[ Regressions ] linux next 20201008: blk_update_request: I/O error, dev sda, sector 0 op 0x1:(WRITE) flags 0x800 phys_seg 0 prio class 0

[ Regressions ] linux next 20201008: blk_update_request: I/O error, dev sda, sector 0 op 0x1:(WRITE) flags 0x800 phys_seg 0 prio class 0

[Naresh Kamboju]

There are two major regressions noticed on linux next tag 20201008.
I will bisect this problem and get back to you.

1) qemu_i386 and qemu_x86 boot failed due to mount rootfs failing [1].

        Starting Remount Root and Kernel File Systems...
[    1.750740] ata1.00: WARNING: zero len r/w req
[    1.751423] ata1.00: WARNING: zero len r/w req
[    1.752361] ata1.00: WARNING: zero len r/w req
[    1.753400] ata1.00: WARNING: zero len r/w req
[    1.754447] ata1.00: WARNING: zero len r/w req
[    1.755529] ata1.00: WARNING: zero len r/w req
[    1.756630] sd 0:0:0:0: [sda] tag#0 FAILED Result:
hostbyte=DID_ERROR driverbyte=DRIVER_OK cmd_age=0s
[    1.758622] sd 0:0:0:0: [sda] tag#0 CDB: Synchronize Cache(10) 35
00 00 00 00 00 00 00 00 00
[    1.760576] blk_update_request: I/O error, dev sda, sector 0 op
0x1:(WRITE) flags 0x800 phys_seg 1 prio class 0
[    1.761534] Buffer I/O error on dev sda, logical block 0, lost sync
page write
[    1.764158] EXT4-fs (sda): I/O error while writing superblock


2) the devices boot pass but mkfs failed on x86_64, i386, arm64
Juno-r2 devices [2].

mkfs -t ext4 /dev/disk/by-id/ata-TOSHIBA_MG03ACA100_37O9KGL0F
[   72.159789] ata3.00: WARNING: zero len r/w req
[   72.164287] ata3.00: WARNING: zero len r/w req
[   72.168774] ata3.00: WARNING: zero len r/w req
[   72.168777] ata3.00: WARNING: zero len r/w req
[   72.168779] ata3.00: WARNING: zero len r/w req
[   72.168781] ata3.00: WARNING: zero len r/w req
[   72.168786] sd 2:0:0:0: [sda] tag#5 FAILED Result:
hostbyte=DID_ERROR driverbyte=DRIVER_OK cmd_age=0s
[   72.168788] sd 2:0:0:0: [sda] tag#5 CDB: Synchronize Cache(10) 35
00 00 00 00 00 00 00 00 00
[   72.168791] blk_update_request: I/O error, dev sda, sector 0 op
0x1:(WRITE) flags 0x800 phys_seg 0 prio class 0

Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>

metadata:
  git branch: master
  git repo: https://gitlab.com/Linaro/lkft/mirrors/next/linux-next
  git commit: e4fb79c771fbe2e6fcb3cffa87d5823a9bbf3f10
  git describe: next-20201008
  make_kernelversion: 5.9.0-rc8
  kernel-config:
https://builds.tuxbuild.com/pOW-FELX2VUycejkuyiKZg/kernel.config


steps to reproduce:
--------------------------
1) qemu boot command:

/usr/bin/qemu-system-x86_64 -cpu host -enable-kvm -nographic -net
nic,model=virtio,macaddr=DE:AD:BE:EF:66:06 -net tap -m 1024 -monitor
none -kernel bzImage --append "root=/dev/sda  rootwait
console=ttyS0,115200" -hda
rpb-console-image-lkft-intel-corei7-64-20200723162342-41.rootfs.ext4
-m 4096 -smp 4 -nographic

2) boot x86_64 with linux next 20201008 tag kernel and attach SDD drive.

mkfs -t ext4 /dev/<drive-partition>

Full log links,
[1 ]https://lkft.validation.linaro.org/scheduler/job/1823906#L688
[2] https://lkft.validation.linaro.org/scheduler/job/1823938#L2065


-- 
Linaro LKFT
https://lkft.linaro.org

Re: [ Regressions ] linux next 20201008: blk_update_request: I/O error, dev sda, sector 0 op 0x1:(WRITE) flags 0x800 phys_seg 0 prio class 0

[Naresh Kamboju]

On Thu, 8 Oct 2020 at 23:41, Naresh Kamboju <naresh.kamboju@linaro.org> wrote:
>
> There are two major regressions noticed on linux next tag 20201008.
> I will bisect this problem and get back to you.

Reverting scsi: patch set on  linux next tag 20201008 fixed reported problems.
git revert --no-edit 653eb7c99d84..ed7fb2d018fd


>
> 1) qemu_i386 and qemu_x86 boot failed due to mount rootfs failing [1].
>
>         Starting Remount Root and Kernel File Systems...
> [    1.750740] ata1.00: WARNING: zero len r/w req
> [    1.751423] ata1.00: WARNING: zero len r/w req
> [    1.752361] ata1.00: WARNING: zero len r/w req
> [    1.753400] ata1.00: WARNING: zero len r/w req
> [    1.754447] ata1.00: WARNING: zero len r/w req
> [    1.755529] ata1.00: WARNING: zero len r/w req
> [    1.756630] sd 0:0:0:0: [sda] tag#0 FAILED Result:
> hostbyte=DID_ERROR driverbyte=DRIVER_OK cmd_age=0s
> [    1.758622] sd 0:0:0:0: [sda] tag#0 CDB: Synchronize Cache(10) 35
> 00 00 00 00 00 00 00 00 00
> [    1.760576] blk_update_request: I/O error, dev sda, sector 0 op
> 0x1:(WRITE) flags 0x800 phys_seg 1 prio class 0
> [    1.761534] Buffer I/O error on dev sda, logical block 0, lost sync
> page write
> [    1.764158] EXT4-fs (sda): I/O error while writing superblock
>
>
> 2) the devices boot pass but mkfs failed on x86_64, i386, arm64
> Juno-r2 devices [2].
>
> mkfs -t ext4 /dev/disk/by-id/ata-TOSHIBA_MG03ACA100_37O9KGL0F
> [   72.159789] ata3.00: WARNING: zero len r/w req
> [   72.164287] ata3.00: WARNING: zero len r/w req
> [   72.168774] ata3.00: WARNING: zero len r/w req
> [   72.168777] ata3.00: WARNING: zero len r/w req
> [   72.168779] ata3.00: WARNING: zero len r/w req
> [   72.168781] ata3.00: WARNING: zero len r/w req
> [   72.168786] sd 2:0:0:0: [sda] tag#5 FAILED Result:
> hostbyte=DID_ERROR driverbyte=DRIVER_OK cmd_age=0s
> [   72.168788] sd 2:0:0:0: [sda] tag#5 CDB: Synchronize Cache(10) 35
> 00 00 00 00 00 00 00 00 00
> [   72.168791] blk_update_request: I/O error, dev sda, sector 0 op
> 0x1:(WRITE) flags 0x800 phys_seg 0 prio class 0
>
> Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
>
> metadata:
>   git branch: master
>   git repo: https://gitlab.com/Linaro/lkft/mirrors/next/linux-next
>   git commit: e4fb79c771fbe2e6fcb3cffa87d5823a9bbf3f10
>   git describe: next-20201008
>   make_kernelversion: 5.9.0-rc8
>   kernel-config:
> https://builds.tuxbuild.com/pOW-FELX2VUycejkuyiKZg/kernel.config
>
>
> steps to reproduce:
> --------------------------
> 1) qemu boot command:
>
> /usr/bin/qemu-system-x86_64 -cpu host -enable-kvm -nographic -net
> nic,model=virtio,macaddr=DE:AD:BE:EF:66:06 -net tap -m 1024 -monitor
> none -kernel bzImage --append "root=/dev/sda  rootwait
> console=ttyS0,115200" -hda
> rpb-console-image-lkft-intel-corei7-64-20200723162342-41.rootfs.ext4
> -m 4096 -smp 4 -nographic
>
> 2) boot x86_64 with linux next 20201008 tag kernel and attach SDD drive.
>
> mkfs -t ext4 /dev/<drive-partition>
>
> Full log links,
> [1 ]https://lkft.validation.linaro.org/scheduler/job/1823906#L688
> [2] https://lkft.validation.linaro.org/scheduler/job/1823938#L2065
>
>
> --
> Linaro LKFT
> https://lkft.linaro.org

Re: [ Regressions ] linux next 20201008: blk_update_request: I/O error, dev sda, sector 0 op 0x1:(WRITE) flags 0x800 phys_seg 0 prio class 0

[Jens Axboe]

On 10/8/20 2:05 PM, Naresh Kamboju wrote:
> On Thu, 8 Oct 2020 at 23:41, Naresh Kamboju <naresh.kamboju@linaro.org> wrote:
>>
>> There are two major regressions noticed on linux next tag 20201008.
>> I will bisect this problem and get back to you.
> 
> Reverting scsi: patch set on  linux next tag 20201008 fixed reported problems.
> git revert --no-edit 653eb7c99d84..ed7fb2d018fd

Just for everyones edification, that would be these 9 patches from the
SCSI tree:

Christoph Hellwig (9):
      scsi: core: Don't export scsi_device_from_queue()
      scsi: core: Remove scsi_init_cmd_errh
      scsi: core: Move command size detection out of the fast path
      scsi: core: Use rq_dma_dir in scsi_setup_cmnd()
      scsi: core: Rename scsi_prep_state_check() to scsi_device_state_check()
      scsi: core: Rename scsi_mq_prep_fn() to scsi_prepare_cmd()
      scsi: core: Clean up allocation and freeing of sgtables
      scsi: core: Remove scsi_setup_cmnd() and scsi_setup_fs_cmnd()
      scsi: core: Only start the request just before dispatching

-- 
Jens Axboe

Re: [ Regressions ] linux next 20201008: blk_update_request: I/O error, dev sda, sector 0 op 0x1:(WRITE) flags 0x800 phys_seg 0 prio class 0

[Christoph Hellwig]

On Thu, Oct 08, 2020 at 02:17:41PM -0600, Jens Axboe wrote:
> Just for everyones edification, that would be these 9 patches from the
> SCSI tree:

I sent the fixes out a bit ago and Cced the reporters..

Re: [ Regressions ] linux next 20201008: blk_update_request: I/O error, dev sda, sector 0 op 0x1:(WRITE) flags 0x800 phys_seg 0 prio class 0

[Martin K. Petersen]

Christoph,

> On Thu, Oct 08, 2020 at 02:17:41PM -0600, Jens Axboe wrote:
>> Just for everyones edification, that would be these 9 patches from the
>> SCSI tree:
>
> I sent the fixes out a bit ago and Cced the reporters..

I do not have any libata-connected devices in the SCSI test setup so
things worked fine for me yesterday. I have a retired Nehalem server in
the rack which has a couple of ATA 500GB disk drives in it. I'll try to
see if I can add that to my test pool. Just for good measure.

In any case the fixes are now in my for-next branch:

b6ba9b0e201a scsi: core: Set sc_data_direction to DMA_NONE for no-transfer commands
9120ac54cce6 scsi: sr: Initialize ->cmd_len

-- 
Martin K. Petersen	Oracle Linux Engineering