Re: Changes in config.param between 2.0.0 and 2.0.1
Michael Beach <[email protected]> Fri, 16 Aug 2002 15:38:51 +1000
| Newsgroups | gmane.linux.ngpt.user |
|---|---|
| Message-ID | <0208161538510H.09629@gilgamesh> |
On Friday 16 August 2002 00:39, [email protected] wrote: > On 08/15/2002 at 03:42:49 PM ZE10, Michael Beach <[email protected]> wrote: > > Hi all, I recently downloaded NGPT 2.0.1, configured it and was surprised > > to > > > find that it wouldn't compile. After a bit of digging around I noticed > > that > > > during the configuration process, 2.0.1 was making the decision > > > > decision on mctx implementation... mcsc/sc/mc > > > > whereas 2.0.0 behaved differently, giving > > > > decision on mctx implementation... sjlj/ssjlj/sas > > > > BTW, this is all on a pretty much stock SuSE 7.2 system, with the only > > significany difference being a 2.4.19 kernel with the patches required by > > NGPT applied. The configure script identifies the system as > > > > Platform: i686-suse-linux-gnu2.4glibc2.2 > > While your solution will work, we recommend that you upgrade glibc to 2.2.4 > or higher. The mcsc/sc/mc method of context switch is far superior both in > performance and reliability than the earlier method. The 2.0.2 version of > NGPT will require use of glibc 2.2.4 or higher. I'll take your word as for which one is better. My point is that if sjlj/ssjlj/sas did in fact work, and the configure code contained all of the machinery to detect it, why override that detection and force the use of mcsc/sc/mc, despite the fact that it wasn't detected. If on the other hand sjlj/ssjlj/sas really does have such serious problems, and you don't want to use it, then why bother checking for it? Just check for mcsc/sc/mc, and if you don't find it get the configure script to inform the user of the problem. Better than just blindly assuming that it's there and having the subsequent build fail with some mysterious looking errors. Hmm, sorry if all of the above sounds to negative. I hope you will take such criticism in the positive way it is intended, as I'm really very pleased with NGPT overall. It's already saved my bacon big-time! Regards M.Beach > > Regards, > Bill Abt > Senior Software Engineer > Next Generation POSIX Threading for Linux > IBM Cambridge, MA, USA 02142 > Ext: +(00)1 603-673-3440 > T/L: 253-9938 > Cell: +(00)1 617-803-7514 > [email protected] or [email protected] > http://oss.software.ibm.com/developerworks/opensource/pthreads > > _______________________________________________ > pthreads-users mailing list > [email protected] > http://www-124.ibm.com/developerworks/oss/mailman/listinfo/pthreads-users