Messages in this thread | | | Date | Fri, 25 Jul 2008 08:49:14 -0700 | From | Mike Travis <> | Subject | Re: [RFC 00/15] x86_64: Optimize percpu accesses |
| |
Jeremy Fitzhardinge wrote: > H. Peter Anvin wrote: >> Did the suspected linker bug issue ever get resolved? > > I don't believe so. I think Mike is getting very early crashes > depending on some combination of gcc, linker and kernel config. Or > something. > > This fragility makes me very nervous. It seems hard enough to get this > stuff working with current tools; making it work over the whole range of > supported tools looks like its going to be hard. > > J
FYI, I think it was a combination of errors that was causing my problems. In any case, I've successfully compiled and booted Ingo's pesky config config-Tue_Jul__1_16_48_45_CEST_2008.bad with gcc's 4.2.0, 4.2.3 and 4.2.4 on both Intel and AMD boxes. (As well as a variety of other configs.)
I think Hugh's change adding "text" to the BUILD_IRQ macro might also have helped, since interrupts seemed to have always been involved in the panics. That might explain the variable-ness of the gcc version's.
Thanks, Mike
| |