| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
| |
months that I broke it before the 5.5 release.
confirmed as not being required by ports by sthen@, ajacoutot@, dcoppa@
|
| |
|
|
|
|
| |
rescinded 22 July 1999. Proofed by myself and Theo.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Fix bug where altivec context was not freed on process exit.
Fix bug where vscr was not correctly saved/restored.
replace asm statement was macros which expand to the same
asm code or to .long XXX which evaluates to the same instruction
since in-tree gas does not support altivec.
Enable ALTIVEC support by default on macppc, still conditional
for other powerpc ports.
|
|
|
|
|
| |
in ptrace.h, not in md header files.
Protect vector functions/variables with _KERNEL.
|
|
|
|
|
|
|
| |
the 'as' in openbsd source tree does not yet support altivec instructions.
The pieces to enable it have been put in macppc/conf/GENERIC and
macppc/conf/Makefile.macppc in comments.
Once 'as' is updated the kernel option should be removed.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
access to the registers. The format of the registers returned in ptrace,
/proc and in the coredump header match what gdb was expecting.
Floating point registers are not yet copied into the structure.
The location of the fpr needs to be found and copied. <- TODO
Changes to trap.c was to add some interm debugging. code is commented out.
powerpc does not currently have a MID_* type, since it uses ELF instead
of a.out, one was never added. Currently the coredump is done in form
of MID_NONE. What is the correct approach for this?? Elf format coredumps...
maybe?
|
|
The reg.h file may not be complete, currently the port does not support
ptrace.
|