Skip to contentSkip to conversation

venus: fix multiple encoder crash

PATCH In review Mansur Alisha Shaik 2 participants 2 messages May 1, 2020 – Jun 2, 2020

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 venus: fix multiple encoder crash

Currently we are considering the instances which are available in core->inst list for load calculation in min_loaded_core() function, but this is incorrect because by the time we call decide_core() for second instance, the third instance not filled yet codec_freq_data pointer.

Solve this by considering the instances whose session has started.

---
Changes in V2:
- As per Alex and Jeffrey comments, elaborated problem and addressed review comments.

 drivers/media/platform/qcom/venus/pm_helpers.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/media/platform/qcom/venus/pm_helpers.c +4 -0
index abf9315..531e7a4 100644
@@ -496,6 +496,10 @@ min_loaded_core(struct venus_inst *inst, u32 *min_coreid, u32 *min_load)
 	list_for_each_entry(inst_pos, &core->instances, list) {
 		if (inst_pos == inst)
 			continue;
+
+		if (inst_pos->state != INST_START)
+			continue;
+
 		vpp_freq = inst_pos->clk_data.codec_freq_data->vpp_freq;
 		coreid = inst_pos->clk_data.core_id;
 
Signature · 1 line
2.7.4
Tested-by on v2 1/1
#1 lore raw

Hi,

Quoted 9 lines · Mansur Alisha Shaik

Currently we are considering the instances which are available in core->inst list for load calculation in min_loaded_core() function, but this is incorrect because by the time we call decide_core() for second instance, the third instance not filled yet codec_freq_data pointer.

Solve this by considering the instances whose session has started.

Quoted 8 lines

---
Changes in V2:
- As per Alex and Jeffrey comments, elaborated problem and addressed review comments.

 drivers/media/platform/qcom/venus/pm_helpers.c | 4 ++++
 1 file changed, 4 insertions(+)

This fixes the same crash I reported here:

https://lore.kernel.org/r/20200601150314.RFC.1.I1e40623bbe8fa43ff1415fc273cba66503b9b048@changeid

Thus:

Fixes eff82f79c562 ("media: venus: introduce core selection")

I still have the same reservations I expressed in the patch I posed about whether this is truly safe from a locking point of view, but certainly it puts us in a better state than we are today.

-Doug