Skip to contentSkip to conversation

Linux 2.4.13-ac1

DISCUSSION John Weber 4 participants 6 messages Oct 26, 2001 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

John Weber

author
#1 lore raw

Where is ac7 ?

Quoted 31 lines · Alan Cox

2.4.13-ac1
o Merge with Linus 2.4.13

2.4.12-ac7

o       Fix i2o_proc locking thinko                     (me)
o       Aiptek tablet driver                            (Chris Atenasio)
o       Update ARM920T support                          (Russell King)
o       Add some additional PCI idents                  (Tim Hockin)
o       Further SA1100/CERF ARM updates                 (Russell King)
o       Reduce printk noise in amd flash driver         (David Woodhouse)
o       Add missing schedule() to jffs2 gc thread       (David Woodhouse)
o       Add license tag to NTFS                         (Richard Russon)
o       PCMCIA suspend fix for SA1100 ARM               (Russell King)
o       Increase close wait tracking for ip_conntrack   (Darrell Escola)
o       PCI hot plug support                            (Greg Kroah-Hartmann)
o       Compaq FC driver update                         (Steve Cameron)
o       Fix UP APIC without IOAPIC build fail           (Mikael Pettersson)
o       Intel 82092 PCI/PCMCIA bridge support           (Arjan van de Ven)
        | Based heavily on the pcmcia package
o       Include buffer size extension in dhcp frames    (Andreas Steinmetz)
o       Fix u14f/u34f build problem                     (Andreas Steinmetz)
o       Fix trident.c build on Alpha                    (me)
o       Various other Alpha build fixes                 (Jeff Garzik)
o       Simplify serverworks mtrr check                 (Dave Jones)
o       ARM configuration updates                       (Russell King)
o       Update natsemi driver                           (Jeff Garzik)
o       Computone ip2 update                            (Michael Warfield)
o       Fix CONFIG_SIMNOW                               (Andi Kleen)
o       Update 8139too docs                             (Jeff Garzik)
o       Make autofs4 return symlink lengths
#2 lore raw
In reply to a message not in the archive
Quoted 4 lines · Alan Cox

2.4.13-ac1
o Merge with Linus 2.4.13

Does this mean the ac tree now uses the AA VM, or is this a merge with everything but the VM, like the earlier 2.4.1x-ac trees?

Thanks,

Chris

Signature · 7 lines
Christopher S. Swingley             phone: 907-474-2689
Computer / Network Manager          email: [email protected]
IARC -- Frontier Program            GPG and PGP keys at my web page:
University of Alaska Fairbanks      www.frontier.iarc.uaf.edu/~cswingle

 "They that can give up essential liberty to obtain a little temporary 
  safety deserve neither liberty nor safety."  -- Ben Franklin
#3 lore raw
Quoted 2 lines

Does this mean the ac tree now uses the AA VM, or is this a merge with everything but the VM, like the earlier 2.4.1x-ac trees?

It still uses buffer cache based raw disk access (so things like DVD players actually work ok) and the Riel VM. The way things are panning out I suspect 2.4.14ac* may well be a point I switch to the Andrea/Marcelo/Linus VM.

Alan

#4 lore raw
In reply to Christopher S. Swingley
Quoted 4 lines · Alan Cox
Quoted 2 lines

Does this mean the ac tree now uses the AA VM, or is this a merge with everything but the VM, like the earlier 2.4.1x-ac trees?

It still uses buffer cache based raw disk access (so things like DVD players

btw, AFIK the fact dvd player can hang some millisecond across a close/open cycle, isn't really because of the blkdev in pagecache but simply because the <2.4.10 buffer cache layer wasn't able to do proper readahead on the blkdev. Now we do readahead properly and so in turn the the lack of media-change trust of the vfs shows up. So as far I can tell the right fix have no influence on the blkdev in pagecache, but it only consists in resurrecting the media-change detection with a per-device bitflag whitelist. I cannot see other source of stalls across a close/open cycle.

Andrea

#5 lore raw
Quoted 7 lines

simply because the <2.4.10 buffer cache layer wasn't able to do proper readahead on the blkdev. Now we do readahead properly and so in turn the the lack of media-change trust of the vfs shows up. So as far I can tell the right fix have no influence on the blkdev in pagecache, but it only consists in resurrecting the media-change detection with a per-device bitflag whitelist. I cannot see other source of stalls across a close/open cycle.

I'm not currently sure if the impact is from the cost of the page cache flushing or the invalidate/re-read it triggers. There probably are two or three seeks on the DVD if the data is invalidated so that would make sense.

Alan

#6 lore raw
In reply to Andrea Arcangeli
Quoted 11 lines · Alan Cox
Quoted 7 lines

simply because the <2.4.10 buffer cache layer wasn't able to do proper readahead on the blkdev. Now we do readahead properly and so in turn the the lack of media-change trust of the vfs shows up. So as far I can tell the right fix have no influence on the blkdev in pagecache, but it only consists in resurrecting the media-change detection with a per-device bitflag whitelist. I cannot see other source of stalls across a close/open cycle.

I'm not currently sure if the impact is from the cost of the page cache flushing or the invalidate/re-read it triggers. There probably are two or three seeks on the DVD if the data is invalidated so that would make sense.

plus we also need to wait completion of the readahead synchronously in the cache flushing at close, the larger the readahead, the larger the stall.

I doubt the cpu cost of truncate_inode_pages as originally suspected is the culprit, more likely the fact we have to wait for a larger I/O to complete plus the disk seek due the fact we throw away the readahead, infact invalidate_buffers in-cpu-core cost in -ac could be lot worse considering it has to walk a list that contains all kind of buffers not only belonging to the interesting device to flush.

But anyways the media change detection will avoid to throw away the readahead and in turn it will avoid the seek, it will avoid the synchronous I/O completion wait in truncate_inode_pages, and it will avoid the truncate_inode_pages in-cpu-core cost cost as well, so it should most certainly fix the problem.

Andrea