Am 19.09.2011 um 08:00 schrieb David Gibson <da...@gibson.dropbear.id.au>:
> On Sat, Sep 17, 2011 at 05:08:29PM +0000, Blue Swirl wrote: >> On Wed, Sep 14, 2011 at 8:43 AM, Alexander Graf <ag...@suse.de> wrote: >>> From: David Gibson <da...@gibson.dropbear.id.au> >>> >>> This patch implements support for the CFAR SPR on POWER7 (Come From >>> Address Register), which snapshots the PC value at the time of a branch or >>> an rfid. The latest powerpc-next kernel also catches it and can show it in >>> xmon or in the signal frames. >>> >>> This works well enough to let recent kernels boot (which otherwise oops >>> on the CFAR access). It hasn't been tested enough to be confident that the >>> CFAR values are actually accurate, but one thing at a time. >> >> This looks accurate at least for the cases covered. >> >> A higher performance implementation could be to only update the >> register lazily when the SPR is read, in most other times CFAR would >> be only stored to DisasContext. > > Uh, yeah. I think I'll leave that to someone who actually understands > TCG, rather than just bluffing it. CFAR is only written on branches at which point we'd have to flush everything from disascontext to env anyways :). So putting it there doesn't buy us anything. Alex
