[Intel-xe] [PATCH v2] drm/xe: use variable instead of multiple function calls

[Ruhl, Michael J]

This looks reasonable to me.

Reviewed-by: Michael J. Ruhl michael.j.ruhl at intel.com<mailto:michael.j.ruhl at intel.com>

M


From: Dani Liberman <dliberman at habana.ai>
Sent: Tuesday, October 31, 2023 5:10 PM
To: Ruhl, Michael J <michael.j.ruhl at intel.com>; intel-xe at lists.freedesktop.org
Subject: Re: [Intel-xe] [PATCH v2] drm/xe: use variable instead of multiple function calls

On 31/10/2023 22:09, Ruhl, Michael J wrote:



-----Original Message-----

From: Intel-xe <intel-xe-bounces at lists.freedesktop.org><mailto:intel-xe-bounces at lists.freedesktop.org> On Behalf Of Dani

Liberman

Sent: Tuesday, October 31, 2023 1:12 PM

To: intel-xe at lists.freedesktop.org<mailto:intel-xe at lists.freedesktop.org>

Subject: [Intel-xe] [PATCH v2] drm/xe: use variable instead of multiple function

calls



Using function calls with negative logic was a bit confusing,

positive logic is more readable.



v2:

- Update commit message



Cc: Matthew Brost <matthew.brost at intel.com><mailto:matthew.brost at intel.com>

Cc: Ville Syrjala <ville.syrjala at linux.intel.com><mailto:ville.syrjala at linux.intel.com>

Signed-off-by: Dani Liberman <dliberman at habana.ai><mailto:dliberman at habana.ai>

Reviewed-by: Matthew Brost <matthew.brost at intel.com><mailto:matthew.brost at intel.com>

---

drivers/gpu/drm/xe/xe_exec.c | 15 ++++++++-------

1 file changed, 8 insertions(+), 7 deletions(-)



diff --git a/drivers/gpu/drm/xe/xe_exec.c b/drivers/gpu/drm/xe/xe_exec.c

index 28e84a0bbeb0..2de6c2c05078 100644

--- a/drivers/gpu/drm/xe/xe_exec.c

+++ b/drivers/gpu/drm/xe/xe_exec.c

@@ -145,7 +145,7 @@ int xe_exec_ioctl(struct drm_device *dev, void *data,

struct drm_file *file)

  struct xe_sched_job *job;

  struct dma_fence *rebind_fence;

  struct xe_vm *vm;

- bool write_locked;

+ bool write_locked, vm_with_dma_fences;

  ktime_t end = 0;

  int err = 0;



@@ -196,8 +196,9 @@ int xe_exec_ioctl(struct drm_device *dev, void *data,

struct drm_file *file)

          }

  }



+ vm_with_dma_fences = !xe_vm_no_dma_fences(vm);

retry:



Are you 100% certain that this value will never change on retry?



M

It depends on vm flags that are getting set on vm create, so yes.



Dani





- if (!xe_vm_no_dma_fences(vm) && xe_vm_userptr_check_repin(vm)) {

+ if (vm_with_dma_fences && xe_vm_userptr_check_repin(vm)) {

          err = down_write_killable(&vm->lock);

          write_locked = true;

  } else {

@@ -279,7 +280,7 @@ int xe_exec_ioctl(struct drm_device *dev, void *data,

struct drm_file *file)

  }



  /* Wait behind munmap style rebinds */

- if (!xe_vm_no_dma_fences(vm)) {

+ if (vm_with_dma_fences) {

          err = drm_sched_job_add_resv_dependencies(&job->drm,

                                                  &vm->resv,



DMA_RESV_USAGE_KERNEL);

@@ -292,7 +293,7 @@ int xe_exec_ioctl(struct drm_device *dev, void *data,

struct drm_file *file)

  if (err)

          goto err_put_job;



- if (!xe_vm_no_dma_fences(vm)) {

+ if (vm_with_dma_fences) {

          err = down_read_interruptible(&vm->userptr.notifier_lock);

          if (err)

                  goto err_put_job;

@@ -307,7 +308,7 @@ int xe_exec_ioctl(struct drm_device *dev, void *data,

struct drm_file *file)

   * the job and let the DRM scheduler / backend clean up the job.

   */

  xe_sched_job_arm(job);

- if (!xe_vm_no_dma_fences(vm)) {

+ if (vm_with_dma_fences) {

          /* Block userptr invalidations / BO eviction */

          dma_resv_add_fence(&vm->resv,

                            &job->drm.s_fence->finished,

@@ -330,14 +331,14 @@ int xe_exec_ioctl(struct drm_device *dev, void

*data, struct drm_file *file)

  xe_sched_job_push(job);

  xe_vm_reactivate_rebind(vm);



- if (!err && !xe_vm_no_dma_fences(vm)) {

+ if (!err && vm_with_dma_fences) {

          spin_lock(&xe->ttm.lru_lock);

          ttm_lru_bulk_move_tail(&vm->lru_bulk_move);

          spin_unlock(&xe->ttm.lru_lock);

  }



err_repin:

- if (!xe_vm_no_dma_fences(vm))

+ if (vm_with_dma_fences)

          up_read(&vm->userptr.notifier_lock);

err_put_job:

  if (err)

--

2.34.1




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-xe/attachments/20231101/0bdb9c55/attachment-0001.htm>