On Thu, 2007-12-20 at 02:00 -0800, Roland McGrath wrote: > > On first glance looks good to me, I'll test it. Only why do you rename > > set_single_step and clear_single_step? Is that required for the utrace > > patches? > > It's for the series of patches linked from http://lwn.net/Articles/259841/ > If you do the renaming now, then when those get merged you can just remove > the whole bunch of code for PTRACE_SINGLESTEP et al and call ptrace_request. > (And yes, it will also leave things clean and ready for something new akin > to utrace to come along later without your having to do anything more about it.) Ok, test has been successful. The old kernel did indeed do a zero-step on an ignored SIGWINCH. Your patch fixes this and like the old code a signal handler is stopped on the first instrution. The gdb testsuite completes without new failures (although there are quite a few existing failures). To get the kernel compiled you need one small addition to the patch: diff -urpN linux-2.6/include/asm-s390/ptrace.h linux-2.6-patched/include/asm-s390/ptrace.h --- linux-2.6/include/asm-s390/ptrace.h 2007-12-20 12:17:31.000000000 +0100 +++ linux-2.6-patched/include/asm-s390/ptrace.h 2007-12-20 10:37:23.000000000 +0100 @@ -469,6 +469,7 @@ struct user_regs_struct * These are defined as per linux/ptrace.h, which see. */ #define arch_has_single_step() (1) +struct task_struct; extern void user_enable_single_step(struct task_struct *); extern void user_disable_single_step(struct task_struct *); The task_struct is not defined yet in ptrace.h. I'll add the patch to the git390 and include it to the patch-set for 2.6.25 merge window. Thanks Roland and have a merry christmas. -- blue skies, Martin. "Reality continues to ruin my life." - Calvin. - To unsubscribe from this list: send the line "unsubscribe linux-s390" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
