Hi, On Fri, Mar 2, 2018 at 10:01 AM, Will Deacon <will.dea...@arm.com> wrote: > On Thu, Mar 01, 2018 at 11:38:03AM -0800, Douglas Anderson wrote: >> This is the equivalent of commit 001bf455d206 ("ARM: 8428/1: kgdb: Fix >> registers on sleeping tasks") but for arm64. Nuff said. > > It's a pity that 001bf455d206 doesn't explain *why* past_pt_regs doesn't > work.
Ah, sorry. I can add that to the description but Mark Rutland also explained it. task_pt_regs are the userspace registers. That's not what kgdb needs. > Anyway, does this mean you're actually using kgdb on arm64? Does the rest of > it appear to work? Yeah, I've been using it for years. I had this patch on my list to upstream but it conflicted with commit 0d15ef677839 ("arm64: kgdb: Match pstate size with gdbserver protocol") and I never had time to figure out if that mattered, so it's been sitting at http://crosreview.com/371942 forever. NOTE: in general I haven't been using single step and breakpoints since those have always been a bit iffy on serial kgdb and especially since we have read-only text on our system, but maybe I'll try that now that I'm running with a newer kernel. -Doug