| Message ID | 20190910063608.10081-1-xiubli@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | nbd: remove the duplicated code | expand |
On 09/10/2019 01:36 AM, xiubli@redhat.com wrote: > From: Xiubo Li <xiubli@redhat.com> > > The followed code will do the same check, and this part is redandant. > > Signed-off-by: Xiubo Li <xiubli@redhat.com> > --- > drivers/block/nbd.c | 3 --- > 1 file changed, 3 deletions(-) > > diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c > index 478aa86fc1f2..8c10ab51a086 100644 > --- a/drivers/block/nbd.c > +++ b/drivers/block/nbd.c > @@ -1046,9 +1046,6 @@ static int nbd_reconnect_socket(struct nbd_device *nbd, unsigned long arg) > for (i = 0; i < config->num_connections; i++) { > struct nbd_sock *nsock = config->socks[i]; > > - if (!nsock->dead) > - continue; > - Was this check to used to speed up reconnects? For example, if a send was stuck waiting on socket memory to free up in the network layer, then the above check would allow us to skip past those sockets without having to wait on the socket that was trying to send. > mutex_lock(&nsock->tx_lock); > if (!nsock->dead) { > mutex_unlock(&nsock->tx_lock); >
On 2019/9/10 23:56, Mike Christie wrote: > On 09/10/2019 01:36 AM, xiubli@redhat.com wrote: >> From: Xiubo Li <xiubli@redhat.com> >> >> The followed code will do the same check, and this part is redandant. >> >> Signed-off-by: Xiubo Li <xiubli@redhat.com> >> --- >> drivers/block/nbd.c | 3 --- >> 1 file changed, 3 deletions(-) >> >> diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c >> index 478aa86fc1f2..8c10ab51a086 100644 >> --- a/drivers/block/nbd.c >> +++ b/drivers/block/nbd.c >> @@ -1046,9 +1046,6 @@ static int nbd_reconnect_socket(struct nbd_device *nbd, unsigned long arg) >> for (i = 0; i < config->num_connections; i++) { >> struct nbd_sock *nsock = config->socks[i]; >> >> - if (!nsock->dead) >> - continue; >> - > Was this check to used to speed up reconnects? For example, if a send > was stuck waiting on socket memory to free up in the network layer, then > the above check would allow us to skip past those sockets without having > to wait on the socket that was trying to send. > Yeah, in this case it really could help a little bit. Or maybe in the network layer when allocating new memories but it needs to do the memory reclaim, which will be stuck for a long time ? Thanks BRs Xiubo > >> mutex_lock(&nsock->tx_lock); >> if (!nsock->dead) { >> mutex_unlock(&nsock->tx_lock); >>
diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c index 478aa86fc1f2..8c10ab51a086 100644 --- a/drivers/block/nbd.c +++ b/drivers/block/nbd.c @@ -1046,9 +1046,6 @@ static int nbd_reconnect_socket(struct nbd_device *nbd, unsigned long arg) for (i = 0; i < config->num_connections; i++) { struct nbd_sock *nsock = config->socks[i]; - if (!nsock->dead) - continue; - mutex_lock(&nsock->tx_lock); if (!nsock->dead) { mutex_unlock(&nsock->tx_lock);