Re: [PATCH 01/10] staging: ks7010: use linux circular buffer header macros to handle tx and rx queues
On Mon, Apr 23, 2018 at 3:19 PM, Greg KH wrote: > On Mon, Apr 23, 2018 at 03:02:20PM +0200, Sergio Paracuellos wrote: >> On Mon, Apr 23, 2018 at 2:30 PM, Greg KH wrote: >> > On Mon, Apr 16, 2018 at 12:29:23PM +0200, Sergio Paracuellos wrote: >> >> This commit replace current custom implementation of some circular >> >> buffer head and tail logic in favour of the use of macros defined >> >> in linux circ_buf.h header. Queue related inline function names >> >> have been review also. >> >> >> >> Signed-off-by: Sergio Paracuellos >> >> --- >> >> drivers/staging/ks7010/ks7010_sdio.c | 49 >> >> ++-- >> >> 1 file changed, 30 insertions(+), 19 deletions(-) >> > >> > Not all patches in this series would apply. Can you please rebase and >> > resend the remaining ones? >> >> I will. I think previous one to this series is not being applied at >> all. Is it posible? The one with cover-letter subject "cleanups >> continue" which contains 14 patches. It has a mail thread with Dan >> about circular buffer stuff to fix and no resend the series because is >> the wrong one is the last patch. > > I have no idea, sorry, all of these patches are long-gone from my patch > queue. If I haven't applied something, rebase and resend. I'm dealing > with over a thousand staging driver patches right now, I can't stop and > "think" about what I already reviewed :) ack! II'll resend those too with the remaining ones which problems to apply in all the series. > > thanks, > > greg k-h Thanks, Sergio Paracuellos ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
Re: [PATCH 01/10] staging: ks7010: use linux circular buffer header macros to handle tx and rx queues
On Mon, Apr 23, 2018 at 03:02:20PM +0200, Sergio Paracuellos wrote: > On Mon, Apr 23, 2018 at 2:30 PM, Greg KH wrote: > > On Mon, Apr 16, 2018 at 12:29:23PM +0200, Sergio Paracuellos wrote: > >> This commit replace current custom implementation of some circular > >> buffer head and tail logic in favour of the use of macros defined > >> in linux circ_buf.h header. Queue related inline function names > >> have been review also. > >> > >> Signed-off-by: Sergio Paracuellos > >> --- > >> drivers/staging/ks7010/ks7010_sdio.c | 49 > >> ++-- > >> 1 file changed, 30 insertions(+), 19 deletions(-) > > > > Not all patches in this series would apply. Can you please rebase and > > resend the remaining ones? > > I will. I think previous one to this series is not being applied at > all. Is it posible? The one with cover-letter subject "cleanups > continue" which contains 14 patches. It has a mail thread with Dan > about circular buffer stuff to fix and no resend the series because is > the wrong one is the last patch. I have no idea, sorry, all of these patches are long-gone from my patch queue. If I haven't applied something, rebase and resend. I'm dealing with over a thousand staging driver patches right now, I can't stop and "think" about what I already reviewed :) thanks, greg k-h ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
Re: [PATCH 01/10] staging: ks7010: use linux circular buffer header macros to handle tx and rx queues
On Mon, Apr 23, 2018 at 2:30 PM, Greg KH wrote: > On Mon, Apr 16, 2018 at 12:29:23PM +0200, Sergio Paracuellos wrote: >> This commit replace current custom implementation of some circular >> buffer head and tail logic in favour of the use of macros defined >> in linux circ_buf.h header. Queue related inline function names >> have been review also. >> >> Signed-off-by: Sergio Paracuellos >> --- >> drivers/staging/ks7010/ks7010_sdio.c | 49 >> ++-- >> 1 file changed, 30 insertions(+), 19 deletions(-) > > Not all patches in this series would apply. Can you please rebase and > resend the remaining ones? I will. I think previous one to this series is not being applied at all. Is it posible? The one with cover-letter subject "cleanups continue" which contains 14 patches. It has a mail thread with Dan about circular buffer stuff to fix and no resend the series because is the wrong one is the last patch. > > thanks, > > greg k-h Thanks,, Sergio Paracuellos ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
Re: [PATCH 01/10] staging: ks7010: use linux circular buffer header macros to handle tx and rx queues
On Mon, Apr 16, 2018 at 12:29:23PM +0200, Sergio Paracuellos wrote: > This commit replace current custom implementation of some circular > buffer head and tail logic in favour of the use of macros defined > in linux circ_buf.h header. Queue related inline function names > have been review also. > > Signed-off-by: Sergio Paracuellos > --- > drivers/staging/ks7010/ks7010_sdio.c | 49 > ++-- > 1 file changed, 30 insertions(+), 19 deletions(-) Not all patches in this series would apply. Can you please rebase and resend the remaining ones? thanks, greg k-h ___ devel mailing list de...@linuxdriverproject.org http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
[PATCH 01/10] staging: ks7010: use linux circular buffer header macros to handle tx and rx queues
This commit replace current custom implementation of some circular
buffer head and tail logic in favour of the use of macros defined
in linux circ_buf.h header. Queue related inline function names
have been review also.
Signed-off-by: Sergio Paracuellos
---
drivers/staging/ks7010/ks7010_sdio.c | 49 ++--
1 file changed, 30 insertions(+), 19 deletions(-)
diff --git a/drivers/staging/ks7010/ks7010_sdio.c
b/drivers/staging/ks7010/ks7010_sdio.c
index 9c591e0..fedcbb3 100644
--- a/drivers/staging/ks7010/ks7010_sdio.c
+++ b/drivers/staging/ks7010/ks7010_sdio.c
@@ -10,6 +10,7 @@
* published by the Free Software Foundation.
*/
+#include
#include
#include
#include
@@ -111,11 +112,10 @@ static inline void inc_txqtail(struct ks_wlan_private
*priv)
priv->tx_dev.qtail = (priv->tx_dev.qtail + 1) % TX_DEVICE_BUFF_SIZE;
}
-static inline unsigned int cnt_txqbody(struct ks_wlan_private *priv)
+static inline bool txq_has_space(struct ks_wlan_private *priv)
{
- unsigned int tx_cnt = priv->tx_dev.qtail - priv->tx_dev.qhead;
-
- return (tx_cnt + TX_DEVICE_BUFF_SIZE) % TX_DEVICE_BUFF_SIZE;
+ return (CIRC_SPACE(priv->tx_dev.qhead, priv->tx_dev.qtail,
+ TX_DEVICE_BUFF_SIZE) > 0);
}
static inline void inc_rxqhead(struct ks_wlan_private *priv)
@@ -128,11 +128,22 @@ static inline void inc_rxqtail(struct ks_wlan_private
*priv)
priv->rx_dev.qtail = (priv->rx_dev.qtail + 1) % RX_DEVICE_BUFF_SIZE;
}
-static inline unsigned int cnt_rxqbody(struct ks_wlan_private *priv)
+static inline bool rxq_has_space(struct ks_wlan_private *priv)
+{
+ return (CIRC_SPACE(priv->rx_dev.qhead, priv->rx_dev.qtail,
+ RX_DEVICE_BUFF_SIZE) > 0);
+}
+
+static inline unsigned int txq_count(struct ks_wlan_private *priv)
{
- unsigned int rx_cnt = priv->rx_dev.qtail - priv->rx_dev.qhead;
+ return CIRC_CNT_TO_END(priv->tx_dev.qhead, priv->tx_dev.qtail,
+ TX_DEVICE_BUFF_SIZE);
+}
- return (rx_cnt + RX_DEVICE_BUFF_SIZE) % RX_DEVICE_BUFF_SIZE;
+static inline unsigned int rxq_count(struct ks_wlan_private *priv)
+{
+ return CIRC_CNT_TO_END(priv->rx_dev.qhead, priv->rx_dev.qtail,
+ RX_DEVICE_BUFF_SIZE);
}
/* Read single byte from device address into byte (CMD52) */
@@ -258,11 +269,11 @@ static void _ks_wlan_hw_power_save(struct ks_wlan_private
*priv)
atomic_read(&priv->psstatus.status),
atomic_read(&priv->psstatus.confirm_wait),
atomic_read(&priv->psstatus.snooze_guard),
- cnt_txqbody(priv));
+ txq_count(priv));
if (atomic_read(&priv->psstatus.confirm_wait) ||
atomic_read(&priv->psstatus.snooze_guard) ||
- cnt_txqbody(priv)) {
+ txq_has_space(priv)) {
queue_delayed_work(priv->wq, &priv->rw_dwork, 0);
return;
}
@@ -308,7 +319,7 @@ static int enqueue_txdev(struct ks_wlan_private *priv,
unsigned char *p,
goto err_complete;
}
- if ((TX_DEVICE_BUFF_SIZE - 1) <= cnt_txqbody(priv)) {
+ if ((TX_DEVICE_BUFF_SIZE - 1) <= txq_count(priv)) {
netdev_err(priv->net_dev, "tx buffer overflow\n");
ret = -EOVERFLOW;
goto err_complete;
@@ -366,7 +377,7 @@ static void tx_device_task(struct ks_wlan_private *priv)
struct tx_device_buffer *sp;
int ret;
- if (cnt_txqbody(priv) <= 0 ||
+ if (!txq_has_space(priv) ||
atomic_read(&priv->psstatus.status) == PS_SNOOZE)
return;
@@ -385,7 +396,7 @@ static void tx_device_task(struct ks_wlan_private *priv)
(*sp->complete_handler)(priv, sp->skb);
inc_txqhead(priv);
- if (cnt_txqbody(priv) > 0)
+ if (txq_has_space(priv))
queue_delayed_work(priv->wq, &priv->rw_dwork, 0);
}
@@ -413,7 +424,7 @@ int ks_wlan_hw_tx(struct ks_wlan_private *priv, void *p,
unsigned long size,
result = enqueue_txdev(priv, p, size, complete_handler, skb);
spin_unlock(&priv->tx_dev.tx_dev_lock);
- if (cnt_txqbody(priv) > 0)
+ if (txq_has_space(priv))
queue_delayed_work(priv->wq, &priv->rw_dwork, 0);
return result;
@@ -424,12 +435,12 @@ static void rx_event_task(unsigned long dev)
struct ks_wlan_private *priv = (struct ks_wlan_private *)dev;
struct rx_device_buffer *rp;
- if (cnt_rxqbody(priv) > 0 && priv->dev_state >= DEVICE_STATE_BOOT) {
+ if (rxq_has_space(priv) && priv->dev_state >= DEVICE_STATE_BOOT) {
rp = &priv->rx_dev.rx_dev_buff[priv->rx_dev.qhead];
hostif_receive(priv, rp->data, rp->size);
inc_rxqhead(priv);
- if (cnt_rxqbody(priv) > 0)
+ if (rxq_has_space(priv))
tasklet_schedule(&pr
