Re: linux-next: manual merge of the block tree with the vfs-idmapping tree

[Jens Axboe]

On 5/9/23 6:55 PM, Stephen Rothwell wrote:
> Hi all,
> 
> Today's linux-next merge of the block tree got a conflict in:
> 
>   fs/pipe.c
> 
> between commit:
> 
>   2b10649c2316 ("pipe: enable handling of IOCB_NOWAIT")
> 
> from the vfs-idmapping tree and commit:
> 
>   3f6ded8dd89d ("pipe: check for IOCB_NOWAIT alongside O_NONBLOCK")
> 
> from the block tree.
> 
> The former added
> 
> 	const bool nonblock = iocb->ki_flags & IOCB_NOWAIT;
> 
> and then did
> 
> -		if (filp->f_flags & O_NONBLOCK) {
> +		if (filp->f_flags & O_NONBLOCK || nonblock) {
> 
> while the latter just did
> 
> -		if (filp->f_flags & O_NONBLOCK) {
> +		if (filp->f_flags & O_NONBLOCK || iocb->ki_flags & IOCB_NOWAIT) {
> 
> so I just used the former though I suspect that the former may be a
> previous version of these changes?).
> 
> I fixed it up (see above) and can carry the fix as necessary. This
> is now fixed as far as linux-next is concerned, but any non trivial
> conflicts should be mentioned to your upstream maintainer when your tree
> is submitted for merging.  You may also want to consider cooperating
> with the maintainer of the conflicting tree to minimise any particularly
> complex conflicts.

Yes, since Linus didn't like the original approach, it was redone
and that particular patch was originally forgotten and then redone
without the 'nonblock' variable. So you should just ignore the
old version, and Christian should just drop that branch from his
for-next tree as it's dead.

-- 
Jens Axboe