Signed-off-by: Patrick Boettcher <pboettcher@kernellabs.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
This small patch removes superfluous DTV_CMDs from dvb_frontend.c which were added in the initially when ISBD-T support was added.
They were there unnoticed even though compilers should have warning about those duplicates. Finally they did and now we can remove them.
Thanks to Dan Carpenter <dan.carpenter@oracle.com> for pointing that out.
Signed-off-by: Patrick Boettcher <pboettcher@kernellabs.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
queue_setup callback has been extended with struct v4l2_format *fmt
parameter in 2d86401c2c commit. This patch adds this parameter to
s5p-jpeg driver.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
Rewrite a duplicated test to test the correct value
The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)
// <smpl>
@@
expression E;
@@
(
* E
|| ... || E
|
* E
&& ... && E
)
// </smpl>
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
Rewrite a duplicated test to test the correct value
The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)
// <smpl>
@@
expression E;
@@
(
* E
|| ... || E
|
* E
&& ... && E
)
// </smpl>
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
This patch fixes the mismatch between control IDs (CID) and controls
for hflip, vflip and rotate.
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
|
|
This driver does an unconditional read of io space during module init which
causes a bad dereference on ARM. It looks to me like this is an x86 only
drivers, so restrict it to only compile on x86.
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Cc: Denis Turischev <denis@compulab.co.il>
|
|
pit_expect_msb() returns success wrongly in the below SMI scenario:
a. pit_verify_msb() has not yet seen the MSB transition.
b. we are close to the MSB transition though and got a SMI immediately after
returning from pit_verify_msb() which didn't see the MSB transition. PIT MSB
transition has happened somewhere during SMI execution.
c. returned from SMI and we noted down the 'tsc', saw the pit MSB change now and
exited the loop to calculate 'deltatsc'. Instead of noting the TSC at the MSB
transition, we are way off because of the SMI. And as the SMI happened
between the pit_verify_msb() and before the 'tsc' is recorded in the
for loop, 'delattsc' (d1/d2 in quick_pit_calibrate()) will be small and
quick_pit_calibrate() will not notice this error.
Depending on whether SMI disturbance happens while computing d1 or d2, we will
see the TSC calibrated value smaller or bigger than the expected value. As a
result, in a cluster we were seeing a variation of approximately +/- 20MHz in
the calibrated values, resulting in NTP failures.
[ As far as the SMI source is concerned, this is a periodic SMI that gets
disabled after ACPI is enabled by the OS. But the TSC calibration happens
before the ACPI is enabled. ]
To address this, change pit_expect_msb() so that
- the 'tsc' is the TSC in between the two reads that read the MSB
change from the PIT (same as before)
- the 'delta' is the difference in TSC from *before* the MSB changed
to *after* the MSB changed.
Now the delta is twice as big as before (it covers four PIT accesses,
roughly 4us) and quick_pit_calibrate() will loop a bit longer to get
the calibrated value with in the 500ppm precision. As the delta (d1/d2)
covers four PIT accesses, actual calibrated result might be closer to
250ppm precision.
As the loop now takes longer to stabilize, double MAX_QUICK_PIT_MS to 50.
SMI disturbance will showup as much larger delta's and the loop will take
longer than usual for the result to be with in the accepted precision. Or will
fallback to slow PIT calibration if it takes more than 50msec.
Also while we are at this, remove the calibration correction that aims to
get the result to the middle of the error bars. We really don't know which
direction to correct into, so remove it.
Reported-and-tested-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1326843337.5291.4.camel@sbsiddha-mobl2
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
|
Ohad Ben-Cohen
Linus Torvalds
H. Peter Anvin
Al Viro
David Howells
Mimi Zohar
Stefan Berger
Stephen Rothwell
Laxman Dewangan
Tony Luck
Thomas Renninger
Matt Carlson
Ben Pfaff
Eric Dumazet
Benjamin Larsson
Vladimir LAVALLADE
Sonic Zhang
Jerry Huang
Lin Ming
Laurent Pinchart
Sachin Kamat
Dan Carpenter
Hans Verkuil
Jiri Slaby
Patrick Boettcher
Marek Szyprowski
Kamil Debski
Julia Lawall
Grant Likely