| Message ID | 1709704406-22501-1-git-send-email-quic_dikshita@quicinc.com (mailing list archive) |
|---|---|
| State | Not Applicable |
| Headers | show |
| Series | media: venus: avoid multiple core dumps | expand |
On 06/03/2024 05:53, Dikshita Agarwal wrote: > Core dump is generated whenever there is system error reported > by firmware. Right now, multiple such dumps are generated if > recovery fails in first attempt, since the sys error handler is > invoked again for every failed recovery. > To avoid it, add conditional check to generate core dump only > once during every system error notification from firmware. > > Signed-off-by: Dikshita Agarwal <quic_dikshita@quicinc.com> > --- Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
On 06/03/2024 05:53, Dikshita Agarwal wrote: > Core dump is generated whenever there is system error reported > by firmware. Right now, multiple such dumps are generated if > recovery fails in first attempt, since the sys error handler is > invoked again for every failed recovery. > To avoid it, add conditional check to generate core dump only > once during every system error notification from firmware. > > Signed-off-by: Dikshita Agarwal <quic_dikshita@quicinc.com> > --- From my non-filter mail addr Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
diff --git a/drivers/media/platform/qcom/venus/core.c b/drivers/media/platform/qcom/venus/core.c index a712dd4..cc77900 100644 --- a/drivers/media/platform/qcom/venus/core.c +++ b/drivers/media/platform/qcom/venus/core.c @@ -67,6 +67,7 @@ static void venus_event_notify(struct venus_core *core, u32 event) mutex_lock(&core->lock); set_bit(0, &core->sys_error); + set_bit(0, &core->dump_core); list_for_each_entry(inst, &core->instances, list) inst->ops->event_notify(inst, EVT_SESSION_ERROR, NULL); mutex_unlock(&core->lock); @@ -109,7 +110,10 @@ static void venus_sys_error_handler(struct work_struct *work) venus_shutdown(core); - venus_coredump(core); + if (test_bit(0, &core->dump_core)) { + venus_coredump(core); + clear_bit(0, &core->dump_core); + } pm_runtime_put_sync(core->dev); diff --git a/drivers/media/platform/qcom/venus/core.h b/drivers/media/platform/qcom/venus/core.h index 4a63326..c04ea62 100644 --- a/drivers/media/platform/qcom/venus/core.h +++ b/drivers/media/platform/qcom/venus/core.h @@ -233,6 +233,7 @@ struct venus_core { u32 minor; u32 rev; } venus_ver; + unsigned long dump_core; }; struct vdec_controls {
Core dump is generated whenever there is system error reported by firmware. Right now, multiple such dumps are generated if recovery fails in first attempt, since the sys error handler is invoked again for every failed recovery. To avoid it, add conditional check to generate core dump only once during every system error notification from firmware. Signed-off-by: Dikshita Agarwal <quic_dikshita@quicinc.com> --- drivers/media/platform/qcom/venus/core.c | 6 +++++- drivers/media/platform/qcom/venus/core.h | 1 + 2 files changed, 6 insertions(+), 1 deletion(-)