On Fri, Sep 03 2021, Mark Kettenis <[email protected]> wrote:
>> From: Jeremie Courreges-Anglas <[email protected]>
>> Date: Fri, 03 Sep 2021 14:32:26 +0200
>>
>> Two changes that would be useful:
>> - enable PT_*FPREGS, Mark has already done the job.
>> - hide PT_STEP since hardware support seems missing (the spec only talks
>> about single stepping support in "Debug mode", which is not "Machine" or
>> "Supervisor" mode). Since we don't emulate it (like mips64 for
>> example) it doesn't make sense to provide the define.
>> I doubt that
>> hiding PT_STEP [will] magically fix devel/gdb runtime support though.
Actually it does.
(gdb) b main
Breakpoint 1 at 0x1d00: file /usr/src/usr.bin/du/du.c, line 69.
(gdb) r
Starting program: /usr/obj/usr.bin/du/du
Breakpoint 1, main (argc=<optimized out>, argv=<optimized out>) at
/usr/src/usr.bin/du/du.c:69
69 if (pledge("stdio rpath", NULL) == -1)
(gdb)
The details are a bit gory but somehow egdb manages to cope.
>>
>> Thoughts? ok?
>
> ok kettenis@
Thank you Mark!
>> Index: ptrace.h
>> ===================================================================
>> RCS file: /d/cvs/src/sys/arch/riscv64/include/ptrace.h,v
>> retrieving revision 1.2
>> diff -u -p -p -u -r1.2 ptrace.h
>> --- ptrace.h 12 May 2021 01:20:52 -0000 1.2
>> +++ ptrace.h 3 Sep 2021 12:25:54 -0000
>> @@ -16,10 +16,10 @@
>> * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
>> */
>>
>> +#if 0
>> #define PT_STEP (PT_FIRSTMACH + 0)
>> +#endif
>> #define PT_GETREGS (PT_FIRSTMACH + 1)
>> #define PT_SETREGS (PT_FIRSTMACH + 2)
>> -#if 0 // XXX ptrace fpreg support
>> #define PT_GETFPREGS (PT_FIRSTMACH + 3)
>> #define PT_SETFPREGS (PT_FIRSTMACH + 4)
>> -#endif
>>
>>
>> --
>> jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
>>
>>
>
--
jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
