On Sun, May 31, 2015 at 8:40 AM, Reyk Floeter <[email protected]> wrote:
> On Sat, May 30, 2015 at 11:18:59PM -0400, sven falempin wrote: > > The patch is in a turned off machine, > > after doing the patch > > i contacted directly reyk for the matter, > > i d like to push the issue further by not accepting > > the socket while the connect didnt succeed when possible. > > > > (The pastebin expire) > > > > Please be a bit patient. > > > > Yes, I saw your email but you didn't send me a diff either. It was > just an email explaining a concept. I have to admit that I'm > sceptical about it as it involves a new and very special syscall. > > But maybe it is better if you share your diff directly on tech@. > > Reyk > > > > > On Sat, May 30, 2015 at 12:35 PM, Theo de Raadt <[email protected] > > > > wrote: > > > > > > it is possible to remove the inflight var by opening the socket > before > > > > accepting. > > > > Thus there is no need for guessing if the peer socket can be open, > > > > > > Where is your diff to demonstrate this? > > > > > > Show your work. (Or we'll have to make an assumption). > > > > > > > The diff is here: http://pastebin.com/SuVRapZy I cut to deep in the code and remove the FD_RESERVE, i didnt fix this for a few raeson: - allocation is done before accept, which is not good - I d like to fully use divert-to and not accepting the pending connection before connecting the peer, 2 way are possible * accessing the socket info before acepting * an accept4 that does not send a ACK and a new function to send it The last would make an <invisible proxy> (tm)(copyright). I wont be available for a few days. note that the former code was ignoring error in <preaccept> and that preaccept serves no obvious purpose: My religion is against non web client for mail, so pastebin for \t to be kept. --------------------------------------------------------------------------------------------------------------------- () ascii ribbon campaign - against html e-mail /\
