17:36 <deathmist> bamse: ping, would you mind tagging qrtr, pd-mapper, rmtfs & tqftpserv repos?

18:52 <bamse> deathmist: sorry, forgot about it again...i don't mind, but i would prefer to look at a bug i found recently in tqftpserv before just tagging that

18:52 <bamse> deathmist: to avoid forgetting it i've created a ticket in our jira thing and assigned to myself...

18:56 <Tooniis[m]> there is currently this issue where slimbus-qcom-ngd-ctrl doesn't work if it's probed after adsp is powered on. It's because it relies on getting a ssr event to start, but it never gets its ssr notifier gets registered after adsp is powered on.

18:57 <Tooniis[m]> to fix it, I made some changes in remoteproc/qcom_common.c to keep track of the last ssr event issued and added a function to get this state, then made slimbus-qcom-ngd-ctrl call that function and start right away if it finds that adsp is already on

18:58 <Tooniis[m]> https://gitlab.com/msm8996-mainline/linux-msm8996/-/commit/e2e9610402c3f44da59c381447ca0a8df105a724

18:58 <Tooniis[m]> https://gitlab.com/msm8996-mainline/linux-msm8996/-/commit/275ac4efff0b4d8eb36f20dbe50825570e0bb16b

18:58 <Tooniis[m]> but I am not sure if this is the right approach

18:58 <Tooniis[m]> any comments?

18:59 <bamse> Tooniis[m]: i was worried that this would become an issue when i reviewed the ssr notification patches...

19:00 <bamse> Tooniis[m]: i don't like sprinkling getters for apis that are callback driven in nature, but i haven't spent enough thought on it yet...

19:02 <Tooniis[m]> I also thought of calling the notifier once when registered with the current state, but I am not sure if that might have some side effects

19:02 <bamse> Tooniis[m]: i fear all available options comes with side effects :/

19:03 <bamse> Tooniis[m]: but i haven't thought the details through

19:03 <bamse> Tooniis[m]: do you have the patch in your mailbox that broke this? (i.e. introduced the ssr notifier dependency)

19:04 <Tooniis[m]> I can look for it in the git history

19:04 <bamse> Tooniis[m]: would you be able to reply to that patch describing the problem, and make sure that you have Bjorn Andersson and Siddharth Gupta among the recipients?

19:05 <bamse> Siddharth must have run into this downstream...

19:05 <Tooniis[m]> I will look for it then

20:02 <steev> Tooniis[m]: what does that look like in dmesg output? i think i see that (or similar?) on one of my c630's but not the other

20:04 <steev> Tooniis[m]: something along these lines? https://bpa.st/25YA

20:08 <Tooniis[m]> <steev "Tooniis: what does that look lik"> nothing on dmesg

20:09 <Tooniis[m]> it just does nothing

20:09 <steev> ahh, okay.

20:10 <steev> you mentioned ssr notifications and i remembered my dmesg output complaining about ssr

20:59 <steev> Tooniis[m]: well, for what its worth, i pulled in those two patches here to my 5.13 tree and that does get rid of my null pointer