Problem intr11 livelocked

Jonathan Buschmann jonthn at agmact.com
Tue Mar 8 12:23:28 PST 2005


Chuck Tuffli a écrit :
On Tue, Mar 08, 2005 at 10:32:50AM -0800, Matthew Dillon wrote:

   Ok.  You submitted a similar patch January 31st.   Joerg was
   going to do something with it back then but must have forgotten :-)


This is just another in the list of PCI ids the Intel interrupt code
should handle.

   I'll commit it.  It probably isn't entirely correct but eventually we
   are going to have to bring in the latest ATA driver from FreeBSD anyhow,
   so I am not going to worry about it changes that might not entirely work,
   but wouldn't have worked at all before for those device ID's anyway.
					-Matt


I completely agree. I took a quick try at porting the code from 5.x,
but the extensive use of mutexes, semaphores, and locking made porting
exceed my pain threshold. Besides, my SCSI focused management would
look at me funny :)
With regard to the patch, the 5.x ATA code does something similar
(handles all ICH* ATA devices the same) but in a more maintainable
way.
I hope that the ATA code from Fbsd 5.X will be ported to Dfly because 
without that i can't install Dfly on my PC.

JonTHn





More information about the Bugs mailing list