On 2011-02-17, Dmitry V. Levin <[email protected]> wrote:
> On Wed, Feb 16, 2011 at 06:10:55PM +0000, Grant Edwards wrote:
> [...]
>> Would people be agreeable to adding argument type info (or argument
>> format function pointers) to the sysent table? In addition to making
>> the fd/path match function simpler, it might allow the "printargs"
>> function to handle many more cases.
>
> Such an extension is definitely useful. The only drawback is that
> someone would have to update linux/*/syscallent*.h files.
Updating all 19 architectures is probably beyond what I can justify,
and right now I only have the ability to test on IA32. With some
additional work I could setup systems to test on AMD64 and ARM, but
that would have to be done in my spare time.
I could envision a first pass where only information about location of
file descriptor and path args is added and the system call display
functions are pretty much left as-is (using hard-wired parameter
decoding rather than table-driven, but with the addition of
"printfd()" calls similar to "printpath()" calls).
> For some syscalls, e.g. indirect_ipccall syscalls like msgrcv(2), and
> syscalls with "long long" arguments where printllval() is used to
> print them, even the number of arguments differ from one architecture
> to another.
--
Grant Edwards grant.b.edwards Yow! I'm also pre-POURED
at pre-MEDITATED and
gmail.com pre-RAPHAELITE!!
------------------------------------------------------------------------------
The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
Pinpoint memory and threading errors before they happen.
Find and fix more than 250 security defects in the development cycle.
Locate bottlenecks in serial and parallel code that limit performance.
http://p.sf.net/sfu/intel-dev2devfeb
_______________________________________________
Strace-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/strace-devel
