Hi,
On 08/05/2015 05:00 PM, Cyril Hrubis wrote:
> Hi!
>> +int safe_bind(const char *file, const int lineno, void (cleanup_fn)(void),
>> + int socket, const struct sockaddr *address,
>> + socklen_t address_len)
>> +{
>> + int err, ret, i;
>> +
>> + for (i = 0; i < 30; i++) {
30 sec still may not be enough, I would set it to 2 minutes.
>> + ret = bind(socket, address, address_len);
>> + err = errno;
>> +
>> + if (!ret)
>> + return 0;
>> +
>> + if (err != EADDRINUSE) {
>> + tst_brkm(TBROK | TERRNO, cleanup_fn,
Also, we don't need ret and err vars, we could do:
if (!bind(socket, address, address_len))
return 0;
if (errno != EADDRINUSE)
tst_brkm(TBROK | TERRNO, ....);
>> + "%s:%d: bind(%d, %p, %d) failed", file, lineno,
>> + socket, address, address_len);
> I think that it would be a bit nicer if we created a helper to translate
> the socket structure to it's string representation (and use it whenever
> we deal with struct sockaddr passed from user). We would have to switch
> on sa_family and then print either ip address or path (for unix
> sockets), etc. But that would get us bind(8, '127.0.0.1:1024', ...)
> failed instead of the raw pointer to the structure which is kind of
> useless.
>
>> + }
>> +
>> + tst_resm(TINFO, "bind('%p') failed with %s, try %2i...",
>> + address, tst_strerrno(err), i+1);
> Hmm, why do we do tst_strerno(err) here when
> the only value for errno can be EADDRINUSE
> (since we exit with tst_brkm() otherwise)
> Why not just hardcode EADDRINUSE to the
> message?
Also, we shouldn't print the message every second while waiting, could be:
if ((i + 1) % 10 == 0)
tst_resm(TINFO, "address is in use, waited %3i sec", i + 1);
Best regards,
Alexey
------------------------------------------------------------------------------
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list
