Re: [RFC] readdir mess

[Linus Torvalds]

On Tue, 12 Aug 2008, Al Viro wrote:
> 
> you've just lost e.g. -EIO for getdents().  And if you bail out on
> non-zero return value from vfs_readdir(), you are back to -EINVAL
> on full buffer.

Btw, this whole sentence, and the one from your next email seems to really 
show a fundamental misunderstanding of the whole readdir() error handling:

> PS: we might get away with both, if we used _positive_ values as well.
> E.g. have getdents() filldir return 1 if we are out of buffer *and*
> have ->previous != NULL (and -EINVAL if we are out of buffer on the
> first call)...  And have some other positive constant for "->readdir()
> didn't feel like going all the way to the end of directory".

We *must* handle partial returns by returning "success". And we do, 
except for our _confusion_ about ->readdir() returning error and that 
somehow "overriding" the fact that it already returned non-errors earlier 
through the callback.

All your blathering about "positive values as well" seems to ttoally 
misunderstand how readdir() works. We absolutely do *not* need positive 
return values, because the fact is, the only positive return value we ever 
need is the "we already filled _earlier_ buffers". And that's the one 
that we already do.

The fact is, NO ERROR VALUE CAN POSSIBLY MATTER if we already returned one 
or more entries to getdents/readdir(). We should return a success value.

			Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/