Add upstream patch to disable ASPM L0s for SA8775P#2176
Conversation
Kavinaya99
requested review from
anujm1,
koenkooi,
lumag,
ndechesne,
ricardosalveti,
sbanerjee-quic and
vkraleti
as code owners
| Subject: [PATCH] Revert "PCI: dwc: Remove MSI/MSIX capability for Root Port if | ||
| iMSI-RX is used as MSI controller" | ||
|
|
||
| Upstream-Status: Inappropriate [platform-specific regression workaround] |
There was a problem hiding this comment.
Misplaced tag (all tags should be at the end of the commit message).
Please get this sorted out upstream and then provide a backport here.
There was a problem hiding this comment.
Sure, uploaded the new patch with the modifications
|
https://lore.kernel.org/all/20260419093934.1223027-1-shengchao.guo@oss.qualcomm.com/ should solve this. You may bring this change as a "Submitted:" patch to unblock Leamans testing. |
@vkraleti please don't give bad advices. See the responses from Mani. The patch is invalid. |
Add an upstream-submitted PCIe patch that disables ASPM L0s for the SA8775P Root Port configuration. This avoids Data Link Layer Replay Timer Timeout errors observed on SA8775P/Lemans platforms during early boot. Signed-off-by: Kavinaya S <kavinaya@qti.qualcomm.com>
Agree this isn’t a right change. However, we need to unblock linux-yocto-dev testing on Lemans. The last successful testing was on 6.19, about two months ago. Since then, we haven’t been able to validate Lemans on any newer upstream kernels. At this point, it looks like we have only two viable options:
Are there any better suggestions? |
|
BTW, qcom-next took the revert approach to unblock. |
|
@vkraleti Neither of the patches is correct. The revert simply enables unsupported capability, possibly causing MSI interrupts to be missed. What is causing AER erorrs? Could we try lowering PCIe link speed instead? |
Kavinaya99
changed the title
3 participants
Add upstream patch to disable ASPM L0s for SA8775P