Re: [kvm-devel] [patch 0/3] separate thread for IO handling V3
Marcelo Tosatti wrote: This version fixes the vmdk problems found by the regression testing. Dor, regarding the option to disable the IO thread, it would require duplicating most of the changed code. For now I believe its better to get the patch into a state where its considered stable enough for inclusion. Please rerun the regression tests. Thanks. The regression tests are happy. There's still something wrong. When I start Windows with these patches applied, Windows startup consumes ~50 sec of cpu time, compared to ~12 sec without. Pinning qemu into cpu 0 seems to fix this. I'll investigate some more. -- error compiling committee.c: too many arguments to function - Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace ___ kvm-devel mailing list kvm-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/kvm-devel
Re: [kvm-devel] [patch 0/3] separate thread for IO handling V3
Avi Kivity wrote: There's still something wrong. When I start Windows with these patches applied, Windows startup consumes ~50 sec of cpu time, compared to ~12 sec without. Pinning qemu into cpu 0 seems to fix this. I'll investigate some more. Changing GUI_REFRESH_INTERVAL to 1000 (from 30) reduces cpu usage back to normal. So I'm guessing there's some bad interaction between the iothread getting the dirty log and vcpu 0. -- error compiling committee.c: too many arguments to function - Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace ___ kvm-devel mailing list kvm-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/kvm-devel
Re: [kvm-devel] [patch 0/3] separate thread for IO handling V3
Avi Kivity wrote: Avi Kivity wrote: There's still something wrong. When I start Windows with these patches applied, Windows startup consumes ~50 sec of cpu time, compared to ~12 sec without. Pinning qemu into cpu 0 seems to fix this. I'll investigate some more. Changing GUI_REFRESH_INTERVAL to 1000 (from 30) reduces cpu usage back to normal. So I'm guessing there's some bad interaction between the iothread getting the dirty log and vcpu 0. Turns out we held slots_lock while in guest mode (which we can be in for unbounded time). Dropping the lock restored behavior to normal, so I'm applying the patches. -- error compiling committee.c: too many arguments to function - Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace ___ kvm-devel mailing list kvm-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/kvm-devel
[kvm-devel] [patch 0/3] separate thread for IO handling V3
This version fixes the vmdk problems found by the regression testing. Dor, regarding the option to disable the IO thread, it would require duplicating most of the changed code. For now I believe its better to get the patch into a state where its considered stable enough for inclusion. Please rerun the regression tests. Thanks. -- - Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace ___ kvm-devel mailing list kvm-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/kvm-devel