I get these errors every now and then i wonder if i can solve this issues out but i dont know what the problem is. Is it the driver for sata support or is it something else?

I got
00:0c.0 Mass storage controller: Silicon Image, Inc. SiI 3512 [SATALink/SATARaid] Serial ATA Controller (rev 01)
00:0d.0 Mass storage controller: Silicon Image, Inc. SiI 3112 [SATALink/SATARaid] Serial ATA Controller (rev 02)

and get there errors

Mar 29 22:30:21 tkan kernel: ata2.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen
Mar 29 22:30:21 tkan kernel: ata2.00: cmd 25/00:00:e7:19:5e/00:01:25:00:00/e0 tag 0 cdb 0x0 data 131072 in
Mar 29 22:30:21 tkan kernel: res 40/00:8f:58:1a:5e/00:00:25:00:00/e0 Emask 0x4 (timeout)
Mar 29 22:30:21 tkan kernel: ata2.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen
Mar 29 22:30:21 tkan kernel: ata2.00: cmd 25/00:00:e7:1a:5e/00:01:25:00:00/e0 tag 0 cdb 0x0 data 131072 in
Mar 29 22:30:21 tkan kernel: res ff/ff:ff:ff:ff:ff/ff:ff:ff:ff:ff/ff Emask 0x2 (HSM violation)
Mar 29 22:30:52 tkan kernel: ata2.00: qc timeout (cmd 0xec)
Mar 29 22:30:52 tkan kernel: ata2.00: failed to IDENTIFY (I/O error, err_mask=0x4)
Mar 29 22:30:52 tkan kernel: ata2.00: revalidation failed (errno=-5)
Mar 29 22:30:52 tkan kernel: ata2: failed to recover some devices, retrying in 5 secs

Thanks in advance

what kernel are you using ? 2.6.x ? what config file are you using - stock or custom, have you gone through the config file and made sure that the Silicon Image, Inc. SiI is there and is built as a module or built in? i would choose built in anyway

edit: i have had an issue with nfs on slack12 seems the stock kernel doesn't like the nfs module as it complains about invalid format or something, - i think a recompile should sort your issues out, if not get hold of me on irc and i will sort it for you (much like old times on the 2.4.x kernels )

If boot args like "acpi=off noapic nolapic" work for you it's a HW problem (and I know it is). If you want to see problems cards like the SiI 3512 cause (or vendors implementing protocols badly or not disclosing with or w/o NDA) you only have to check LQ or the LKML.

he had it working under a 2.4 kernel i compiled for him a year or so back, so its more than likely hes misconfig'd something

SATA things changed radically between 2.4 and 2.6 AFAIK.

true - i would have thought support would have got better

While Libata is still maturing, device support gets better each time developers get their hands on specs one way or another. While some vendors may choose to disclose voluntarily or under NDA or conform to known standards anyway, others saturate the market with cheap, malfunctioning crap that implements protocols partially, like the SiI 3512 does. The Linux side of things should not be blamed for the fsckup SiL is.

its a hardware issue he has - running sata 3.0gbps drives on a 1.5gbps card, limiting the drives to 1.5gbps hopefully will sort it for him