[openib-general] [PATCH] install.sh: Cause less pain to SRP users who didn't RTFM

[ishai]

Hi Roland,

SRP was tested on a 32-bit operating system running on a 32-bit platform
and on 64-bit OS and there are no known problems.
In the interoperability tests done in UNH-IOL on September we found out
that SRP on a 32-bit operating system running on a 64-bit platform
causes crashes. (It was tested on RHEL4-U3).
Since we did not have enough time to solve this problem until the 
release and since we think that this combination (32-bit OS on 64-bit 
platform) is less common, we treat this issue as low priority.

The remark in the release notes indicates that SRP does not work on this
combination.

Ishai

Roland Dreier wrote:

> > We seem to get panics during multithreaded IO on the initiator.  The
> > panics don't seem to always point to any SRP code, but might be a
> > symptom of memory corruption.  It only seems to show up on 32-bit
> > kernels.  Our distro is a derivative of Fedora.  There were a few more
> > things we wanted to consider, but we stopped debugging when we saw an
> > indication in the release notes that it's not a supported configuration.
>
>I'm not sure who declared it "unsupported" and I would really like to
>know what issue(s) led to that declaration.  Your report is the first
>I've heard of anything like this, and I have to say that it seems
>pretty implausible that running a 32-bit kernel on 64-bit-capable
>hardware would be the source of problems -- if there is an issue then
>I would expect it to be something to do with the 32-bit kernel.
>
>In any case I definitely consider 32-bit kernels as something I
>support, so if you could post a real bug report (what specific kernel
>version, if you are running out-of-tree drivers (like OFED), host
>server details, SRP target details, how to reproduce, etc) for your
>problems with 32-bit kernels then I will try to debug things.
>
> - R.
>
>_______________________________________________
>openib-general mailing list
>openib-general at openib.org
>http://openib.org/mailman/listinfo/openib-general
>
>To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
>  
>