Re: 2.6.25-rc1 ahci regression

[Anton Vorontsov <cbouatmailru@gmail.com>]

On Thu, Jun 03, 2010 at 04:11:57PM +0800, Luming Yu wrote:
> On Thu, Jun 3, 2010 at 3:54 PM, Anton Vorontsov <cbouatmailru@gmail.com> wrote:
[...]
> >> I happened to notice that the commit 365cfa1ed5 caused the following
> >> boot error with 2.6.25-rc1
[...]
> > Can you provide full boot log (dmesg) please?
[...]

I don't see any AHCI devices probed.

[...]
> ------------[ cut here ]------------
> WARNING: at fs/sysfs/dir.c:451 sysfs_add_one+0xc9/0x130()
> Hardware name: X8DTN
> sysfs: cannot create duplicate filename
> '/devices/pci0000:00/0000:00:1c.4/0000:0f:00.0/slot'

This looks suspicious.

> usb 1-5: New USB device found, idVendor=14dd, idProduct=0002

(...this is async USB stuff, doesn't matter...)

> Modules linked in:
> Pid: 1, comm: swapper Not tainted 2.6.35-rc1 #1
> Call Trace:
>  [<ffffffff8105ed8f>] warn_slowpath_common+0x7f/0xc0
>  [<ffffffff8105ee86>] warn_slowpath_fmt+0x46/0x50
>  [<ffffffff811ad969>] sysfs_add_one+0xc9/0x130
>  [<ffffffff811ae0f8>] sysfs_do_create_link+0xd8/0x1a0
>  [<ffffffff811ae1f3>] sysfs_create_link+0x13/0x20
>  [<ffffffff81236466>] pci_create_sysfs_dev_files+0x3d6/0x5f0

Anyway, I just tried to move code around from libahci.c
back to ahci.c by hand, to verify that there are actually
no code changes, and there isn't.

So, I wonder, have you actually bisected the thing correctly?
Does "git reset --hard 365cfa1ed5^" actually help and does
"git reset --hard 365cfa1ed5" actually break things?