On Sun, Mar 9, 2014 at 11:21 PM, Mark Cave-Ayland <mark.cave-ayl...@ilande.co.uk> wrote: > On 05/03/14 18:11, Richard Henderson wrote: > >> Rescued from a branch from September. I can't recall if I ever >> actually posted these back then, so I'll wait a week before asking >> for a pull. >> >> Re-tested on v8plus. The machine doesn't have 64-bit libraries >> installed, so I'm a bit stuck with that. >> >> >> r~ >> >> >> Richard Henderson (11): >> tcg-sparc: Fix ld64 for 32-bit mode >> tcg-sparc: Fix tlb read >> tcg-sparc: Tidy call+jump patterns >> tcg-sparc: Use intptr_t as appropriate >> tcg-sparc: Don't handle remainder >> tcg-sparc: Dont handle constant arguments to ext32 ops >> tcg-sparc: Improve tcg_out_movi >> tcg-sparc: Use TCGMemOp within qemu_ldst routines >> tcg-sparc: Tidy tcg_out_tlb_load interface >> tcg-sparc: Convert to new ldst helpers >> tcg-sparc: Convert to new ldst opcodes >> >> tcg/sparc/tcg-target.c | 599 >> +++++++++++++++++++++++++------------------------ >> tcg/sparc/tcg-target.h | 6 +- >> 2 files changed, 304 insertions(+), 301 deletions(-) > > > Hi Richard, > > The basic patchset applies fine to git master although I got the following > warning from git-am: > > Applying: tcg-sparc: Fix tlb read > /home/build/src/qemu/git/qemu/.git/rebase-apply/patch:49: trailing > whitespace. > > All my SPARC32 and SPARC64 images run with no regressions, and there is a > perceivable slight increase in performance here.
> Tested-by: Mark Cave-Ayland <mark.cave-ayl...@ilande.co.uk> Hi Mark, I think a better test case would be looking for regressions in x86/x86-64 images. I'd suspect someone would launch QEMU on a SPARC machine to emulate some other platform. (Solaris containers would be more efficient for running SPARC on SPARC) Artyom -- Regards, Artyom Tarasenko linux/sparc and solaris/sparc under qemu blog: http://tyom.blogspot.com/search/label/qemu
