Re: bdflush synchronous IO on prerelease-diff

2001-01-03 Thread Linus Torvalds
Mind checking out the current prerelease-diff? It fixes this, and cleans up some remaining things (now that we don't task-switch into bdflush for all our cleanup, 'nrefill' should really be much lower to get smoother behavior). Linus - To unsubscribe from this list: send the li

Re: bdflush synchronous IO on prerelease-diff

2001-01-03 Thread Marcelo Tosatti
On Wed, 3 Jan 2001, Marcelo Tosatti wrote: > > Hi Linus, > > I've noticed you changed bdflush to do synchronous IO on page_launder(). > > That seems to be a performance problem, since kflushd will have to wait > for dirty buffers to get synced instead looping on the inactive dirty > list mor

bdflush synchronous IO on prerelease-diff

2001-01-03 Thread Marcelo Tosatti
Hi Linus, I've noticed you changed bdflush to do synchronous IO on page_launder(). That seems to be a performance problem, since kflushd will have to wait for dirty buffers to get synced instead looping on the inactive dirty list more often. Here is a patch to change this. --- linux.orig/f