> Please send a patch that creates a linux/m68k/syscallent.h with correct
> contents, unless m68k really matches i386 exactly. Please also compose
> ChangeLog entries in the canonical format you see there and send those
> ahead of the patch.
Thanks for reminding me. m68k does not match i386 exactly, nor does it
slightly resemble i386 anymore beyond syscall no. 218 (the earlier
differences were due to chown/lchown being swapped because historically
m68k chose the opposite default behavior from i386 WRT chown of symlinks,
at least that's my vague memory from the old days).
The differences have grown too big to be incorporated into the i386 table,
so I'll separate out the m68k table entirely as you suggest. This means
I'll also send a patch clearing all m68k bits from the i386 table, OK?
> > A point I'm not clear on: is the inclusion of __NR_exit_group or
> > SYS_exit as real syscall no in the defs really necessary? powerpc also
> > calls the same function on both exit and exit_group, but doesn't pass the
> > real syscall number to the strace handler function.
>
> Sorry, I haven't figured out what you are talking about here. The
> __NR_exit_group and SYS_* definitions need to be correct (what the kernel
The definitions are correct.
> really uses), and the table entries for those should use sys_exit.
> Otherwise, the internal process tracking will not work right because it
> won't be treating the exit and exit_group syscalls with the right magic.
OK, that's what I was wondering.
> I've put in the change to define __NR_exit_group for M68K.
> That much is clearly correct.
Do I need to change anything else in order to get a new
linux/m68k/syscallent.h used by the build?
Michael
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
