* [PATCH V1] accel/amdxdna: Destroy the context when hwctx restart fails
@ 2026-11-17 16:57 Lizhi Hou
2026-11-17 20:54 ` Max Zhen
0 siblings, 1 reply; 2+ messages in thread
From: Lizhi Hou @ 2026-11-17 16:57 UTC (permalink / raw)
To: ogabbay, quic_jhugo, dri-devel, mario.limonciello,
karol.wachowski, max.zhen
Cc: Lizhi Hou, linux-kernel, sonal.santan
aie2_hwctx_restart() creates the firmware context first, then maps the
host buffers and configures the CUs. When any of the later steps fails,
the error is propagated to the caller but the context created at the
beginning is left behind: hwctx->fw_ctx_id stays valid, the mailbox
channel stays allocated and started, and ndev->hwctx_num is not
decremented. The hardware context is reported as failed yet remains
alive, and because aie2_sched_job_run() only tests hwctx->priv->mbox_chann
the scheduler keeps submitting jobs to a context which has no host buffer
mapped and no CU configured. Neither aie2_hwctx_resume() nor
aie2_sched_job_timedout() looks at the return value, so the context is
never cleaned up and the firmware context slot is leaked until the
hardware context is destroyed.
Tear the context down on the error path.
Fixes: aac243092b70 ("accel/amdxdna: Add command execution")
Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
---
drivers/accel/amdxdna/aie2_ctx.c | 10 +++++++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/drivers/accel/amdxdna/aie2_ctx.c b/drivers/accel/amdxdna/aie2_ctx.c
index 74d7019268cb..11f5421e497a 100644
--- a/drivers/accel/amdxdna/aie2_ctx.c
+++ b/drivers/accel/amdxdna/aie2_ctx.c
@@ -122,7 +122,7 @@ static int aie2_hwctx_restart(struct amdxdna_dev *xdna, struct amdxdna_hwctx *hw
heap->mem.size);
if (ret) {
XDNA_ERR(xdna, "Map host buf failed, ret %d", ret);
- goto out;
+ goto destroy;
}
xa_for_each_range(&hwctx->client->dev_heap_xa, heap_id, heap, 1,
@@ -132,16 +132,20 @@ static int aie2_hwctx_restart(struct amdxdna_dev *xdna, struct amdxdna_hwctx *hw
heap->mem.size);
if (ret) {
XDNA_ERR(xdna, "Add heap %ld failed ret %d", heap_id, ret);
- goto out;
+ goto destroy;
}
}
ret = aie2_config_cu(hwctx, NULL);
if (ret) {
XDNA_ERR(xdna, "Config cu failed, ret %d", ret);
- goto out;
+ goto destroy;
}
+ return 0;
+
+destroy:
+ aie2_hwctx_stop(xdna, hwctx, NULL);
out:
XDNA_DBG(xdna, "%s restarted, ret %d", hwctx->name, ret);
return ret;
--
2.34.1
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH V1] accel/amdxdna: Destroy the context when hwctx restart fails
2026-11-17 16:57 [PATCH V1] accel/amdxdna: Destroy the context when hwctx restart fails Lizhi Hou
@ 2026-11-17 20:54 ` Max Zhen
0 siblings, 0 replies; 2+ messages in thread
From: Max Zhen @ 2026-11-17 20:54 UTC (permalink / raw)
To: Lizhi Hou, ogabbay, quic_jhugo, dri-devel, mario.limonciello,
karol.wachowski
Cc: linux-kernel, sonal.santan
On 11/17/2026 Tue 08:57, Lizhi Hou wrote:
> aie2_hwctx_restart() creates the firmware context first, then maps the
> host buffers and configures the CUs. When any of the later steps fails,
> the error is propagated to the caller but the context created at the
> beginning is left behind: hwctx->fw_ctx_id stays valid, the mailbox
> channel stays allocated and started, and ndev->hwctx_num is not
> decremented. The hardware context is reported as failed yet remains
> alive, and because aie2_sched_job_run() only tests hwctx->priv->mbox_chann
> the scheduler keeps submitting jobs to a context which has no host buffer
> mapped and no CU configured. Neither aie2_hwctx_resume() nor
> aie2_sched_job_timedout() looks at the return value, so the context is
> never cleaned up and the firmware context slot is leaked until the
> hardware context is destroyed.
>
> Tear the context down on the error path.
>
> Fixes: aac243092b70 ("accel/amdxdna: Add command execution")
> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
Reviewed-by: Max Zhen <max.zhen@amd.com>
> ---
> drivers/accel/amdxdna/aie2_ctx.c | 10 +++++++---
> 1 file changed, 7 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/accel/amdxdna/aie2_ctx.c b/drivers/accel/amdxdna/aie2_ctx.c
> index 74d7019268cb..11f5421e497a 100644
> --- a/drivers/accel/amdxdna/aie2_ctx.c
> +++ b/drivers/accel/amdxdna/aie2_ctx.c
> @@ -122,7 +122,7 @@ static int aie2_hwctx_restart(struct amdxdna_dev *xdna, struct amdxdna_hwctx *hw
> heap->mem.size);
> if (ret) {
> XDNA_ERR(xdna, "Map host buf failed, ret %d", ret);
> - goto out;
> + goto destroy;
> }
>
> xa_for_each_range(&hwctx->client->dev_heap_xa, heap_id, heap, 1,
> @@ -132,16 +132,20 @@ static int aie2_hwctx_restart(struct amdxdna_dev *xdna, struct amdxdna_hwctx *hw
> heap->mem.size);
> if (ret) {
> XDNA_ERR(xdna, "Add heap %ld failed ret %d", heap_id, ret);
> - goto out;
> + goto destroy;
> }
> }
>
> ret = aie2_config_cu(hwctx, NULL);
> if (ret) {
> XDNA_ERR(xdna, "Config cu failed, ret %d", ret);
> - goto out;
> + goto destroy;
> }
>
> + return 0;
> +
> +destroy:
> + aie2_hwctx_stop(xdna, hwctx, NULL);
> out:
> XDNA_DBG(xdna, "%s restarted, ret %d", hwctx->name, ret);
> return ret;
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-11-17 20:54 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-11-17 16:57 [PATCH V1] accel/amdxdna: Destroy the context when hwctx restart fails Lizhi Hou
2026-11-17 20:54 ` Max Zhen
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®