Matthew Dillon wrote: > The issue with dup2() was a race against open() or close() > I believe, where dup2() could potentially dup into a > descriptor that open() was about to use. Unfortunately, it > does appear that dup() has the same issue. > > fdalloc() does not reserve the descriptor number it > returns, it simply finds a free slot and says 'this > index is a free slot'. Even in the latest -current, > fdalloc() releases the fdp lock when it goes to > MALLOC so the race appears to still be present.
The correct way to deal with this is to make the descriptor slot allocation happen at a higher layer, and pass the address of it in as an agrument to the code. This will incidently let you pass in a static or other non-FD table associated descriptor in place to all these calls, which will have the effect of permitting you to do "descriptor" based file I/O from kernel space much, much easier. If the work would be accepted as a delta against -stable, I would be willing to do it, since I need kernel file I/O, and the problem hasn't been solving itself for about six years now. -- Terry To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
