Bug#565225: don't assume RESUME partition will never change

[antoine]

Package: initramfs-tools
Version: 0.93.4
Severity: normal

With swap on lvm, /etc/initramfs-tools/conf.d/resume containing UUID prevented
the system to resume changing back to dev/mapper (/dev/mapper/vg0-swap) allowed
my system to resume from hibernate  As my problem appears linked to this /etc
/initramfs-tools/conf.d/resume config, and as other previous posts are
referring to it, I reported here - if you need me to open a separate bug,
please don't hesitate !


-- Package-specific info:
-- /proc/cmdline
root=/dev/mapper/vg0-root ro quiet vga=775

-- /proc/filesystems
cramfs
ext4
ext2
fuseblk

-- lsmod
Module  Size  Used by
af_packet  18194  4 
sco 8283  2 
bridge 46509  0 
stp 1504  1 bridge
bnep   10474  2 
rfcomm 34217  0 
l2cap  31981  6 bnep,rfcomm
bluetooth  49547  6 sco,bnep,rfcomm,l2cap
acpi_cpufreq6139  1 
cpufreq_ondemand7973  8 
cpufreq_powersave918  0 
cpufreq_stats   2960  0 
freq_table  2347  3 acpi_cpufreq,cpufreq_ondemand,cpufreq_stats
cpufreq_performance  922  0 
cpufreq_conservative 9084  0 
fuse   58612  1 
ext2   63284  1 
loop   14603  0 
firewire_sbp2  12925  0 
pcmcia 20818  0 
arc41306  2 
snd_hda_codec_idt  50507  1 
ecb 1905  2 
snd_hda_intel  21810  5 
snd_hda_codec  66685  2 snd_hda_codec_idt,snd_hda_intel
uvcvideo   59498  0 
iwlagn117278  0 
snd_hwdep   5876  1 snd_hda_codec
snd_pcm68170  3 snd_hda_intel,snd_hda_codec
videodev   37339  1 uvcvideo
snd_timer  18173  1 snd_pcm
v4l1_compat13994  2 uvcvideo,videodev
v4l2_compat_ioctl3210177  1 videodev
iwlcore99686  1 iwlagn
joydev  9448  0 
i2c_i8018342  0 
yenta_socket   20564  1 
snd52531  15 
snd_hda_codec_idt,snd_hda_intel,snd_hda_codec,snd_hwdep,snd_pcm,snd_timer
tifm_7xx1   4218  0 
dell_laptop 2321  0 
nvidia   9514631  45 
rtc_cmos8926  0 
rsrc_nonstatic  8119  1 yenta_socket
tpm_tis 8077  0 
tifm_core   4772  1 tifm_7xx1
dcdbas  5608  1 dell_laptop
soundcore   5679  1 snd
i2c_core   17177  3 videodev,i2c_i801,nvidia
rtc_core   13189  1 rtc_cmos
mac80211  166636  2 iwlagn,iwlcore
pcmcia_core28360  3 pcmcia,yenta_socket,rsrc_nonstatic
tpm10269  1 tpm_tis
tpm_bios4761  1 tpm
psmouse52070  0 
serio_raw   4200  0 
rtc_lib 1833  1 rtc_core
snd_page_alloc  6825  2 snd_hda_intel,snd_pcm
cfg80211  131532  3 iwlagn,iwlcore,mac80211
processor  24406  9 acpi_cpufreq
ac  2080  0 
battery 4934  0 
evdev   8468  30 
dell_wmi2779  0 
rfkill 14676  5 bluetooth,dell_laptop,cfg80211
wmi 5957  1 dell_wmi
ext4  237049  6 
mbcache 5578  2 ext2,ext4
jbd2   53085  1 ext4
crc16   1287  2 l2cap,ext4
sg 20824  0 
sr_mod 13891  0 
cdrom  34036  1 sr_mod
sd_mod 29480  3 
usbhid 35815  0 
hid75077  1 usbhid
sdhci_pci   6483  0 
sdhci  15729  1 sdhci_pci
mmc_core   52292  1 sdhci
ahci   34999  2 
firewire_ohci  23305  0 
led_class   2561  1 sdhci
firewire_core  42882  2 firewire_sbp2,firewire_ohci
ehci_hcd   36007  0 
video  17514  0 
crc_itu_t   1291  1 firewire_core
output  1740  1 video
usbcore   142960  4 uvcvideo,usbhid,ehci_hcd
libata154451  1 ahci
nls_base6665  1 usbcore
tg3   120823  0 
scsi_mod  136715  5 firewire_sbp2,sg,sr_mod,sd_mod,libata
button  4538  0 
thermal11647  0 
fan 2386  0 
dm_mirror  13026  0 
dm_region_hash  7096  1 dm_mirror
dm_log  8448  2 dm_mirror,dm_region_hash
dm_mod 62070  23 dm_mirror,dm_log

-- /etc/kernel-img.conf
# Kernel image management overrides
# See kernel-img.conf(5) for details
do_symlinks = yes
do_bootloader = no
do_initrd = yes
link_in_boot = no
postinst_hook = update-grub
postrm_hook   = update-grub

-- /etc/initramfs-tools/initramfs.conf
MODULES=most
BUSYBOX=y
KEYMAP=n
BOOT=local
DEVICE=eth0
NFSROOT=auto


-- System Information:
Deb

Bug#565225: don't assume RESUME partition will never change

[maximilian attems]

On Wed, 13 Jan 2010, Ben Hutchings wrote:

> Or even to store the UUID of the swap partition.
> 
> maks, is this another one for the libata transition script?

yes, indeed it is.



-- 
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20100224173747.gh11...@stro.at

Bug#565225: don't assume RESUME partition will never change

[Ben Hutchings]

On Thu, 2010-01-14 at 07:20 +0800, jida...@jidanni.org wrote:
> Package: initramfs-tools
> Version: 0.93.4
> 
> How can one just _assume_ that nobody will ever adjust their disk
> partitions ever?
[...]
> Anyway it makes much more sense to check where the swap space is located
> upon each time /boot/initrd.img.* is made. -- And perhaps even more like
> each boot or shutdown or ... suspend.
[...]

Or even to store the UUID of the swap partition.

maks, is this another one for the libata transition script?

Ben.

-- 
Ben Hutchings
The generation of random numbers is too important to be left to chance.
- Robert Coveyou


signature.asc
Description: This is a digitally signed message part

Bug#565225: don't assume RESUME partition will never change

[jidanni]

Package: initramfs-tools
Version: 0.93.4

How can one just _assume_ that nobody will ever adjust their disk
partitions ever?

In /var/lib/dpkg/info/initramfs-tools.preinst
you do
 # First time install.  Can we autodetect the RESUME partition?

So then forever and ever
/etc/initramfs-tools/conf.d/resume and thus
$ zcat /boot/initrd.img-2.6.*-trunk-686|cpio -i --to-stdout conf/conf.d/resume
contain this never changing resume partition from years ago when we
first bought the computer.

OK, now let's look at what happens when the user boots.
He will see
Loading... please wait
readlink: No such file or directory

Which of course also zip way off the screen, and don't end up in any
logs.

Now he will wonder "gee, those guys are missing the "readlink" command".
Bug#564812

And when one day he does actually try to resume... who knows what will
happen.

Anyway it makes much more sense to check where the swap space is located
upon each time /boot/initrd.img.* is made. -- And perhaps even more like
each boot or shutdown or ... suspend.

Or upon boot warn the user that it was detected that his swap partition
is not where it used to be, and he should now "remove and reinstall
initramfs-tools"! (as apparently there is no script command to fix it
short of editing the file by hand.) ... (And don't just let the message
fly off the screen beyond where Shift PgUp can reach, and not kept in
any log file.)

Also man initramfs-tools:
 resume
   On install initramfs-tools tries to autodetect the resume parti-
   tion. On success the RESUME variable is written to
   /etc/initramfs-tools/conf.d/resume. The boot variable noresume
   overrides it.
Here again it is like the author _just assumes_ that the RESUME part of
the computer is the only part of the computer that never could change.



-- 
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org