Binary package hint: kernel-image-2.6.28-11-generic-di
On a sony laptop with a 250 Go HD WD2500BEVS (5400 rpm) running Ubuntu 9.04 - Kernel 2.6.28-11.42-generic, suspend and resume works fine, but after resuming the whole system becomes extremely slow and completely unusable. The hard disk becomes absolutely unresponsive and very slow, as it can be showned by applying the hdparm command :
Before suspend, "hdparm -tT /dev/sda" gives :
Timing cached reads: 1900 MB in 2.00 seconds = 950.18 MB/sec
Timing buffered disk reads: 162 MB in 3.01 seconds = 53.83 MB/sec
after suspend, the same command gives :
Timing cached reads: 1156 MB in 2.00 seconds = 578.00 MB/sec
Timing buffered disk reads: 8 MB in 3.20 seconds = 2.50 MB/sec
This issue seems to be related to the use of the ahci driver for managing the disk, and also probably to a IRQ conflict (the SATA and the ethernet controllers share in fact the same IRQ). A workaround is to use the ata_piix driver instead of ahci, but because the ahci driver is built-in and therefore cant be blacklisted, this workaround requires to recompile the kernel with the ahci driver as module. With a kernel 2.6.28-11.42 recompiled in this way, the disk doesn't slow down any more after resuming :
With the ata_piix driver in use (in a recompiled kernel), "hdparm -tT /dev/sda" gives before suspend :
Timing cached reads: 1732 MB in 2.00 seconds = 866.17 MB/sec
Timing buffered disk reads: 164 MB in 3.01 seconds = 54.52 MB/sec
After suspend, the same command gives :
Timing cached reads: 1838 MB in 2.00 seconds = 919.26 MB/sec
Timing buffered disk reads: 164 MB in 3.00 seconds = 54.61 MB/sec
Another problem related to the first one is that the system doesn't restart properly, it hangs at "The system will now restart". With the ata_piix driver, the restart is fine.
ProblemType: Bug
Architecture: i386
CurrentDmesg:
[ 30.322677] r8169: eth0: link up
[ 30.323819] ADDRCONF(NETDEV_UP): eth1: link is not ready
[ 41.220016] eth0: no IPv6 routers present
DistroRelease: Ubuntu 9.04
HibernationDevice: RESUME=UUID=a455774b-d44d-4827-96d4-d7a3fb0cd26f
MachineType: Sony Corporation VGN-A417M
Package: linux-image-2.6.28-11-generic 2.6.28-11.42
ProcCmdLine: root=UUID=c039a0be-449d-4966-a1b3-07ce4489b67c ro quiet splash locale=fr_FR crashkernel=384M-2G:64M@16M,2G-:128M@16M
ProcEnviron:
PATH=(custom, user)
LANG=fr_FR.UTF-8
SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.28-11.42-generic
SourcePackage: linux
I can verify that I get the same behavior on a Vaio VGN-S5XP with a an Intel Corporation 82801FBM SATA controller and a FUJITSU MHV2100BH 100GB disk. On the latest Jaunty kernel (2.6.28-
The sar utility reports excessive iowait use between 70% and 90%.
I have reverted to kernel 2.6.27-14-generic, where resume works as expected.