qat: Improve error recovery flows
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
linux (Ubuntu) |
Fix Released
|
Undecided
|
Thibf | ||
Noble |
Fix Released
|
Undecided
|
Thibf |
Bug Description
[Impact]
This set improves the error recovery flows in the QAT drivers and
adds a mechanism to test it through an heartbeat simulator.
This is an upstream patch set applied to linux-next and scheduled for 6.9.
Link to the upstream submission:
https:/
We should apply this set to the Noble 6.8 kernel,
in order to experience less issues with qat and improve maintainability.
An added commit is required to update the configuration.
[Test case]
Unload and reload the module to verify that qat recover
and log issues properly. Use the added error injection mechanism
to verify the recovery flow.
[Fix]
Apply the following commits (from linux-next):
2ecd43413d76 Documentation: qat: fix auto_reset section
7d42e097607c crypto: qat - resolve race condition during AER recovery
c2304e1a0b80 crypto: qat - change SLAs cleanup flow at shutdown
9567d3dc7609 crypto: qat - improve aer error reset handling
750fa7c20e60 crypto: qat - limit heartbeat notifications
f5419a4239af crypto: qat - add auto reset on error
2aaa1995a94a crypto: qat - add fatal error notification
4469f9b23468 crypto: qat - re-enable sriov after pf reset
ec26f8e6c784 crypto: qat - update PFVF protocol for recovery
758a0087db98 crypto: qat - disable arbitration before reset
ae508d7afb75 crypto: qat - add fatal error notify method
e2b67859ab6e crypto: qat - add heartbeat error simulator
[Regression potential]
We may experience qat regression when crashing or restarting the module.
description: | updated |
Changed in linux (Ubuntu Noble): | |
status: | New → In Progress |
description: | updated |
Changed in linux (Ubuntu Noble): | |
assignee: | nobody → Thibf (thibf) |
summary: |
- qat: Improve error/reset handling + qat: Improve error recovery flows |
Changed in linux (Ubuntu Noble): | |
status: | In Progress → Fix Committed |
Changed in linux (Ubuntu Noble): | |
status: | Fix Committed → Fix Released |
This patchset was included in noble/linux 6.8.0-16.16 which has not been promoted to the -release pocket yet:
linux | 6.8.0-11.11 | noble | source
linux | 6.8.0-20.20 | noble-proposed | source
Therefore I'm resetting the status to fix committed.