Merge tag 'nand/for-4.20' of git://git.infradead.org/linux-mtd into mtd/next

NAND core changes:
- Two batchs of cleanups of the NAND API, including:
  * Deprecating a lot of interfaces (now replaced by ->exec_op()).
  * Moving code in separate drivers (JEDEC, ONFI), in private files
    (internals), in platform drivers, etc.
  * Functions/structures reordering.
  * Exclusive use of the nand_chip structure instead of the MTD one
    all across the subsystem.
- Addition of the nand_wait_readrdy/rdy_op() helpers.

Raw NAND controllers drivers changes:
- Various coccinelle patches.
- Marvell:
  * Use regmap_update_bits() for syscon access.
  * More documentation.
  * BCH failure path rework.
  * More layouts to be supported.
  * IRQ handler complete() condition fixed.
- Fsl_ifc:
  * SRAM initialization fixed for newer controller versions.
- Denali:
  * Fix licenses mismatch and use a SPDX tag.
  * Set SPARE_AREA_SKIP_BYTES register to 8 if unset.
- Qualcomm:
  * Do not include dma-direct.h.
- Docg4:
  * Removed.
- Ams-delta:
  * Use of a GPIO lookup table
  * Internal machinery changes.

Raw NAND chip drivers changes:
- Toshiba:
  * Add support for Toshiba memory BENAND
  * Pass a single nand_chip object to the status helper.
- ESMT:
  * New driver to retrieve the ECC requirements from the 5th ID byte.

1 files changed, 2 insertions, 1 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index 5518e623fed2..51b5e977ca88 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -164,8 +164,10 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 		return r;
 	}
 
+	need_ctx_switch = ring->current_ctx != fence_ctx;
 	if (ring->funcs->emit_pipeline_sync && job &&
 	    ((tmp = amdgpu_sync_get_fence(&job->sched_sync, NULL)) ||
+	     (amdgpu_sriov_vf(adev) && need_ctx_switch) ||
 	     amdgpu_vm_need_pipeline_sync(ring, job))) {
 		need_pipe_sync = true;
 		dma_fence_put(tmp);
@@ -196,7 +198,6 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	}
 
 	skip_preamble = ring->current_ctx == fence_ctx;
-	need_ctx_switch = ring->current_ctx != fence_ctx;
 	if (job && ring->funcs->emit_cntxcntl) {
 		if (need_ctx_switch)
 			status |= AMDGPU_HAVE_CTX_SWITCH;