This is a major sync up, up to ~Jewel. The highlights are:
- per-session request trees (vs a global per-client tree)
- per-session locking (vs a global per-client rwlock)
- homeless OSD session
- no ad-hoc global per-client lists
- support for pool quotas
- foundation for watch/notify v2 support
- foundation for map check (pool deletion detection) support
The switchover is incomplete: lingering requests can be setup and
teared down but aren't ever reestablished. This functionality is
restored with the introduction of the new lingering infrastructure
(ceph_osd_linger_request, linger_work, etc) in a later commit.
libceph: protect osdc->osd_lru list with a spinlock
OSD client is getting moved from the big per-client lock to a set of
per-session locks. The big rwlock would only be held for read most of
the time, so a global osdc->osd_lru needs additional protection.
create_osd() is called way too deep in the stack to be able to error
out in a sane way; a failing create_osd() just messes everything up.
The current req_notarget list solution is broken - the list is never
traversed as it's not entirely clear when to do it, I guess.
If we were to start traversing it at regular intervals and retrying
each request, we wouldn't be far off from what __GFP_NOFAIL is doing,
so allocate OSD sessions with __GFP_NOFAIL, at least until we come up
with a better fix.
Separate osdmap handling from decoding and iterating over a bag of maps
in a fresh MOSDMap message. This sets up the scene for the updated OSD
client.
Of particular importance here is the addition of pi->was_full, which
can be used to answer "did this pool go full -> not-full in this map?".
This is the key bit for supporting pool quotas.
We won't be able to downgrade map_sem for much longer, so drop
downgrade_write().
Linus Torvalds [Wed, 25 May 2016 22:54:35 +0000 (15:54 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Pull vfs xattr regression fixes from Al Viro.
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
make xattr_resolve_handlers() safe to use with NULL ->s_xattr
xattr: Fail with -EINVAL for NULL attribute names
Linus Torvalds [Wed, 25 May 2016 22:38:56 +0000 (15:38 -0700)]
Merge tag 'acpi-4.7-rc1-more' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull ACPI fix from Rafael Wysocki:
"Additional ACPI update for v4.7-rc1
Just one fix for incorrect async_synchronize_cookie() usage in the
ACPI battery driver (Chris Wilson)"
* tag 'acpi-4.7-rc1-more' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
ACPI / battery: Correctly serialise with the pending async probe
Both homeless OSD sessions and watch/notify v2, introduced in later
commits, require periodic ticks which don't depend on ->num_requests.
Schedule the initial tick from ceph_osdc_init() and reschedule from
handle_timeout() unconditionally.
libceph: move schedule_delayed_work() in ceph_osdc_init()
ceph_osdc_stop() isn't called if ceph_osdc_init() fails, so we end up
with handle_osds_timeout() running on invalid memory if any one of the
allocations fails. Call schedule_delayed_work() after everything is
setup, just before returning.
libceph: redo callbacks and factor out MOSDOpReply decoding
If you specify ACK | ONDISK and set ->r_unsafe_callback, both
->r_callback and ->r_unsafe_callback(true) are called on ack. This is
very confusing. Redo this so that only one of them is called:
->r_unsafe_callback(true), on ack
->r_unsafe_callback(false), on commit
or
->r_callback, on ack|commit
Decode everything in decode_MOSDOpReply() to reduce clutter.
libceph: drop msg argument from ceph_osdc_callback_t
finish_read(), its only user, uses it to get to hdr.data_len, which is
what ->r_result is set to on success. This gains us the ability to
safely call callbacks from contexts other than reply, e.g. map check.
Ilya Dryomov [Wed, 25 May 2016 22:29:52 +0000 (00:29 +0200)]
libceph: switch to calc_target(), part 2
The crux of this is getting rid of ceph_osdc_build_request(), so that
MOSDOp can be encoded not before but after calc_target() calculates the
actual target. Encoding now happens within ceph_osdc_start_request().
Also nuked is the accompanying bunch of pointers into the encoded
buffer that was used to update fields on each send - instead, the
entire front is re-encoded. If we want to support target->name_len !=
base->name_len in the future, there is no other way, because oid is
surrounded by other fields in the encoded buffer.
Encoding OSD ops and adding data items to the request message were
mixed together in osd_req_encode_op(). While we want to re-encode OSD
ops, we don't want to add duplicate data items to the message when
resending, so all call to ceph_osdc_msg_data_add() are factored out
into a new setup_request_data().
Replace __calc_request_pg() and most of __map_request() with
calc_target() and start using req->r_t.
ceph_osdc_build_request() however still encodes base_oid, because it's
called before calc_target() is and target_oid is empty at that point in
time; a printf in osdc_show() also shows base_oid. This is fixed in
"libceph: switch to calc_target(), part 2".
Introduce ceph_osd_request_target, containing all mapping-related
fields of ceph_osd_request and calc_target() for calculating mappings
and populating it.
Knowning just acting set isn't enough, we need to be able to record up
set as well to detect interval changes. This means returning (up[],
up_len, up_primary, acting[], acting_len, acting_primary) and passing
it around. Introduce and switch to ceph_osds to help with that.
Rename ceph_calc_pg_acting() to ceph_pg_to_up_acting_osds() and return
both up and acting sets from it.
Rename ceph_oloc_oid_to_pg() to ceph_object_locator_to_pg(). Emphasise
that returned is raw PG and return -ENOENT instead of -EIO if the pool
doesn't exist.
eversion_t is version+epoch in userspace and is encoded in that order.
ceph_eversion is defined as epoch+version in rados.h, yet we memcpy it
in __send_request(). Reoder ceph_eversion fields.
Currently ceph_object_id can hold object names of up to 100
(CEPH_MAX_OID_NAME_LEN) characters. This is enough for all use cases,
expect one - long rbd image names:
- a format 1 header is named "<imgname>.rbd"
- an object that points to a format 2 header is named "rbd_id.<imgname>"
We operate on these potentially long-named objects during rbd map, and,
for format 1 images, during header refresh. (A format 2 header name is
a small system-generated string.)
Lift this 100 character limit by making ceph_object_id be able to point
to an externally-allocated string. Apart from being able to work with
almost arbitrarily-long named objects, this allows us to reduce the
size of ceph_object_id from >100 bytes to 64 bytes.
libceph: change how osd_op_reply message size is calculated
For a message pool message, preallocate a page, just like we do for
osd_op. For a normal message, take ceph_object_id into account and
don't bother subtracting CEPH_OSD_SLAB_OPS ceph_osd_ops.
libceph: move message allocation out of ceph_osdc_alloc_request()
The size of ->r_request and ->r_reply messages depends on the size of
the object name (ceph_object_id), while the size of ceph_osd_request is
fixed. Move message allocation into a separate function that would
have to be called after ceph_object_id and ceph_object_locator (which
is also going to become variable in size with RADOS namespaces) have
been filled in:
req = ceph_osdc_alloc_request(...);
<fill in req->r_base_oid>
<fill in req->r_base_oloc>
ceph_osdc_alloc_messages(req);
Ilya Dryomov [Mon, 16 May 2016 11:18:57 +0000 (13:18 +0200)]
rbd: get/put img_request in rbd_img_request_submit()
By the time we get to checking for_each_obj_request_safe(img_request)
terminating condition, all obj_requests may be complete and img_request
ref, that rbd_img_request_submit() takes away from its caller, may be
put. Moving the next_obj_request cursor is then a use-after-free on
img_request.
It's totally benign, as the value that's read is never used, but
I think it's still worth fixing.
Linus Torvalds [Wed, 25 May 2016 22:29:21 +0000 (15:29 -0700)]
Merge tag 'pm-4.7-rc1-more' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull more power management updates from Rafael Wysocki:
"These are two stable-candidate fixes (PM core, cpuidle) and a bunch of
cpufreq cleanups.
Specifics:
- Stable-candidate cpuidle fix to make it check the right variable
when deciding whether or not to enable interrupts on the local CPU
so as to avoid enabling iterrupts too early in some cases if the
system has both coupled and per-core idle states (Daniel Lezcano).
- Stable-candidate PM core fix to make it handle failures at the
"late suspend" stage of device suspend consistently for all devices
regardless of whether or not async suspend/resume is enabled for
them (Rafael Wysocki).
- Cleanups in the cpufreq core, the schedutil governor and the
intel_pstate driver (Rafael Wysocki, Pankaj Gupta, Viresh Kumar)"
* tag 'pm-4.7-rc1-more' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
PM / sleep: Handle failures in device_suspend_late() consistently
cpufreq: schedutil: Improve prints messages with pr_fmt
cpuidle: Fix cpuidle_state_is_coupled() argument in cpuidle_enter()
cpufreq: simplified goto out in cpufreq_register_driver()
cpufreq: governor: CPUFREQ_GOV_STOP never fails
cpufreq: governor: CPUFREQ_GOV_POLICY_EXIT never fails
intel_pstate: Simplify conditional in intel_pstate_set_policy()
Al Viro [Wed, 25 May 2016 21:36:19 +0000 (17:36 -0400)]
do "fold checks into iterate_and_advance()" right
the only case when we should skip the iterate_and_advance() guts
is when nothing's left in the iterator, _not_ just when requested
amount is 0. Said guts will do nothing in the latter case anyway;
the problem we tried to deal with in the aforementioned commit is
that when there's nothing left *and* the amount requested is 0,
we might end up deferencing one iovec too many; the value we fetch
from there is discarded in that case, but theoretically it might
oops if the iovec array ends exactly at the end of page with the
next page not mapped.
Bailing out on zero size requested had an unexpected side effect -
zero-length segment in the beginning of iovec array ended up
throwing do_loop_readv_writev() into infinite spin; we do not
advance past the empty segment at all. Reproducer is trivial:
echo '#include <sys/uio.h>' >a.c
echo 'main() {char c; struct iovec v[] = {{&c,0},{&c,1}}; readv(0,v,2);}' >>a.c
cc a.c && ./a.out </proc/uptime
which should end up with the process not hanging. Probably ought to
go into LTP or xfstests...
Commit 98e9cb57 improved the xattr name checks in xattr_resolve_name but
didn't update the NULL attribute name check appropriately, so NULL
attribute names lead to NULL pointer dereferences. Turn that into
-EINVAL results instead.
Zhao Lei [Tue, 17 May 2016 09:37:38 +0000 (17:37 +0800)]
btrfs: scrub: Set bbio to NULL before calling btrfs_map_block
We usually call btrfs_put_bbio() when btrfs_map_block() failed,
btrfs_put_bbio() works right whether bbio is a valid value, or NULL.
But there is a exception, in some case, btrfs_map_block() will return
fail without touching *bbio(keeping its original value), and if bbio
was not initialized yet, invalid memory accessing will happened.
Above case is in scrub_missing_raid56_pages(), and similar case in
scrub_raid56_parity().
Merge branches 'pm-cpufreq', 'pm-cpuidle' and 'pm-core'
* pm-cpufreq:
cpufreq: schedutil: Improve prints messages with pr_fmt
cpufreq: simplified goto out in cpufreq_register_driver()
cpufreq: governor: CPUFREQ_GOV_STOP never fails
cpufreq: governor: CPUFREQ_GOV_POLICY_EXIT never fails
intel_pstate: Simplify conditional in intel_pstate_set_policy()
* pm-cpuidle:
cpuidle: Fix cpuidle_state_is_coupled() argument in cpuidle_enter()
* pm-core:
PM / sleep: Handle failures in device_suspend_late() consistently
Mark Bloch [Wed, 18 May 2016 13:42:43 +0000 (16:42 +0300)]
IB/IPoIB: Allow setting the device address
In IB networks, and specifically in IPoIB/rdmacm traffic, the device
address of an IPoIB interface is used as a means to exchange information
between nodes needed for communication.
Currently an IPoIB interface will always be created with a device
address based on its node GUID without a way to change that.
This change adds the ability to set the device address of an IPoIB
interface by value. We use the set mac address ndo to do that.
The flow should be broken down to two:
1) The GID value is already in the GID table,
in this case the interface will be able to set carrier up.
2) The GID value is not yet in the GID table,
in this case the interface won't try to join the multicast group
and will wait (listen on GID_CHANGE event) until the GID is inserted.
In order to track those changes, we add a new flag:
* IPOIB_FLAG_DEV_ADDR_SET.
When set, it means the dev_addr is a based on a value in the gid
table. this bit will be cleared upon a dev_addr change triggered
by the user and set after validation.
Per IB spec the port GUID can't change if the module is loaded.
port GUID is the basis for GID at index 0 which is the basis for
the default device address of a ipoib interface.
The issue is that there are devices that don't follow the spec,
they change the port GUID while HCA is powered on, so in order
not to break userspace applications. We need to check if the
user wanted to control the device address and we assume that
if he sets the device address back to be based on GID index 0,
he no longer wishs to control it.
In order to track this, we add an additional flag:
* IPOIB_FLAG_DEV_ADDR_CTRL
When setting the device address, there is no validation of the upper
twelve bytes of the device address (flags, qpn, subnet prefix) as those
bytes are not under the control of the user.
Erez Shitrit [Wed, 25 May 2016 19:02:07 +0000 (22:02 +0300)]
IB/ipoib: Support SendOnlyFullMember MCG for SendOnly join
Check (via an SA query) if the SM supports the new option for SendOnly
multicast joins.
If the SM supports that option it will use the new join state to create
such multicast group.
If SendOnlyFullMember is supported, we wouldn't use faked FullMember state
join for SendOnly MCG, use the correct state if supported.
This check is performed at every invocation of mcast_restart task, to be
sure that the driver stays in sync with the current state of the SM.
Erez Shitrit [Wed, 25 May 2016 19:02:06 +0000 (22:02 +0300)]
IB/core: Support new type of join-state for multicast
There are four types for MCG, FullMember, NonMember, SendOnlyNonMember,
and the new added type: SendOnlyFullMember.
Add support for the new SendOnlyFullMember join state.
The new type allows host to send join request as sendonly, it will cause the
group to be created but without getting packets from this multicast back to the
host.
Erez Shitrit [Wed, 25 May 2016 19:02:05 +0000 (22:02 +0300)]
IB/SA Agent: Add support for SA agent get ClassPortInfo
New SA query function to return the ClassPortInfo struct from the SA.
If the SM supports FullMemberSendOnly mode for MCG's, it sets a
capability bit in the capability_mask2 field of the response.
Erez Shitrit [Wed, 25 May 2016 19:02:04 +0000 (22:02 +0300)]
IB/core: Introduce capabilitymask2 field in ClassPortInfo mad
Change struct ib_class_port_info to conform to IB Spec 1.3
That in order to get specific capability mask from ClassPortInfo mad.
>From the IB Spec, ClassPortInfo section:
"CapabilityMask2 Bits 0-26: Additional class-specific capabilities...
RespTimeValue the rest 5 bits"
The new struct now has one field for capabilitymask2 (previously was the
reserved field) and the resp_time field.
And it fixes up qib and srpt, use of the field repurposed to be used as
capabilitymask2:
IB/qib: Change pma_get_classportinfo
IB/srpt: Adjust the use of ib_class_port_info
Mark Brown [Wed, 25 May 2016 18:18:00 +0000 (19:18 +0100)]
Merge tag 'asoc-v4.7' into asoc-linus
ASoC: Updates for v4.7
The updates this time around are almost all driver code:
- Further slow progress on the topology code.
- Substantial updates and improvements for the da7219, es8328, fsl-ssi
Intel and rcar drivers.
# gpg: Signature made Mon 16 May 2016 12:08:43 BST using RSA key ID 5D5487D0
# gpg: Good signature from "Mark Brown <[email protected]>"
# gpg: aka "Mark Brown <[email protected]>"
# gpg: aka "Mark Brown <[email protected]>"
# gpg: aka "Mark Brown <[email protected]>"
# gpg: aka "Mark Brown <[email protected]>"
# gpg: aka "Mark Brown <[email protected]>"
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg: There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 3F25 68AA C269 98F9 E813 A1C5 C3F4 36CA 30F5 D8EB
# Subkey fingerprint: ADE6 68AA 6757 18B5 9FE2 9FEA 24D6 8B72 5D54 87D0
Liu Bo [Sat, 14 May 2016 00:06:59 +0000 (17:06 -0700)]
Btrfs: free sys_array eb as soon as possible
While reading sys_chunk_array in superblock, btrfs creates a temporary
extent buffer. Since we don't use it after finishing reading
sys_chunk_array, we don't need to keep it in memory.
Linus Torvalds [Wed, 25 May 2016 17:40:15 +0000 (10:40 -0700)]
Merge tag 'pwm/for-4.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/thierry.reding/linux-pwm
Pull pwm updates from Thierry Reding:
"This set of changes introduces an atomic API to the PWM subsystem.
This is influenced by the DRM atomic API that was introduced a while
back, though it is obviously a lot simpler. The fundamental idea
remains the same, though: drivers provide a single callback to
implement the atomic configuration of a PWM channel.
As a side-effect the PWM subsystem gains the ability for initial state
retrieval, so that the logical state mirrors that of the hardware.
Many use-cases don't care about this, but for others it is essential.
These new features require changes in all users, which these patches
take care of. The core is transitioned to use the atomic callback if
available and provides a fallback mechanism for other drivers.
Changes to transition users and drivers to the atomic API are
postponed to v4.8"
* tag 'pwm/for-4.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/thierry.reding/linux-pwm: (30 commits)
pwm: Add information about polarity, duty cycle and period to debugfs
pwm: Switch to the atomic API
pwm: Update documentation
pwm: Add core infrastructure to allow atomic updates
pwm: Add hardware readout infrastructure
pwm: Move the enabled/disabled info into pwm_state
pwm: Introduce the pwm_state concept
pwm: Keep PWM state in sync with hardware state
ARM: Explicitly apply PWM config extracted from pwm_args
drm: i915: Explicitly apply PWM config extracted from pwm_args
input: misc: pwm-beeper: Explicitly apply PWM config extracted from pwm_args
input: misc: max8997: Explicitly apply PWM config extracted from pwm_args
backlight: lm3630a: explicitly apply PWM config extracted from pwm_args
backlight: lp855x: Explicitly apply PWM config extracted from pwm_args
backlight: lp8788: Explicitly apply PWM config extracted from pwm_args
backlight: pwm_bl: Use pwm_get_args() where appropriate
fbdev: ssd1307fb: Use pwm_get_args() where appropriate
regulator: pwm: Use pwm_get_args() where appropriate
leds: pwm: Use pwm_get_args() where appropriate
input: misc: max77693: Use pwm_get_args() where appropriate
...
Since the patch "NFS: Allow multiple commit requests in flight per file"
we can run multiple simultaneous commits on the same inode. This
introduced a race over collecting pages to commit that made it possible
to call nfs_init_commit() with an empty list - which causes crashes like
the one below.
The fix is to catch this race and avoid calling nfs_init_commit and
initiate_commit when there is no work to do.
- Use cfg_size to avoid probing extended config space (Alexey
Kardashevskiy)
* tag 'vfio-v4.7-rc1' of git://github.com/awilliam/linux-vfio:
vfio_pci: Test for extended capabilities if config space > 256 bytes
vfio_iommu_spapr_tce: Remove unneeded iommu_group_get_iommudata
vfio/pci: Add test for BAR restore
vfio/pci: Hide broken INTx support from user
Linus Torvalds [Wed, 25 May 2016 16:37:50 +0000 (09:37 -0700)]
Merge tag 'drm-4.7-rc1-headers-fix' of git://people.freedesktop.org/~airlied/linux
Pull header warning fix from Dave Airlie:
"Here is the C++ guards warning fix from Arnd"
[ Background: there are 'extern "C" { }' guards in include/uapi for the
GPU headers.
They should arguably be wrapped somehow, but as it is they caused
checkpatch to warn because it would trigger on the 'extern' and think
it's exporting a function or variable from the kernel to user space.
This just fixes checkpatch. Whether we wrap the C++ guards some way
in the future will be an independent issue. ]
* tag 'drm-4.7-rc1-headers-fix' of git://people.freedesktop.org/~airlied/linux:
headers_check: don't warn about c++ guards
Linus Torvalds [Wed, 25 May 2016 16:27:52 +0000 (09:27 -0700)]
Merge branch 'parisc-4.7-1' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux
Pull parisc updates from Helge Deller:
- Add native high-resolution timing code for sched_clock() and other
timing functions based on the processor internal cr16 cycle counters
- Add syscall tracepoint support
- Add regset support
- Speed up get_user() and put_user() functions
- Updated futex.h to match generic implementation (John David Anglin)
- A few smaller ftrace build fixes
- Fixed thuge-gen kernel self test to utilize architectured MAP_HUGETLB
value
- Added parisc architecture to seccomp_bpf kernel self test
- Various typo fixes (Andrea Gelmini)
* 'parisc-4.7-1' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
parisc: Whitespace cleanups in unistd.h
parisc: Use long jump to reach ftrace_return_to_handler()
parisc: Fix typo in fpudispatch.c
parisc: Fix typos in eisa_eeprom.h
parisc: Fix typo in ldcw.h
parisc: Fix typo in pdc.h
parisc: Update futex.h to match generic implementation
parisc: Merge ftrace C-helper and assembler functions into .text.hot section
selftests/thuge-gen: Use platform specific MAP_HUGETLB value
parisc: Add native high-resolution sched_clock() implementation
parisc: Add ARCH_TRACEHOOK and regset support
parisc: Add 64bit get_user() and put_user() for 32bit kernel
parisc: Simplify and speed up get_user() and put_user()
parisc: Add syscall tracepoint support
Janosch Frank [Wed, 18 May 2016 11:26:25 +0000 (13:26 +0200)]
tools: kvm_stat: Add comments
A lot of the code works with the perf events about which only sparse
documentation was available until 2012. Having that information now,
we can clarify what is done in the code.
Janosch Frank [Wed, 18 May 2016 11:26:24 +0000 (13:26 +0200)]
tools: kvm_stat: Introduce pid monitoring
Having stats for single VMs can help to determine the problem of a VM
without the need of running other tools like perf.
The tracepoints already allowed pid level monitoring, but kvm_stat
didn't have support for it till now. Support for the newly implemented
debugfs vm monitoring was also implemented.
Janosch Frank [Wed, 18 May 2016 11:26:23 +0000 (13:26 +0200)]
KVM: Create debugfs dir and stat files for each VM
This patch adds a kvm debugfs subdirectory for each VM, which is named
after its pid and file descriptor. The directories contain the same
kind of files that are already in the kvm debugfs directory, but the
data exported through them is now VM specific.
This makes the debugfs kvm data a convenient alternative to the
tracepoints which already have per VM data. The debugfs data is easy
to read and low overhead.
kvm_stat script is failing to execute on powerpc :
# ./kvm_stat
Traceback (most recent call last):
File "./kvm_stat", line 825, in <module>
main()
File "./kvm_stat", line 813, in main
providers = get_providers(options)
File "./kvm_stat", line 778, in get_providers
providers.append(TracepointProvider())
File "./kvm_stat", line 416, in __init__
self.filters = get_filters()
File "./kvm_stat", line 315, in get_filters
if ARCH.exit_reasons:
AttributeError: 'ArchPPC' object has no attribute 'exit_reasons'
This is because, its trying to access a non-defined attribute.
Also, the IOCTL number of RESET is incorrect for powerpc. The correct
number has been added.
Janosch Frank [Wed, 18 May 2016 11:26:21 +0000 (13:26 +0200)]
tools: Add kvm_stat vm monitor script
This tool displays kvm vm exit statistics to ease vm monitoring. It
takes its data from the kvm debugfs files or the vm tracepoints and
outputs them as a curses ui or simple text.
It was moved from qemu, as it is dependent on the kernel whereas qemu
works with a large number of kernel versions, some of which may break
the script.
Roman Kagan [Wed, 18 May 2016 14:48:20 +0000 (17:48 +0300)]
kvm:vmx: more complete state update on APICv on/off
The function to update APICv on/off state (in particular, to deactivate
it when enabling Hyper-V SynIC) is incomplete: it doesn't adjust
APICv-related fields among secondary processor-based VM-execution
controls. As a result, Windows 2012 guests get stuck when SynIC-based
auto-EOI interrupt intersected with e.g. an IPI in the guest.
In addition, the MSR intercept bitmap isn't updated every time "virtualize
x2APIC mode" is toggled. This path can only be triggered by a malicious
guest, because Windows didn't use x2APIC but rather their own synthetic
APIC access MSRs; however a guest running in a SynIC-enabled VM could
switch to x2APIC and thus obtain direct access to host APIC MSRs
(CVE-2016-4440).
Jeff Mahoney [Tue, 24 May 2016 17:47:44 +0000 (13:47 -0400)]
drm/amd: add Kconfig dependency for ACP on DRM_AMDGPU
The DRM_AMD_ACP option doesn't have any dependencies and selects
MFD_CORE, which results in MFD_CORE=y. Since the code is only called
from DRM_AMDGPU, it should depend on it. Adding the dependency results
in MFD_CORE being selected as a module again if amdgpu is also a module.
Mario Kleiner [Tue, 24 May 2016 16:12:43 +0000 (18:12 +0200)]
drm/amdgpu: Fix hdmi deep color support.
When porting the hdmi deep color detection code from
radeon-kms to amdgpu-kms apparently some kind of
copy and paste error happened, attaching an else
branch to the wrong if statement.
The result is that hdmi deep color mode is always
disabled, regardless of gpu and display capabilities and
user wishes, as the code mistakenly thinks that the display
doesn't provide the required max_tmds_clock limit and falls
back to 8 bpc.
This patch fixes deep color support, as tested on a
R9 380 Tonga Pro + suitable display, and should be
backported to all kernels with amdgpu-kms support.
Peter Zijlstra [Mon, 23 May 2016 09:19:07 +0000 (11:19 +0200)]
sched/core: Fix remote wakeups
Commit:
b5179ac70de8 ("sched/fair: Prepare to fix fairness problems on migration")
... introduced a bug: Mike Galbraith found that it introduced a
performance regression, while Paul E. McKenney reported lost
wakeups and bisected it to this commit.
The reason is that I mis-read ttwu_queue() such that I assumed any
wakeup that got a remote queue must have had the task migrated.
Since this is not so; we need to transfer this information between
queueing the wakeup and actually doing the wakeup. Use a new
task_struct::sched_flag for this, we already write to
sched_contributes_to_load in the wakeup path so this is a hot and
modified cacheline.
Dave Airlie [Wed, 25 May 2016 02:36:20 +0000 (12:36 +1000)]
Merge tag 'imx-drm-fixes-2016-05-24' of git://git.pengutronix.de/git/pza/linux into drm-next
imx-drm probing fix
Commit 950b410dd1ab ("gpu: ipu-v3: Fix imx-ipuv3-crtc module autoloading")
broke probing of the imx-drm driver in the non-modular case because the
unset dev->of_node during probing of imx-ipuv3-crtc would cause the
component matching to fail. This patch patch instead matches against
an of_node pointer stored in platform data, allowing dev->of_node to
be left unset for the platform probed imx-ipuv3-crtc devices.
* tag 'imx-drm-fixes-2016-05-24' of git://git.pengutronix.de/git/pza/linux:
drm/imx: Match imx-ipuv3-crtc components using device node in platform data
Linus Torvalds [Tue, 24 May 2016 22:50:58 +0000 (15:50 -0700)]
Merge tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Pull ARM SoC fixes from Arnd Bergmann:
"This is a first set of bug fixes on top of what was merged for 4.7.
Two patches for lpc32xx address a harmless build warning that was just
introduced, one patch for the mediatek soc driver fixes a warning for
arm64, and the pxa changes are minor cleanups that should have been
part of the original pull requests but that I forgot to apply to the
cleanup-fixes branch earlier"
Linus Torvalds [Tue, 24 May 2016 22:46:06 +0000 (15:46 -0700)]
Merge tag 'armsoc-late' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Pull ARM SoC late DT updates from Arnd Bergmann:
"This is a collection of a few late fixes and other misc stuff that had
dependencies on things being merged from other trees.
The Renesas R-Car power domain handling, and the Nvidia Tegra USB
support both hand notable changes that required changing the DT
binding in a way that only provides compatibility with old DT blobs on
new kernels but not vice versa. As a consequence, the DT changes are
based on top of the driver changes and are now in this branch.
For NXP i.MX and Samsung Exynos, the changes in here depend on other
changes that got merged through the clk maintainer tree"
* tag 'armsoc-late' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (35 commits)
ARM: dts: exynos: Add support of Bus frequency using VDD_INT for exynos5422-odroidxu3
ARM: dts: exynos: Add bus nodes using VDD_INT for Exynos542x SoC
ARM: dts: exynos: Add NoC Probe dt node for Exynos542x SoC
ARM: dts: exynos: Add support of bus frequency for exynos4412-trats/odroidu3
ARM: dts: exynos: Expand the voltage range of buck1/3 regulator for exynos4412-odroidu3
ARM: dts: exynos: Add support of bus frequency using VDD_INT for exynos3250-rinato
ARM: dts: exynos: Add exynos4412-ppmu-common dtsi to delete duplicate PPMU nodes
ARM: dts: exynos: Add bus nodes using VDD_MIF for Exynos4210
ARM: dts: exynos: Add bus nodes using VDD_INT for Exynos4x12
ARM: dts: exynos: Add bus nodes using VDD_MIF for Exynos4x12
ARM: dts: exynos: Add bus nodes using VDD_INT for Exynos3250
ARM: dts: exynos: Add DMC bus frequency for exynos3250-rinato/monk
ARM: dts: exynos: Add DMC bus node for Exynos3250
ARM: tegra: Enable XUSB on Nyan
ARM: tegra: Enable XUSB on Jetson TK1
ARM: tegra: Enable XUSB on Venice2
ARM: tegra: Add Tegra124 XUSB controller
ARM: tegra: Move Tegra124 to the new XUSB pad controller binding
ARM: dts: r8a7794: Use SYSC "always-on" PM Domain
ARM: dts: r8a7793: Use SYSC "always-on" PM Domain
...
Linus Torvalds [Tue, 24 May 2016 22:24:37 +0000 (15:24 -0700)]
Merge tag 'asm-generic-4.7' of git://git.kernel.org/pub/scm/linux/kernel/git/arnd/asm-generic
Pull asm-generic cleanup from Arnd Bergmann:
"I have only one patch for asm-generic in this release, this one is
from James Hogan and updates the generic system call table for
renameat2 so we don't need to provide both renameat and renameat2 in
newly added architectures"
* tag 'asm-generic-4.7' of git://git.kernel.org/pub/scm/linux/kernel/git/arnd/asm-generic:
asm-generic: Drop renameat syscall from default list
Linus Torvalds [Tue, 24 May 2016 21:39:20 +0000 (14:39 -0700)]
Merge tag 'nfsd-4.7' of git://linux-nfs.org/~bfields/linux
Pull nfsd updates from Bruce Fields:
"A very quiet cycle for nfsd, mainly just an RDMA update from Chuck
Lever"
* tag 'nfsd-4.7' of git://linux-nfs.org/~bfields/linux:
sunrpc: fix stripping of padded MIC tokens
svcrpc: autoload rdma module
svcrdma: Generalize svc_rdma_xdr_decode_req()
svcrdma: Eliminate code duplication in svc_rdma_recvfrom()
svcrdma: Drain QP before freeing svcrdma_xprt
svcrdma: Post Receives only for forward channel requests
svcrdma: Remove superfluous line from rdma_read_chunks()
svcrdma: svc_rdma_put_context() is invoked twice in Send error path
svcrdma: Do not add XDR padding to xdr_buf page vector
svcrdma: Support IPv6 with NFS/RDMA
nfsd: handle seqid wraparound in nfsd4_preprocess_layout_stateid
Remove unnecessary allocation
Linus Torvalds [Tue, 24 May 2016 19:55:26 +0000 (12:55 -0700)]
Merge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4
Pull ext4 updates from Ted Ts'o:
"Fix a number of bugs, most notably a potential stale data exposure
after a crash and a potential BUG_ON crash if a file has the data
journalling flag enabled while it has dirty delayed allocation blocks
that haven't been written yet. Also fix a potential crash in the new
project quota code and a maliciously corrupted file system.
In addition, fix some DAX-specific bugs, including when there is a
transient ENOSPC situation and races between writes via direct I/O and
an mmap'ed segment that could lead to lost I/O.
Finally the usual set of miscellaneous cleanups"
* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (23 commits)
ext4: pre-zero allocated blocks for DAX IO
ext4: refactor direct IO code
ext4: fix race in transient ENOSPC detection
ext4: handle transient ENOSPC properly for DAX
dax: call get_blocks() with create == 1 for write faults to unwritten extents
ext4: remove unmeetable inconsisteny check from ext4_find_extent()
jbd2: remove excess descriptions for handle_s
ext4: remove unnecessary bio get/put
ext4: silence UBSAN in ext4_mb_init()
ext4: address UBSAN warning in mb_find_order_for_block()
ext4: fix oops on corrupted filesystem
ext4: fix check of dqget() return value in ext4_ioctl_setproject()
ext4: clean up error handling when orphan list is corrupted
ext4: fix hang when processing corrupted orphaned inode list
ext4: remove trailing \n from ext4_warning/ext4_error calls
ext4: fix races between changing inode journal mode and ext4_writepages
ext4: handle unwritten or delalloc buffers before enabling data journaling
ext4: fix jbd2 handle extension in ext4_ext_truncate_extend_restart()
ext4: do not ask jbd2 to write data for delalloc buffers
jbd2: add support for avoiding data writes during transaction commits
...
Arnd Bergmann [Thu, 19 May 2016 08:31:31 +0000 (10:31 +0200)]
ARM: lpc32xx: fix NR_IRQS confict
With the change to sparse IRQs, the lpc32xx platform gets a warning about
conflicting macros:
In file included from arch/arm/mach-lpc32xx/irq.c:31:0:
arch/arm/mach-lpc32xx/include/mach/irqs.h:115:0: warning: "NR_IRQS" redefined
#define NR_IRQS 96
arch/arm/include/asm/irq.h:9:0: note: this is the location of the previous definition
#define NR_IRQS NR_IRQS_LEGACY
One such instance was in the old irq driver that is now removed by
the previous patch, but any other file including mach/irqs.h still
has the issue. Since none of them use this constant, we can just
remove the old definition.
Mark Bloch [Thu, 19 May 2016 14:12:36 +0000 (17:12 +0300)]
IB/core: Add IP to GID netlink offload
There is an assumption that rdmacm is used only between nodes
in the same IB subnet, this why ARP resolution can be used to turn
IP to GID in rdmacm.
When dealing with IB communication between subnets this assumption
is no longer valid. ARP resolution will get us the next hop device
address and not the peer node's device address.
To solve this issue, we will check user space if it can provide the
GID of the peer node, and fail if not.
We add a sequence number to identify each request and fill in the GID
upon answer from userspace.
Mark Bloch [Thu, 19 May 2016 14:12:35 +0000 (17:12 +0300)]
IB/core: Register SA ibnl client during ib_core initialization
Move SA ibnl client registration to ib_core module init.
This will allow us to register a single client to handle
all RDMA_NL_LS operations and make it SA independent.
Mark Bloch [Thu, 19 May 2016 14:12:34 +0000 (17:12 +0300)]
IB/netlink: Add a new local service operation
This commits adds a new RDMA local service operation:
- IP to GID resolution.
The client request would include the ifindex of the outgoing interface
and would place in an attribute (LS_NLA_TYPE_IPV4 or LS_NLA_TYPE_IPV6)
the destnation IP.
The local service would answer with a message that has the attribute:
- LS_NLA_TYPE_DGID - The destination GID.
Linus Torvalds [Tue, 24 May 2016 18:12:32 +0000 (11:12 -0700)]
Merge tag 'spi-v4.7' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Pull spi updates from Mark Brown:
"Another quiet release for SPI, almost entirely driver specific changes
with the diffstat dominated by two new drivers which are about two
thirds of it in terms of lines of code:
- new drivers for PIC32 standard and SQI controllers
- the Cadence driver has had runtime PM support added and quite a few
fixes and cleanups
- flash-specific accelerated path support now has a feature query
interface
- the pxa2xx driver has been moved to use the core DMA mapping support"
* tag 'spi-v4.7' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi: (48 commits)
spi: pic32-sqi: Fix linker error, undefined reference to `bad_dma_ops'
spi: dw-pci: Spelling s/paltforms/platforms/g
spi: pic32-sqi: Remove pic32_sqi_setup and pic32_sqi_cleanup
spi: Fix simple typo s/impelment/implement
spi: rockchip: potential NULL dereference on error
spi: zynqmp: disable clocks in error paths
spi: Drop unnecessary dependencies on relaxed I/O accessors
spi: qup: Add spi_master_put in remove function
spi: qup: Handle clocks in pm_runtime suspend and resume
spi: st-ssc4: Fix missing spi_master_put in spi_st_probe error paths
spi: st-ssc4: Allow compile test build
spi: omap2-mcspi: Use dma_request_chan() for requesting DMA channel
spi: davinci: Use dma_request_chan() for requesting DMA channel
spi: pic32: Fix checking return value of devm_ioremap_resource
spi: spi-fsl-dspi: Update DT binding documentation
spi: Drop duplicate code to set master->dev.parent
spi: pic32: Set proper bits_per_word_mask
spi: return error if kmap'd buffers passed to spi_map_buf()
spi: core: add hook flash_read_supported to spi_master
spi: pic32-sqi: silence array overflow warning
...
Linus Torvalds [Tue, 24 May 2016 18:00:20 +0000 (11:00 -0700)]
Merge tag 'for-linus-20160523' of git://git.infradead.org/linux-mtd
Pull MTD updates from Brian Norris:
"First cycle with Boris as NAND maintainer! Many (most) bullets stolen
from him.
Generic:
- Migrated NAND LED trigger to be a generic MTD trigger
NAND:
- Introduction of the "ECC algorithm" concept, to avoid overloading
the ECC mode field too much more
- Replaced the nand_ecclayout infrastructure with something a little
more flexible (finally!) and future proof
- Rework of the OMAP GPMC and NAND drivers; the TI folks pulled some
of this into their own tree as well
- Prepare the sunxi NAND driver to receive DMA support
- Handle bitflips in erased pages on GPMI revisions that do not
support this in hardware.
SPI NOR:
- Start using the spi_flash_read() API for SPI drivers that support
it (i.e., SPI drivers with special memory-mapped flash modes)
And other small scattered improvments"
* tag 'for-linus-20160523' of git://git.infradead.org/linux-mtd: (155 commits)
mtd: spi-nor: support GigaDevice gd25lq64c
mtd: nand_bch: fix spelling of "probably"
mtd: brcmnand: respect ECC algorithm set by NAND subsystem
gpmi-nand: Handle ECC Errors in erased pages
Documentation: devicetree: deprecate "soft_bch" nand-ecc-mode value
mtd: nand: add support for "nand-ecc-algo" DT property
mtd: mtd: drop NAND_ECC_SOFT_BCH enum value
mtd: drop support for NAND_ECC_SOFT_BCH as "soft_bch" mapping
mtd: nand: read ECC algorithm from the new field
mtd: nand: fsmc: validate ECC setup by checking algorithm directly
mtd: nand: set ECC algorithm to Hamming on fallback
staging: mt29f_spinand: set ECC algorithm explicitly
CRIS v32: nand: set ECC algorithm explicitly
mtd: nand: atmel: set ECC algorithm explicitly
mtd: nand: davinci: set ECC algorithm explicitly
mtd: nand: bf5xx: set ECC algorithm explicitly
mtd: nand: omap2: Fix high memory dma prefetch transfer
mtd: nand: omap2: Start dma request before enabling prefetch
mtd: nandsim: add __init attribute
mtd: nand: move of_get_nand_xxx() helpers into nand_base.c
...
Linus Torvalds [Tue, 24 May 2016 17:22:34 +0000 (10:22 -0700)]
Merge tag 'for-linus-4.7-rc0-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip
Pull xen bug fixes from David Vrabel.
* tag 'for-linus-4.7-rc0-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen: use same main loop for counting and remapping pages
xen/events: Don't move disabled irqs
xen/x86: actually allocate legacy interrupts on PV guests
Xen: don't warn about 2-byte wchar_t in efi
xen/gntdev: reduce copy batch size to 16
xen/x86: don't lose event interrupts
Linus Torvalds [Tue, 24 May 2016 16:46:45 +0000 (09:46 -0700)]
Merge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Pull virtio updates from Michael Tsirkin:
"Looks like a quiet cycle for virtio. There's a new inorder option for
the ringtest tool, and a bugfix for balloon for ppc platforms when
using virtio 1 mode"
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
ringtest: pass buf != NULL
virtio_balloon: fix PFN format for virtio-1
virtio: add inorder option
Linus Torvalds [Tue, 24 May 2016 16:19:38 +0000 (09:19 -0700)]
Merge tag 'microblaze-4.7-rc1' of git://git.monstr.eu/linux-2.6-microblaze
Pull Microblaze updates from Michal Simek:
- Wire-up new syscalls
- Fix link error
* tag 'microblaze-4.7-rc1' of git://git.monstr.eu/linux-2.6-microblaze:
microblaze: pci: export isa_io_base to fix link errors
microblaze: Wire up userfaultfd, membarrier, mlock2 syscalls
Fixes: b955150ea784 ('RDMA/cxgb3: When a user QP is marked in error, also mark the CQs in error') Signed-off-by: Honggang Li <[email protected]> Reviewed-by: Leon Romanovsky <[email protected]> Reviewed-by: Steve Wise <[email protected]> Signed-off-by: Doug Ledford <[email protected]>
Jiri Kosina [Tue, 24 May 2016 14:38:15 +0000 (16:38 +0200)]
bcache: bch_allocator_thread() is not freezable
bch_allocator_thread() is calling try_to_freeze(), but that's just an
expensive no-op given the fact that the thread is not marked freezable.
Bucket allocator has to be up and running to the very last stages of the
suspend, as the bcache I/O that's in flight (think of writing an
hibernation image to a swap device served by bcache).
projects / linux.git / log