On Sun, Aug 21, 2005 at 09:22:28AM -0500, vinai wrote: > On Sun, August 21, 2005 6:47, Derrik Pates wrote: > > > On Sun, Aug 21, 2005 at 12:27:51AM -0400, Rick Thomas wrote: > >> Does it help to remember that SMP-safe is irrelevant to OldWorld > >> Pmacs, since there are none-such? > > > > Actually, that's not true - I have a 7500 with an aftermarket > > 604e/180 dual CPU card, and I'm sure there was at least one 9600 > > model that had dual processors from the factory, so MP-safety most > > certainly is an issue. Besides, he said he was testing it on an SMP > > box anyway - it'd be hard to get a SWIM3 floppy controller on a > > NewWorld, since the core chipsets in the NewWorlds don't have 'em. > > In addition to Apple's machines (like the 9500 180/MP like Rogerio > mentioned) that came SMP-ready, and CPU upgrade cards, there were > also clones at this time that were also SMP-capable. I remember a > 4-way 604e non-Apple Power Mac. Since all of this arose from the > same line of motherboard and chipsets, I would guess these machines > also have SWIM3 controllers for the floppy drives. > > But just to clarify: I presently do NOT have SMP on my 8500. I have > a single G3 CPU upgrade card. I've compiled a kernel with SMP set, > and tried to use SMP-compatible locking primitives in my 1st attempt > to make the SWIM3 driver SMP safe. But I did not use these properly. > > My attempt here is to try to do several things in one job: get my 8500 > and all its hardware working, learning about device driver programming, > and making code SMP-safe. > > If desireable, I can send the patch to the current 2.6.12 line to get > the driver to compile cleanly. All I did was to replace the older > interrupt calling routines with their updated equivalent.
Yes, please do, as a bug report to linux-2.6 package in sid. Friendly, Sven Luther -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
