Re: [PATCH 1/1] ptrace/x86: fix the TIF_FORCED_TF logic in handle_signal()

2015-02-23 Thread Oleg Nesterov
On 02/17, Andres Freund wrote: > > Could this patch please be picked up? I very regularly hit the problems > caused due to this in gdb (just single step out of a system call that > returns due to EINTR to reproduce and then single step some more...). > > I've first spent an embarassing amount of ti

Re: [PATCH 1/1] ptrace/x86: fix the TIF_FORCED_TF logic in handle_signal()

2015-02-16 Thread Andres Freund
Hi, Could this patch please be picked up? I very regularly hit the problems caused due to this in gdb (just single step out of a system call that returns due to EINTR to reproduce and then single step some more...). I've first spent an embarassing amount of time trying to figure out what's wrong,

Re: [PATCH 1/1] ptrace/x86: fix the TIF_FORCED_TF logic in handle_signal()

2014-11-27 Thread Oleg Nesterov
ping ;) Should I resend? This fixes the real (although not that serious) bug and nobody objected. On 11/03, Oleg Nesterov wrote: > > When the TIF_SINGLESTEP tracee dequeues a signal, handle_signal() > clears TIF_FORCED_TF and X86_EFLAGS_TF but leaves TIF_SINGLESTEP set. > > If the tracer does PT

Re: [PATCH 1/1] ptrace/x86: fix the TIF_FORCED_TF logic in handle_signal()

2014-11-05 Thread Pedro Alves
On 11/04/2014 11:55 PM, Oleg Nesterov wrote: > On 11/03, Pedro Alves wrote: >> >> Thanks a lot Oleg. > > thanks for the detailed report ;) > >> Question - shouldn't ptrace tests be put in >> tools/testing/selftests/ptrace/ in the kernel tree nowadays? > > Oh, I do not know. Personally I am not s

Re: [PATCH 1/1] ptrace/x86: fix the TIF_FORCED_TF logic in handle_signal()

2014-11-04 Thread Oleg Nesterov
On 11/03, Pedro Alves wrote: > > Thanks a lot Oleg. thanks for the detailed report ;) > Question - shouldn't ptrace tests be put in > tools/testing/selftests/ptrace/ in the kernel tree nowadays? Oh, I do not know. Personally I am not sure that selftests/ptrace/ makes sense and I did not even kno

Re: [PATCH 1/1] ptrace/x86: fix the TIF_FORCED_TF logic in handle_signal()

2014-11-03 Thread Pedro Alves
Thanks a lot Oleg. Question - shouldn't ptrace tests be put in tools/testing/selftests/ptrace/ in the kernel tree nowadays? Thanks, Pedro Alves On 11/03/2014 08:13 PM, Oleg Nesterov wrote: > When the TIF_SINGLESTEP tracee dequeues a signal, handle_signal() > clears TIF_FORCED_TF and X86_EFLAGS_T

[PATCH 1/1] ptrace/x86: fix the TIF_FORCED_TF logic in handle_signal()

2014-11-03 Thread Oleg Nesterov
When the TIF_SINGLESTEP tracee dequeues a signal, handle_signal() clears TIF_FORCED_TF and X86_EFLAGS_TF but leaves TIF_SINGLESTEP set. If the tracer does PTRACE_SINGLESTEP again, enable_single_step() sets X86_EFLAGS_TF but not TIF_FORCED_TF. This means that the subsequent PTRACE_CONT doesn't not