[Ocfs2-devel] [PATCH] ocfs2: Use memalloc_nofs_save instead of memalloc_noio_save
Joseph Qi
joseph.qi at linux.alibaba.com
Thu Mar 26 18:28:35 PDT 2020
On 2020/3/27 04:02, Matthew Wilcox wrote:
> From: "Matthew Wilcox (Oracle)" <willy at infradead.org>
>
> OCFS2 doesn't mind if memory reclaim makes I/Os happen; it just cares
> that it won't be reentered, so it can use memalloc_nofs_save() instead
> of memalloc_noio_save().
>
> Signed-off-by: Matthew Wilcox (Oracle) <willy at infradead.org>
Reviewed-by: Joseph Qi <joseph.qi at linux.alibaba.com>
> ---
> fs/ocfs2/cluster/tcp.c | 24 ++++++++++--------------
> 1 file changed, 10 insertions(+), 14 deletions(-)
>
> diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
> index 9261c1f06a9f..2c512b40a940 100644
> --- a/fs/ocfs2/cluster/tcp.c
> +++ b/fs/ocfs2/cluster/tcp.c
> @@ -1570,15 +1570,13 @@ static void o2net_start_connect(struct work_struct *work)
> struct sockaddr_in myaddr = {0, }, remoteaddr = {0, };
> int ret = 0, stop;
> unsigned int timeout;
> - unsigned int noio_flag;
> + unsigned int nofs_flag;
>
> /*
> - * sock_create allocates the sock with GFP_KERNEL. We must set
> - * per-process flag PF_MEMALLOC_NOIO so that all allocations done
> - * by this process are done as if GFP_NOIO was specified. So we
> - * are not reentering filesystem while doing memory reclaim.
> + * sock_create allocates the sock with GFP_KERNEL. We must
> + * prevent the filesystem from being reentered by memory reclaim.
> */
> - noio_flag = memalloc_noio_save();
> + nofs_flag = memalloc_nofs_save();
> /* if we're greater we initiate tx, otherwise we accept */
> if (o2nm_this_node() <= o2net_num_from_nn(nn))
> goto out;
> @@ -1683,7 +1681,7 @@ static void o2net_start_connect(struct work_struct *work)
> if (mynode)
> o2nm_node_put(mynode);
>
> - memalloc_noio_restore(noio_flag);
> + memalloc_nofs_restore(nofs_flag);
> return;
> }
>
> @@ -1810,15 +1808,13 @@ static int o2net_accept_one(struct socket *sock, int *more)
> struct o2nm_node *local_node = NULL;
> struct o2net_sock_container *sc = NULL;
> struct o2net_node *nn;
> - unsigned int noio_flag;
> + unsigned int nofs_flag;
>
> /*
> - * sock_create_lite allocates the sock with GFP_KERNEL. We must set
> - * per-process flag PF_MEMALLOC_NOIO so that all allocations done
> - * by this process are done as if GFP_NOIO was specified. So we
> - * are not reentering filesystem while doing memory reclaim.
> + * sock_create_lite allocates the sock with GFP_KERNEL. We must
> + * prevent the filesystem from being reentered by memory reclaim.
> */
> - noio_flag = memalloc_noio_save();
> + nofs_flag = memalloc_nofs_save();
>
> BUG_ON(sock == NULL);
> *more = 0;
> @@ -1934,7 +1930,7 @@ static int o2net_accept_one(struct socket *sock, int *more)
> if (sc)
> sc_put(sc);
>
> - memalloc_noio_restore(noio_flag);
> + memalloc_nofs_restore(nofs_flag);
> return ret;
> }
>
>
More information about the Ocfs2-devel
mailing list