Re: SiI3112 SATA controller problems - status

From: Derek Ragona <derek_at_computinginnovations.com>
Date: Mon, 06 Oct 2003 01:55:53 -0500
Søren,

The SATA seems better, but is still not working.  I tried the Oct 5 
snapshot, doing a clean install.  I CVSup'd and did
make buildworld
make buildkernel KERNCONF=GENERIC
make installkernel KERNCONF=GENERIC
reboot to single user
mount -a
swapon -a
mergemaster -p
make installworld

System has the same DMA write problems.  In this recent case it was 
installing termcap.  I tried the Oct 1 snapshot and it was the same, it 
errored in installworld installing tar.

The SATA is stabler as I can buildworld, but not stable enough to handle 
the installworld copies.

Here is what I copied from the console after it went into the debugger:
spec_getpages I/O Read failure(error=5)
vm_fault: pager read error
panic: initiate_write_inodeblock_ufs2: already started

Let me know if you want more information.  I will have to reinstall to get 
a dmesg, if you want a new one.

         -Derek


At 04:09 PM 9/30/2003 -0500, Derek Ragona wrote:
>Great news to hear Søren.
>
>Do you know when a snapshot with the fix will be available?  Since I 
>cannot complete a buildworld using the SATA drive without the fix.
>
>Thanks for the fix.
>
>         -Derek
>
>
>At 10:06 PM 9/30/2003 +0200, Soren Schmidt wrote:
>
>>I have now found a way to reproduce the DMA problems some setups seems
>>to be having with the SiI3112A SATA chip.
>>
>>Now, if I use real SATA drives, (and I mean real SATA drives, the
>>WD Raptor fx is not a real SATA device but just a PATA device
>>with a build in PATA->SATA converter chip) it works just fine
>>no matter what I try.
>>
>>If however I use an older PATA drive and a certain PATA->SATA "dongle"
>>(in this case 1 out of 4 samples) I can reproduce the problem pretty
>>easily.
>>
>>This suggests to me that we are looking at a timing problem of sorts
>>the question is where.
>>
>>I'll work on the problem as time permits...
>>
>>-Søren
>>_______________________________________________
>>freebsd-current_at_freebsd.org mailing list
>>http://lists.freebsd.org/mailman/listinfo/freebsd-current
>>To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"
>_______________________________________________
>freebsd-current_at_freebsd.org mailing list
>http://lists.freebsd.org/mailman/listinfo/freebsd-current
>To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"
Received on Sun Oct 05 2003 - 21:55:58 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:24 UTC