Re: Bug#351685: udev storm: pcmcia CF adapter device file constantly deleted, recreated
reassign 351685 linux-2.6
thanks
Hi,
Joey Hess:
> I inserted a pcmcia CF adapter containing a partition with a FAT filesystem
> and tried to mount it, but /dev/hde1 was not there, though /dev/hde was.
> I eventually realised that hde1 was being constantly deleted and
> recreated over and over.
I don't think this is a bug in pcmciautils. It's quite a simple
package which doesn't do much. It is much more likely that the bug is
in the kernel, although I guess it could be udev's fault
too. Reassigning to the kernel for now.
> ls showed it there and missing, randomly.
>
> root@dragon:/>while :; do ls /dev/hde*; done
> /dev/hde
> /dev/hde
> /dev/hde
> /dev/hde
> /dev/hde
> /dev/hde
> /dev/hde /dev/hde1
> ls: /dev/hde1: No such file or directory
> /dev/hde
> /dev/hde
> /dev/hde
> /dev/hde
> /dev/hde /dev/hde1
> /dev/hde /dev/hde1
> /dev/hde /dev/hde1
> ls: /dev/hde1: No such file or directory
> /dev/hde
> /dev/hde
> /dev/hde
> /dev/hde
> /dev/hde
> /dev/hde
>
> I manually used mknod to create the device file in /tmp, and then was
> able to mount and access it with no difficulty.
>
> Here's the kernel log:
>
> Feb 6 12:27:23 dragon kernel: pccard: PCMCIA card inserted into slot 0
> Feb 6 12:27:23 dragon kernel: pcmcia: registering new device pcmcia0.0
> Feb 6 12:27:24 dragon kernel: Probing IDE interface ide2...
> Feb 6 12:27:24 dragon kernel: hde: Renesas XX.V.3.5.0.0, CFA DISK drive
> Feb 6 12:27:25 dragon kernel: ide2 at 0x100-0x107,0x10e on irq 3
> Feb 6 12:27:25 dragon kernel: hde: max request size: 128KiB
> Feb 6 12:27:25 dragon kernel: hde: 2002896 sectors (1025 MB) w/1KiB Cache, CHS=1987/16/63
> Feb 6 12:27:25 dragon kernel: hde: cache flushes not supported
> Feb 6 12:27:25 dragon kernel: hde: hde1
> Feb 6 12:27:25 dragon kernel: ide-cs: hde: Vcc = 3.3, Vpp = 0.0
> Feb 6 12:27:25 dragon kernel: hde: hde1
> Feb 6 12:29:16 dragon kernel: hde: hde1
> Feb 6 12:29:21 dragon last message repeated 37 times
> Feb 6 12:29:46 dragon kernel: hde: hde1
> Feb 6 12:30:16 dragon last message repeated 148 times
> Feb 6 12:31:17 dragon last message repeated 386 times
> Feb 6 12:32:01 dragon last message repeated 300 times
> Feb 6 12:32:01 dragon kernel: hde: hde1
> Feb 6 12:32:06 dragon last message repeated 28 times
> Feb 6 12:32:06 dragon kernel: hde: hde1
> Feb 6 12:32:37 dragon last message repeated 145 times
> Feb 6 12:33:38 dragon last message repeated 349 times
> Feb 6 12:34:36 dragon last message repeated 442 times
>
> Apparently each of these "hde: hde1" printk's corresponds to a deletion
> and recretion of the device node. After I mounted the device, used it for
> a while, and unmounted it, this activity had stopped, now /dev/hde1 is stable.
>
> If I eject and re-insert the card, the same problem occurs. If I eject
> the card while this storm is ongoing, my system freezes hard with the
> pcmcia activity LED showing constant pcmcia activity.
>
> I don't really know if this is a pcmciautils, udev, kernel, or all three
> bug. Using the 2.4 kernel and pcmcia-cs, this card works fine.
>
> -- System Information:
> Debian Release: testing/unstable
> APT prefers unstable
> APT policy: (500, 'unstable'), (1, 'experimental')
> Architecture: i386 (i686)
> Shell: /bin/sh linked to /bin/bash
> Kernel: Linux 2.6.15-1-686
> Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
>
> Versions of packages pcmciautils depends on:
> ii libc6 2.3.5-12 GNU C Library: Shared libraries an
> ii libsysfs1 1.3.0-7 interface library to sysfs
> ii module-init-tools 3.2.2-1 tools for managing Linux kernel mo
>
> Versions of packages pcmciautils recommends:
> ii udev 0.084-1 /dev/ and hotplug management daemo
>
> -- no debconf information
>
> Here's what the kernel has to say about my pcmcia at boot:
>
> Yenta: CardBus bridge found at 0000:00:0c.0 [10cf:10c6]
> Yenta: adjusting diagnostic: 41 -> 61
> Yenta: Enabling burst memory read transactions
> Yenta: Using CSCINT to route CSC interrupts to PCI
> Yenta: Routing CardBus interrupts to PCI
> Yenta TI: socket 0000:00:0c.0, mfunc 0x00c61d22, devctl 0x46
> Yenta: ISA IRQ mask 0x0cf8, PCI irq 9
> Socket status: 30000006
>
--
Pelle
Reply to: