[ofa-general] RE: [ewg] OFED 1.2 and backwards binary compatibility

[Steffen Persvold]

Aha!
 
Thanks so much, I will look into this. 
 
Cheers,
Steffen Persvold
Technical Director Americas
tel. 508-281-7100 x401
fax. 508-281-7171

http://www.scali.com/
Scaling the Linux datacenter

________________________________

From: Roland Dreier [mailto:rdreier at cisco.com]
Sent: Fri 6/8/2007 9:59 AM
To: Steffen Persvold
Cc: EWG; OpenFabrics General
Subject: Re: [ofa-general] RE: [ewg] OFED 1.2 and backwards binary compatibility



 > 1.0 vs. 1.1 is all good. That works. I'm talking about 1.1/1.0 vs 1.2, that's not working. The diffset below is between 1.1 and 1.2.

Sorry for being confusing.  I was talking about the libibverbs
version.  OFED 1.0 and 1.1 both included libibverbs 1.0, and OFED 1.2
includes libibverbs 1.1.

 > What we're doing is using dlopen()/dlsym() to dynamically open the library so that we have no library dependencies (this allows us to runtime wise check if ofed is installed or other IB stacks). This apparently breaks.

Yes, you are basically implementing a broken dynamic linker yourself.
For this to work you will need to use dlvsym() and request all symbols
with version IBVERBS_1.0.  There may be a slight performance penalty
on libibverbs 1.1 (OFED 1.2) because you will be going through
compatibility wrappers.

 > I don't find any "post_send_wrapper_1_0" nor "post_send_wrapper_1_1" symbols in my libraries ?? :

Right, they're internal symbols.  Take a look at the libibverbs source
if you're curious about how it works.

 - R.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/ewg/attachments/20070608/e3fc2a7e/attachment.html>