* Re: [Intel-gfx] [PATCH] drm/i915: Improve debug print in vm_fault_ttm
2022-09-22 12:09 [PATCH] drm/i915: Improve debug print in vm_fault_ttm Nirmoy Das
@ 2022-09-22 14:24 ` kernel test robot
2022-09-22 16:25 ` kernel test robot
2022-09-22 16:38 ` Matthew Auld
2 siblings, 0 replies; 5+ messages in thread
From: kernel test robot @ 2022-09-22 14:24 UTC (permalink / raw)
To: Nirmoy Das, intel-gfx; +Cc: kbuild-all, matthew.auld, dri-devel
Hi Nirmoy,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on drm-tip/drm-tip]
url: https://github.com/intel-lab-lkp/linux/commits/Nirmoy-Das/drm-i915-Improve-debug-print-in-vm_fault_ttm/20220922-201041
base: git://anongit.freedesktop.org/drm/drm-tip drm-tip
config: x86_64-randconfig-a011 (https://download.01.org/0day-ci/archive/20220922/202209222231.mIrsqiu7-lkp@intel.com/config)
compiler: gcc-11 (Debian 11.3.0-5) 11.3.0
reproduce (this is a W=1 build):
# https://github.com/intel-lab-lkp/linux/commit/e1a426a9e14837ada7e883d20af7c9abdf59823c
git remote add linux-review https://github.com/intel-lab-lkp/linux
git fetch --no-tags linux-review Nirmoy-Das/drm-i915-Improve-debug-print-in-vm_fault_ttm/20220922-201041
git checkout e1a426a9e14837ada7e883d20af7c9abdf59823c
# save the config file
mkdir build_dir && cp config build_dir/.config
make W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash drivers/gpu/drm/i915/
If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <lkp@intel.com>
All warnings (new ones prefixed by >>):
In file included from include/drm/drm_mm.h:51,
from include/drm/ttm/ttm_bo_driver.h:33,
from drivers/gpu/drm/i915/gem/i915_gem_ttm.c:8:
drivers/gpu/drm/i915/gem/i915_gem_ttm.c: In function 'vm_fault_ttm':
>> drivers/gpu/drm/i915/gem/i915_gem_ttm.c:1037:38: warning: format '%p' expects argument of type 'void *', but argument 4 has type 'int' [-Wformat=]
1037 | drm_dbg(dev, "Unable to make resource CPU accessible(err = %pe)\n", err);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~
| |
| int
include/drm/drm_print.h:461:63: note: in definition of macro 'drm_dbg'
461 | drm_dev_dbg((drm) ? (drm)->dev : NULL, DRM_UT_DRIVER, fmt, ##__VA_ARGS__)
| ^~~
drivers/gpu/drm/i915/gem/i915_gem_ttm.c:1037:85: note: format string is defined here
1037 | drm_dbg(dev, "Unable to make resource CPU accessible(err = %pe)\n", err);
| ~^
| |
| void *
| %d
vim +1037 drivers/gpu/drm/i915/gem/i915_gem_ttm.c
986
987 static vm_fault_t vm_fault_ttm(struct vm_fault *vmf)
988 {
989 struct vm_area_struct *area = vmf->vma;
990 struct ttm_buffer_object *bo = area->vm_private_data;
991 struct drm_device *dev = bo->base.dev;
992 struct drm_i915_gem_object *obj;
993 intel_wakeref_t wakeref = 0;
994 vm_fault_t ret;
995 int idx;
996
997 obj = i915_ttm_to_gem(bo);
998 if (!obj)
999 return VM_FAULT_SIGBUS;
1000
1001 /* Sanity check that we allow writing into this object */
1002 if (unlikely(i915_gem_object_is_readonly(obj) &&
1003 area->vm_flags & VM_WRITE))
1004 return VM_FAULT_SIGBUS;
1005
1006 ret = ttm_bo_vm_reserve(bo, vmf);
1007 if (ret)
1008 return ret;
1009
1010 if (obj->mm.madv != I915_MADV_WILLNEED) {
1011 dma_resv_unlock(bo->base.resv);
1012 return VM_FAULT_SIGBUS;
1013 }
1014
1015 if (i915_ttm_cpu_maps_iomem(bo->resource))
1016 wakeref = intel_runtime_pm_get(&to_i915(obj->base.dev)->runtime_pm);
1017
1018 if (!i915_ttm_resource_mappable(bo->resource)) {
1019 int err = -ENODEV;
1020 int i;
1021
1022 for (i = 0; i < obj->mm.n_placements; i++) {
1023 struct intel_memory_region *mr = obj->mm.placements[i];
1024 unsigned int flags;
1025
1026 if (!mr->io_size && mr->type != INTEL_MEMORY_SYSTEM)
1027 continue;
1028
1029 flags = obj->flags;
1030 flags &= ~I915_BO_ALLOC_GPU_ONLY;
1031 err = __i915_ttm_migrate(obj, mr, flags);
1032 if (!err)
1033 break;
1034 }
1035
1036 if (err) {
> 1037 drm_dbg(dev, "Unable to make resource CPU accessible(err = %pe)\n", err);
1038 dma_resv_unlock(bo->base.resv);
1039 ret = VM_FAULT_SIGBUS;
1040 goto out_rpm;
1041 }
1042 }
1043
1044 if (drm_dev_enter(dev, &idx)) {
1045 ret = ttm_bo_vm_fault_reserved(vmf, vmf->vma->vm_page_prot,
1046 TTM_BO_VM_NUM_PREFAULT);
1047 drm_dev_exit(idx);
1048 } else {
1049 ret = ttm_bo_vm_dummy_page(vmf, vmf->vma->vm_page_prot);
1050 }
1051
1052 if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
1053 goto out_rpm;
1054
1055 /* ttm_bo_vm_reserve() already has dma_resv_lock */
1056 if (ret == VM_FAULT_NOPAGE && wakeref && !obj->userfault_count) {
1057 obj->userfault_count = 1;
1058 mutex_lock(&to_gt(to_i915(obj->base.dev))->lmem_userfault_lock);
1059 list_add(&obj->userfault_link, &to_gt(to_i915(obj->base.dev))->lmem_userfault_list);
1060 mutex_unlock(&to_gt(to_i915(obj->base.dev))->lmem_userfault_lock);
1061 }
1062
1063 if (wakeref & CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND)
1064 intel_wakeref_auto(&to_gt(to_i915(obj->base.dev))->userfault_wakeref,
1065 msecs_to_jiffies_timeout(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND));
1066
1067 i915_ttm_adjust_lru(obj);
1068
1069 dma_resv_unlock(bo->base.resv);
1070
1071 out_rpm:
1072 if (wakeref)
1073 intel_runtime_pm_put(&to_i915(obj->base.dev)->runtime_pm, wakeref);
1074
1075 return ret;
1076 }
1077
--
0-DAY CI Kernel Test Service
https://01.org/lkp
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [Intel-gfx] [PATCH] drm/i915: Improve debug print in vm_fault_ttm
2022-09-22 12:09 [PATCH] drm/i915: Improve debug print in vm_fault_ttm Nirmoy Das
2022-09-22 14:24 ` [Intel-gfx] " kernel test robot
@ 2022-09-22 16:25 ` kernel test robot
2022-09-22 16:38 ` Matthew Auld
2 siblings, 0 replies; 5+ messages in thread
From: kernel test robot @ 2022-09-22 16:25 UTC (permalink / raw)
To: Nirmoy Das, intel-gfx; +Cc: kbuild-all, matthew.auld, dri-devel
Hi Nirmoy,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on drm-tip/drm-tip]
url: https://github.com/intel-lab-lkp/linux/commits/Nirmoy-Das/drm-i915-Improve-debug-print-in-vm_fault_ttm/20220922-201041
base: git://anongit.freedesktop.org/drm/drm-tip drm-tip
config: x86_64-defconfig (https://download.01.org/0day-ci/archive/20220923/202209230001.og3h9EMY-lkp@intel.com/config)
compiler: gcc-11 (Debian 11.3.0-5) 11.3.0
reproduce (this is a W=1 build):
# https://github.com/intel-lab-lkp/linux/commit/e1a426a9e14837ada7e883d20af7c9abdf59823c
git remote add linux-review https://github.com/intel-lab-lkp/linux
git fetch --no-tags linux-review Nirmoy-Das/drm-i915-Improve-debug-print-in-vm_fault_ttm/20220922-201041
git checkout e1a426a9e14837ada7e883d20af7c9abdf59823c
# save the config file
mkdir build_dir && cp config build_dir/.config
make W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash
If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <lkp@intel.com>
All errors (new ones prefixed by >>):
In file included from include/drm/drm_mm.h:51,
from include/drm/ttm/ttm_bo_driver.h:33,
from drivers/gpu/drm/i915/gem/i915_gem_ttm.c:8:
drivers/gpu/drm/i915/gem/i915_gem_ttm.c: In function 'vm_fault_ttm':
>> drivers/gpu/drm/i915/gem/i915_gem_ttm.c:1037:38: error: format '%p' expects argument of type 'void *', but argument 4 has type 'int' [-Werror=format=]
1037 | drm_dbg(dev, "Unable to make resource CPU accessible(err = %pe)\n", err);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~
| |
| int
include/drm/drm_print.h:461:63: note: in definition of macro 'drm_dbg'
461 | drm_dev_dbg((drm) ? (drm)->dev : NULL, DRM_UT_DRIVER, fmt, ##__VA_ARGS__)
| ^~~
drivers/gpu/drm/i915/gem/i915_gem_ttm.c:1037:85: note: format string is defined here
1037 | drm_dbg(dev, "Unable to make resource CPU accessible(err = %pe)\n", err);
| ~^
| |
| void *
| %d
cc1: all warnings being treated as errors
vim +1037 drivers/gpu/drm/i915/gem/i915_gem_ttm.c
986
987 static vm_fault_t vm_fault_ttm(struct vm_fault *vmf)
988 {
989 struct vm_area_struct *area = vmf->vma;
990 struct ttm_buffer_object *bo = area->vm_private_data;
991 struct drm_device *dev = bo->base.dev;
992 struct drm_i915_gem_object *obj;
993 intel_wakeref_t wakeref = 0;
994 vm_fault_t ret;
995 int idx;
996
997 obj = i915_ttm_to_gem(bo);
998 if (!obj)
999 return VM_FAULT_SIGBUS;
1000
1001 /* Sanity check that we allow writing into this object */
1002 if (unlikely(i915_gem_object_is_readonly(obj) &&
1003 area->vm_flags & VM_WRITE))
1004 return VM_FAULT_SIGBUS;
1005
1006 ret = ttm_bo_vm_reserve(bo, vmf);
1007 if (ret)
1008 return ret;
1009
1010 if (obj->mm.madv != I915_MADV_WILLNEED) {
1011 dma_resv_unlock(bo->base.resv);
1012 return VM_FAULT_SIGBUS;
1013 }
1014
1015 if (i915_ttm_cpu_maps_iomem(bo->resource))
1016 wakeref = intel_runtime_pm_get(&to_i915(obj->base.dev)->runtime_pm);
1017
1018 if (!i915_ttm_resource_mappable(bo->resource)) {
1019 int err = -ENODEV;
1020 int i;
1021
1022 for (i = 0; i < obj->mm.n_placements; i++) {
1023 struct intel_memory_region *mr = obj->mm.placements[i];
1024 unsigned int flags;
1025
1026 if (!mr->io_size && mr->type != INTEL_MEMORY_SYSTEM)
1027 continue;
1028
1029 flags = obj->flags;
1030 flags &= ~I915_BO_ALLOC_GPU_ONLY;
1031 err = __i915_ttm_migrate(obj, mr, flags);
1032 if (!err)
1033 break;
1034 }
1035
1036 if (err) {
> 1037 drm_dbg(dev, "Unable to make resource CPU accessible(err = %pe)\n", err);
1038 dma_resv_unlock(bo->base.resv);
1039 ret = VM_FAULT_SIGBUS;
1040 goto out_rpm;
1041 }
1042 }
1043
1044 if (drm_dev_enter(dev, &idx)) {
1045 ret = ttm_bo_vm_fault_reserved(vmf, vmf->vma->vm_page_prot,
1046 TTM_BO_VM_NUM_PREFAULT);
1047 drm_dev_exit(idx);
1048 } else {
1049 ret = ttm_bo_vm_dummy_page(vmf, vmf->vma->vm_page_prot);
1050 }
1051
1052 if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
1053 goto out_rpm;
1054
1055 /* ttm_bo_vm_reserve() already has dma_resv_lock */
1056 if (ret == VM_FAULT_NOPAGE && wakeref && !obj->userfault_count) {
1057 obj->userfault_count = 1;
1058 mutex_lock(&to_gt(to_i915(obj->base.dev))->lmem_userfault_lock);
1059 list_add(&obj->userfault_link, &to_gt(to_i915(obj->base.dev))->lmem_userfault_list);
1060 mutex_unlock(&to_gt(to_i915(obj->base.dev))->lmem_userfault_lock);
1061 }
1062
1063 if (wakeref & CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND)
1064 intel_wakeref_auto(&to_gt(to_i915(obj->base.dev))->userfault_wakeref,
1065 msecs_to_jiffies_timeout(CONFIG_DRM_I915_USERFAULT_AUTOSUSPEND));
1066
1067 i915_ttm_adjust_lru(obj);
1068
1069 dma_resv_unlock(bo->base.resv);
1070
1071 out_rpm:
1072 if (wakeref)
1073 intel_runtime_pm_put(&to_i915(obj->base.dev)->runtime_pm, wakeref);
1074
1075 return ret;
1076 }
1077
--
0-DAY CI Kernel Test Service
https://01.org/lkp
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] drm/i915: Improve debug print in vm_fault_ttm
2022-09-22 12:09 [PATCH] drm/i915: Improve debug print in vm_fault_ttm Nirmoy Das
2022-09-22 14:24 ` [Intel-gfx] " kernel test robot
2022-09-22 16:25 ` kernel test robot
@ 2022-09-22 16:38 ` Matthew Auld
2022-09-23 7:27 ` Das, Nirmoy
2 siblings, 1 reply; 5+ messages in thread
From: Matthew Auld @ 2022-09-22 16:38 UTC (permalink / raw)
To: Nirmoy Das, intel-gfx; +Cc: Chris Wilson, dri-devel
On 22/09/2022 13:09, Nirmoy Das wrote:
> Print the error code returned by __i915_ttm_migrate()
> for better debuggability.
>
> References: https://gitlab.freedesktop.org/drm/intel/-/issues/6889
> Signed-off-by: Nirmoy Das <nirmoy.das@intel.com>
> ---
> drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
> index e3fc38dd5db0..9619c0fe1025 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
> @@ -1034,7 +1034,7 @@ static vm_fault_t vm_fault_ttm(struct vm_fault *vmf)
> }
>
> if (err) {
> - drm_dbg(dev, "Unable to make resource CPU accessible\n");
> + drm_dbg(dev, "Unable to make resource CPU accessible(err = %pe)\n", err);
Yeah, looks useful. I think for that bug the object is just too large
for the mappable part of lmem, so this just gives -2big or similar on
small-bar systems. I presume that the test needs to be updated to
account for the cpu_size or so.
With the kernel test robot warning fixed:
Acked-by: Matthew Auld <matthew.auld@intel.com>
I looked at the GEM_BUG_ON(rq->reserved_space > ring->space), and I
think the issue is maybe with emit_pte() using the ring->space to
manually figure out the number of dwords it can emit (instead of the
usual ring_begin()), which I guess works, but if we are unlucky and get
interrupted (like with a very well timed sigbus here), while waiting for
more ring space and end up bailing early, we might have trampled over
the reserved_space when submitting the request. I guess normally the
next ring_begin() would take care of the reserved_space, like when
constructing the actual copy packet.
> dma_resv_unlock(bo->base.resv);
> ret = VM_FAULT_SIGBUS;
> goto out_rpm;
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] drm/i915: Improve debug print in vm_fault_ttm
2022-09-22 16:38 ` Matthew Auld
@ 2022-09-23 7:27 ` Das, Nirmoy
0 siblings, 0 replies; 5+ messages in thread
From: Das, Nirmoy @ 2022-09-23 7:27 UTC (permalink / raw)
To: Matthew Auld, intel-gfx; +Cc: Chris Wilson, dri-devel
On 9/22/2022 6:38 PM, Matthew Auld wrote:
> On 22/09/2022 13:09, Nirmoy Das wrote:
>> Print the error code returned by __i915_ttm_migrate()
>> for better debuggability.
>>
>> References: https://gitlab.freedesktop.org/drm/intel/-/issues/6889
>> Signed-off-by: Nirmoy Das <nirmoy.das@intel.com>
>> ---
>> drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
>> b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
>> index e3fc38dd5db0..9619c0fe1025 100644
>> --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
>> @@ -1034,7 +1034,7 @@ static vm_fault_t vm_fault_ttm(struct vm_fault
>> *vmf)
>> }
>> if (err) {
>> - drm_dbg(dev, "Unable to make resource CPU accessible\n");
>> + drm_dbg(dev, "Unable to make resource CPU accessible(err
>> = %pe)\n", err);
>
> Yeah, looks useful. I think for that bug the object is just too large
> for the mappable part of lmem, so this just gives -2big or similar on
> small-bar systems. I presume that the test needs to be updated to
> account for the cpu_size or so.
Yeah, can't think of any other case. The test need to be updated, going
to send out igt fixes for this.
>
> With the kernel test robot warning fixed:
> Acked-by: Matthew Auld <matthew.auld@intel.com>
Thanks, I will resend a updated one.
>
> I looked at the GEM_BUG_ON(rq->reserved_space > ring->space), and I
> think the issue is maybe with emit_pte() using the ring->space to
> manually figure out the number of dwords it can emit (instead of the
> usual ring_begin()), which I guess works, but if we are unlucky and
> get interrupted (like with a very well timed sigbus here), while
> waiting for more ring space and end up bailing early, we might have
> trampled over the reserved_space when submitting the request. I guess
> normally the next ring_begin() would take care of the reserved_space,
> like when constructing the actual copy packet.
I am not so familiar with the code but sounds logical.
Nirmoy
>
>> dma_resv_unlock(bo->base.resv);
>> ret = VM_FAULT_SIGBUS;
>> goto out_rpm;
^ permalink raw reply [flat|nested] 5+ messages in thread