Bug#669006: ata_piix poor performance on ich5

2012-05-02 Thread Teemu
: : Teemu Järvinen wrote:
: :  At first I noticed that NFS performance is low when tried to access this
: :  server, and any serious reading attempt leads nfsd consuming all the
: :  available cpu time. Later it became obvious that the local filesystem
: :  is to blame, it behaves like there is no DMA.

 The problem solved, faulty hd eventually revealed itself by
 generating enough i/o-errors.

: [1.756838] ata2.00: ATA-8: WDC WD5000AACS-00G8B1, 05.04C05, max UDMA/133

 After replacing this disc, the speed issues are history.

 The cpu burning is related to NFSv4, after reconfigured to use NFSv3
 my file server is working smoothly.

-- 
 Teemu



--
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20120502125300.ga5...@tambu.hopto.org



Bug#669006: ata_piix poor performance on ich5

2012-04-17 Thread Teemu
On Mon, Apr 16, 2012 at 11:17:23AM -0500, Jonathan Nieder wrote:
: Teemu Järvinen wrote:
:  At first I noticed that NFS performance is low when tried to access this
:  server, and any serious reading attempt leads nfsd consuming all the
:  available cpu time. Later it became obvious that the local filesystem
:  is to blame, it behaves like there is no DMA.
: 
: Please attach full dmesg output from booting and reproducing the problem.

 I'll attach my dmesg

: Also, just to rule out a userspace (udev, etc) problem: if you install
: the squeeze kernel on your wheezy/sid system, it doesn't reproduce the
: problem, right?

 Well, I might be barking at the wrong tree here?

 I'm not able to use the exactly same kernel I was using with squeeze,
 it was a custom 2.6.32.8 and I'm not able to fiddle with it as it
 seems to not support my usb-keyboard I'm using to poke with it now.

 With squeeze's 2.6.32-5-486 tried simple dd if=/dev/zero of=foo (512 MB)
 which does not consume all the cpu time. There's a little difference
 in performance:
 3.2.0-2-486 33.9 MB/s
 2.6.32-5-486 45.9 MB/s

 However, didn't try this (yet) with 2.6.32 as I didn't have my X
 running and didn't know that my dd-test was not heavy enough...
 When copying about 1.3 GB over NFS, the cpu load is at 100% -
 Pretty sure this was not the case when running squeeze.

-- 
 Teemu

[0.00] Initializing cgroup subsys cpuset
[0.00] Initializing cgroup subsys cpu
[0.00] Linux version 3.2.0-2-486 (Debian 3.2.12-1) 
(debian-kernel@lists.debian.org) (gcc version 4.6.3 (Debian 4.6.3-1) ) #1 Tue 
Mar 20 18:45:21 UTC 2012
[0.00] BIOS-provided physical RAM map:
[0.00]  BIOS-e820:  - 0009f800 (usable)
[0.00]  BIOS-e820: 0009f800 - 000a (reserved)
[0.00]  BIOS-e820: 000f - 0010 (reserved)
[0.00]  BIOS-e820: 0010 - 7fff (usable)
[0.00]  BIOS-e820: 7fff - 7fff3000 (ACPI NVS)
[0.00]  BIOS-e820: 7fff3000 - 8000 (ACPI data)
[0.00]  BIOS-e820: fec0 - 0001 (reserved)
[0.00] Notice: NX (Execute Disable) protection missing in CPU!
[0.00] DMI 2.2 present.
[0.00] DMI:/IS7-E2 (Intel i865-ICH5), BIOS 6.00 PG 06/01/2004
[0.00] e820 update range:  - 0001 (usable) 
== (reserved)
[0.00] e820 remove range: 000a - 0010 (usable)
[0.00] last_pfn = 0x7fff0 max_arch_pfn = 0x10
[0.00] MTRR default type: uncachable
[0.00] MTRR fixed ranges enabled:
[0.00]   0-9 write-back
[0.00]   A-B uncachable
[0.00]   C-CCFFF write-protect
[0.00]   CD000-E uncachable
[0.00]   F-F write-through
[0.00] MTRR variable ranges enabled:
[0.00]   0 base 0 mask F8000 write-back
[0.00]   1 disabled
[0.00]   2 disabled
[0.00]   3 disabled
[0.00]   4 disabled
[0.00]   5 disabled
[0.00]   6 disabled
[0.00]   7 disabled
[0.00] x86 PAT enabled: cpu 0, old 0x7040600070406, new 0x7010600070106
[0.00] found SMP MP-table at [c00f5bc0] f5bc0
[0.00] initial memory mapped : 0 - 0180
[0.00] Base memory trampoline at [c009c000] 9c000 size 12288
[0.00] init_memory_mapping: -377fe000
[0.00]  00 - 40 page 4k
[0.00]  40 - 003740 page 2M
[0.00]  003740 - 00377fe000 page 4k
[0.00] kernel direct mapping tables up to 377fe000 @ 17fb000-180
[0.00] RAMDISK: 7fd31000 - 7fff
[0.00] Allocated new RAMDISK: 3753f000 - 377fd884
[0.00] Move RAMDISK from 7fd31000 - 7ffef883 to 
3753f000 - 377fd883
[0.00] ACPI: RSDP 000f7750 00014 (v00 IntelR)
[0.00] ACPI: RSDT 7fff3000 0002C (v01 IntelR AWRDACPI 42302E31 AWRD 
)
[0.00] ACPI: FACP 7fff3040 00074 (v01 IntelR AWRDACPI 42302E31 AWRD 
)
[0.00] ACPI: DSDT 7fff30c0 0445C (v01 INTELR AWRDACPI 1000 MSFT 
010E)
[0.00] ACPI: FACS 7fff 00040
[0.00] ACPI: APIC 7fff7540 00068 (v01 IntelR AWRDACPI 42302E31 AWRD 
)
[0.00] ACPI: Local APIC address 0xfee0
[0.00] 1159MB HIGHMEM available.
[0.00] 887MB LOWMEM available.
[0.00]   mapped low ram: 0 - 377fe000
[0.00]   low ram: 0 - 377fe000
[0.00] Zone PFN ranges:
[0.00]   DMA  0x0010 - 0x1000
[0.00]   Normal   0x1000 - 0x000377fe
[0.00]   HighMem  0x000377fe - 0x0007fff0
[0.00] Movable zone start PFN for each node
[0.00] early_node_map[2] active PFN ranges
[0.00] 0: 0x0010 - 0x009f
[0.00] 0: 0x0100 - 0x0007fff0
[0.00] On node 0 

Bug#669006: ata_piix poor performance on ich5

2012-04-16 Thread Jonathan Nieder
Teemu Järvinen wrote:

 At first I noticed that NFS performance is low when tried to access this
 server, and any serious reading attempt leads nfsd consuming all the
 available cpu time. Later it became obvious that the local filesystem
 is to blame, it behaves like there is no DMA.

Please attach full dmesg output from booting and reproducing the problem.

Also, just to rule out a userspace (udev, etc) problem: if you install
the squeeze kernel on your wheezy/sid system, it doesn't reproduce the
problem, right?

Thanks,
Jonathan



--
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20120416161723.GP5813@burratino