Skip to contentSkip to conversation

Problem with ide-cs

DISCUSSION Larry Finger 2 participants 3 messages Jul 30, 2011

lore
Share & tools
lore
#0 lore raw

With kernel 3.0, my external hard drive attached to a PCMCIA card reports "irq 11: nobody cared (try booting with the "irqpoll" option)". The driver in question is ide-cs. When I try irqpoll, the device works correctly. Naturally, the interrupt is shared with yenta, the PCMCIA adapter driver. In addition, that same interrupt is used by a BCM4318 Cardbus device.

I plan on preparing a Bugzilla entry for this problem, but I'm not quite ready. What I would like to know if there is any documentation on what kinds of changes are needed to fix the problem.

Thanks,

Larry

#1 lore raw

Larry,

Quoted 10 lines · Larry Finger

With kernel 3.0, my external hard drive attached to a PCMCIA card reports "irq 11: nobody cared (try booting with the "irqpoll" option)". The driver in question is ide-cs. When I try irqpoll, the device works correctly. Naturally, the interrupt is shared with yenta, the PCMCIA adapter driver. In addition, that same interrupt is used by a BCM4318 Cardbus device.

I plan on preparing a Bugzilla entry for this problem, but I'm not quite ready. What I would like to know if there is any documentation on what kinds of changes are needed to fix the problem.

Let me guess: it worked fine in 2.6.39, or .38? Output of "lspci -vv", "dmesg" for a working and a non-working kernel would be good starting points for trying to debug this issue.

Best,
Dominik

#2 lore raw
Quoted 10 lines · Dominik Brodowski

Larry,

Quoted 7 lines · Larry Finger

With kernel 3.0, my external hard drive attached to a PCMCIA card reports "irq 11: nobody cared (try booting with the "irqpoll" option)". The driver in question is ide-cs. When I try irqpoll, the device works correctly. Naturally, the interrupt is shared with yenta, the PCMCIA adapter driver. In addition, that same interrupt is used by a BCM4318 Cardbus device.

Quoted 7 lines
Quoted 3 lines

I plan on preparing a Bugzilla entry for this problem, but I'm not quite ready. What I would like to know if there is any documentation on what kinds of changes are needed to fix the problem.

Let me guess: it worked fine in 2.6.39, or .38? Output of "lspci -vv", "dmesg" for a working and a non-working kernel would be good starting points for trying to debug this issue.

It worked fine in 2.6.37 and is a regression. The bug is #40362.

I tried to bisect between 2.6.38 and 2.6.37, but didn't really get anywhere. As noted in the bug report, 38 and 39 lock the computer when I plug in. Only 3.0 comes back with the nobody cared message.

Larry