[Bugme-new] [Bug 4163] New: HPT3xx (370A) DMA issues and the HighPoint open source(?) driver

[bugme-daemon at osdl.org]

http://bugme.osdl.org/show_bug.cgi?id=4163

           Summary: HPT3xx (370A) DMA issues and the HighPoint open
                    source(?) driver
    Kernel Version: 2.6.10 vanilla
            Status: NEW
          Severity: normal
             Owner: bzolnier at elka.pw.edu.pl
         Submitter: grin at grin.hu


Distribution: Debian/sid
Hardware Environment: i386 (K7), VIA KT400 chipset, HPT370A soft"raid" crap
Software Environment: nothing special

Problem Description:
The HPT370 driver have a long-long-long standing problem with DMA, it has been 
buggy under 2.4.xx and it's still dead fish under 2.6.10. The effect is that 
after a short while I get dma timeouts, lost interrupts and other ugly things, 
which basically renders the interface dead and the module stuck. Reboot helps (I 
didn't try to force remove it). Switching off DMA (before the errors) usually 
solves the problem and the same time generates at least 90% CPU load. 

Problem is present with one master, or master+slave or any other possible 
combinations on the 2 channels.

http://www.highpoint-tech.com/USA/bios_hpt370_370a.htm contains a linux driver 
source which seem to work, no DMA problems. I do not know whether anyone checked 
it (not found on lkml.org achive), whether it's proper open source or not, but I 
wanted to tell that it works and comes with sources, and the kernel one seems 
not to like this particular controller.

Or maybe a user error, but I tried to check it well before submitting. As far as 
I am able to...

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.