On 4/24/2013 3:33 PM, Carlos O'Donell wrote:
In my opinion, debian unstable is now "stable". Probably, this is because it has been pretty much frozen except for bug fixes for several months. It taken many months but the SMP stability problems are pretty much fixed. Helge's machine was running stablyOn Wed, Apr 24, 2013 at 9:48 AM, John David Anglin <dave.anglin@bell.net> wrote:On 4/24/2013 9:18 AM, Carlos O'Donell wrote:On Sat, Apr 20, 2013 at 9:48 AM, Helge Deller <deller@gmx.de> wrote:So, sorry for forgetting to mention GENTOO LINUX on the website! I've fixed it now (please wait one hour until the cron job activates the changes).I'm using Gentoo for all my hppa hacking right now since I wanted to focus on toolchain issues rather than distribution issues :/I have to say one learns a lot about toolchain issues building a distribution...Absolutely. That doesn't detract from needing a stable environment to build and test with, which gentoo provides.
at a load average of ~250 in hackbench testing. It used to crash reliably. It is great that gentoo supported users through a rocky period.
Debian 2.13-38. Nothing much has changed since your patches were integrated.Have you played with mysql-5.5? Testsuite appears to show some glibc problems (pthread_cond_timedwait). Thoughts?What glibc version?
Kernel was 3.9-r7+.
We use the generic pthread_cond_timedwait implementation. The only thought I have is that pthread_cond_timedwait in master might try to use PI-aware mutexes and we haven't had much testing of the PI support in hppa. So there might be some issues to review there. I'll run the mysql-5.5 testsuite to try validate some of the cleanup I have planned.
That would be great. Dave -- John David Anglin dave.anglin@bell.net