CQE incremental fixes part 2 #6419
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
Author
|
From my testing, the loop that previously failed in 20 minutes survived an overnight run. There's one positive response in the issue, and one pending. |
Contributor
Author
|
This PR now fixes two definite bugs in the CQ implementation so the default can switch back to enabled for supported cards. During testing there was a mild panic as the other "known good" A2 card from Integral would reliably break the filesystem if trimmed. However, that's unrelated to CQ - it just doesn't erase properly. |
Posted write tracking introduced in the commit below raced with re-use of the requests between completion and submission, potentially causing underflow of the pending write count. Fixes: e6c1e86 ("mmc: restrict posted write counts for SD cards in CQ mode") Signed-off-by: Jonathan Bell <[email protected]>
__mmc_start_request turns CQE off before issuing the actual command so this additional call is redundant. This reverts commit 8e40644.
Cards with manufacture dates in 2019 and 2020 have been seen in the wild that hang indefinitely if issued a cache flush command in CQ mode. Signed-off-by: Jonathan Bell <[email protected]>
Only CMD38 with Arg=0x1 (Discard) is supported when in CQ mode, so turn it off before issuing a non-discard erase op. Signed-off-by: Jonathan Bell <[email protected]>
This reverts commit 1b92c93. Signed-off-by: Jonathan Bell <[email protected]>
Recent Integral cards end up with corrupt sectors after a flash erase. This covers sizes for the A2 range, which can't be differentiated from the A1 range which might not have the same issue. Signed-off-by: Jonathan Bell <[email protected]>
P33M
force-pushed
the
sd_cqe_fixes_pt2
branch
from
26a071c to
d0abbb2
Compare
popcornmix
added a commit
to raspberrypi/firmware
that referenced
this pull request
Oct 18, 2024
See: raspberrypi/linux#6421 kernel: Fixes for video-mux with Pi5 See: raspberrypi/linux#6410 kernel: OV5647 link frequency control See: raspberrypi/linux#6423 kernel: CQE incremental fixes part 2 See: raspberrypi/linux#6419
popcornmix
added a commit
to raspberrypi/rpi-firmware
that referenced
this pull request
Oct 18, 2024
See: raspberrypi/linux#6421 kernel: Fixes for video-mux with Pi5 See: raspberrypi/linux#6410 kernel: OV5647 link frequency control See: raspberrypi/linux#6423 kernel: CQE incremental fixes part 2 See: raspberrypi/linux#6419
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
If #6418 fixes the generic case of a hang on boot with an A2 card, then this PR should be merged.
The 2nd commit should be a no-op, as I placed an assert in cqe_off() for cq_host->qcnt != 0 and it never fired. As it turns out, mmcblk /does/ serialise async requests if a synchronous one arrives, and the existing claim mechanism should prevent other contexts from running in parallel. Retuning is a bit special as it is invoked prior to starting either a cqe or regular request, but it appears to have been thought about.
The third commit should disable CQ on older Sandisk cards such as https://forums.raspberrypi.com/viewtopic.php?t=367459&start=25#p2208495 (and I have one of these).