From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040312 Description of problem: Hello, When I insert my pcmcia cf card adapter, I get the below warning in /var/log/messages: Apr 9 10:58:40 localhost cardmgr[1745]: socket 0: ATA/IDE Fixed Disk Apr 9 10:58:42 localhost kernel: hde: SAMSUNG CF/ATA, CFA DISK drive Apr 9 10:58:42 localhost kernel: ide2 at 0x100-0x107,0x10e on irq 5 Apr 9 10:58:43 localhost kernel: hde: max request size: 128KiB Apr 9 10:58:43 localhost kernel: hde: 253952 sectors (130 MB) w/1KiB Cache, CHS =496/16/32 Apr 9 10:58:43 localhost udev[3193]: creating device node '/udev/hde' Apr 9 10:58:43 localhost kernel: hde: hde1 < hde5 > Apr 9 10:58:43 localhost udev[3202]: creating device node '/udev/hde1' Apr 9 10:58:43 localhost udev[3211]: creating device node '/udev/hde5' Apr 9 10:58:43 localhost kernel: Module ide_cs cannot be unloaded due to unsafe usage in include/linux/module.h:516 Apr 9 10:58:43 localhost kernel: ide-cs: hde: Vcc = 3.3, Vpp = 0.0 Apr 9 10:58:43 localhost cardmgr[1745]: executing: './ide start hde 2>&1' Apr 9 10:58:43 localhost cardmgr[1745]: + ./ide: line 1: /sbin/ide_info: No suc h file or directory This is true: /sbin/ide_info doesn't exist, while it was available in FC1 in the kernel-pcmcia-cs package. This is not blocking, as I can still mount my CF card afterwards. However, as the /etc/pcmcia/ide script requires /sbin/ide_info, it should be available (or /etc/pcmcia/ide should be modified). Thanks for everything. Cheers, Michael. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Insert a pcmcia storage card (or pcmcia/cf adaptor) 2. or look at the /etc/pcmcia/ide script Additional info:
Hello, Just to confirm that this issue is still there in FC2 T3. Cheers, Michael.
Changing version to test3
Looks like it just needs adding to the spec file. PCMCIA uses it to try and do smart stuff about what and where to mount disks.
/sbin/ide_info is still missing on May 10, after installing all available updates, while /etc/pcmcia/ide still requires this script. Forget me not... ;-) Michael.
Nag, Nag, Nag. I was about to bugzilla this too, fortunately it was easy to find this report. Philip
Hello, This bug is still there in FC2 final. Don't forget... :-) Michael.
Still present in FC3 test 1. /sbin/ide_info is in fact explicitly removed by the spec file, which says "scsi_info and ide_info are obsoleted by hdparm". Seems to me that either this needs rethinking, or 1. /etc/pcmcia/ide needs to use hdparm instead of ide_info 2. the pcmcia-cs pkg needs to get a dependency on the hdparm pkg 3. possibly hdparm needs a new option to output the needed information in an easily digestible format.
Created attachment 102508 [details] /etc/pcmcia/ide with ide_info replaced by hdparm and sed This is a possible example of using sed and hdparm instead of the missing ide_info command. It could be formatted better, but it works.
Well.... Removable device configuration in all honesty shouldn't be being done in the PCMCIA scripts anyway - this should be done in udev, g-v-m, etc.
*** Bug 123376 has been marked as a duplicate of this bug. ***
*** Bug 135240 has been marked as a duplicate of this bug. ***
Problem still exists in FC3rc5. The fix mentioned in comment #8 works great. An icon appears on the desktop and the volume is mounted automatically when a PCMCIA ATA flash card is inserted.
Mounting of disks should be done through hal and fstab-sync in FC3.
Bill is right about mounting, but the bug is not about mounting. The bug is about an error reported to /var/log/messages. I propose we accept the first segment of the patch (well, Paul forgot to make a patch, but it's no big deal). I reject the updfstab bits because, in my testing, hal updates fstab just fine. I am able to mount the device in in Nautilus.
Created attachment 114377 [details] Candidate #1
I suppose one question here is how much of /etc/pcmcia/ide functionality should realistically be maintained; it a different scheme from the 'normal' HAL/hotplug/udev framework.
Created attachment 114439 [details] Candidate #2 - throw it all away
I don't care either way. I have verified that with the candidate patch #2 everything works, the volume can be mounted and umounted from Nautius, "cardctl eject" works with unmounted and when mounted (in which case HAL unmounts fine). It's a pity to throw away Paul's handiwork, but I need to have Michael's complaint fixed. [Some 2.6.11 kernels oops on eject, I have one with Bartolomej's fixes. Please, do not attempt to mutate this bug into a kernel bug.]
*** Bug 152140 has been marked as a duplicate of this bug. ***
This bug is still present on FC4 (system upgraded from FC2->3->4
An this bug is still present on FC3 fully updated via yum. Furthermore, I can't even manually mount the device; it shows up, then I try to mount, and it disappears, with mount complaining "mount: special device /dev/hde1 does not exist"
I think an update must have broke mounting compact flash recently. I just received the same error as comment #21 today. I have successfully used the compact flash in the pcmcia slot until now with minimal hassle. How about someone just fix the problem? This bug has turned into a sstanding joke. Philip
Some more funny info: FC3 - fully updated After inserting CF -> /sbin/ide_info missing. Applied patch #1 After inserting CF -> no errors. /media/idedisk added #mount /media/idedisk mount: special device /dev/hde1 does not exist #grep /dev/hde1 /etc/fstab /dev/hde1 /media/idedisk auto pamconsole,exec,noauto,noatime,sync,managed 0 0 changed fs type auto to vfat #grep /dev/hde1 /etc/fstab /dev/hde1 /media/idedisk auto pamconsole,exec,noauto,noatime,sync,managed 0 0 and now mount /media/idedisk works !
sorry for a typo - in the last comment at the last grep there should be of course: #grep /dev/hde1 /etc/fstab /dev/hde1 /media/idedisk vfat pamconsole,exec,noauto,noatime,sync,managed 0 0
I've looked into creating a /etc/hotplug/block.agent to auto-mount/unmount the device as it's made available to the system. However, there's a major problem (FC3, full updates as of 2005/08/14). Issuing a command like 'hdparm -i /dev/hdc' or "mount /dev/hdc1 /mnt/cf" causes the block script to be rerun (remove and then add events) for hdc1. That's very bad, as it means that during the "normal" use of a device, it's going to be loaded and unloaded dozens of times. Now in theory, I might ignore hdc events that aren't preceeded by a pcmcia insert, but that's awfully restrictive. what about usb drives? (actually, I think they show up as scsi), but there are other possibilities too. firewire? (hm, there is a ieee1394.agent already). in any case, for now, hotplug just doesn't seem viable to me.
(In reply to comment #25) > I've looked into creating a /etc/hotplug/block.agent to auto-mount/unmount the > device as it's made available to the system. However, there's a major problem > (FC3, full updates as of 2005/08/14). Issuing a command like 'hdparm -i > /dev/hdc' or "mount /dev/hdc1 /mnt/cf" causes the block script to be rerun > (remove and then add events) for hdc1. That's very bad, as it means that during > the "normal" use of a device, it's going to be loaded and unloaded dozens of > times. > > Now in theory, I might ignore hdc events that aren't preceeded by a pcmcia > insert, but that's awfully restrictive. what about usb drives? (actually, I > think they show up as scsi), but there are other possibilities too. firewire? > (hm, there is a ieee1394.agent already). > > in any case, for now, hotplug just doesn't seem viable to me. correction. I'm seeing the same bug as mentioned in #21 "can't mount" and the cause seems to be that hotplug is removing /dev/hdc1 when I try and mount it! that seems to be separate from this issue, and I will hunt around for that bug and add to it instead of adding further to this bug. I will revisit this when I can once again mount my cf cards...
so I think I've figured this out. mkdir /var/log/hotplug touch /var/log/hotplug/events # make sure that hotplug.d/default/default.agent logs things to the above log file mv /etc/hotplug/block.agent /etc/hotplug/block.agent.O # turn off pcmcia ide script mv /etc/pcmcia/ide /etc/pcmcia/ide.O cat >/etc/pcmcia/ide <<'_EOF_' #!/bin/sh exit 0 _EOF_ chmod +x /etc/pcmcia/ide cardctl eject tail -f /var/log/hotplug/events& cardctl insert You should see in the log: add for pcmcia remove for module /module/ide_cs remove for module drivers /bus/pcmcia/drivers/ide-cs add for module /module/ide_cs add for module drivers /bus/pcmcia/drivers/ide-cs (pause as pcmcia scripts run) add for ide add for block (hdc) add for block (hdc1) remove for block (hdc1) add for block (hdc1) and I'm not trying to mount the drive or anything. hotplug just seems to excessively thrash on the adding and deleting of hdc1. here's some more strangeness: # ls -al /dev/hdc1 ; mount -v /dev/hdc1 /mnt/cf ; ls -al /dev/hdc1 brw-rw---- 1 root disk 22, 1 Aug 14 11:30 /dev/hdc1 mount: you didn't specify a filesystem type for /dev/hdc1 I will try all types mentioned in /etc/filesystems or /proc/filesystems Trying vfat mount: special device /dev/hdc1 does not exist ls: /dev/hdc1: No such file or directory and simultaneously you will see a remove/add for hdc1 pair. so the act of mounting causes a remove? That sure is strange. however, mount specifying -t vfat seems to work, but only because it seems to beat hotplug to the punch. note that the trailing ls still fails to find the device, because it's been removed by the same hotplug remove/add pair that you saw above: # ls -al /dev/hdc1 ; mount -v -t vfat /dev/hdc1 /mnt/cf ; ls -al /dev/hdc1 brw-rw---- 1 root disk 22, 1 Aug 14 11:33 /dev/hdc1 /dev/hdc1 on /mnt/cf type vfat (rw) ls: /dev/hdc1: No such file or directory so I guess this is currently more of a hotplug bug? can someone point me in the right direction?
found it. hotplug seems to be broken as of a recent fc3: bug # 130232
This is still broken with a fully updated FC3 as of 2005-10-29. *poke* *poke* :)
Poke poke what, Eli? You didn't tell me if patch #2 suits you, so no action could be taken. Nothing was heard from Michael either.
Pete, I missed that you needed that tested; sorry. With patch #2: insert CF card into PCMCIA slot, then # mount /dev/hde1 /media/card mount: special device /dev/hde1 does not exist # tail /var/log/messages Oct 30 19:37:17 pegasus last message repeated 2 times Oct 30 19:38:26 pegasus cardmgr[2958]: socket 0: ATA/IDE Fixed Disk Oct 30 19:38:26 pegasus kernel: hde: SanDisk SDCFB-128, CFA DISK drive Oct 30 19:38:27 pegasus kernel: ide2 at 0x100-0x107,0x10e on irq 3 Oct 30 19:38:27 pegasus kernel: hde: max request size: 128KiB Oct 30 19:38:27 pegasus kernel: hde: 250880 sectors (128 MB) w/1KiB Cache, CHS=980/8/32 Oct 30 19:38:27 pegasus kernel: hde: cache flushes not supported Oct 30 19:38:27 pegasus kernel: hde: hde1 Oct 30 19:38:27 pegasus kernel: ide-cs: hde: Vcc = 3.3, Vpp = 0.0 Oct 30 19:38:39 pegasus kernel: hde: hde1 However, # mount -t vfat /dev/hde1 /media/card/ does work.. it also worked before the patched /etc/pcmcia/ide though... What else do I need to check? Thanks for looking at this.
Eli - the mount issue is separate from this bug. it's a kernel/hotplug problem, as described in bug # 130232
As of FC4 with pcmcia-cs-3.2.8-4.12 /etc/pcmcia/ide line 34 eval `/sbin/ide_info /dev/$DEVICE` || usage As /sbin/ide_info does not exist, ide exits ungracefully. This means I cannot hotplug anything. Should this bug report not be updated to FC4 and actioned ?
Andrew, what do you mean by "cannot hotplug anything"? Please provide a more substantial report with exact commands you typed and unedited messages (please do not try to "save someone's time" by cutting, attach whole instead). Hotplug works just fine for me. The legacy scripts are not used anymore. In fact, FC5 will come out without pcmcia-cs at all.
I wasn't saving people's time. The script I cited, at the line given, has a reference to a program that is no longer there. The program dies at that line. I didn't go into specifics as what didn't hotplug, as nothing hotplugs. ie anything mountable that I plug in to the PCMCIA slot, does not mount. But then, as long as it works for you, I guess there isn't a problem.
*** Bug 160604 has been marked as a duplicate of this bug. ***
pcmcia-cs is obsolete and has been removed from Fedora Core. It is not present in FC5 or FC6, the only fully supported versions of Fedora Core. (FC3 and FC4 are supported by Fedora Legacy for security fixes only.) It has been replaced by pcmciautils. Due to the change in codebase, can you please retest this bug against pcmciautils in FC5 or FC6? If the bug is still present there, please change the Component to pcmciautils and the version appropriately. Thanks! If there is no response within several weeks, the bug will be closed.
Per the previous comment, closing bug since pcmcia-cs is no longer supported. If a similar bug still exists in pcmciautils, the replacement for pcmcia-cs, please reopen the bug or file a new one.