Confirming on another Intel 9-series system (different board vendor,
same SATA controller ID).
Hardware: ASRock Z97 Pro4, Intel 9 Series AHCI (00:1f.2) [8086:8c82],
Samsung SSD 750 EVO 120GB (fw MAT01B6Q), plain ext4 root.
Kernel: 7.0.0-42.42 (also the build in proposed2 - not fixed).
Without workaround (max_sectors_kb=4096):
- failed WRITE/READ FPDMA QUEUED (mostly 4 MiB), SError { UnrecovData
HostInt Handshk }, link down to 1.5 Gbps, ext4 I/O errors, root fs
remounted read-only
- 347 DMAR faults in one boot, all on [00:1f.2], e.g.:
DMAR: [DMA Read NO_PASID] Request device [00:1f.2] fault addr 0xc7800000
[fault reason 0x0c] non-zero reserved fields in PTE
- SMART CRC_Error_Count 52 -> 135
- triggered even by "apt update"
Also tested, did not fix it:
- libata.force=1:noncq (errors continue as WRITE DMA EXT failures)
- ahci.mobile_lpm_policy=1 (7.0 defaults to lpm-pol 3 here, 6.8 to
lpm-pol 0; forcing max_performance reduced errors but did not stop them)
- new SATA cable
With libata.force=max_sec=2560 (max_sectors_kb=1280):
- reproducer from the description (8 GiB, bs=4M, buffered, fsync=16):
err=0, 154 MiB/s
- 0 failed commands, 0 DMAR faults, CRC unchanged
Kernel 6.8.0-147 on the same hardware: no errors under the same load.
** Attachment added: "Log files"
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2171415/+attachment/6013450/+files/dmesg-7.0-logs.zip
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2171415
Title:
[Regression] Intel H97 SATA I/O errors and DMAR faults with kernel
7.0, avoided by libata.force=max_sec=2560
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2171415/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
