Skip to contentSkip to conversation

HD access sluggish in 2.6.0

DISCUSSION Yun Zhou 3 participants 3 messages Dec 19, 2003

lore
Share & tools
lore

Yun Zhou

author
#0 lore raw

I'm using a system with 1HD (DC WD600BB-00CAA1 60 GB) using kernel 2.6.0. Whenever the system uses the disk extensively (copying a file, untarring, etc.), it grinds to a near halt, with the CPU usage at about 100%, even for a simple copying operation.

This problem is not present when using 2.4.22, nor any of the 2.4 series kernels that I've tried. Does anyone know what is causing this?

Thanks in advance!

#1 lore raw
Quoted 4 lines · Yun Zhou

I'm using a system with 1HD (DC WD600BB-00CAA1 60 GB) using kernel 2.6.0. Whenever the system uses the disk extensively (copying a file, untarring, etc.), it grinds to a near halt, with the CPU usage at about 100%, even for a simple copying operation.

Please show the output of hdparm /dev/hda:
/dev/hda:

 multcount    = 16 (on)
 IO_support   =  0 (default 16-bit)
 unmaskirq    =  0 (off)
 using_dma    =  1 (on)
 keepsettings =  0 (off)
 readonly     =  0 (off)
 readahead    = 256 (on)
 geometry     = 65535/16/63, sectors = 117210240, start = 0

You will probably see using_dma = 0 there. This in turn probably means you did not compile in the right IDE driver. Try hdparm -d1 /dev/hda, and/or recompile your kernel.

Quoted 2 lines

This problem is not present when using 2.4.22, nor any of the 2.4 series kernels that I've tried. Does anyone know what is causing this?

You did not specify any details of your kernel compile so it is hard to tell.

Signature · 2 lines
http://www.PowerDNS.com      Open source, database driven DNS Software 
http://lartc.org           Linux Advanced Routing & Traffic Control HOWTO
#2 lore raw
In reply to Yun Zhou
Quoted 7 lines · Yun Zhou

I'm using a system with 1HD (DC WD600BB-00CAA1 60 GB) using kernel 2.6.0. Whenever the system uses the disk extensively (copying a file, untarring, etc.), it grinds to a near halt, with the CPU usage at about 100%, even for a simple copying operation.
This problem is not present when using 2.4.22, nor any of the 2.4 series kernels that I've tried. Does anyone know what is causing this? Thanks in advance!

This is typical of IDE disks where the driver is pessimistically using PIO. A dmesg and .config would help, plus maybe readprofile(1) results (you'll have to boot with profile=1 on the kernel command-line.) just in case it's actually something else.

-- wli