Attention is currently required from: Nico Huber, Thomas Heijligen, Angel Pons. Anastasia Klimchuk has posted comments on this change. ( https://review.coreboot.org/c/flashrom/+/56103 )
Change subject: spi_master: Use new API to register shutdown function ......................................................................
Patch Set 7:
(4 comments)
Patchset:
PS6:
This also fixes a lot resource leakage in case register_shutdown() […]
Yes that's a very good thing, I can boast in commit message, thanks :)
Programmers which have tests and affected here: linux_spi, dediprog, mec1308 and ene_lpc (since they are not removed yet), also realtek has test which is under review CB:56911 but I verified it passes on the top of this patch.
I have test which makes register_shutdown to fail, it is buried somewhere in my local repo, I never sent it. I can return back to it given that CB:56753 works. Totally makes sense to test error paths. Although it is very likely to take longer [to write proper test for error paths] than it seems to me now.
Patchset:
PS7: I rebased and updated commit message, and lost all the votes.
File ch341a_spi.c:
https://review.coreboot.org/c/flashrom/+/56103/comment/2ea2c6cd_072b87d7 PS6, Line 511: return 0;
Just raising an eyebrow: this changes unrelated behaviour. Maybe worth to […]
This is right! Added to commit message.
File dediprog.c:
https://review.coreboot.org/c/flashrom/+/56103/comment/fa345d0b_fd2a3d71 PS6, Line 1276: /* shutdown function does cleanup */
Still somewhat true, but less obvious what is meant.
This is not a final state for dediprog, I want it to return register_spi_master() like the others. But I recalled there is CB:51706, I think we need to discuss and decide with the author of that patch what to do.