zstas has quit [Remote host closed the connection]
zstas has joined #linux-msm
valpackett has quit [Read error: Connection reset by peer]
valpackett has joined #linux-msm
f_ is now known as funderscore
jhovold has joined #linux-msm
pespin has joined #linux-msm
msumro has joined #linux-msm
<MainudeenSumro[m]>
Still no update can you guys help me fix modem
zstas has quit [Remote host closed the connection]
<loki666>
Hi, since 6.15-rc1 on Odin2 (SM8550), I'm having some kind of stuttering/flickering on GPU. It's not freezing...
<loki666>
Could someone help me find patches that could make this behavior, so I can try to revert et investigate ? Thanks
msumro has quit [Quit: Connection closed for inactivity]
<z3ntu>
bamse: Since I see you working on mdt-loader bits, if you want a small challenge for a quiet day, this old msm8226 HTC One Mini 2 has some funky wcnss firmware that seems to defy some existing expectations. Basically we need this hack to boot wcnss on this phone https://github.com/msm8226-mainline/linux/commit/1fd99ec3dd923ab0626e56a1d6e3be9f50326894
<z3ntu>
got some notes and logs here https://public.lucaweiss.eu/tmp/memul-wcnss-notes.txt , but basically iirc the complete wcnss.mdt needs to be loaded into memory, not just what we think we need to copy based on some headers
<z3ntu>
Personally I'm happy enough with having found this hack to be able to use wcnss on it, and I need to carry this tree probably indefinitely anyways, and apart from some other super old boards I don't imagine fixing or working around this will be useful for other boards fwiw :D
Caterpillar has joined #linux-msm
jhovold has quit [Ping timeout: 480 seconds]
<bamse>
z3ntu: so p_filesz of 0x168 is completely wrong(?)
<bamse>
in addition to the mdt being padded with 0xff, so the equality check doesn't work
<z3ntu>
bamse apparently, not sure which part of it is technically wrong, but somehow HTCs build system must've spit out the file, and given how the downstream driver works, maybe it just didn't matter?
<bamse>
z3ntu: iirc the ELF header is typically lying about the offset of the hash segment...so i guess there was a buggy version of the tool that didn't put the correct size either(?)
pespin has quit [Remote host closed the connection]
<bamse>
z3ntu: i wonder if case 1 and 3 in that statement in that function handles split binaries and case 2 solely deals with non-split firmware...
<bamse>
z3ntu: if so this would have been better written as if (split) 3 || 1 else 2
<bamse>
z3ntu: but i am affraid of poking that code, as i don't have the means to test all these variants
Caterpillar has quit [Quit: Konversation terminated!]