Re: 3.2.11: PCI Express card cannot be re-detected withing cca 60sec timeframe

[Yinghai Lu]

On Mon, Apr 16, 2012 at 1:53 PM, Yinghai Lu <yinghai@xxxxxxxxxx> wrote:
> On Mon, Apr 16, 2012 at 12:09 PM, Martin Mokrejs
> <mmokrejs@xxxxxxxxxxxxxxxxxx> wrote:
>>>>> On Wed, Mar 21, 2012 at 3:46 PM, Martin Mokrejs
>>>>   I had some Ooopses with 3.4-rc2 but now I have tested 3.4-rc3. Still, linux
>>>> does NOT realize that I have removed an ExpressCard from the slot. It realizes
>>>> that only when I pluging the card back again. Actually, this is a case of a
>>>> NEC chipset base USB3.0 card.
>>>>
>>>>  In contrast, kernel DOES recognize that I unplugged a FireWire card from the
>>>> same slot.
>>>>
>>>>  This is a Dell Vostro 3550 laptop. Below is what syslog logged with my comment.
>>>> Empty lines between the log entries are inserted whe I either inserted or removed
>>>> a card. Probably a bit more will be in the attached dmesg file.
>
> so the interrupt for removal get generated, but pcie_isr() is delayed
> when another surprise is generated...
>

please try to revert
commit 486b10b9f43500741cd63a878d0ef23cd87fc66d
Author: Kenji Kaneshige <kaneshige.kenji@xxxxxxxxxxxxxx>
Date:   Mon Nov 7 20:56:50 2011 +0900

    PCI: pciehp: Handle push button event asynchronously

    Use non-ordered workqueue for attention button events.

    Attention button events on each slot can be handled asynchronously. So
    we should use non-ordered workqueue. This patch also removes ordered
    workqueue in pciehp as a result.

to see if there is any difference.

Yinghai

Attachment: commit-486b10b
Description: Binary data