Skip to contentSkip to conversation

Fix dubious bitwise and/or usage.

PATCH In review Alexey Zaytsev 3 participants 11 messages Jan 9, 2009 – Jan 10, 2009

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

cover Fix dubious bitwise and/or usage.

Hello.

Sorry to bother the lkml with such trivial stuff, but I could not get any response from either Jiri Kosina ot trivial@ when I sent the patches directly to them. Maybe a linux-trivial@ mailing list would be a good idea..

These patches fix a number of cases where &/| were by mistake used instead of &&/||. In every single case the semantics stays the same, so none is a real bug.

---

Alexey Zaytsev (5):
Fix dubious bitwise 'and' usage spotted by sparse. Fix dubious bitwise 'and' usage spotted by sparse. Fix dubious bitwise 'and' usage spotted by sparse. Fix dubious bitwise 'or' usage spotted by sparse. Fix dubious bitwise 'or' usage spotted by sparse.

 drivers/isdn/mISDN/l1oip_codec.c               |    2 +-
 drivers/media/video/gspca/m5602/m5602_s5k4aa.c |    2 +-
 drivers/net/wireless/ath9k/rc.c                |    2 +-
 drivers/usb/wusbcore/security.c                |    2 +-
 mm/page_alloc.c                                |    2 +-
 5 files changed, 5 insertions(+), 5 deletions(-)

--

#1 lore raw

PATCH 1/5 Fix dubious bitwise 'or' usage spotted by sparse.

It doesn't change the semantics, but it looks like the logical 'or' was meant to be used here.

---
 mm/page_alloc.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
mm/page_alloc.c +1 -1
index d8ac014..6923237 100644
@@ -306,7 +306,7 @@ static void destroy_compound_page(struct page *page, unsigned long order)
 	for (i = 1; i < nr_pages; i++) {
 		struct page *p = page + i;
 
-		if (unlikely(!PageTail(p) |
+		if (unlikely(!PageTail(p) ||
 				(p->first_page != page)))
 			bad_page(page);
 		__ClearPageTail(p);
#2 lore raw

PATCH 2/5 Fix dubious bitwise 'or' usage spotted by sparse.

In reply to Alexey Zaytsev

It doesn't change the semantics, but it looks like the logical 'or' was meant to be used here.

---
 drivers/isdn/mISDN/l1oip_codec.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/isdn/mISDN/l1oip_codec.c +1 -1
index a2dc457..43b3795 100644
@@ -330,7 +330,7 @@ l1oip_4bit_alloc(int ulaw)
 	/* alloc conversion tables */
 	table_com = vmalloc(65536);
 	table_dec = vmalloc(512);
-	if (!table_com | !table_dec) {
+	if (!table_com || !table_dec) {
 		l1oip_4bit_free();
 		return -ENOMEM;
 	}
#3 lore raw

PATCH 3/5 Fix dubious bitwise 'and' usage spotted by sparse.

In reply to Alexey Zaytsev

It doesn't change the semantics, but it looks like the logical 'and' was meant to be used here.

---
 drivers/media/video/gspca/m5602/m5602_s5k4aa.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/media/video/gspca/m5602/m5602_s5k4aa.c +1 -1
index 14b1eac..c3ebcca 100644
@@ -114,7 +114,7 @@ int s5k4aa_read_sensor(struct sd *sd, const u8 address,
 	if (err < 0)
 		goto out;
 
-	for (i = 0; (i < len) & !err; i++) {
+	for (i = 0; (i < len) && !err; i++) {
 		err = m5602_read_bridge(sd, M5602_XB_I2C_DATA, &(i2c_data[i]));
 
 		PDEBUG(D_CONF, "Reading sensor register "
#4 lore raw

Re: [PATCH 3/5] Fix dubious bitwise 'and' usage spotted by sparse.

Quoted 11 lines · Alexey Zaytsev

It doesn't change the semantics, but it looks like the logical 'and' was meant to be used here.

---
 drivers/media/video/gspca/m5602/m5602_s5k4aa.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/media/video/gspca/m5602/m5602_s5k4aa.c +0 -0
index 14b1eac..c3ebcca 100644

drivers/media/video/gspca/m5602/m5602_s5k4aa.c int s5k4aa_read_sensor(struct sd *sd, const u8 address, · line 120

Quoted 8 lines
drivers/media/video/gspca/m5602/m5602_s5k4aa.c +1 -1
@@ -114,7 +114,7 @@ int s5k4aa_read_sensor(struct sd *sd, const u8 address,
 	if (err < 0)
 		goto out;
 
-	for (i = 0; (i < len) & !err; i++) {
+	for (i = 0; (i < len) && !err; i++) {
 		err = m5602_read_bridge(sd, M5602_XB_I2C_DATA, &(i2c_data[i]));
 
 		PDEBUG(D_CONF, "Reading sensor register "

Hi Alexey,

this one doesn't apply to Linus' tree anymore, the whole s5k4aa_read_sensor() function is not there.

Signature · 2 lines
Jiri Kosina
SUSE Labs
#5 lore raw

PATCH 4/5 Fix dubious bitwise 'and' usage spotted by sparse.

In reply to Alexey Zaytsev

It doesn't change the semantics, but it looks like the logical 'and' was meant to be used here.

---
 drivers/net/wireless/ath9k/rc.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/net/wireless/ath9k/rc.c +1 -1
index cca2fc5..e8eb20d 100644
@@ -609,7 +609,7 @@ ath_rc_get_nextvalid_txrate(const struct ath_rate_table *rate_table,
 
 static int ath_rc_valid_phyrate(u32 phy, u32 capflag, int ignore_cw)
 {
-	if (WLAN_RC_PHY_HT(phy) & !(capflag & WLAN_RC_HT_FLAG))
+	if (WLAN_RC_PHY_HT(phy) && !(capflag & WLAN_RC_HT_FLAG))
 		return FALSE;
 	if (WLAN_RC_PHY_DS(phy) && !(capflag & WLAN_RC_DS_FLAG))
 		return FALSE;
#6 lore raw

Re: [PATCH 4/5] Fix dubious bitwise 'and' usage spotted by sparse.

Quoted 13 lines · Alexey Zaytsev

It doesn't change the semantics, but it looks like the logical 'and' was meant to be used here.

---
 drivers/net/wireless/ath9k/rc.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/net/wireless/ath9k/rc.c +0 -0
index cca2fc5..e8eb20d 100644

drivers/net/wireless/ath9k/rc.c ath_rc_get_nextvalid_txrate(const struct ath_rate_table *rate_table, · line 615

Quoted 8 lines
drivers/net/wireless/ath9k/rc.c +1 -1
@@ -609,7 +609,7 @@ ath_rc_get_nextvalid_txrate(const struct ath_rate_table *rate_table,
 
 static int ath_rc_valid_phyrate(u32 phy, u32 capflag, int ignore_cw)
 {
-	if (WLAN_RC_PHY_HT(phy) & !(capflag & WLAN_RC_HT_FLAG))
+	if (WLAN_RC_PHY_HT(phy) && !(capflag & WLAN_RC_HT_FLAG))
 		return FALSE;
 	if (WLAN_RC_PHY_DS(phy) && !(capflag & WLAN_RC_DS_FLAG))
 		return FALSE;

This one needs a small fixup, as the buggy code got moved around to ath_rc_valid_phyrate().

Signature · 2 lines
Jiri Kosina
SUSE Labs
#7 lore raw

PATCH 5/5 Fix dubious bitwise 'and' usage spotted by sparse.

In reply to Alexey Zaytsev

It doesn't change the semantics, but clearly
the logical 'and' was meant to be used here.

---
 drivers/usb/wusbcore/security.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/usb/wusbcore/security.c +1 -1
index a101cad..8f953ab 100644
@@ -626,7 +626,7 @@ void wusbhc_gtk_rekey(struct wusbhc *wusbhc)
 		struct wusb_dev *wusb_dev;
 
 		wusb_dev = wusbhc->port[p].wusb_dev;
-		if (!wusb_dev || !wusb_dev->usb_dev | !wusb_dev->usb_dev->authenticated)
+		if (!wusb_dev || !wusb_dev->usb_dev || !wusb_dev->usb_dev->authenticated)
 			continue;
 
 		usb_fill_control_urb(wusb_dev->set_gtk_urb, wusb_dev->usb_dev,
#8 lore raw

Re: [PATCH 5/5] Fix dubious bitwise 'and' usage spotted by sparse.

2009/1/9 Alexey Zaytsev <[email protected]>:

Quoted 14 lines

It doesn't change the semantics, but clearly
the logical 'and' was meant to be used here.

---
 drivers/usb/wusbcore/security.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/usb/wusbcore/security.c +0 -0
index a101cad..8f953ab 100644

drivers/usb/wusbcore/security.c void wusbhc_gtk_rekey(struct wusbhc *wusbhc) · line 632

Quoted 8 lines
drivers/usb/wusbcore/security.c +1 -1
@@ -626,7 +626,7 @@ void wusbhc_gtk_rekey(struct wusbhc *wusbhc)
               struct wusb_dev *wusb_dev;
 
               wusb_dev = wusbhc->port[p].wusb_dev;
-               if (!wusb_dev || !wusb_dev->usb_dev | !wusb_dev->usb_dev->authenticated)
+               if (!wusb_dev || !wusb_dev->usb_dev || !wusb_dev->usb_dev->authenticated)
                       continue;
 
               usb_fill_control_urb(wusb_dev->set_gtk_urb, wusb_dev->usb_dev,

Hi Alexey,

This one is a logical 'or'... your changelog says 'and'!

Signature · 3 lines
Tom Spink
Douglas William Jerrold  - "The only athletic sport I ever mastered
was backgammon."
#9 lore raw
In reply to Alexey Zaytsev
Quoted 4 lines · Alexey Zaytsev

Sorry to bother the lkml with such trivial stuff, but I could not get any response from either Jiri Kosina ot trivial@ when I sent the patches directly to them. Maybe a linux-trivial@ mailing list would be a good idea..

Sorry, I have just been busy with too many other things during the merge window.

Quoted 8 lines

---

Alexey Zaytsev (5):
Fix dubious bitwise 'and' usage spotted by sparse. Fix dubious bitwise 'and' usage spotted by sparse. Fix dubious bitwise 'and' usage spotted by sparse. Fix dubious bitwise 'or' usage spotted by sparse. Fix dubious bitwise 'or' usage spotted by sparse.

I have applied all the patches with exception for the one patching drivers/media/video/gspca/m5602/m5602_s5k4aa.c, as it doesn't apply any more.

Thanks,

Signature · 2 lines
Jiri Kosina
SUSE Labs
#10 lore raw
Quoted 9 lines · Jiri Kosina
Quoted 4 lines · Alexey Zaytsev

Sorry to bother the lkml with such trivial stuff, but I could not get any response from either Jiri Kosina ot trivial@ when I sent the patches directly to them. Maybe a linux-trivial@ mailing list would be a good idea..

Sorry, I have just been busy with too many other things during the merge window.

Sorry for the disturbance then. I was not sure if you received the patches.

Quoted 10 lines
Quoted 9 lines

---

Alexey Zaytsev (5):
Fix dubious bitwise 'and' usage spotted by sparse. Fix dubious bitwise 'and' usage spotted by sparse. Fix dubious bitwise 'and' usage spotted by sparse. Fix dubious bitwise 'or' usage spotted by sparse. Fix dubious bitwise 'or' usage spotted by sparse.

Quoted 3 lines

I have applied all the patches with exception for the one patching drivers/media/video/gspca/m5602/m5602_s5k4aa.c, as it doesn't apply any more.

Thank you. The bugs seem to be leaving by themselves. ;)