In the OPTEE_SMC_CALL_GET_OS_REVISION request, the previously reserved
parameter a2 is now documented as being an optional build identifier
(such as an SCM revision or commit ID, for instance).
A new structure optee_smc_call_get_os_revision_result is introduced to
be used when querying the secure OS version, instead of re-using the
struct defined for OPTEE_SMC_CALLS_REVISION.
Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org>
Reviewed-by: Matthias Brugger <mbruger@suse.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
The privileged dev id range is [TEE_NUM_DEVICES / 2, TEE_NUM_DEVICES).
The non-privileged dev id range is [0, TEE_NUM_DEVICES / 2).
So when finding a slot for them, need to use different max value.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Jens Wiklander <jens.wiklander@linaro.org>
The dvb_attach()/dvb_detach() methods are ugly hacks designed
to keep using the I2C low-level API. The proper way is to
do I2C bus bindings instead.
Several modules were already converted to use it. Yet,
it is painful to use it, as lots of code need to be
duplicated.
Make it easier by providing two new helper functions:
- dvb_module_probe()
- dvb_module_release()
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
With CONFIG_KASAN, the init function uses a large amount of kernel stack:
drivers/media/usb/em28xx/em28xx-dvb.c: In function 'em28xx_dvb_init.part.4':
drivers/media/usb/em28xx/em28xx-dvb.c:2061:1: error: the frame size of 3232 bytes is larger than 2048 bytes [-Werror=frame-larger-than=]
Using gcc-7 with -fsanitize-address-use-after-scope makes this even worse:
drivers/media/usb/em28xx/em28xx-dvb.c: In function 'em28xx_dvb_init':
drivers/media/usb/em28xx/em28xx-dvb.c:2069:1: error: the frame size of 4280 bytes is larger than 3072 bytes [-Werror=frame-larger-than=]
By splitting out each part of the switch/case statement that has its own local
variables into a separate function, no single one of them uses more than 500 bytes,
and with a noinline_for_stack annotation we can ensure that they are not merged
back together.
[mchehab@s-opensource.com: fix conflict with changeset
be7fd3c3a8 ("media: em28xx: Hauppauge DualHD second tuner functionality")]
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Mauro Carvalho Chehab <>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Make sure to not disallow the chips on adapters that are not capable
of reading the device id, but also make sure to check the device id
before writing to the chip.
Tested-by: Adrian Fiergolski <adrian.fiergolski@cern.ch>
Signed-off-by: Peter Rosin <peda@axentia.se>
Can be used during probe to double check that the probed device is
what is expected.
Loosely based on code from Adrian Fiergolski <adrian.fiergolski@cern.ch>.
Tested-by: Adrian Fiergolski <adrian.fiergolski@cern.ch>
Reviewed-by: Wolfram Sang <wsa@the-dreams.de>
Signed-off-by: Peter Rosin <peda@axentia.se>
If the I2C speed is too slow, it should wait more for an
answer.
While here, change disconnected type from char to unsigned
int, just like all other bitmask fields there at em28xx
struct.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
There are several em28xx static structs that can now be constified.
That caused a significant reduction at data segment:
Before:
text data bss dec hex filename
85017 59588 576 145181 2371d drivers/media/usb/em28xx/em28xx.o
After:
text data bss dec hex filename
112345 32292 576 145213 2373d drivers/media/usb/em28xx/em28xx.o
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Writing at the device's struct is evil, as two em28xx devices
may be using it. So, stop abusing it, storing the values
inside struct em28xx_dev.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The logic there should be called on two places. Also,
ideally, it should not be modifying the device struct.
So, change the logic accordingly.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
While coherent memory is cheap on x86, it may cause performance
impacts on other archs. As we don't have any good reason to
use it, let's change the logic by allocating memory via kmalloc()
and letting the USB core to do the DMA mapping and memory free
for us.
While here, also fixes an issue that it was not de-allocating
memories if something gets wrong during memory block
allocation.
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Both lgdt33606a_release and lgdt3306a_remove kfree state, but _release is
called first, then _remove operates on states members before kfree'ing it.
This can lead to random oops/GPF/etc on USB disconnect.
Signed-off-by: Brad Love <brad@nextdimension.cc>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
When used as an i2c device there is a module usage count mismatch on
removal, preventing the driver from being used thereafter. dvb_attach
increments the usage count so it is properly balanced on removal.
On disconnect of Hauppauge SoloHD/DualHD before:
lsmod | grep lgdt3306a
lgdt3306a 28672 -1
i2c_mux 16384 1 lgdt3306a
On disconnect of Hauppauge SoloHD/DualHD after:
lsmod | grep lgdt3306a
lgdt3306a 28672 0
i2c_mux 16384 1 lgdt3306a
Signed-off-by: Brad Love <brad@nextdimension.cc>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
As configured currently, modulation in the driver is set to auto detect,
no matter what the user sets modulation to. This leads to both QAM64
and QAM256 having the same effect. QAM AUTO is explicitly added here for
compatibility with scanning software who can use AUTO instead of doing
essentially the same scan twice.
Also included is a module option to enforce a specific QAM modulation if
desired. The true modulation is read before calculating the snr.
Changes are backwards compatible with current behaviour.
Signed-off-by: Brad Love <brad@nextdimension.cc>
Reviewed-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Add some register updates required for stable viewing
on Cablevision in NY. Does not adversely affect other providers.
Changes since v1:
- Change upper case hex to lower case.
Signed-off-by: Brad Love <brad@nextdimension.cc>
Reviewed-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
If release is part of frontend ops then it is called in the
course of dvb_frontend_detach. The process also decrements
the module usage count. The problem is if the lgdt3306a
driver is reached via i2c_new_device, then when it is
eventually destroyed remove is called, which further
decrements the module usage count to negative. After this
occurs the driver is in a bad state and no longer works.
Also fixed by NULLing out the release callback is a double
kfree of state, which introduces arbitrary oopses/GPF.
This problem is only currently reachable via the em28xx driver.
On disconnect of Hauppauge SoloHD before:
lsmod | grep lgdt3306a
lgdt3306a 28672 -1
i2c_mux 16384 1 lgdt3306a
On disconnect of Hauppauge SoloHD after:
lsmod | grep lgdt3306a
lgdt3306a 28672 0
i2c_mux 16384 1 lgdt3306a
Signed-off-by: Brad Love <brad@nextdimension.cc>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Add a missing device to the driver table.
Signed-off-by: Brad Love <brad@nextdimension.cc>
Reviewed-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This platform was only one phone Jack.
Add dummy lineout verb to fix automute mode disable.
This just the workaround.
[ More background information:
since the platform has only a headphone jack without speaker, the
driver doesn't create the auto-mute control. Meanwhile we do update
the headset mode via the automute hook in the driver, thus with this
setup, the headset won't be updated any longer.
By adding a dummy line-out pin here, the auto-mute is added by the
driver, and the headset update is triggered properly.
Note that this is different from the other
ALC274_FIXUP_DELL_AIO_LINEOUT_VERB, which has the real line-out pin,
while this quirk adds a dummy line-out pin. -- tiwai ]
Signed-off-by: Kailang Yang <kailang@realtek.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
error->device_info.has_guc, which we check in capture_uc_state, is set
in capture_gen_state, so the latter needs to be performed first.
v2: rebased
Reported-by: Vinay Belgaumkar <vinay.belgaumkar@intel.com>
Fixes: 7d41ef3479 (drm/i915: Add Guc/HuC firmware details to error state)
Cc: Vinay Belgaumkar <vinay.belgaumkar@intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180305222122.3547-3-daniele.ceraolospurio@intel.com
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Add additional pids to driver list
Signed-off-by: Brad Love <brad@nextdimension.cc>
Reviewed-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
some of the static functions used from capture() have the "i915_"
prefix while other don't; most of them take i915 as a parameter, but one
of them derives it internally from error->i915. Let's be consistent by
avoiding prefix for static functions and by getting i915 from
error->i915. While at it, s/dev_priv/i915 in functions that don't
perform register reads.
v2: take i915 from error->i915 (Michal), s/dev_priv/i915,
update commit message
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Reviewed-by: Michel Thierry <michel.thierry@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180305222122.3547-2-daniele.ceraolospurio@intel.com
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Maximum 4 em28xx boards is too low, this can be maxed out by two devices.
This allows all the dvb adapters in the system to be em28xx if so desired.
Signed-off-by: Brad Love <brad@nextdimension.cc>
Reviewed-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hauppauge em28xx bulk devices exhibit continuity errors and corrupted
packets, when run in VMWare virtual machines. Unknown if other
manufacturers bulk models exhibit the same issue. KVM/Qemu is unaffected.
According to documentation the maximum packet multiplier for em28xx in bulk
transfer mode is 256 * 188 bytes. This changes the size of bulk transfers
to maximum supported value and have a bonus beneficial alignment.
Before:
After:
This sets up USB to expect just as many bytes as the em28xx is set to emit.
Successful usage under load afterwards natively and in both VMWare
and KVM/Qemu virtual machines.
Signed-off-by: Brad Love <brad@nextdimension.cc>
Reviewed-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Set appropriate bulk/ISOC transfer multiplier on capture start.
This sets ISOC transfer to USB endpoint configuration
This sets bulk transfer to 48128 bytes (188 * 256)
The bulk multiplier is maximum allowed according to Empia.
Signed-off-by: Brad Love <brad@nextdimension.cc>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
em28xx_duplicate_dev() is static. This were supposed to be
merged on the last patch, but somehow, I forgot "-a" when
I called git commit --amend.
Fixes: be7fd3c3a8 ("media: em28xx: Hauppauge DualHD second tuner functionality")
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Implement use of secondary TS port on em28xx.
Adds has_dual_ts field, allows secondary demod/tuner to be
added to a single em28xx device.
Hauppauge DualHD models are configured to use this feature.
[mchehab@s-opensource.com: em28xx_duplicate_dev() should be static]
Signed-off-by: Brad Love <brad@nextdimension.cc>
Reviewed-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Replace usages of the locally defined ABS() macro with calls to the
canonical abs() from kernel.h and remove the old definitions of ABS()
This change was originally motivated by two local definitions of the
ABS (absolute value) macro that fail to parenthesize their parameter
properly. This can lead to a bad expansion for low-precedence
expression arguments.
For example: ABS(1-2) currently expands to ((1-2) < 0 ? (-1-2) : (1-2))
which evaluates to -3. But the correct expansion would be
((1-2) < 0 ? -(1-2) : (1-2)) which evaluates to 1.
Signed-off-by: Dan Gopstein <dgopstein@nyu.edu>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
when a system call is interrupted we might call the critical section
cleanup handler that re-does some of the operations. When we are between
.Lsysc_vtime and .Lsysc_do_svc we might also redo the saving of the
problem state registers r0-r7:
.Lcleanup_system_call:
[...]
0: # update accounting time stamp
mvc __LC_LAST_UPDATE_TIMER(8),__LC_SYNC_ENTER_TIMER
# set up saved register r11
lg %r15,__LC_KERNEL_STACK
la %r9,STACK_FRAME_OVERHEAD(%r15)
stg %r9,24(%r11) # r11 pt_regs pointer
# fill pt_regs
mvc __PT_R8(64,%r9),__LC_SAVE_AREA_SYNC
---> stmg %r0,%r7,__PT_R0(%r9)
The problem is now, that we might have already zeroed out r0.
The fix is to move the zeroing of r0 after sysc_do_svc.
Reported-by: Farhan Ali <alifm@linux.vnet.ibm.com>
Fixes: 7041d28115 ("s390: scrub registers on kernel entry and KVM exit")
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Bamvor changed his mail so let's updat his mail address
everywhere.
Cc: Bamvor Jian Zhang <bamv2005@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
The Parfait (version 2.1.0) static code analysis tool found the
following NULL pointer dereference problem.
- drivers/gpu/drm/drm_drv.c
Any calls to drm_minor_get_slot() could result in the return of a NULL
pointer when an invalid DRM device type is encountered. The
return of NULL was removed with BUG() from drm_minor_get_slot().
Signed-off-by: Joe Moriarty <joe.moriarty@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20180220191157.100960-3-joe.moriarty@oracle.com
The Parfait (version 2.1.0) static code analysis tool found the
following NULL pointer derefernce problem.
- drivers/gpu/drm/drm_vblank.c
Null pointer checks were added to return values from calls to
drm_crtc_from_index(). There is a possibility, however minute, that
crtc->index may not be found when trying to find the struct crtc
from it's assigned index given in drm_crtc_init_with_planes().
3 return checks for NULL where added with a call to
WARN_ON(!crtc).
Signed-off-by: Joe Moriarty <joe.moriarty@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20180220191157.100960-2-joe.moriarty@oracle.com
In XenGT, ioreq copy is used to trap mmio write and ppgtt write. Both
of them are memory write, ioreq handler couldn't distinguish them. So
ioreq handler probe the ppgtt write handler, if it is succuess, this
ioreq is ppgtt write, otherwise it is mmio write.
So ppgtt write handler should return an error at the failure of finding
page track, it is fatal to implement ioreq handler in XenGT.
Signed-off-by: Xiong Zhang <xiong.y.zhang@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
page_track_handler take lock at the beginning, the lock should be released
at the failure of finding page track. Otherwise deadlock will happen.
Fixes: e502a2af4c ("drm/i915/gvt: Provide generic page_track infrastructure for write-protected page")
Signed-off-by: Xiong Zhang <xiong.y.zhang@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
- Be more robust when drawing arrows in the annotation TUI, avoiding a
segfault when jump instructions have as a target addresses in functions
other that the one currently being annotated. The full fix will come in
the following days, when jumping to other functions will work as call
instructions (Arnaldo Carvalho de Melo)
- Allow asking for the maximum allowed sample rate in 'top' and
'record', i.e. 'perf record -F max' will read the
kernel.perf_event_max_sample_rate sysctl and use it (Arnaldo Carvalho de Melo)
- When the user specifies a freq above kernel.perf_event_max_sample_rate,
Throttle it down to that max freq, and warn the user about it, add as
well --strict-freq so that the previous behaviour of not starting the
session when the desired freq can't be used can be selected (Arnaldo Carvalho de Melo)
- Find 'call' instruction target symbol at parsing time, used so far in
the TUI, part of the infrastructure changes that will end up allowing
for jumps to navigate to other functions, just like 'call'
instructions. (Arnaldo Carvalho de Melo)
- Use xyarray dimensions to iterate fds in 'perf stat' (Andi Kleen)
- Ignore threads for which the current user hasn't permissions when
enabling system-wide --per-thread (Jin Yao)
- Fix some backtrace perf test cases to use 'perf record' + 'perf script'
instead, till 'perf trace' starts using ordered_events or equivalent
to avoid symbol resolving artifacts due to reordering of
PERF_RECORD_MMAP events (Jiri Olsa)
- Fix crash in 'perf record' pipe mode, it needs to allocate the ID
array even for a single event, unlike non-pipe mode (Jiri Olsa)
- Make annoying fallback message on older kernels with newer 'perf top'
binaries trying to use overwrite mode and that not being present
in the older kernels (Kan Liang)
- Switch last users of old APIs to the newer perf_mmap__read_event()
one, then discard those old mmap read forward APIs (Kan Liang)
- Fix the usage on the 'perf kallsyms' man page (Sangwon Hong)
- Simplify cgroup arguments when tracking multiple events (weiping zhang)
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
-----BEGIN PGP SIGNATURE-----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=TNRi
-----END PGP SIGNATURE-----
Merge tag 'perf-core-for-mingo-4.17-20180305' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core
Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo:
- Be more robust when drawing arrows in the annotation TUI, avoiding a
segfault when jump instructions have as a target addresses in functions
other that the one currently being annotated. The full fix will come in
the following days, when jumping to other functions will work as call
instructions (Arnaldo Carvalho de Melo)
- Allow asking for the maximum allowed sample rate in 'top' and
'record', i.e. 'perf record -F max' will read the
kernel.perf_event_max_sample_rate sysctl and use it (Arnaldo Carvalho de Melo)
- When the user specifies a freq above kernel.perf_event_max_sample_rate,
Throttle it down to that max freq, and warn the user about it, add as
well --strict-freq so that the previous behaviour of not starting the
session when the desired freq can't be used can be selected (Arnaldo Carvalho de Melo)
- Find 'call' instruction target symbol at parsing time, used so far in
the TUI, part of the infrastructure changes that will end up allowing
for jumps to navigate to other functions, just like 'call'
instructions. (Arnaldo Carvalho de Melo)
- Use xyarray dimensions to iterate fds in 'perf stat' (Andi Kleen)
- Ignore threads for which the current user hasn't permissions when
enabling system-wide --per-thread (Jin Yao)
- Fix some backtrace perf test cases to use 'perf record' + 'perf script'
instead, till 'perf trace' starts using ordered_events or equivalent
to avoid symbol resolving artifacts due to reordering of
PERF_RECORD_MMAP events (Jiri Olsa)
- Fix crash in 'perf record' pipe mode, it needs to allocate the ID
array even for a single event, unlike non-pipe mode (Jiri Olsa)
- Make annoying fallback message on older kernels with newer 'perf top'
binaries trying to use overwrite mode and that not being present
in the older kernels (Kan Liang)
- Switch last users of old APIs to the newer perf_mmap__read_event()
one, then discard those old mmap read forward APIs (Kan Liang)
- Fix the usage on the 'perf kallsyms' man page (Sangwon Hong)
- Simplify cgroup arguments when tracking multiple events (weiping zhang)
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Due to an oversight when refactoring siginfo_t si_pkey has been in the
wrong position since 4.16-rc1. Add an explicit check of the offset of
every user space field in siginfo_t and compat_siginfo_t to make a
mistake like this hard to make in the future.
I have run this code on 4.15 and 4.16-rc1 with the position of si_pkey
fixed and all of the fields show up in the same location.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
The change moving addr_lsb into the _sigfault union failed to take
into account that _sigfault._addr_bnd._lower being a pointer forced
the entire union to have pointer alignment. In practice this only
mattered for the offset of si_pkey which is why this has taken so long
to discover.
To correct this change _dummy_pkey and _dummy_bnd to have pointer type.
Reported-by: kernel test robot <shun.hao@intel.com>
Fixes: b68a68d3dc ("signal: Move addr_lsb into the _sigfault union for clarity")
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
All of the conflicts were cases of overlapping changes.
In net/core/devlink.c, we have to make care that the
resouce size_params have become a struct member rather
than a pointer to such an object.
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a new debugfs entry kvmgt_nr_cache_entries under vgpu which shows
the number of entry in dma cache.
$ cat /sys/kernel/debug/gvt/vgpu1/kvmgt_nr_cache_entries
10101
v3: fix compiling error for some configuration. (Xiong Zhang <xiong.y.zhang@intel.com>)
v2: keep debugfs layout flat.
Signed-off-by: Changbin Du <changbin.du@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
The implementation of current kvmgt implicitly setup dma mapping at MPT
API gfn_to_mfn. First this design against the API's original purpose.
Second, there is no unmap hit in this design. The result is that the
dma mapping keep growing larger and larger. For mutl-vm case, they will
consume IOMMU IOVA low 4GB address space quickly and so tons of rbtree
entries crated in the IOMMU IOVA allocator. Finally, single IOVA
allocation can take as long as ~70ms. Such latency is intolerable.
To address both above issues, this patch introduced two new MPT API:
o dma_map_guest_page - setup dma map for guest page
o dma_unmap_guest_page - cancel dma map for guest page
The kvmgt implements these 2 API. And to reduce dma setup overhead for
duplicated pages (eg. scratch pages), two caches are used: one is for
mapping gfn to struct gvt_dma, another is for mapping dma addr to
struct gvt_dma.
With these 2 new API, the gtt now is able to cancel dma mapping when page
table is invalidated. The dma mapping is not in a gradual increase now.
v2: follow the old logic for VFIO_IOMMU_NOTIFY_DMA_UNMAP at this point.
Cc: Hang Yuan <hang.yuan@intel.com>
Cc: Xiong Zhang <xiong.y.zhang@intel.com>
Signed-off-by: Changbin Du <changbin.du@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Fix below error with minor code refactor.
CHECK drivers/gpu/drm/i915//gvt/handlers.c
drivers/gpu/drm/i915//gvt/handlers.c:203 sanitize_fence_mmio_access() error: 'vgpu' dereferencing possible ERR_PTR()
Reviewed-by: Zhi Wang <zhi.a.wang@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Fix check error at
CHECK drivers/gpu/drm/i915//gvt/kvmgt.c
drivers/gpu/drm/i915//gvt/kvmgt.c:455 intel_vgpu_create() error: we previously assumed 'vgpu' could be null (see line 454)
For failed vgpu create, just show error return in failure message.
Reviewed-by: Zhi Wang <zhi.a.wang@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>