Re: [Valgrind-developers] adding system call support

[Bob F. <bfr...@si...>]

On Sat, 12 Jun 2004, Tom Hughes wrote:
>>
>> Surely something must be done in valgrind.  Prior to my updates,
>> valgrind simply asserted since the timer_* system calls were not
>> supported.  What must be done to keep valgrind from asserting while
>> allowing librt to emulate the real thing?
>
> What was the assertion? I found that I just got a warning about
> system call 266 not being implemented (that's clock_getres) and it
> then fell back to the old technique that was used with 2.4 kernels
> that didn't have the timer system calls.

Unfortunately I did not record the exact assertion message and don't 
have an unmodified valgrind immediately available.  However, the 
assertion was from valgrind and was due to support for timer_create 
not being implemented.  It said not to "panic" since it is easy to add 
system calls to valgrind.

In Linux 2.6.3 timer_create is system call number 259:

#define __NR_timer_create       259
#define __NR_timer_settime      (__NR_timer_create+1)
#define __NR_timer_gettime      (__NR_timer_create+2)
#define __NR_timer_getoverrun   (__NR_timer_create+3)
#define __NR_timer_delete       (__NR_timer_create+4)
#define __NR_clock_settime      (__NR_timer_create+5)
#define __NR_clock_gettime      (__NR_timer_create+6)
#define __NR_clock_getres       (__NR_timer_create+7)
#define __NR_clock_nanosleep    (__NR_timer_create+8)

Bob
======================================
Bob Friesenhahn
bfr...@si...
http://www.simplesystems.org/users/bfriesen