[PATCH] Bug 815 [was: [PATCH] Bug 836]

[Yann E. MORIN]

On Sunday 11 June 2006 185, Bernhard Fischer wrote:
> On Sun, Jun 11, 2006 at 06:48:20PM +0200, Bernhard Fischer wrote:
> >Shouldn't the DEFAULT_ROOT_LOGIN_PATH be in the env of the process
> >execing insmod/rmmod? Putting the absolute path in there sounds a bit
> >odd to me, no?
> To be more elaborate.
> Imagine my box has system-specific stuff in /system/, in a flat model
> (libs, sbin's, bin's, config-files all flat out in /system/).
> The superuser's path (and eventually library path) is set to /system/
> See how hardcoding /sbin/ isn't a good idea and how it's the superuser's
> duty to set the path to system binaries properly..
> In my POV #815 is a config issue of that specific user and nothing we
> should (and can!) take care of..

Presented this way, I have no option but to agree. :-)

I made further tests as to why this should/shouldn't work. So my setup was
pretty simple: have init=/bin/ash and call modprobe. That fails, because, by
default, ash does not export PATH (wether that is sane or not I don't bother,
presonnaly think it _is_). So modprobe has no PATH in its env, and thus can
not call insmod (or rmmod, or whatever, if that matters). Explictly exporting
PATH before calling modprobe solves the problem.

I guess we shall not commit the patch, and forget it. I'll ask the reporter
more details on his/her setup before suggesting to close this issue.

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +0/33 662376056 | Software  Designer | \ / CAMPAIGN     |   ^                |
| --==< °_° >==-- °---.----------------:  X  AGAINST      |  /e\  There is no  |
| web: ymorin.free.fr | SETI at home 3808 | / \ HTML MAIL    |  """  conspiracy.  |
°---------------------°----------------°------------------°--------------------°