|
From: lemmel <le...@gm...> - 2007-03-09 12:02:47
|
I encountered the : --14751-- WARNING: unhandled syscall: 124 --14751-- You may be able to write your own handler. --14751-- Read the file README_MISSING_SYSCALL_OR_IOCTL. and wished to add this syscall, and then checkout the 3.3.0 version of valgrind, and looked for the README_MISSING_SYSCALL_OR_IOCTL. But the steps doesn't seem to bu updated for the current version. How can I add this syscall ? |
|
From: Tom H. <to...@co...> - 2007-03-09 12:15:16
|
In message <200703091302.45841@PowerMail>
le...@gm... wrote:
> I encountered the :
>
> --14751-- WARNING: unhandled syscall: 124
> --14751-- You may be able to write your own handler.
> --14751-- Read the file README_MISSING_SYSCALL_OR_IOCTL.
>
> and wished to add this syscall, and then checkout the 3.3.0 version of
> valgrind, and looked for the README_MISSING_SYSCALL_OR_IOCTL. But the
> steps doesn't seem to bu updated for the current version. How can I
> add this syscall ?
The documentation doesn't look hugely out of date - what was the
problem you encountered with it?
BTW, please enter a bug for the missing system call (and another one
for any problems with the documentation).
Make sure you indicate what platform you're talking about for the
system call, as the number is useless without knowing the platform.
Tom
--
Tom Hughes (to...@co...)
http://www.compton.nu/
|
|
From: lemmel <le...@gm...> - 2007-03-09 12:57:14
|
> The documentation doesn't look hugely out of date - what was the > problem you encountered with it? In the paragraph "Writing your own syscall wrappers", there is : Copy this entry to coregrind/vki_unistd-$(VG_PLATFORM).h. file that I didn't found (find ./ -iname "*vki_unistd-*" return nothing). > BTW, please enter a bug for the missing system call (and another one > for any problems with the documentation). I wanted to perform the source modification, for I'll be able to achieve my tests, and help a little the valgrind team. If I just filled a bug report, I'll have neither one of those. |
|
From: Julian S. <js...@ac...> - 2007-03-09 13:11:53
|
> In the paragraph "Writing your own syscall wrappers", there is : > Copy this entry to coregrind/vki_unistd-$(VG_PLATFORM).h. > file that I didn't found (find ./ -iname "*vki_unistd-*" return nothing). True; that should be include/vki/vki-scnums-$(VG_PLATFORM).h now. > > BTW, please enter a bug for the missing system call (and another one > > for any problems with the documentation). > > I wanted to perform the source modification, for I'll be able to achieve my > tests, and help a little the valgrind team. That would be nice. If you tell us what architecture you are using then maybe we could suggest a similar syscall which you can use the wrapper as a starting point. > If I just filled a bug report, I'll have neither one of those. Even just filing good bug reports is very useful. J |
|
From: Tom H. <to...@co...> - 2007-03-09 13:40:27
|
In message <200...@ac...>
Julian Seward <js...@ac...> wrote:
>> > BTW, please enter a bug for the missing system call (and another one
>> > for any problems with the documentation).
>>
>> I wanted to perform the source modification, for I'll be able to achieve my
>> tests, and help a little the valgrind team.
>
> That would be nice. If you tell us what architecture you are using then
> maybe we could suggest a similar syscall which you can use the wrapper as
> a starting point.
Well as 124 is adjtimex on three platforms (and getsid on the other,
but that is implemented) I am guessing it is adjtimex.
Which just happens to have an old commented out wrapper already that
will just need a cleanup and it can be re-enabled.
>> If I just filled a bug report, I'll have neither one of those.
>
> Even just filing good bug reports is very useful.
Indeed - especially as a patch can be attached once it is written ;-)
Tom
--
Tom Hughes (to...@co...)
http://www.compton.nu/
|
|
From: lemmel <le...@gm...> - 2007-03-09 14:26:24
|
> Well as 124 is adjtimex on three platforms (and getsid on the other, > but that is implemented) I am guessing it is adjtimex. Good guess :-) (i486) > Which just happens to have an old commented out wrapper already that > will just need a cleanup and it can be re-enabled. I'll look for this. I am not sure to have enough time to do it now. If not, I'll try to provide a patch monday or tuesday. > >> If I just filled a bug report, I'll have neither one of those. > > > > Even just filing good bug reports is very useful. Ok. > Indeed - especially as a patch can be attached once it is written ;-) :-D |