mbox series

[0/1,SRU,F/OEM-5.6/G] UBUNTU: SAUCE: Fix disk probing under SATA controller behind VMD

Message ID 20200908080334.1106239-1-vicamo.yang@canonical.com
Headers show
Series UBUNTU: SAUCE: Fix disk probing under SATA controller behind VMD | expand

Message

You-Sheng Yang Sept. 8, 2020, 8:03 a.m. UTC
BugLink: https://bugs.launchpad.net/bugs/1894778

[Impact]

When booting with a certain platforms with boot disk attached to SATA
bus behind Intel VMD controller, disk probing may fail with following
error messages left in dmesg:

  [ 6.163286] ata1.00: qc timeout (cmd 0xec)
  [ 6.165630] ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4)

[Fix]

Proposed kernel patch https://patchwork.kernel.org/patch/11758345/

[Test Case]

Check dmesg/lsblk for disk probe.

For pci MSI address, check lspci output:

  $ lspci -vvnn
  ....
      Capabilities: [80] MSI: Enable+ Count=1/1 maskable- 64bit-
          Address: fee00000  Data: 0000

When it fails, the address is fee00000. And with a patched kernel:

  $ lspci -vvnn
  ....
      Capabilities: [80] MSI: Enable+ Count=1/1 maskable- 64bit-
          Address: fee01000  Data: 0000

[Regression Potential]
Low.  For previous NVMe based platforms, this patch brings no effective
change for NVMe devices because they will still stay in fast-interrupt
list.

Jon Derrick (1):
  PCI: vmd: Add AHCI to fast interrupt list

 drivers/pci/controller/vmd.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

Comments

Timo Aaltonen Sept. 8, 2020, 8:39 a.m. UTC | #1
On 8.9.2020 11.03, You-Sheng Yang wrote:
> BugLink: https://bugs.launchpad.net/bugs/1894778
> 
> [Impact]
> 
> When booting with a certain platforms with boot disk attached to SATA
> bus behind Intel VMD controller, disk probing may fail with following
> error messages left in dmesg:
> 
>   [ 6.163286] ata1.00: qc timeout (cmd 0xec)
>   [ 6.165630] ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4)
> 
> [Fix]
> 
> Proposed kernel patch https://patchwork.kernel.org/patch/11758345/
> 
> [Test Case]
> 
> Check dmesg/lsblk for disk probe.
> 
> For pci MSI address, check lspci output:
> 
>   $ lspci -vvnn
>   ....
>       Capabilities: [80] MSI: Enable+ Count=1/1 maskable- 64bit-
>           Address: fee00000  Data: 0000
> 
> When it fails, the address is fee00000. And with a patched kernel:
> 
>   $ lspci -vvnn
>   ....
>       Capabilities: [80] MSI: Enable+ Count=1/1 maskable- 64bit-
>           Address: fee01000  Data: 0000
> 
> [Regression Potential]
> Low.  For previous NVMe based platforms, this patch brings no effective
> change for NVMe devices because they will still stay in fast-interrupt
> list.
> 
> Jon Derrick (1):
>   PCI: vmd: Add AHCI to fast interrupt list
> 
>  drivers/pci/controller/vmd.c | 8 +++-----
>  1 file changed, 3 insertions(+), 5 deletions(-)
> 

applied to oem-5.6, thanks
Paolo Pisati Sept. 10, 2020, 9:24 a.m. UTC | #2
On Tue, Sep 08, 2020 at 04:03:33PM +0800, You-Sheng Yang wrote:
> BugLink: https://bugs.launchpad.net/bugs/1894778

Did it land in any upstream tree?