the preempted request has been
completed.
v2:
- refine description, add completed check and notification in
execlists_cancel_port_requests. (Chris)
v3:
- use ternary confitional, remove local variable. (Tvrtko)
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
Signed-off-by: Weinan Li <
the preempted request has been
completed.
v2:
- refine description, add completed check and notification in
execlists_cancel_port_requests. (Chris)
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
Signed-off-by: Weinan Li <weinan.z...@intel.com>
Signed-off-by: Zhenyu Wang <zhen...@
There is one corner case missing schedule out notification for GVT-g in
__unwind_incomplete_requests, it may cause vgpu no response. Add
notification when ensure one request has been completed during doing
unwind.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
Signed-off-by: Zhenyu Wang
switch.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/gvt/cmd_parser.c | 19 +++
drivers/gpu/drm/i915/gvt/mmio_context.c | 33 ++---
2 files changed, 37 insertions(+), 15 deletions(-)
diff --git a/drivers/gpu/drm/i9
Refine trace_render_mmio to show the vm id before and after vgpu switch,
tag host id as '0', this patch will be used in the future patch for refine
mocs switch policy.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/gvt/mmio_context.c | 4 ++--
drivers/gpu/drm/i9
mmio save/restore, it will reduce the CPU utilization and
performance while there is multi VMs with heavy work load.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/gvt/mmio_context.c | 196 ++--
1 file changed, 85 insertions(+), 111 del
Load host render mocs registers once for delta update of mocs switch, it
reduces mmio read times obviously, then brings performance improvement
during multi-vms switch.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/gvt/mmio_context.
Weinan Li (4):
drm/i915/gvt: refine trace_render_mmio
drm/i915/gvt: optimize for vGPU mmio switch
drm/i915/gvt: refine mocs save restore policy
drm/i915/gvt: load host render mocs once in mocs switch
drivers/gpu/drm/i915/gvt/cmd_parser.c | 19 +++
drivers/gpu/drm/i915/gvt/mmio_context.c
switch.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/gvt/cmd_parser.c | 19 +++
drivers/gpu/drm/i915/gvt/render.c | 33 ++---
2 files changed, 37 insertions(+), 15 deletions(-)
diff --git a/drivers/gpu/drm/i9
Load host render mocs registers once for delta update of mocs switch, it reduces
mmio read times obviously, then brings performance improvement during multi-vms
switch.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/gvt/render.
Merge switch_mmio_to_vgpu and switch_mmio_to_host, use delta update for
mocs save restore, deal host mocs value as fixed, it won't be changed after
initialization. These can save vgpu switch time to reduce CPU utilization
and improve GPU performance in GVT-g with multi-VMs.
Weinan Li (3):
drm
mmio save/restore, it will reduce the CPU utilization and
performance while there is multi VMs with heavy work load.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/gvt/render.c | 212 --
drivers/gpu/drm/i915/gvt/trace.h | 15
".
Cc: Zhi Wang <zhi.a.w...@intel.com>
Signed-off-by: Weinan Li <weinan.z...@intel.com>
Signed-off-by: Xiong Zhang <xiong.y.zh...@intel.com>
---
drivers/gpu/drm/i915/gvt/handlers.c | 45 +
1 file changed, 5 insertions(+), 40 deletions(-)
virtual HWSP caps check function.
v6 : Comments refine.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahti...@linux.intel.com>
---
drivers/gpu/drm/i915/i915_pvinfo.h | 1 +
drivers/gpu/drm/i915/i91
virtual HWSP caps check function.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahti...@linux.intel.com>
---
drivers/gpu/drm/i915/i915_pvinfo.h | 1 +
drivers/gpu/drm/i915/i915_vgpu.h | 6 +
Let GVT-g VM read the CSB and CSB write pointer from virtual HWSP, not all
the host support this feature, need to check the BIT(3) of caps in PVINFO.
v3 : Remove unnecessary comments.
v4 : Separate VM enable patch with GVT-g implementation patch due to code
dependency
Signed-off-by: Weinan Li
v2 : clean merge confict
v3 :
remove unnecessary comments
add address audit in HWSP address update
Weinan Li (2):
drm/i915/gvt: update CSB and CSB write pointer in virtual HWSP
drm/i915: enable to read CSB and CSB write pointer from HWSP in GVT-g
VM
drivers/gpu/drm/i915/gvt
address audit in HWSP address update.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
---
drivers/gpu/drm/i915/gvt/execlist.c | 17 ++
drivers/gpu/drm/i915/gvt/gvt.h | 1 +
drivers/gpu/drm/i915/gvt/han
Let GVT-g VM read the CSB and CSB write pointer from virtual HWSP, not all
the host support this feature, need to check the BIT(3) of caps in PVINFO.
v3 : Remove unnecessary comments.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
Cc: Joo
Let GVT-g VM read the CSB and CSB write pointer from virtual HWSP, not all
the host support this feature, need to check the BIT(3) of caps in PVINFO.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
---
drivers/gpu/drm/i915/i915_vgpu.
V2: clean merge confict.
Weinan Li (2):
drm/i915/gvt: update CSB and CSB write pointer in virtual HWSP
drm/i915: enable to read CSB and CSB write pointer from HWSP in GVT-g
VM
drivers/gpu/drm/i915/gvt/execlist.c| 16 +
drivers/gpu/drm/i915/gvt/gvt.h | 1
-by: Weinan Li <weinan.z...@intel.com>
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
---
drivers/gpu/drm/i915/gvt/execlist.c | 16 ++
drivers/gpu/drm/i915/gvt/gvt.h | 1 +
drivers/gpu/drm/i915/gvt/handlers.c | 42 +++--
drivers/gpu/drm/i91
-by: Weinan Li <weinan.z...@intel.com>
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
---
drivers/gpu/drm/i915/gvt/execlist.c | 14 +
drivers/gpu/drm/i915/gvt/gvt.h | 1 +
drivers/gpu/drm/i915/gvt/handlers.c | 41 +++--
drivers/gpu/drm/i91
Let GVT-g VM read the CSB and CSB write pointer from virtual HWSP, not all
the host support this feature, need to check the BIT(3) of caps in PVINFO.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
---
drivers/gpu/drm/i915/i915
l.com>
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahti...@linux.intel.com>
Cc: Zhenyu Wang <zhen...@linux.intel.com>
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/i915_gem.c | 4 ++--
drivers/gpu/drm/i915/i9
<ch...@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahti...@linux.intel.com>
Cc: Zhenyu Wang <zhen...@linux.intel.com>
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/i915_gem.c | 4 ++--
drivers/gpu/drm/i915/i915_gem_gtt.h | 1 +
driv
c: Joonas Lahtinen <joonas.lahti...@linux.intel.com>
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/i915_gem.c | 4 ++--
drivers/gpu/drm/i915/i915_gem_gtt.h | 1 +
drivers/gpu/drm/i915/i915_vgpu.c| 44 ++---
3 files
teardown to balloon and deballoon to make sure the reserved
stays correct. Code style refine.
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahti...@linux.intel.com>
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/i915_gem.c | 4
teardown to balloon and deballoon to make sure the reserved
stays correct. Code style refine.
Cc: Chris Wilson <ch...@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahti...@linux.intel.com>
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/i915_gem.c | 4
the reserved size
in ggtt by balloon.
v3: remain aper_size as total, adjust aper_available_size exclude reserved
and pinned. UMD driver need to adjust the max allocation size according to
the available aperture size but not total size.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
d
, we
also need to exclude the reserved part in GTT.
v2: add 'reserved' in struct i915_address_space to record the reserved size
in ggtt by balloon.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/i915_gem.c | 6 ++
drivers/gpu/drm/i915/i915_gem_context
the reserved part by balloon.
I915_GEM_CONTEXT_GETPARAM ioctl query the I915_CONTEXT_PARAM_GTT_SIZE, we
also need to exclude the reserved part in GTT.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/i915_gem.c | 7 +++
drivers/gpu/drm/i915/i915_gem_context.
Add pcode read/write emulation in gvt for BDW.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/gvt/handlers.c | 33 ++---
1 file changed, 18 insertions(+), 15 deletions(-)
diff --git a/drivers/gpu/drm/i915/gvt/handlers.c
b/drivers/g
skl_pcode_try_request() call sandybridge_pcode_read(), check both return
status and value simultanously, ensure it got correct value without error.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/intel_pm.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Having converted the force_wake_get/_put routines for a vGPU to be no-op,
we can use the common mmio accessors and remove our specialised routines
that simply skipped the calls to control force_wake.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
Reviewed-by: Chris Wilson <ch
(making our vgpu specific mmio routines redundant and to be
removed in the next patch).
Signed-off-by: Weinan Li <weinan.z...@intel.com>
Reviewed-by: Chris Wilson <ch...@chris-wilson.co.uk>
---
drivers/gpu/drm/i915/intel_uncore.c | 13 +
1 file changed, 13 insertions(+)
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/intel_uncore.c | 58 -
1 file changed, 58 deletions(-)
diff --git a/drivers/gpu/drm/i915/intel_uncore.c
b/drivers/gpu/drm/i915/intel_uncore.c
index 9fad4de..e9046fa 100644
--- a/d
t;. Unnecessary MMIO access in guest
waste much CPU cost. Since we full virtualize the MMIO, just noop the
forcewake get/put.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/intel_uncore.c | 13 +
1 file changed, 13 insertions(+)
diff --git a/drivers/gpu/
No need to do posting read when vgpu actived.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/i915_drv.h | 10 +++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 5
t;. Unnecessary MMIO access in guest
waste much CPU cost. Since we full virtualize the MMIO, just ignore the
forcewake get/put in low level.
Signed-off-by: Weinan Li <weinan.z...@intel.com>
---
drivers/gpu/drm/i915/intel_uncore.c | 78 ++---
1 file changed, 20 i
40 matches
Mail list logo