On Mon, Jun 14, 2021 at 02:26:48PM +0200, Peter Krempa wrote: > On Mon, Jun 14, 2021 at 14:14:47 +0200, Michal Prívozník wrote: > > On 6/14/21 1:31 PM, Tim Wiederhake wrote: > > > On Mon, 2021-06-14 at 13:06 +0200, Michal Privoznik wrote: > > >> In a few occasions in tests we pass INT_MAX to > > >> virFileReadLimFD(). This is not safe because virFileReadAll() > > >> will call virFileReadLimFD() under the hood which takes the limit > > >> and adds 1 to it. > > > > > > Calling virFileReadAll with "INT_MAX - 1" looks funny. Is it possible > > > to check for "maxlen >= INT_MAX" in virFileReadLimFD instead? > > > > Actually, I don't understand why we need to add 1 in the first place. > > I'll push the other two patches and send v2 for this that removes the +1. > > It's so that it guarantees that a file of 'maxlen' length is read > completely and the terminating '\0' is in the resulting string. > > Removing the '+ 1' would change this kind of semantics, which may > require audit of all callers. Isnt it just a matter of semantics of 'maxlen'. Your description is saying the semantics for 'maxlen' are the total length of file plus a possible trailing null. Can we just define 'maxlen' to mean the total length of the file, not including an extra trailing null ? ie so that 'maxlen' is essentially equal to strlen(buf) in the case where the file has no embedded nuls Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
