Your message dated Thu, 26 Oct 2006 06:17:22 -0700 with message-id <[EMAIL PROTECTED]> and subject line Bug#295465: fixed in lilo 1:22.6.1-8 has caused the attached Bug report to be marked as done.
This means that you claim that the problem has been dealt with. If this is not the case it is now your responsibility to reopen the Bug report if necessary, and/or fix the problem forthwith. (NB: If you are a system administrator and have no idea what I am talking about this indicates a serious mail system misconfiguration somewhere. Please contact me immediately.) Debian bug tracking system administrator (administrator, Debian Bugs database)
--- Begin Message ---Package: lilo Version: 1:22.6.1-4 Severity: important Justification: breaks upgrades Tags: patch sid sarge This is actually the same bug as #222098, but I cannot reopen it as it has been archived. In any case, this bug has not been fixed. Lilo's changelog says: * Added suggestion to use the "inaccessible" tag on the error message that caused a lot of questions from the users. (closes: #222098) and the patch (06_verbose-errmsg) is there in the sources (and is applied properly) However, I'm suffering this issue, just like a number of Debian users posting to mailing lists out there. It turns out that the patch only tries to (partially) fix the issue (it only presents the message in one condition) and it doesn't even fix it at all since it's broken (see below) Here are the details: My /etc/fstab has: /dev/sda1 /media/dsc vfat w,noauto,user 0 0 /dev/sdb1 /media/hdext ntfs ro,noauto,user 0 0 First one is used for my digital camera, second for an external HD. I use the usb mass storage modules: $ lsmod |grep usb usb-storage 65088 0 scsi_mod 95044 7 [sr_mod sg sd_mod ide-scsi sata_promise libata usb-storage] usbkbd 3096 0 (unused) input 3648 0 [keybdev usbkbd] usb-uhci 23344 0 (unused) usbcore 62924 1 [ehci-hcd usb-storage usbkbd usb-uhci] ide-core 108568 12 (autoclean) [ide-scsi usb-storage ide-detect ide-disk piix pdc202xx_new] If lilo is run: # lilo Warning: '/proc/partitions' does not match '/dev' directory structure. Name change: '/dev/scsi/host3/bus0/target0/lun0/disc' -> '/dev/sda' Fatal: VolumeID read error: sector 0 of /dev/sdb not readable # echo $? 1 (This obviously breaks upgrades) The kernel spouts (as seen through dmesg and in the /var/log/messages) this: Feb 16 00:10:46 silicio kernel: Device 08:10 not ready. Feb 16 00:10:46 silicio kernel: I/O error: dev 08:10, sector 0 Notice that there is no 'inaccessible' message at all there. If I add this to lilo.conf the problem goes away: # For USB drives disk=/dev/sda inaccessible disk=/dev/sdb inaccessible (I had to google a bit before finding this was the proper solution) Reviewing the patch you introduced I notice two things: 1.- The error I get is generated by a different line of where you put the message in. 714 if (read(fd, &buf, sizeof(buf)) != sizeof(buf)) 715 die("VolumeID read error: sector 0 of %s not readable", dev. 715 name); 2.- The message you introduced in the patch is after the die() statement so it will _never_ get executed: 231 else if ((dev->fd = open(name,flags)) < 0) 232 { 233 die("open %s: %s",name,strerror(errno)); 234 fprintf (errstd, "\nMaybe you missed to put\n" 235 "\tdisk=%s inaccessible\n" 236 "on your lilo.conf\n", name); 237 } An attached patch fixes both of these issues, hopefully you will include it an new release. Please set this as a medium or higher priorit since this bug might break upgrades (or prevent new kernels from being properly installed in new systems that have USB mass storage devices...) In any case I don't really understand why lilo would need to touch _at_all_ a user-mounted device that is actually not mounted in the system. Is there any reason why Lilo should bother with these? Try googling for the "Fatal: VolumeID read error: sector 0" and you'll see a lot of users complaining about this bug in forums. Some of them are actually Debian users.. :-( Actually, I believe you answered one of those recently: http://lists.debian.org/debian-user/2005/01/msg01614.html Also, even if you fix the code, couldn't you add the above lines (commented out) to the default lilo.conf with an explanation of when they might be needed? Removable USB media is quite common this days (from memory stick/SD/whatever readers to digital cameras to portable hard drives) and this is a problem that many users will stumble upon.... Maybe something like this: ---------------------------------------------------------------- (...) #disk=/dev/sda # bios=0x80 # If you are using removable USB drivers (with mass-storage) # you will need to tell LILO to not use these devices even # if defined in /etc/fstab and referenced in /proc/partitions. # Adjust these lines to your devices: # # disk=/dev/sda inaccessible # disk=/dev/sdb inaccessible ---------------------------------------------------------------- Thanks Javier--- device.c.err 2005-02-16 00:22:41.000000000 +0100 +++ device.c 2005-02-16 00:29:47.000000000 +0100 @@ -230,10 +230,11 @@ if (flags == O_BYPASS) dev->fd = -1; else if ((dev->fd = open(name,flags)) < 0) { + fprintf (errstd, "Cannot proceed maybe you need to add " + "this to your lilo.conf:\n" + "\tdisk=%s inaccessible\n" + "(real error shown below)\n", name); die("open %s: %s",name,strerror(errno)); - fprintf (errstd, "\nMaybe you missed to put\n" - "\tdisk=%s inaccessible\n" - "on your lilo.conf\n", name); } dev->name = stralloc(name); return dev->fd; @@ -711,8 +712,13 @@ die("VolumeID set/get bad device %04X\n", device); fd = dev_open(&dev, device, option ? O_RDWR : O_RDONLY); - if (read(fd, &buf, sizeof(buf)) != sizeof(buf)) + if (read(fd, &buf, sizeof(buf)) != sizeof(buf)) { + fprintf (errstd, "Cannot proceed maybe you need to add " + "this to your lilo.conf:\n" + "\tdisk=%s inaccessible\n" + "(real error shown below)\n", dev.name); die("VolumeID read error: sector 0 of %s not readable", dev.name); + } if (option==ID_SET) { make_backup(NULL, 0, &buf, device, "master disk volume ID record");
signature.asc
Description: Digital signature
--- End Message ---
--- Begin Message ---Source: lilo Source-Version: 1:22.6.1-8 We believe that the bug you reported is fixed in the latest version of lilo, which is due to be installed in the Debian FTP archive: lilo-doc_22.6.1-8_all.deb to pool/main/l/lilo/lilo-doc_22.6.1-8_all.deb lilo_22.6.1-8.diff.gz to pool/main/l/lilo/lilo_22.6.1-8.diff.gz lilo_22.6.1-8.dsc to pool/main/l/lilo/lilo_22.6.1-8.dsc lilo_22.6.1-8_i386.deb to pool/main/l/lilo/lilo_22.6.1-8_i386.deb A summary of the changes between this version and the previous one is attached. Thank you for reporting the bug, which will now be closed. If you have further comments please address them to [EMAIL PROTECTED], and the maintainer will reopen the bug report if appropriate. Debian distribution maintenance software pp. Andrés Roldán <[EMAIL PROTECTED]> (supplier of updated lilo package) (This message was generated automatically at their request; if you believe that there is a problem with it please contact the archive administrators by mailing [EMAIL PROTECTED]) -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Format: 1.7 Date: Wed, 25 Oct 2006 15:08:22 +0000 Source: lilo Binary: lilo-doc lilo Architecture: source i386 all Version: 1:22.6.1-8 Distribution: unstable Urgency: high Maintainer: Andrés Roldán <[EMAIL PROTECTED]> Changed-By: Andrés Roldán <[EMAIL PROTECTED]> Description: lilo - LInux LOader - The Classic OS loader can load Linux and others lilo-doc - Documentation for LILO (LInux LOader) Closes: 295465 333278 338070 340632 351542 356214 378923 379518 Changes: lilo (1:22.6.1-8) unstable; urgency=high . * debian/patches/06_geometry.dpatch: - Patch from Nathanael Nerode <[EMAIL PROTECTED]> to make sure that disks marked 'inaccessible' do not have their partition tables checked and do not have 'stat' run. (Closes: #379518) * debian/patches/06_verbose-errmsg.dpatch: - Applied changes suggested by Javier Fernández-Sanguino Pen~a <[EMAIL PROTECTED]> because some parts of the patch were never executed. (Closes: #295465) * debian/{liloconfig,lilo.templates}: - Applied changes suggested by Artur R. Czechowski <[EMAIL PROTECTED]> to use debconf in liloconfig. (Closes: #351542) * debian/patches/17_dont_scan_udev.dpatch: - Patch from Noel Maddy <[EMAIL PROTECTED]> to avoid checking the /dev/.udev directory when scanning /dev. (Closes: #378923) * debian/po/{sv,ru,nl,pt_PT}.po: - Added and updated translations. Thanks guys. (Closes: #333278, #338070, #340632, #356214) Files: f07e0cfeb8980d36862c8612e5656637 731 base optional lilo_22.6.1-8.dsc 6e12e81aeebe76622896045b295e397d 192022 base optional lilo_22.6.1-8.diff.gz 05b889dda67804cc74f082035d95b02b 349106 base optional lilo_22.6.1-8_i386.deb d3fbacbff2569894ecc43c367ea47348 225442 doc optional lilo-doc_22.6.1-8_all.deb -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) iD8DBQFFQK/x2OByS7KTlusRAtoRAJ0VoQPUi0Yef9L2LADOW8wGv0nmLwCfWPfZ ecs4UGhn+bOMyaLsIgp3hpM= =EjE4 -----END PGP SIGNATURE-----
--- End Message ---
