Aspire one disk issue on current
 

Hi everyone,
yesterday I installed current on my Acer Aspire one 110. It installed flawlessly and everything seems to be working fine, except for an issue i haven't seen before. The SSD disk seems incapable of doing two things at the same time... I came to that conclusion after altering my firefox settings. Firefox would freeze everytime the hdd led was blinking, so in order to find out what was wrong i disabled caching in firefox. That fixed the problem as far as firefox is concerned but i still cannot have the sdd used by two processes at any time.
Any ideas?

Check if DMA is enabled, replace sdX with your drive.

hdparm -I /dev/sdX | grep dma

and

hdparm -I /dev/sdX | grep DMA

Hdparm
 

hdparm reports udma4 (there is a * next to it and i'm assuming that's what it means).
By the way hdparm -Tt return 642.20MB/s and 38.65MB/s
The weird part is that the disk seems to be working fine with slack 12.2 I was using before...

Hi,
If the system was working with 12.2, what was the reason to move to -current?

'-current' is for testing before it is released as stable.

Are you using ext4 by any chance?

Also, are your drive(s) listed as /dev/hd$ or /dev/sd$. The aspire one uses the SATA interface, and should be listed as /dev/sd$.

Why current
 

The aspire serves as a "testing mule" i used to run slack 12.2 but last month i installed Ubuntu 9.04 in order to check it out (as expected i didn't like it, but it did have the same disk issue.. could it have something to do with the kernel?) and then since i want to try out the new KDE i installed current...

disturbed1, my disks are recognized as /dev/sdX... and i'm using ext2 in order to minimize the wear of the SSD

The big kernel regression started being noticed ~2.6.18. Those that know think it happened around the time CFQ was made the default scheduler. But can't pin it down, because it depends on how the moon, stars, and sun are aligned along with the temperature of the sea ;)

I personally haven't noticed any major regressions from 2.6.26 -2.6.30 besides some minor AGP issues that have been fixed.

Ubuntu uses 2.6.28, Slackware 12.2 uses 2.6.27, current is on 2.6.30. You could try a different scheduler.

If you weren't using ext2 on 12.2, that could be your issue as well.

Maybe compile the same kernel from 12.2 to see if that fixes your issue. Then do a bisect if you have the time, or at least dig through the kernel change logs to see what changes apply to your hardware.

Yesterday I switched /home from the SSD to an SD card and the problem has diminished significantly, but I am yet to find an explanation, I'll try switching /tmp to a ramdisk since it is something suggested from a lot of people in order to limit the SSD wear and see how it goes.
Should i expect any problems from a "session only" /tmp ?

Hi,

You can set the ram allocation for the tmpfs with the size option. Setup through your '/etc/fstab'; example post.