Re: 2.4 fix for write throttling on x86 >1G

2005-03-11 Thread Marcelo Tosatti
On Fri, Mar 11, 2005 at 09:53:09PM +0100, Andrea Arcangeli wrote: > Hello Marcelo, > > On Fri, Mar 11, 2005 at 01:04:13PM -0300, Marcelo Tosatti wrote: > > Out of curiosity, that was SuSE not mainline ? > > yep. > > > Do we really want to limit dirty cache to low mem on HIGHIO capable > >

Re: 2.4 fix for write throttling on x86 >1G

2005-03-11 Thread Andrea Arcangeli
Hello Marcelo, On Fri, Mar 11, 2005 at 01:04:13PM -0300, Marcelo Tosatti wrote: > Out of curiosity, that was SuSE not mainline ? yep. > Do we really want to limit dirty cache to low mem on HIGHIO capable > machines? I'm afraid doing so might hurt performance on such systems. > > I think it

Re: 2.4 fix for write throttling on x86 >1G

2005-03-11 Thread Marcelo Tosatti
Hi Andrea! On Fri, Mar 11, 2005 at 07:10:35AM +0100, Andrea Arcangeli wrote: > Hello Marcelo, > > I've got a fix for you on 2.4. I got reports of stalls with heavy writes > on 2.4. Out of curiosity, that was SuSE not mainline ? > There was a mistake in nr_free_buffer_pages. That function is

Re: 2.4 fix for write throttling on x86 1G

2005-03-11 Thread Marcelo Tosatti
Hi Andrea! On Fri, Mar 11, 2005 at 07:10:35AM +0100, Andrea Arcangeli wrote: Hello Marcelo, I've got a fix for you on 2.4. I got reports of stalls with heavy writes on 2.4. Out of curiosity, that was SuSE not mainline ? There was a mistake in nr_free_buffer_pages. That function is

Re: 2.4 fix for write throttling on x86 1G

2005-03-11 Thread Andrea Arcangeli
Hello Marcelo, On Fri, Mar 11, 2005 at 01:04:13PM -0300, Marcelo Tosatti wrote: Out of curiosity, that was SuSE not mainline ? yep. Do we really want to limit dirty cache to low mem on HIGHIO capable machines? I'm afraid doing so might hurt performance on such systems. I think it might

Re: 2.4 fix for write throttling on x86 1G

2005-03-11 Thread Marcelo Tosatti
On Fri, Mar 11, 2005 at 09:53:09PM +0100, Andrea Arcangeli wrote: Hello Marcelo, On Fri, Mar 11, 2005 at 01:04:13PM -0300, Marcelo Tosatti wrote: Out of curiosity, that was SuSE not mainline ? yep. Do we really want to limit dirty cache to low mem on HIGHIO capable machines? I'm

2.4 fix for write throttling on x86 >1G

2005-03-10 Thread Andrea Arcangeli
Hello Marcelo, I've got a fix for you on 2.4. I got reports of stalls with heavy writes on 2.4. There was a mistake in nr_free_buffer_pages. That function is definitely meant _not_ to take highmem into account (dirty cache cannot spread over highmem in 2.4 [even when on top of fs]). For unknown

2.4 fix for write throttling on x86 1G

2005-03-10 Thread Andrea Arcangeli
Hello Marcelo, I've got a fix for you on 2.4. I got reports of stalls with heavy writes on 2.4. There was a mistake in nr_free_buffer_pages. That function is definitely meant _not_ to take highmem into account (dirty cache cannot spread over highmem in 2.4 [even when on top of fs]). For unknown