Interesting observation of JMicron SATA controller

[Elliott Mitchell <ehem+xen@xxxxxxx>]

Firmly wearing my user hat here, so sending to xen-users.  I'm unsure
where to send this as I'm unsure who needs to made aware.


Apparently the JMicron AHCI SATA controllers have poor reputations with
users of Linux.  Recently I was able to experiment a bit more and made
what seems a rather interesting observation.

As a stopgap measure I'm using a PCIe card with a JMB585 chip (5 internal
SATA ports, AHCI).  Due to a situation I recently needed to boot a Linux
kernel on this machine without Xen.  Didn't work, really didn't work.
Storage devices attached to the JMB585 were causing timeouts and there
were danger signs from other system storage devices.

Luckily due to having seen hints before, I pulled the JMB585 card and
Linux booted without errors (situation needing Linux without Xen
resolved).  After resolving that, reinstalled card and booted Linux on
Xen as this machine normally does.  With Xen everything behaves properly.


I don't really know what to make of this.  I'm pretty sure everything is
working correctly with Linux on Xen.  Notably the filesystem features
checksums likely to detect data corruption, much of the data on top
includes checksums too.  All I can confirm is the controller works fine
when Linux is running on Xen, but not when directly on hardware.


-- 
(\___(\___(\______          --=> 8-) EHM <=--          ______/)___/)___/)
 \BS (    |         ehem+sigmsg@xxxxxxx  PGP 87145445         |    )   /
  \_CS\   |  _____  -O #include <stddisclaimer.h> O-   _____  |   /  _/
8A19\___\_|_/58D2 7E3D DDF4 7BA6 <-PGP-> 41D1 B375 37D0 8714\_|_/___/5445