> Subject: Re: [PATCH -next] pnfs: Use GFP_ATOMIC under spin lock
>
> On Wed, 2018-06-06 at 02:02 +0000, Wei Yongjun wrote:
> > A spin lock is taken here so we should use GFP_ATOMIC.
> >
> > Fixes: 2409a976a299 ("pnfs: Add LAYOUTGET to OPEN of a new file")
> > Signed-off-by: Wei Yongjun <weiyongjun1@xxxxxxxxxx>
> > ---
> > fs/nfs/pnfs.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
> > index d93942f..1ff1998 100644
> > --- a/fs/nfs/pnfs.c
> > +++ b/fs/nfs/pnfs.c
> > @@ -2009,7 +2009,7 @@ struct pnfs_layout_segment *
> > struct pnfs_layout_hdr *lo;
> >
> > spin_lock(&ino->i_lock);
> > - lo = pnfs_find_alloc_layout(ino, ctx, GFP_KERNEL);
> > + lo = pnfs_find_alloc_layout(ino, ctx, GFP_ATOMIC);
> > if (!lo)
> > goto out_unlock;
> > if (!test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags))
> >
> Hi Wei,
>
> The ino->i_lock spinlock is temporarily dropped inside
> pnfs_find_alloc_layout() if we have to perform an allocation, so the
> existing code should be correct.
Hi Trond,
Yes, you are right, thanks for review.
Regards
��.n��������+%������w��{.n����z�ޗ�����n�r������&��z�ޗ�zf���h���~����������_��+v���)ߣ�
