Re: [dm-devel] [ANNOUNCE] multipath-tools-0.3.1

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



> multipath/main.c lines 629-640
> 
> 1) mp->pg at this point is NULL (0x0, at least for me), causing segfaults.
It shouldn't.

I can add a safety-net at that point but better you investigate what
goes wrong before that point.

> 2) The code doesn't do anything useful (unless I'm missing something).  It
> sprintf's into p, but never prints p out.
> 
Don't oversight that p = mp->params, which is the string that gets fed
to the device-mapper via ioctl in setup_map().

setup_map(mp)
\_ select iopolicy for this multipath
\_ apply the iopolicy to mp->pathvec (vector of paths),
   store result in mp->pg (vector of vectors of paths)
\_ assemble_map() transforms mp->pg into mp-params string,
   formatted according to device-mapper params style
\_ select between CREATE or UPDATE device map
\_ do the device-mapper driving

normally, mp->pathvec can't be null at that point.
check with a "make DEBUG=3" if an iopolicy is selected,
because I can't see how an iopolicy can loose paths in the
transformation.

regards,
-- 
christophe varoqui <christophe.varoqui@xxxxxxx>


[Index of Archives]     [DM Crypt]     [Fedora Desktop]     [ATA RAID]     [Fedora Marketing]     [Fedora Packaging]     [Fedora SELinux]     [Yosemite Discussion]     [KDE Users]     [Fedora Docs]

  Powered by Linux