From 957cba89001d8d06e681ca493500209c5a151464 Mon Sep 17 00:00:00 2001 From: Thomas Schwinge Date: Sat, 23 Jul 2011 08:50:54 +0200 Subject: IRC. --- .../io_system/clustered_page_faults.mdwn | 40 ++++++++++++++++++++-- 1 file changed, 37 insertions(+), 3 deletions(-) (limited to 'open_issues/performance') diff --git a/open_issues/performance/io_system/clustered_page_faults.mdwn b/open_issues/performance/io_system/clustered_page_faults.mdwn index 37433e06..9e20f8e1 100644 --- a/open_issues/performance/io_system/clustered_page_faults.mdwn +++ b/open_issues/performance/io_system/clustered_page_faults.mdwn @@ -12,7 +12,10 @@ License|/fdl]]."]]"""]] [[community/gsoc/project_ideas/disk_io_performance]]. -IRC, freenode, #hurd, 2011-02-16 +[[!toc]] + + +# IRC, freenode, #hurd, 2011-02-16 exceptfor the kernel, everything in an address space is represented with a VM object @@ -88,9 +91,8 @@ IRC, freenode, #hurd, 2011-02-16 recommend* ok ---- -IRC, freenode, #hurd, 2011-02-16 +# IRC, freenode, #hurd, 2011-02-16 etenil: OSF Mach does have clustered paging BTW; so that's one place to start looking... @@ -103,3 +105,35 @@ IRC, freenode, #hurd, 2011-02-16 can serve as a starting point + + +# IRC, freenode, #hurd, 2011-07-22 + + but concerning clustered pagins/outs, i'm not sure it's a mach + interface limitation + the external memory pager interface does allow multiple pages to + be transfered + isn't it an internal Mach VM problem ? + isn't it simply the page fault handler ? + braunr: are you sure? I was under the impression that changing the + pager interface was among the requirements... + hm... I wonder whether for pageins, it could actually be handled + in the pages instead of Mach... though this wouldn't work for pageouts, + so probably not very helpful + err... in the pagers + antrik: i'm almost sure + but i've be proven wrong many times, so .. + there are two main facts that lead me to think this + 1/ + http://www.gnu.org/software/hurd/gnumach-doc/Memory-Objects-and-Data.html#Memory-Objects-and-Data + says lengths are provided and doesn't mention the limitation + 2/ when reading about UVM, one of the major improvements (between + 10 and 30% of global performance depending on the benchmarks) was + implementing the madvise semantics + and this didn't involve a new pager interface, but rather a new + page fault handler + braunr: hm... the interface indeed looks like it can handle + multiple pages in both directions... perhaps it was at the Hurd level + where the pager interface needs to be modified, not the Mach one?... + antrik: would be nice wouldn't it ? :) + antrik: more probably the page fault handler -- cgit v1.2.3