init: make rootdelay=N consistent with rootwait behaviour
authorPaul Gortmaker <paul.gortmaker@windriver.com>
Fri, 8 Aug 2014 21:19:46 +0000 (14:19 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 8 Aug 2014 22:57:18 +0000 (15:57 -0700)
commit4dfe694f616e00e6fd83e5bbcd7a3c4d7113493d
tree88cfc07a6e22fc75ba7260e83c0bd485353bc3c4
parent6f4535ed7d626c3760243b9c5223baaa358a5fdd
init: make rootdelay=N consistent with rootwait behaviour

Currently rootdelay=N and rootwait behave differently (aside from the
obvious unbounded wait duration) because they are at different places in
the init sequence.

The difference manifests itself for md devices because the call to
md_run_setup() lives between rootdelay and rootwait, so if you try to use
rootdelay=20 to try and allow a slow RAID0 array to assemble, you get
this:

[    4.526011] sd 6:0:0:0: [sdc] Attached SCSI removable disk
[   22.972079] md: Waiting for all devices to be available before autodetect

i.e.  you've achieved nothing other than delaying the probing 20s, when
what you wanted was a 20s delay _after_ the probing for md devices was
initiated.

Here we move the rootdelay code to be right beside the rootwait code, so
that their behaviour is consistent.

It should be noted that in doing so, the actions based on the
saved_root_name[0] and initrd_load() were previously put on hold by
rootdelay=N and now currently will not be delayed.  However, I think
consistent behaviour is more important than matching historical behaviour
of delaying the above two operations.

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
init/do_mounts.c