mutex_unlock(&mm->vgpu->gvt->gtt.ppgtt_mm_lock);
> +
> invalidate_ppgtt_mm(mm);
> } else {
> vfree(mm->ggtt_mm.virtual_ggtt);
> --
Thanks for the fix!
Reviewed-by: Zhenyu Wang
--
Open Sou
t;cfg_space = mem;
>
> - mem = kmalloc(info->mmio_size, GFP_KERNEL);
> + mem = vmalloc(info->mmio_size);
> if (!mem) {
> kfree(path);
> kfree(firmware->cfg_space);
> --
> 2.7.4
>
Looks fine to me. Thanks!
Reviewed-by: Zheny
On 2020.01.17 17:15:36 +0200, Jani Nikula wrote:
> On Mon, 21 Oct 2019, Chris Wilson wrote:
> > Quoting Zhenyu Wang (2019-10-21 06:37:24)
> >> On 2019.10.16 19:39:02 +0100, Chris Wilson wrote:
> >> > Stop trying to escape out of the gvt layer to find the engine that
(Fred)
Gao Fred (1):
drm/i915/gvt: Fix guest boot warning
Tina Zhang (1):
drm/i915/gvt: Pin vgpu dma address before using
Zhenyu Wang (2):
drm/i915/gvt: use vgpu lock for active state setting
drm/i915/gvt: set guest display buffer as readonly
driver
On 2019.12.17 08:26:16 +, Chris Wilson wrote:
> Quoting Zhenyu Wang (2019-12-17 07:13:54)
> > Current GVT allocates kernel context as vGPU submission context.
> > For vGPU destroy, the kernel context needs to be close then released,
> > otherwise context's vm ppgtt re
653>] __vfs_write+0x1b/0x40
[<a72bac26>] vfs_write+0xb1/0x1a0
[<36135673>] ksys_write+0xa7/0xe0
Cc: Chris Wilson
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/gem/i915_gem_context.c | 5 +
drivers/gpu/drm/i915/gem/i915_gem_context.h | 1 +
drivers/gp
/gtt.c
> @@ -1282,7 +1282,7 @@ static int ppgtt_populate_shadow_entry(struct
> intel_vgpu *vgpu,
> return -EINVAL;
> default:
> GEM_BUG_ON(1);
> - };
> + }
>
> /* direct shadow */
> ret = intel_gvt_hypervisor_dma
On 2019.12.05 10:14:35 +0200, Joonas Lahtinen wrote:
> Quoting Zhenyu Wang (2019-12-05 08:11:41)
> >
> > ping..
>
> This was pulled, but my dim push was interrupted so I had to re-push and
> wait for CI results. Looks all good. Thanks for the PR.
>
> Do you hav
ping..
On 2019.12.02 13:17:11 +0800, Zhenyu Wang wrote:
>
> Hi,
>
> More gvt fixes for 5.5-rc. One is MI_ATOMIC cmd parser fix which
> missed 5.4 and another two for CFL non-priv reg access fixes.
>
> Thanks
> --
> The following changes since commit 83faaf074e6d1
address calucation
drm/i915/gvt: Update force-to-nonpriv register whitelist
Zhenyu Wang (1):
drm/i915/gvt: Fix cmd length check for MI_ATOMIC
drivers/gpu/drm/i915/gvt/cmd_parser.c | 6 +++---
drivers/gpu/drm/i915/gvt/handlers.c | 5 +++--
2 files changed, 6 insertions(+), 5 deletions
length (Zhenyu)
--------
Zhenyu Wang (1):
drm/i915/gvt: Fix cmd length check for MI_ATOMIC
drivers/gpu/drm/i915/gvt/cmd_parser.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
--
Open Source Technology Center, Intel ltd.
$gpg --keyserver wwwkeys.pgp.
Hi,
Here's one fix to remove PVINFO read in initial state for extra
warning messages in debug.
Thanks
--
The following changes since commit d9dace9438945e7c13d91e62927c5c6c88a37ee5:
drm/i915/selftests: Add intel_gt_suspend_prepare (2019-11-05 16:06:25 +0200)
are available in the Git reposito
Hi,
Here's one GVT dmabuf reference drop fix from Pan Bian.
Thanks
--
The following changes since commit ee2c5ef8a9d640ee1617ec97b84fe2f634284e51:
drm/i915/dp: Do not switch aux to TBT mode for non-TC ports (2019-11-04
13:24:14 -0800)
are available in the Git repository at:
https://githu
id)
> file_count(dmabuf->file),
> kref_read(&obj->base.refcount));
>
> + i915_gem_object_put(obj);
> +
> return dmabuf_fd;
>
> out_free_dmabuf:
Looks fine to me. Thanks!
Reviewed-by: Zhenyu Wang
--
Open S
On 2019.10.16 19:39:02 +0100, Chris Wilson wrote:
> Stop trying to escape out of the gvt layer to find the engine that we
> initially setup for use with gvt. Record the engines during initialisation
> and use them henceforth.
>
> add/remove: 1/4 grow/shrink: 22/28 up/down: 341/-1410 (-1069)
>
> S
On 2019.10.16 19:39:01 +0100, Chris Wilson wrote:
> Use the local vgpu_lock while preparing workloads to avoid taking the
> obsolete i915->drm.struct_mutex
>
> Signed-off-by: Chris Wilson
> ---
Looks good to me. Sorry for late reply as just be back from vacation.
Review
On 2019.09.09 15:35:10 -0700, Rodrigo Vivi wrote:
> On Mon, Sep 09, 2019 at 02:54:59PM -0700, Vetter, Daniel wrote:
> > On Mon, 2019-09-09 at 10:24 -0700, Rodrigo Vivi wrote:
> > > Hi guys,
> > >
> > > On Fri, Sep 06, 2019 at 01:42:55PM
Hi,
Here's gvt-next-fixes with two recent fixes, one for recent
guest hang regression and another for guest reset fix.
Thanks.
--
The following changes since commit c36beba6b296b3c05a0f29753b04775e5ae23886:
drm/i915: Seal races between async GPU cancellation, retirement and signaling
(2019-0
ucture.
>
> v2: Fix dim PARENTHESIS_ALIGNMENT check warning
>
> Signed-off-by: Xiong Zhang
> ---
Looks fine to me. You need to refresh after deballoon fix merged.
Reviewed-by: Zhenyu Wang
> drivers/gpu/drm/i915/i915_drv.h | 14 ++
&g
On 2019.08.23 09:35:22 +0100, Chris Wilson wrote:
> Quoting Zhang, Xiaolin (2019-08-23 09:07:01)
> > On 08/23/2019 03:58 PM, Chris Wilson wrote:
> > > Quoting Xiaolin Zhang (2019-08-23 07:57:31)
> > >> vgpu ppgtt notification was split into 2 steps, the first step is to
> > >> update PVINFO's pdp r
eballoon space: range [0x%llx - 0x%llx] %llu KiB.\n",
>node->start,
>node->start + node->size,
Searching shows this is pretty old one and also with r-b from Chris,
but be ignored that nobody picked this up..
I think I hit this on
size after copy
Zhenyu Wang (1):
drm/i915/gvt: Fix typo of VBLANK_TIMER_PERIOD
Zhi Wang (1):
drm/i915/gvt: factor out tlb and mocs register offset table
drivers/gpu/drm/i915/gvt/cmd_parser.c | 180 ++--
drivers/gpu/drm/i915/gvt/debugfs.c
Hi,
Here's one gvt fix for 5.3-rc, to fix one use-after-free issue
introduced in last fixes pull.
Thanks
--
The following changes since commit 4187414808095f645ca0661f8dde77617e2e7cb3:
drm/i915/gvt: Adding ppgtt to GVT GEM context after shadow pdps settled.
(2019-07-30 14:30:56 +0800)
are a
if we set it too low we impact the
> qos of the guest, too high and its affects the qos of the host. Keeping
> it at max user though still gives the kernel some leeway to usurp it.
> And who knows one day we may get a scheduler.
yep, appreciate that. ;)
Acked-by: Zhenyu Wang
--
Open
On 2019.08.08 11:44:21 +0100, Chris Wilson wrote:
> Quoting Dan Carpenter (2019-08-08 11:32:36)
> > We can't free "workload" until after the printk or it's a use after
> > free.
> >
> > Fixes: 2089a76ade90 ("drm/i915/gvt: Checking workload's gma earlier")
> > Signed-off-by: Dan Carpenter
>
> Tha
rning for invalid ggtt access
drm/i915/gvt: Don't use ggtt_validdate_range() with size=0
drm/i915/gvt: Checking workload's gma earlier
Zhenyu Wang (1):
drm/i915/gvt: remove duplicate include of trace.h
drivers/gpu/drm/i915/gvt/cmd_parser.c | 10 --
drivers/gp
ping..
On 2019.07.23 17:29:58 +0800, Zhenyu Wang wrote:
>
> Hi,
>
> Here's recent gvt-fixes for 5.3-rc. It includes safety guard
> for possible ggtt access error with related checks, includecheck
> fix and one cache entry size fix for 2MB ppgtt guest.
>
> Thanks.
e() with size=0
drm/i915/gvt: Checking workload's gma earlier
Zhenyu Wang (1):
drm/i915/gvt: remove duplicate include of trace.h
drivers/gpu/drm/i915/gvt/cmd_parser.c | 10 --
drivers/gpu/drm/i915/gvt/fb_decoder.c | 6 +++---
drivers/gpu/drm/i915/gvt/gtt.c |
bits so that we retain the ability to preempt our own
> semaphores.
>
> Signed-off-by: Chris Wilson
> Cc: Zhenyu Wang
> Cc: Xiaolin Zhang
> Cc: Tvrtko Ursulin
> Cc: Mika Kuoppala
> ---
I've tried to run guest with this one for several benchmarks on SKL
and didn
Chris, would you pick up this one? I think we need this first to make more GVT
tests rolling.
On 2019.07.09 20:53:46 -, Patchwork wrote:
> == Series Details ==
>
> Series: drm/i915/execlists: Disable preemption under GVT (rev6)
> URL : https://patchwork.freedesktop.org/series/62533/
> Sta
where in the middle). Under single request submission of gvt
> it is quite able for the preemption event to occur while another context
> is active and so there be a real need to act upon that preemption.
>
> Signed-off-by: Chris Wilson
> Cc: Tvrtko Ursulin
> Cc: Zhenyu Wang
&g
On 2019.07.12 11:06:01 +0300, Joonas Lahtinen wrote:
> + Zhenyu as FYI
>
> Quoting Chris Wilson (2019-07-02 14:56:04)
> > Quoting Michał Winiarski (2019-07-02 12:31:48)
> > > This reverts commit 4395890a48551982549d222d1923e2833dac47cf.
> > >
> > > It's been over a year since this was merged, and
On 2019.06.21 20:13:13 +0100, Chris Wilson wrote:
> Preempt-to-busy uses a GPU semaphore to enforce an idle-barrier across
> preemption, but mediated gvt does not fully support semaphores.
>
Current looks semaphore is still used from emit_fini_breadcrumb which
caused gvt error, gvt may support mem
gpu (Chris)
>
> Signed-off-by: Daniele Ceraolo Spurio
> Cc: Zhenyu Wang
> Cc: Chris Wilson
Looks good to me.
Acked-by: Zhenyu Wang
> ---
> drivers/gpu/drm/i915/i915_drv.c | 2 ++
> drivers/gpu/drm/i915/i915_pvinfo.h | 5 +++--
> drivers/gpu/drm/i915/i915_vgpu
Hi,
Here's one gvt fix for 5.2-rc6 to sanitize reserved register write in
PVINFO which should be discarded. This fixed some guest behavior which
contain some unmerged features that probe through PVINFO register state.
Thanks.
--
The following changes since commit 15e7f52a4596b496ce3da2fa4c1f94c6
iom.
Reviewed-by: Zhenyu Wang
>
> Because there is no need to check these functions, a number of local
> functions can be made to return void to simplify things as nothing can
> fail.
>
> Cc: Zhenyu Wang
> Cc: Zhi Wang
> Cc: Jani Nikula
> Cc: Joonas Lahtinen
>
gpu/drm/i915/i915.ko] undefined!
ERROR: "__tracepoint_oos_change" [drivers/gpu/drm/i915/i915.ko] undefined!
scripts/Makefile.modpost:91: recipe for target '__modpost' failed
Looks we need fix like below.
Subject: [PATCH] drm/i915/gvt: remove duplicate include of trace.h
This
error happened.
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/i915_vgpu.c | 27 +++
1 file changed, 15 insertions(+), 12 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_vgpu.c b/drivers/gpu/drm/i915/i915_vgpu.c
index 94d3992b599d..5aec34db1aaa 100644
--- a/drivers/gp
Hi,
More gvt fixes for 5.2-rc. This fixed one regression when enabling
debug build of i915 guest, guest ring state fix after execution
for hang check, and with two misc fixes from klocwork check.
Thanks
--
The following changes since commit 3035e8cd6c316cb633b45bc9b38052ba2dfd299b:
drm/i915/g
On 2019.05.29 15:07:09 +0300, Joonas Lahtinen wrote:
> Quoting Zhenyu Wang (2019-05-29 12:06:08)
> >
> > Hi,
> >
> > Here's more gvt fixes for 5.2. This includes fixes for recently
> > seen arbitrary DMA map fault with sane gtt entry initialization,
> >
Hi,
Here's more gvt fixes for 5.2. This includes fixes for recently
seen arbitrary DMA map fault with sane gtt entry initialization,
and several enforcement fixes for e.g ggtt range validation, some
mode register handler and one cmd length fix for parser.
Thanks
--
The following changes since co
Hi,
Here's gvt-fixes for 5.2-rc. It contains vgpu reset fix with
proper timeline handling, fixes for guest TRTT setting which
should be handled in context state instead of pushing directly
to hardware and one error return fix.
Thanks.
--
The following changes since commit a491cc8e1597ea25803191c
On 2019.05.17 11:22:23 +0100, Chris Wilson wrote:
> Caught by smatch:
> drivers/gpu/drm/i915//gvt/gtt.c:1106 ppgtt_populate_spt_by_guest_entry()
> error: uninitialized symbol 'ret'.
>
> Signed-off-by: Chris Wilson
> Cc: Zhenyu Wang
> ---
> drivers/gpu/drm
On 2019.05.13 12:22:44 +0300, Dan Carpenter wrote:
> "ret" is uninitialized on this path but it should be -EINVAL.
>
> Fixes: 930c8dfea4b8 ("drm/i915/gvt: Check if get_next_pt_type() always
> returns a valid value")
> Signed-off-by: Dan Carpenter
> ---
> drivers/gpu/drm/i915/gvt/gtt.c | 4 +++-
Hi,
Here's gvt-next-fixes for 5.2-rc, which includes one revert for BXT
regression, one missed context mmio reg after RCS renaming, sanitize
display buffer size calculation and some klocwork warning/error fixes.
Thanks
--
The following changes since commit 447811a686e8da7325516a78069ccfbd139ef1a
On 2019.04.29 11:10:52 +0800, Xiaolin Zhang wrote:
> To enable vgpu pv features, we need to setup a shared memory page
> which will be used for data exchange directly accessed between both
> guest and backend i915 driver to avoid emulation trap cost.
>
> guest i915 will allocate this page memory a
n path safe in the knowledge that the hard work is already
> > done.
> >
> > Signed-off-by: Chris Wilson
> > Cc: Zhenyu Wang
>
> Hi Zhenyu, could you check through this patch and make sure I haven't
> broken gvt in the process?
>
> The end result
n path safe in the knowledge that the hard work is already
> > done.
> >
> > Signed-off-by: Chris Wilson
> > Cc: Zhenyu Wang
>
> Hi Zhenyu, could you check through this patch and make sure I haven't
> broken gvt in the process?
>
> The end result is th
mode
Yan Zhao (1):
drm/i915/gvt: remove the unused sreg
Zhao Yakui (2):
drm/i915/gvt: Refine the snapshort range of I915 MCHBAR to optimize gvt-g
boot time
drm/i915/gvt: Refine the combined intel_vgpu_oos_page struct to save
memory
Zhenyu Wang (1):
Merge tag 'drm-
On 2019.04.14 17:44:10 +0300, Uri Lublin wrote:
> We started looking at the kvmgt driver code when we noticed
> the 'size' calculation in qemu-kvm is wrong.
> The first fix we had is now already committed upstream
> (7f1a93b1f1d1d2603a49a9e4226259db9272f305).
>
> There is a mismatch between th
Hi,
This includes one vGPU display plane size regression fix,
one for preventing use-after-free in ppgtt shadow free function
and another warning fix for iomem access annotation.
Thanks.
--
The following changes since commit cf9ed66671ec5f6cacc7b6efbad9d7c9e5e31776:
drm/i915/gvt: Fix kerneldo
combined intel_vgpu_oos_page struct to save
memory
Zhenyu Wang (1):
Merge tag 'drm-intel-next-2019-03-28' into gvt-next
drivers/gpu/drm/i915/gvt/cmd_parser.c | 30 +-
drivers/gpu/drm/i915/gvt/display.c | 1 -
drivers/gpu/drm/i915/gvt/gtt.c | 7 ++
drivers/gp
On 2019.04.04 09:11:55 +0100, Chris Wilson wrote:
> Quoting Zhenyu Wang (2019-04-04 08:48:34)
> > On 2019.04.04 08:30:56 +0100, Chris Wilson wrote:
> > > ppgtt_free_all_spt() iterates the radixtree as it is deleting it,
> > > forgoing all protection against the leaves
invalid RCU dereferences.
>
> Fixes: b6c126a39345 ("drm/i915/gvt: Manage shadow pages with radix tree")
> Signed-off-by: Chris Wilson
> Cc: Changbin Du
> Cc: Zhenyu Wang
> Cc: Zhi Wang
> ---
> drivers/gpu/drm/i915/gvt/gtt.c | 12 +---
> 1 file chan
ddr
> drivers/gpu/drm/i915/gvt/kvmgt.c:927:26:got void *aperture_va
>
> Fixes: d480b28a41a6 ("drm/i915/gvt: Fix aperture read/write emulation when
> enable x-no-mmap=on")
> Signed-off-by: Chris Wilson
> Cc: Zhenyu Wang
> Cc: Changbin Du
> Cc: Zhi Wang
> ---
On 2019.04.03 17:12:37 -0700, Rodrigo Vivi wrote:
> On Tue, Apr 02, 2019 at 05:40:37PM +0800, Zhenyu Wang wrote:
> >
> > Hi,
> >
> > Here's gvt-fixes for 5.1-rc4 which includes misc fixes for
> > vGPU display plane size calculation, shadow mm pin count
On 2019.04.03 17:12:37 -0700, Rodrigo Vivi wrote:
> On Tue, Apr 02, 2019 at 05:40:37PM +0800, Zhenyu Wang wrote:
> >
> > Hi,
> >
> > Here's gvt-fixes for 5.1-rc4 which includes misc fixes for
> > vGPU display plane size calculation, shadow mm pin count
Hi,
Here's gvt-fixes for 5.1-rc4 which includes misc fixes for
vGPU display plane size calculation, shadow mm pin count,
error recovery path for workload create and one kerneldoc
fix which I missed to include before.
Thanks.
--
The following changes since commit 26cdaac4793c49357d2c731f2190632ce
On 2019.03.27 09:15:36 +, Chris Wilson wrote:
> Quoting Zhenyu Wang (2019-03-27 09:06:36)
> > This is to disable semaphore usage when on vGPU for now. Unfortunately
> > GVT-g hasn't fully enabled semaphore usage yet, so current guest with
> > semaphore us
it
to turn it on.
Cc: Kevin Tian
Cc: Chris Wilson
Cc: Rodrigo Vivi
Signed-off-by: Zhenyu Wang
---
Note this needs to queue for -next otherwise next kernel i915 as guest
would be broken for GVT.
drivers/gpu/drm/i915/intel_lrc.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
di
On 2019.03.22 08:18:33 +, Chris Wilson wrote:
> Quoting Zhenyu Wang (2019-02-18 08:59:07)
> > On 2019.02.05 20:30:33 +, Chris Wilson wrote:
> > > drivers/gpu/drm/i915/gvt/display.c:457: warning: Function parameter or
> > > member 'connected' not des
g for shadow ppgtt mm reclaim list walk with locking (Zhenyu)
Colin Xu (1):
drm/i915/gvt: Add in context mmio 0x20D8 to gen9 mmio list
Weinan Li (1):
drm/i915/gvt: stop scheduling workload when vgpu is inactive
Zhenyu Wa
h by Bob Paauwe
>
> Signed-off-by: Chris Wilson
> Cc: Bob Paauwe
> Cc: Zhenyu Wang
> Cc: Zhi Wang
> ---
> drivers/gpu/drm/i915/gvt/vgpu.c| 2 +-
> drivers/gpu/drm/i915/i915_drv.c| 2 +-
> drivers/gpu/drm/i915/i915_pvinfo.h | 2 +-
> drivers/gpu/drm/i915/i
Initially found issue with closed context debug check when pin
hw_id for GVT context, looks we should always pin hw_id for that
as GVT context is fixed for each vGPU life cycle, and we'd also
like to get pinned hw_id e.g for perf reason, etc.
Cc: Chris Wilson
Signed-off-by: Zhenyu
On 2019.03.08 08:31:51 +, Chris Wilson wrote:
> Quoting Zhenyu Wang (2019-03-08 07:52:37)
> > Current GVT created context is marked closed as not to be used for
> > host user. But its hw_id should still be used. So this is to relax
> > debug BUG_ON() in __i915_gem_contex
Current GVT created context is marked closed as not to be used for
host user. But its hw_id should still be used. So this is to relax
debug BUG_ON() in __i915_gem_context_pin_hw_id() for GVT contexts
which can use force single submission flag to identify.
Cc: Chris Wilson
Signed-off-by: Zhenyu
On 2019.03.07 15:44:35 -0800, Rodrigo Vivi wrote:
> On Thu, Mar 07, 2019 at 02:08:10PM +0800, Zhenyu Wang wrote:
> >
> > Hi,
> >
> > Here's gvt-fixes for 5.1-rc1.
>
> I'm kind of confused here. this should be -next-fixes right?
>
> $ dim a
On 2019.03.07 15:44:35 -0800, Rodrigo Vivi wrote:
> On Thu, Mar 07, 2019 at 02:08:10PM +0800, Zhenyu Wang wrote:
> >
> > Hi,
> >
> > Here's gvt-fixes for 5.1-rc1.
>
> I'm kind of confused here. this should be -next-fixes right?
>
oops, you're
Xu (1):
drm/i915/gvt: Add in context mmio 0x20D8 to gen9 mmio list
Weinan Li (1):
drm/i915/gvt: stop scheduling workload when vgpu is inactive
Zhenyu Wang (4):
drm/i915/gvt: Fix MI_FLUSH_DW parsing with correct index check
drm/i915/gvt: Don't submit request for error workloa
> 'conncted' description in 'intel_vgpu_emulate_hotplug'
>
> Fixes: 1ca20f33df42 ("drm/i915/gvt: add hotplug emulation")
> Signed-off-by: Chris Wilson
> Cc: Hang Yuan
> Cc: Zhenyu Wang
> Cc: Zhi Wang
> ---
> drivers/gpu/drm/i915/g
Hi,
This should be last gvt-next pull for this round, which adds VFIO edid
region support in GVT, VM manager can use this to specify custom EDID
for VM, which can be used for e.g UI resize, etc.
p.s, Next week will be chinese new year, so team will be offline then.
Thanks.
--
The following chan
Hi,
Here's one fix for gvt to destroy shadow batch and indirect context
properly.
Thanks.
--
The following changes since commit 51b00d8509dc69c98740da2ad07308b630d3eb7d:
drm/i915/gvt: Fix mmap range check (2019-01-15 19:04:45 +0800)
are available in the Git repository at:
https://github.c
e for kernel type (Jani)
Jani Nikula (4):
drm/i915/gvt: remove drmP.h include
drm/i915/gvt: give the cmd parser decode_info a const treatment
drm/i915/gvt: give the cmd parser cmd_info a const treatment
drm/i915/gvt: switch to kernel types
Zhenyu Wang (3):
On 2019.01.21 11:51:41 +0200, Jani Nikula wrote:
> Mixed C99 and kernel types use is getting ugly. Prefer kernel types.
>
> sed -i 's/\buint\(8\|16\|32\|64\)_t\b/u\1/g'
>
> Signed-off-by: Jani Nikula
> ---
Looks good to me.
Acked-by: Zhenyu Wang
Will queue this
)
Colin Xu (1):
drm/i915/gvt: Allow F_CMD_ACCESS on mmio 0x21f0
Hang Yuan (1):
drm/i915/gvt: free VFIO region space in vgpu detach
Zhenyu Wang (1):
drm/i915/gvt: Fix mmap range check
drivers/gpu/drm/i915/gvt/handlers.c | 1 +
drivers/gpu/drm/i915/gvt/hypercall.h | 2
b/drivers/gpu/drm/i915/gvt/dmabuf.c
> index 51ed99a37803..2eb681175fae 100644
> --- a/drivers/gpu/drm/i915/gvt/dmabuf.c
> +++ b/drivers/gpu/drm/i915/gvt/dmabuf.c
> @@ -29,7 +29,6 @@
> */
>
> #include
> -#include
> #include
>
> #include "i915_drv.h"
fixes-2019-01-09
- Fix one race issue between pre-scan of guest workload with submission
Zhenyu Wang (1):
drm/i915/gvt: Fix workload request allocation before request add
drivers/gpu/drm/i915/gvt/scheduler.c
On 2018.12.20 12:33:35 +0100, Gerd Hoffmann wrote:
> On Thu, Dec 20, 2018 at 08:45:09AM +, Zhang, Tina wrote:
> >
> >
> > > -Original Message-
> > > From: intel-gvt-dev [mailto:intel-gvt-dev-boun...@lists.freedesktop.org]
> > > On
> > > Behalf Of Gerd Hoffmann
> > > Sent: Wednesday,
On 2018.12.10 16:41:24 -0800, Rodrigo Vivi wrote:
>
> On Fri, Dec 07, 2018 at 12:36:59PM +0800, Zhenyu Wang wrote:
> >
> > Hi,
> >
> > As I was hoping to possibly merge more new stuff for next kernel e.g
> > CFL support, etc, but seems those're still
The parameter 'void *gvt' is not used and required for hypervisor's
exit call. Even for non-merged Xen hypervisor support. So just remove it.
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/gvt/gvt.c | 2 +-
drivers/gpu/drm/i915/gvt/hypercall.h | 2 +-
drivers/gpu/drm/i
Don't mark hypervisor module's host_init as optional,
but mandatory required.
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/gvt/mpt.h | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/i915/gvt/mpt.h b/drivers/gpu/drm/i915/gvt/mpt.h
index 67
would be cleaned up in device unregister when module unload.
v2:
- Fix kvmgt order after i915 for built-in case
Cc: "Yuan, Hang"
Cc: Alex Williamson
Cc: "He, Min"
Reviewed-by: Yuan, Hang
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/Makefile| 1 +
drivers
job.
v6:
- remove unused variable
Zhenyu Wang (3):
drm/i915/gvt: mandatory require hypervisor's host_init
drm/i915/gvt: remove unused parameter for hypervisor's host_exit call
drm/i915/gvt: Change KVMGT as self load module
drivers/gpu/drm/i915/Makefile| 1 +
drivers/gpu/drm/i9
On 2018.12.07 07:47:59 +, He, Min wrote:
> Zhenyu,
> Overall I think the impact to AcrnGT is not big, we can modify our code to
> adapt
> to the new interfaces.
> But I have some comments embedded.
>
Good!
> > > @@ -467,6 +408,44 @@ int intel_gvt_init_device(struct drm_i915_private
> > *d
Hi,
As I was hoping to possibly merge more new stuff for next kernel e.g
CFL support, etc, but seems those're still not stable enough so better
wait for next cycle, so sorry for the late.
This includes mostly one regression fix for drm-intel-next when we
introduced during previous shadow ctx ppg
unregister when module unload.
v2:
- Fix kvmgt order after i915 for built-in case
Cc: "Yuan, Hang"
Cc: Alex Williamson
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/Makefile| 1 +
drivers/gpu/drm/i915/gvt/Makefile| 1 -
drivers/gpu/drm/i915/gvt/gvt.c |
l new register/unregister
function to initialize hypervisor specific interface for GVT. If
AcrnGT rebase on this, it may need to change initial setup too, so
could you double check?
thanks
On 2018.12.04 10:40:28 +0800, Zhenyu Wang wrote:
> This trys to make 'kvmgt' module as self loa
Hi,
Just one BDW regression fix for tiling mode format return
on vfio gfx dmabuf.
Thanks
--
The following changes since commit 7513edbc096a006f967eaf39088091442e623b83:
drm/i915/gvt: Avoid use-after-free iterating the gtt list (2018-11-21
17:31:56 +0800)
are available in the Git repository
order after i915 for built-in case
Cc: Alex Williamson
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/Makefile| 1 +
drivers/gpu/drm/i915/gvt/Makefile| 1 -
drivers/gpu/drm/i915/gvt/gvt.c | 107 +++
drivers/gpu/drm/i915/gvt/gvt.h |
On 2018.12.03 16:21:04 +0800, Zhenyu Wang wrote:
> On 2018.12.03 15:35:17 +0800, Tina Zhang wrote:
> > Create and initialize vGPU meta framebuffers during vGPU's creation.
> > Each meta framebuffer is an intel_framebuffer. Userspace can get the
> > userspace visib
if guest driver switches to
another plane for display? How could user know that then adjust for
assignment?
And really don't like sysfs to r/w file for id, even if doesn't use
vfio gfx dmabuf interface for guest display object presentation, a
vendor specific vfio device ioctl if possibl
ice open and release. Unused
mdev devices would be cleaned up in device unregister when module unload.
Zhenyu Wang (3):
drm/i915/gvt: mandatory require hypervisor's host_init
drm/i915/gvt: remove unused parameter for hypervisor's host_exit call
drm/i915/gvt: Change KVMGT as self load
The parameter 'void *gvt' is not used and required for hypervisor's
exit call. Even for non-merged Xen hypervisor support. So just remove it.
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/gvt/gvt.c | 2 +-
drivers/gpu/drm/i915/gvt/hypercall.h | 2 +-
drivers/gpu/drm/i
needs to load 'kvmgt' to enable VFIO/mdev interface.
v3:
- Fix module reference handling for device open and release. Unused
mdev devices would be cleaned up in device unregister when module unload.
v2:
- Fix kvmgt order after i915 for built-in case
Cc: Alex Williamson
Signed-off-by:
Don't mark hypervisor module's host_init as optional,
but mandatory required.
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/gvt/mpt.h | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/i915/gvt/mpt.h b/drivers/gpu/drm/i915/gvt/mpt.h
index 67
On 2018.11.30 08:44:23 -0700, Alex Williamson wrote:
> On Fri, 30 Nov 2018 14:51:24 +0800
> Zhenyu Wang wrote:
>
> > This trys to make 'kvmgt' module as self loadable instead of loading
> > by i915/gvt device model. So hypervisor specific module could be
> >
needs to load 'kvmgt' to enable VFIO/mdev interface.
v2:
- Fix kvmgt order after i915 for built-in case
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/Makefile| 1 +
drivers/gpu/drm/i915/gvt/Makefile| 1 -
drivers/gpu/drm/i915/gvt/gvt.c | 107 +++---
needs to load 'kvmgt' to enable VFIO/mdev interface.
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/gvt/gvt.c | 107 +++
drivers/gpu/drm/i915/gvt/gvt.h | 6 +-
drivers/gpu/drm/i915/gvt/hypercall.h | 7 +-
drivers/gpu/drm/i915/gvt/kvmgt.c
Don't mark hypervisor module's host_init as optional,
but mandatory required.
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/gvt/mpt.h | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/i915/gvt/mpt.h b/drivers/gpu/drm/i915/gvt/mpt.h
index 67
The parameter 'void *gvt' is not used and required for hypervisor's
exit call. Even for non-merged Xen hypervisor support. So just remove it.
Signed-off-by: Zhenyu Wang
---
drivers/gpu/drm/i915/gvt/gvt.c | 2 +-
drivers/gpu/drm/i915/gvt/hypercall.h | 2 +-
drivers/gpu/drm/i
201 - 300 of 775 matches
Mail list logo
