Bug 216644
Summary: | Host OS hangs when enabling VMD in UEFI setup | ||
---|---|---|---|
Product: | Drivers | Reporter: | Adrian Huang (ahuang12) |
Component: | PCI | Assignee: | drivers_pci (drivers_pci) |
Status: | NEW --- | ||
Severity: | high | CC: | bjorn, francisco.munoz.ruiz, jonathan.derrick, nirmal.patel |
Priority: | P1 | ||
Hardware: | All | ||
OS: | Linux | ||
Kernel Version: | 6.1-rc2 | Subsystem: | |
Regression: | No | Bisected commit-id: | |
Attachments: |
OS Log (Serial Log)
[dmesg] apply-vmd-Fix-secondary-bus-reset-for-Intel-bridges |
Description
Adrian Huang
2022-10-31 06:05:43 UTC
Created attachment 303124 [details] [dmesg] apply-vmd-Fix-secondary-bus-reset-for-Intel-bridges Tried to apply the upstream discussion: https://lore.kernel.org/all/20220923203757.4918-1-francisco.munoz.ruiz@linux.intel.com/ The patch gets worse (without reverting 6aab5622296b: apply the patch on top of 6.1-rc3). The bus reset spends about 65 seconds until it gives up (timeout = 60 ms --> PCIE_RESET_READY_POLL_MS) And, the original issue is still there. ... [ 9.097319] vmd 0000:64:05.5: PCI host bridge to bus 10000:00 ... [ 11.580708] pci 10000:00:00.0: not ready 1023ms after bus reset; waiting [ 12.668707] pci 10000:00:00.0: not ready 2047ms after bus reset; waiting [ 14.780708] pci 10000:00:00.0: not ready 4095ms after bus reset; waiting [ 19.260716] pci 10000:00:00.0: not ready 8191ms after bus reset; waiting [ 27.964707] pci 10000:00:00.0: not ready 16383ms after bus reset; waiting [ 44.860707] pci 10000:00:00.0: not ready 32767ms after bus reset; waiting [ 81.212716] pci 10000:00:00.0: not ready 65535ms after bus reset; giving up ... [ 104.424169] DMAR: VT-d detected Invalidation Queue Error: Reason f [ 104.424172] DMAR: VT-d detected Invalidation Time-out Error: SID ffff [ 104.431382] DMAR: VT-d detected Invalidation Completion Error: SID ffff [ 104.438919] DMAR: QI HEAD: UNKNOWN qw0 = 0x0, qw1 = 0x0 [ 104.452790] DMAR: QI PRIOR: UNKNOWN qw0 = 0x0, qw1 = 0x0 [ 104.459042] DMAR: Invalidation Time-out Error (ITE) cleared ... I forwarded the original report to the VMD folks and the linux mailing lists: https://lore.kernel.org/r/20221031113924.GA1081013@bhelgaas Most people don't follow bugzilla, so you might want to follow-up to that message and cc: Francisco to make sure he sees it. We are looking into this issue. I will get back to you as soon as I get some update. Does disabling Interrupt remap under T-d setting in BIOS make any difference? Adrian, Did you try what Nirmal suggested? --Francisco. Hi Adrian, I'd like to add that it seems you are using a non-Intel BIOS. Can we rule out a BIOS problem first? --Francisco. (In reply to Francisco Munoz-Ruiz from comment #4) > Adrian, > > Did you try what Nirmal suggested? > > --Francisco. Hi Francisco, Yes, I tried and replied the result on Nov 4: https://lore.kernel.org/all/TYZPR03MB6192E7E96D51489ABF38C4CCB33B9@TYZPR03MB6192.apcprd03.prod.outlook.com/ Adrian, Can you check with your BIOS provider.? Thanks, Francisco. |