patrislav1
commented
7 months ago BTW. this also affects the duplicate label detection:
Before patch
/dev/mmcblk0p1 on /media/boot.vfat type vfat (rw,nosuid,nodev,noexec,relatime,gid=100,fmask=0111,dmask=0000,allow_utime=0022,codepage=437,iocharset=iso8859-1,shortname=mixed,utf8,flush,errors=remount-ro)
/dev/mmcblk1p1 on /media/boot.vfat type vfat (rw,nosuid,nodev,noexec,relatime,gid=100,fmask=0111,dmask=0000,allow_utime=0022,codepage=437,iocharset=iso8859-1,shortname=mixed,utf8,flush,errors=remount-ro)After patch
/dev/mmcblk0p1 on /media/boot.vfat type vfat (rw,nosuid,nodev,noexec,relatime,gid=100,fmask=0111,dmask=0000,allow_utime=0022,codepage=437,iocharset=iso8859-1,shortname=mixed,utf8,flush,errors=remount-ro)
/dev/mmcblk1p1 on /media/boot.vfat_1 type vfat (rw,nosuid,nodev,noexec,relatime,gid=100,fmask=0111,dmask=0000,allow_utime=0022,codepage=437,iocharset=iso8859-1,shortname=mixed,utf8,flush,errors=remount-ro)
There's a race condition between the cleanup step (i.e. removing old/unused mountpoints with
rmdir) and the mount step. Since udev launches multiple instances of the script in parallel, it's possible that one instance tries to mount a device just after another instance has deleted the associated mountpoint. (This is not a theoretical scenario, I had this actually happen on a system booting with multiple SD cards attached).This patch fixes the race condition with a lock in
/var/run, making the script execution sequential.