Bug 6086 - System can freeze when inserting/removing CompactFlash-Cards into PCMCIA slot
Summary: System can freeze when inserting/removing CompactFlash-Cards into PCMCIA slot
Status: CLOSED PATCH_ALREADY_AVAILABLE
Alias: None
Product: Drivers
Classification: Unclassified
Component: PCMCIA (show other bugs)
Hardware: i386 Linux
: P2 normal
Assignee: linux-pcmcia
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-02-16 00:45 UTC by Michael
Modified: 2006-02-17 03:59 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.6.15 (2.6.14?)
Subsystem:
Regression: ---
Bisected commit-id:


Attachments

Description Michael 2006-02-16 00:45:53 UTC
Most recent kernel where this bug did not occur: 2.6.9 (was running with
hotplug, now on udev), have not tested other kernels, however someone else
reported the same problem with 2.6.14 on a mailing list
Distribution: Kanotix
Hardware Environment: Thinkpad X31
Software Environment: Debian SID (udev 0.0084, hal 0.5.6)
Problem Description:
System generates hundreds of kernel messages "kernel: hde: hde1" when a PCCARD
(compact flash) is inserted into the PCMCIA slot. CPU kernel usage goes to 100%.
 
udev generates the device /dev/hde and a few seconds later also /dev/hde1. The
latter, however, is deleted directly. Then, a few seconds later, it's generated
again and directly deleted. This repeats, until the card is removed. Note: It is
possible, that the node is generated and deleted much much more often, maybe a
hundred times per second, which I, however, cannot monitor.

When the card is removed, the system, in most cases, returns to normal (cpu
usage). In some cases, however, the system freezes completely.

Here is the syslog:
Feb 16 09:29:34 LaptopMB kernel: pccard: PCMCIA card inserted into slot 1
Feb 16 09:29:34 LaptopMB kernel: pcmcia: registering new device pcmcia1.0
Feb 16 09:29:34 LaptopMB kernel: hde: CF16W17AE1, CFA DISK drive
Feb 16 09:29:34 LaptopMB kernel: hdf: probing with STATUS(0x50) instead of
ALTSTATUS(0x0a)
Feb 16 09:29:35 LaptopMB kernel: hdf: probing with STATUS(0x00) instead of
ALTSTATUS(0x0a)
Feb 16 09:29:36 LaptopMB kernel: hdf: probing with STATUS(0x50) instead of
ALTSTATUS(0x0a)
Feb 16 09:29:36 LaptopMB kernel: hdf: probing with STATUS(0x00) instead of
ALTSTATUS(0x0a)
Feb 16 09:29:37 LaptopMB kernel: ide2 at 0x4100-0x4107,0x410e on irq 3
Feb 16 09:29:37 LaptopMB kernel: hde: max request size: 128KiB
Feb 16 09:29:37 LaptopMB kernel: hde: 31488 sectors (16 MB) w/1KiB Cache,
CHS=246/4/32
Feb 16 09:29:37 LaptopMB kernel: hde: cache flushes not supported
Feb 16 09:29:37 LaptopMB kernel:  hde: hde1
Feb 16 09:29:37 LaptopMB kernel: ide-cs: hde: Vcc = 3.3, Vpp = 0.0
Feb 16 09:29:38 LaptopMB kernel:  hde: hde1
Feb 16 09:30:09 LaptopMB last message repeated 268 times
Feb 16 09:30:23 LaptopMB last message repeated 148 times
Feb 16 09:30:23 LaptopMB kernel: pccard: card ejected from slot 1

When the system freezes while removing the card, the last message is not in the
syslog.

Steps to reproduce:
Insert a compact flash card in the system described above.
Comment 1 Bartlomiej Zolnierkiewicz 2006-02-16 01:40:24 UTC
should be fixed in 2.6.16-rc3, please try it
Comment 2 Andrew Morton 2006-02-16 02:03:35 UTC
Bart says "udev thing (or rather IDE bug triggered by udev), should be fixed in
2.6.16-rc3".

So, please test 2.6.16-rc3?
Comment 3 Michael 2006-02-16 02:59:11 UTC
Thanks for the VERY fast replies!!!

As for Kanotix there is no 2.6.16-rc3 based kernel available so far, I will
download the source and compile one. I'll report back, as soon as I see, if the
new kernel works. (Will take some time, I guess...)
Comment 4 Michael 2006-02-17 03:59:49 UTC
O.K. I've tested 2.6.16-rc3 and you're right. The problem is fixed here.

Thx,

Michael

Note You need to log in before you can comment on or make changes to this bug.