[openfabrics-ewg] [Fwd: Re: On vacation on Sep-13 till 12-Oct [pre-RC6 build]]

[Michael S. Tsirkin]

Quoting r. Moni Shoua <monis at voltaire.com>:
> Subject: Re: [Fwd: Re: On vacation on Sep-13 till 12-Oct [pre-RC6 build]]
> 
> Or Gerlitz wrote:
> 
> > Aviram Gutman wrote:
> >
> >> Guys,
> >> Can you fix it today?
> >
> >
> > Hi Aviram,
> >
> > Moni Shoua (cc-ed on this email) will post later today over the EWG 
> > mailing list if we impose a limitation of not supporting ofed1.1 build 
> > with iSER as non root user or we provide a fix.
> >
> > Or.
> >
> >> ------------------------------------------------------------------------
> >>
> >> Subject:
> >> Re: [openfabrics-ewg] On vacation on Sep-13 till 12-Oct [pre-RC6 build]
> >> From:
> >> Betsy Zeller <betsy at pathscale.com>
> >> Date:
> >> Mon, 18 Sep 2006 20:56:13 -0700
> >> To:
> >> Aviram Gutman <aviram at dev.mellanox.co.il>
> >>
> >> To:
> >> Aviram Gutman <aviram at dev.mellanox.co.il>
> >> CC:
> >> "Scott Weitzenkamp (sweitzen) <sweitzen"@cisco.com, Jack Morgenstein 
> >> <jackm at mellanox.co.il>, EWG <openfabrics-ewg at openib.org>
> >>
> >>
> >> Aviram - thanks - testing is underway on some distros, but there is a
> >> problem with  on SLES10 with the pre-build - we've opened bug #237 "1.1
> >> RC6-pre1 build failure on SLES 10 in oiscsi" on the OpenFabrics bug
> >> list.
> >> It's currently assigned to ogerlitz at voltaire, but it may be something
> >> Vlad needs to work on. Can you get this turned around so we can test on
> >> SLES10 on Tuesday?
> >>
> >> Thanks, Betsy
> >>
> >> On Sun, 2006-09-17 at 16:41 +0300, Aviram Gutman wrote:
> >>
> >>>    
> >>>
> >>>>>     
> >>>>
> >>> Yes, we can give you "pre-rc6" on Monday
> >>
> >
> >
> >
> Hi,
> We have decided to leave the requirement of "MUST be root to build iSER" 
> as is.

Fair enough. We'll see how simple it is to add the test for OFED scripts.

> Although there is a simple fix for the kernel module there is no such 
> fix for the user mode RPM (which tries to patch /etc/init.d/open-iscsi)

I think it would still be nice to avoid patching the kernel if possible,
reducing the scope of problem - lots of users hate it when you touch their kernel.
If it's really simple, can this be turned around by tomorrow?

-- 
MST