Red Hat Bugzilla – Bug 505648
PCI MSI breaks ESATA (ahci)
Last modified: 2009-09-28 03:33:03 EDT
Description of problem:
After upgrading to F11 (where pci msi is enabled by default). ESATA no longer works (disk does not show up).
Disabling MSI by passing "pci=nomsi" to the kernel fixes that.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
2. plug in a ESATA drive
It fails to detect the drive.
Disk should work.
Created attachment 347643 [details]
Errors in dmesg
When adding a an eSata PCI device to my vm configuration (SB700/SB800 Sata Controller AHCI Mode) receive this error: Kernel; journal commit I/O error. This immediately delets the vm machine from the virt-manager and locks up the host. To return to normal a hard, power off, reset is necessary. There is no way a guest configuration should shutdown the host.
In inspecting the output of mount, I found my eSata disk mounts to the Fedora host at /dev/sdb1. I then went into vm configuration and added a storage disk for this device and seleted SCSI type. With the eSata disk plugged in at vm startup the vm Windows machine saw this disk and installed the driver automatically. I repeated this process only this time selecting IDE disk and it also worked via the eSata connection.
I believe the eSata devices should be plug and play just like USB. If so, how are is eSata to be configured or is this work in progress?
dv7-1230us, Fedora standard release with updates as of 7/13. No testing updates.
(In reply to comment #0)
> Description of problem:
> After upgrading to F11 (where pci msi is enabled by default). ESATA no longer
> works (disk does not show up).
> Disabling MSI by passing "pci=nomsi" to the kernel fixes that.
> Version-Release number of selected component (if applicable):
> How reproducible:
> Steps to Reproduce:
> 1. boot
> 2. plug in a ESATA drive
> 3. fail
> Actual results:
> It fails to detect the drive.
> Expected results:
> Disk should work.
I can not reproduce this with 2.6.30 builds. (i.e seems fixed in 2.6.30)
Paul: Your report seems to be a different issue (not related that MSI breaks eSata)