Re: [PATCH V2 1/2] accel/amdxdna: Detect a driver command without firmware response
From: Max Zhen
Date: Tue Nov 17 2026 - 16:55:01 EST
On 11/17/2026 Tue 13:27, Lizhi Hou wrote:
aie2_hwctx_cfg_debug_bo() and aie2_hwctx_sync_debug_bo() decide successReviewed-by: Max Zhen <max.zhen@xxxxxxx>
solely from cmd->result.
cmd->result is only written by aie2_sched_drvcmd_resp_handler(), which
runs when firmware responds. If the job fails before it reaches
firmware, for example aie2_sync_bo() or aie2_config_debug_bo() fail to
send the mailbox message, no response handler runs. cmd is zero
allocated, so cmd->result stays 0 and the caller treats the failed
command as successful. aie2_hwctx_cfg_debug_bo() then goes on to update
abo->assigned_hwctx, leaving it out of sync with firmware.
Set cmd->result to U32_MAX before submitting, so a command which gets
no response is reported as a failure.
Fixes: 7ea046838021 ("accel/amdxdna: Support firmware debug buffer")
Signed-off-by: Lizhi Hou <lizhi.hou@xxxxxxx>
---
V2:
Use a simplier fix to address sashiko comment.
drivers/accel/amdxdna/aie2_ctx.c | 12 +++++++++---
drivers/accel/amdxdna/amdxdna_ctx.h | 5 +++++
2 files changed, 14 insertions(+), 3 deletions(-)
diff --git a/drivers/accel/amdxdna/aie2_ctx.c b/drivers/accel/amdxdna/aie2_ctx.c
index 11f5421e497aa..32458d05635f9 100644
--- a/drivers/accel/amdxdna/aie2_ctx.c
+++ b/drivers/accel/amdxdna/aie2_ctx.c
@@ -989,10 +989,14 @@ static void aie2_cmd_wait(struct amdxdna_hwctx *hwctx, u64 seq)
struct dma_fence *out_fence = aie2_cmd_get_out_fence(hwctx, seq);
struct amdxdna_dev *xdna = hwctx->client->xdna;
- if (!out_fence) {
- XDNA_ERR(xdna, "Failed to get fence");
+ /*
+ * No fence to wait for. Either seq has already signaled and a later
+ * point pruned it out of the timeline chain, or the syncobj was
+ * reset. Neither means the command failed, so leave the verdict to
+ * the caller, which inspects cmd->result.
+ */
+ if (!out_fence)
return;
- }
mutex_unlock(&xdna->dev_lock);
dma_fence_wait_timeout(out_fence, false, MAX_SCHEDULE_TIMEOUT);
@@ -1022,6 +1026,7 @@ static int aie2_hwctx_cfg_debug_bo(struct amdxdna_hwctx *hwctx, u32 bo_hdl,
goto put_obj;
}
kref_init(&cmd->refcnt);
+ cmd->result = U32_MAX;
if (attach) {
if (abo->assigned_hwctx != AMDXDNA_INVALID_CTX_HANDLE) {
@@ -1095,6 +1100,7 @@ int aie2_hwctx_sync_debug_bo(struct amdxdna_hwctx *hwctx, u32 debug_bo_hdl)
if (!cmd)
return -ENOMEM;
kref_init(&cmd->refcnt);
+ cmd->result = U32_MAX;
cmd->opcode = SYNC_DEBUG_BO;
ret = amdxdna_cmd_submit(client, cmd, AMDXDNA_INVALID_BO_HANDLE,
diff --git a/drivers/accel/amdxdna/amdxdna_ctx.h b/drivers/accel/amdxdna/amdxdna_ctx.h
index 2e6b300d652f8..bf248d3f16ed1 100644
--- a/drivers/accel/amdxdna/amdxdna_ctx.h
+++ b/drivers/accel/amdxdna/amdxdna_ctx.h
@@ -145,6 +145,11 @@ enum amdxdna_job_opcode {
struct amdxdna_drv_cmd {
enum amdxdna_job_opcode opcode;
+ /*
+ * Written by the firmware response handler only. Set it to U32_MAX
+ * before submitting, so a job which never reaches firmware, and thus
+ * gets no response, is not mistaken for a successful command.
+ */
u32 result;
struct kref refcnt;
};