[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: ssh_select possible improvement?
[Thread Prev] | [Thread Next]
- Subject: Re: ssh_select possible improvement?
- From: Norbert Kiesel <nkiesel@xxxxxxxxxxxxxxx>
- Reply-to: libssh@xxxxxxxxxx
- Date: Mon, 20 Sep 2010 09:36:18 -0700
- To: libssh@xxxxxxxxxx
On Tue, 2010-09-21 at 00:20 +0800, Vic Lee wrote: > Hi Andreas, > > I am able to identity this is the exact commit that introduce the > ssh_select() bug: > > http://git.libssh.org/projects/libssh/libssh.git/commit/?h=v0-4&id=22d975a24b69040b28396545e046d2334dd4c386 The only part of the patches that looks fishy is the last one: FD_SET(s->fd,set); - if (s->fd >= *fd_max) { - *fd_max = s->fd + 1; + if (s->fd >= 0 && s->fd != SSH_INVALID_SOCKET) { + *max_fd = s->fd + 1; } } Unless I'm missing something, in non-error-cases *max_fd is now always set to s->fd + 1 whereas this was only done it it's > *max_fd. Or more concretely, shouldn't the last condition read: s->fd >= 0 && s->fd != SSH_INVALID_SOCKET && s->fd >= *max_fd </nk>
ssh_select possible improvement? | Vic Lee <llyzs@xxxxxxx> |
Re: ssh_select possible improvement? | Andreas Schneider <asn@xxxxxxxxxxxx> |
Re: ssh_select possible improvement? | Vic Lee <llyzs@xxxxxxx> |
Re: ssh_select possible improvement? | Vic Lee <llyzs@xxxxxxx> |