Skip to contentSkip to conversation

-stable review

PATCH In review Greg KH 6 participants 20 messages Mar 10, 2005 – Mar 11, 2005

Subsystems (7)
Networking Network drivers Media (V4L/DVB) PCI +3 more subsystems
lore
Share & tools
lore

Greg KH

author
#0 lore raw

cover -stable review

This is the start of the stable review cycle for the 2.6.11.3 release. There are 11 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let us know. If anyone is a maintainer of the proper subsystem, and wants to add a
signed-off-by: line to the patch, please respond with it.

These patches are sent out with a number of different people on the Bcc: line. If you wish to be a reviewer, please email [email protected] to add your name to the list. If you want to be off the reviewer list, also email us.

Responses should be made by Sat, March 12, 23:00 UTC. Anything received after that time, might be too late.

thanks,

the -stable release team (i.e. the ones wearing the joker hat in the corner...)

author
#1 lore raw

PATCH 1/11 fix amd64 2.6.11 oops on modprobe (saa7110)

-stable review patch. If anyone has any objections, please let us know.

------------------

This is a rewrite of the saa7110_write_block function, which was plain broken in the case where the underlying adapter supports I2C_FUNC_I2C. It also includes related fixes which ensure that different parts of the driver agree on the number of registers the chip has.

Signed-off-by Jean Delvare <[email protected]>
Signed-off-by Chris Wright <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>
drivers/media/video/saa7110.c renamed from drivers/media/video/saa7110.c.orig +17 -18
@@ -58,10 +58,12 @@
 #define SAA7110_MAX_INPUT	9	/* 6 CVBS, 3 SVHS */
 #define SAA7110_MAX_OUTPUT	0	/* its a decoder only */
 
-#define	I2C_SAA7110		0x9C	/* or 0x9E */
+#define I2C_SAA7110		0x9C	/* or 0x9E */
+
+#define SAA7110_NR_REG		0x35
 
 struct saa7110 {
-	unsigned char reg[54];
+	u8 reg[SAA7110_NR_REG];
 
 	int norm;
 	int input;

@@ -95,31 +97,28 @@
 		     unsigned int       len)
 {
 	int ret = -1;
-	u8 reg = *data++;
+	u8 reg = *data;		/* first register to write to */
 
-	len--;
+	/* Sanity check */
+	if (reg + (len - 1) > SAA7110_NR_REG)
+		return ret;
 
 	/* the saa7110 has an autoincrement function, use it if
 	 * the adapter understands raw I2C */
 	if (i2c_check_functionality(client->adapter, I2C_FUNC_I2C)) {
 		struct saa7110 *decoder = i2c_get_clientdata(client);
 		struct i2c_msg msg;
-		u8 block_data[54];
 
-		msg.len = 0;
-		msg.buf = (char *) block_data;
+		msg.len = len;
+		msg.buf = (char *) data;
 		msg.addr = client->addr;
-		msg.flags = client->flags;
-		while (len >= 1) {
-			msg.len = 0;
-			block_data[msg.len++] = reg;
-			while (len-- >= 1 && msg.len < 54)
-				block_data[msg.len++] =
-				    decoder->reg[reg++] = *data++;
-			ret = i2c_transfer(client->adapter, &msg, 1);
-		}
+		msg.flags = 0;
+		ret = i2c_transfer(client->adapter, &msg, 1);
+
+		/* Cache the written data */
+		memcpy(decoder->reg + reg, data + 1, len - 1);
 	} else {
-		while (len-- >= 1) {
+		for (++data, --len; len; len--) {
 			if ((ret = saa7110_write(client, reg++,
 						 *data++)) < 0)
 				break;

@@ -192,7 +191,7 @@
 	return 0;
 }
 
-static const unsigned char initseq[] = {
+static const unsigned char initseq[1 + SAA7110_NR_REG] = {
 	0, 0x4C, 0x3C, 0x0D, 0xEF, 0xBD, 0xF2, 0x03, 0x00,
 	/* 0x08 */ 0xF8, 0xF8, 0x60, 0x60, 0x00, 0x86, 0x18, 0x90,
 	/* 0x10 */ 0x00, 0x59, 0x40, 0x46, 0x42, 0x1A, 0xFF, 0xDA,
#2 lore raw

Re: [01/11] fix amd64 2.6.11 oops on modprobe (saa7110)

Quoted 16 lines · Greg KH

-stable review patch. If anyone has any objections, please let us know.

------------------

This is a rewrite of the saa7110_write_block function, which was plain broken in the case where the underlying adapter supports I2C_FUNC_I2C. It also includes related fixes which ensure that different parts of the driver agree on the number of registers the chip has.

Signed-off-by Jean Delvare <[email protected]>
Signed-off-by Chris Wright <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>

drivers/media/video/saa7110.c line 61

Quoted 5 lines
drivers/media/video/saa7110.c renamed from drivers/media/video/saa7110.c.orig +1 -1
@@ -58,10 +58,12 @@
 #define SAA7110_MAX_INPUT	9	/* 6 CVBS, 3 SVHS */
 #define SAA7110_MAX_OUTPUT	0	/* its a decoder only */
 
-#define	I2C_SAA7110		0x9C	/* or 0x9E */
+#define I2C_SAA7110		0x9C	/* or 0x9E */

Not that I really care, but isn't there a rule that a patch "... can not contain any "trivial" fixes in it (spelling changes, whitespace cleanups, etc.)"?

josh

author
#3 lore raw

Re: [01/11] fix amd64 2.6.11 oops on modprobe (saa7110)

Quoted 20 lines · Josh Boyer
Quoted 19 lines · Greg KH

-stable review patch. If anyone has any objections, please let us know.

------------------

This is a rewrite of the saa7110_write_block function, which was plain broken in the case where the underlying adapter supports I2C_FUNC_I2C. It also includes related fixes which ensure that different parts of the driver agree on the number of registers the chip has.

Signed-off-by Jean Delvare <[email protected]>
Signed-off-by Chris Wright <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>
drivers/media/video/saa7110.c renamed from drivers/media/video/saa7110.c.orig +0 -0
@@ -58,10 +58,12 @@
 #define SAA7110_MAX_INPUT	9	/* 6 CVBS, 3 SVHS */
 #define SAA7110_MAX_OUTPUT	0	/* its a decoder only */
Quoted 6 lines

drivers/media/video/saa7110.c line 61

Quoted 2 lines
drivers/media/video/saa7110.c renamed from drivers/media/video/saa7110.c.orig +1 -1
-#define	I2C_SAA7110		0x9C	/* or 0x9E */
+#define I2C_SAA7110		0x9C	/* or 0x9E */

Not that I really care, but isn't there a rule that a patch "... can not contain any "trivial" fixes in it (spelling changes, whitespace cleanups, etc.)"?

Good point. Jean, care to respin the patch?

thanks,

greg k-h

#4 lore raw

Re: [01/11] fix amd64 2.6.11 oops on modprobe (saa7110)

Hi Greg, all,

Quoted 5 lines
Quoted 3 lines

Not that I really care, but isn't there a rule that a patch "... can not contain any "trivial" fixes in it (spelling changes, whitespace cleanups, etc.)"?

Good point. Jean, care to respin the patch?

Sure, sorry for the trouble.

---

This is a rewrite of the saa7110_write_block function, which was plain broken in the case where the underlying adapter supports I2C_FUNC_I2C. It also includes related fixes which ensure that different parts of the driver agree on the number of registers the chip has.

Signed-off-by Jean Delvare <[email protected]>
Signed-off-by Chris Wright <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>
drivers/media/video/saa7110.c renamed from drivers/media/video/saa7110.c.orig +16 -17
@@ -60,8 +60,10 @@
 
 #define	I2C_SAA7110		0x9C	/* or 0x9E */
 
+#define SAA7110_NR_REG		0x35
+
 struct saa7110 {
-	unsigned char reg[54];
+	u8 reg[SAA7110_NR_REG];
 
 	int norm;
 	int input;

@@ -95,31 +97,28 @@
 		     unsigned int       len)
 {
 	int ret = -1;
-	u8 reg = *data++;
+	u8 reg = *data;		/* first register to write to */
 
-	len--;
+	/* Sanity check */
+	if (reg + (len - 1) > SAA7110_NR_REG)
+		return ret;
 
 	/* the saa7110 has an autoincrement function, use it if
 	 * the adapter understands raw I2C */
 	if (i2c_check_functionality(client->adapter, I2C_FUNC_I2C)) {
 		struct saa7110 *decoder = i2c_get_clientdata(client);
 		struct i2c_msg msg;
-		u8 block_data[54];
 
-		msg.len = 0;
-		msg.buf = (char *) block_data;
+		msg.len = len;
+		msg.buf = (char *) data;
 		msg.addr = client->addr;
-		msg.flags = client->flags;
-		while (len >= 1) {
-			msg.len = 0;
-			block_data[msg.len++] = reg;
-			while (len-- >= 1 && msg.len < 54)
-				block_data[msg.len++] =
-				    decoder->reg[reg++] = *data++;
-			ret = i2c_transfer(client->adapter, &msg, 1);
-		}
+		msg.flags = 0;
+		ret = i2c_transfer(client->adapter, &msg, 1);
+
+		/* Cache the written data */
+		memcpy(decoder->reg + reg, data + 1, len - 1);
 	} else {
-		while (len-- >= 1) {
+		for (++data, --len; len; len--) {
 			if ((ret = saa7110_write(client, reg++,
 						 *data++)) < 0)
 				break;

@@ -192,7 +191,7 @@
 	return 0;
 }
 
-static const unsigned char initseq[] = {
+static const unsigned char initseq[1 + SAA7110_NR_REG] = {
 	0, 0x4C, 0x3C, 0x0D, 0xEF, 0xBD, 0xF2, 0x03, 0x00,
 	/* 0x08 */ 0xF8, 0xF8, 0x60, 0x60, 0x00, 0x86, 0x18, 0x90,
 	/* 0x10 */ 0x00, 0x59, 0x40, 0x46, 0x42, 0x1A, 0xFF, 0xDA,
Signature · 1 line
Jean Delvare
#5 lore raw

Re: [stable] Re: [01/11] fix amd64 2.6.11 oops on modprobe (saa7110)

Quoted 9 lines · * Jean Delvare ([email protected]) wrote:

Hi Greg, all,

Quoted 5 lines
Quoted 3 lines

Not that I really care, but isn't there a rule that a patch "... can not contain any "trivial" fixes in it (spelling changes, whitespace cleanups, etc.)"?

Good point. Jean, care to respin the patch?

Sure, sorry for the trouble.

Thanks, updated patch.
-chris

Signature · 1 line
Linux Security Modules     http://lsm.immunix.org     http://lsm.bkbits.net
author
#6 lore raw

PATCH 2/11 cramfs: small stat(2) fix

In reply to Greg KH

-stable review patch. If anyone has any objections, please let us know.

------------------

From: Eric Lammerts <[email protected]>

When I stat(2) a device node on a cramfs, the st_blocks field is bogus (it's derived from the size field which in this case holds the major/minor numbers). This makes du(1) output completely wrong.

Signed-off-by Eric Lammerts <[email protected]>
Signed-off-by Andrew Morton <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>
fs/cramfs/inode.c renamed from fs/cramfs/inode.c~cramfs-small-stat2-fix +1 -0
diff -puN fs/cramfs/inode.c~cramfs-small-stat2-fix fs/cramfs/inode.c
@@ -70,6 +70,7 @@ static struct inode *get_cramfs_inode(st
 			inode->i_data.a_ops = &cramfs_aops;
 		} else {
 			inode->i_size = 0;
+			inode->i_blocks = 0;
 			init_special_inode(inode, inode->i_mode,
 				old_decode_dev(cramfs_inode->size));
 		}

_

author
#7 lore raw

PATCH 3/11 drm missing memset can crash X server..

In reply to Greg KH

-stable review patch. If anyone has any objections, please let us know.

------------------

Egbert Eich reported a bug 2673 on bugs.freedesktop.org and tracked it down to a missing memset in the setversion ioctl, this causes X server crashes...

From: Egbert Eich <[email protected]>

Signed-off-by Dave Airlie <[email protected]>
Signed-off-by Chris Wright <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>
drivers/char/drm/drm_ioctl.c +2 -0
diff -Nru a/drivers/char/drm/drm_ioctl.c b/drivers/char/drm/drm_ioctl.c
@@ -326,6 +326,8 @@
 
 	DRM_COPY_FROM_USER_IOCTL(sv, argp, sizeof(sv));
 
+	memset(&version, 0, sizeof(version));
+
 	dev->driver->version(&version);
 	retv.drm_di_major = DRM_IF_MAJOR;
 	retv.drm_di_minor = DRM_IF_MINOR;
author
#8 lore raw

PATCH 4/11 ppc32: Compilation fixes for Ebony, Luan and Ocotea

In reply to Greg KH

-stable review patch. If anyone has any objections, please let us know.

------------------

From: Matt Porter <[email protected]>

this patch fixes the problem, that the current kernel (linux-2.6.11-rc5) could not be compiled, when "support for early boot texts over serial port" (CONFIG_SERIAL_TEXT_DEBUG=y) is active.

Signed-off-by Gerhard Jaeger <[email protected]>
Signed-off-by Matt Porter <[email protected]>
Signed-off-by Andrew Morton <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>

---

 25-akpm/arch/ppc/platforms/4xx/ebony.h  |    4 ++--
 25-akpm/arch/ppc/platforms/4xx/luan.h   |    6 +++---
 25-akpm/arch/ppc/platforms/4xx/ocotea.h |    4 ++--
 3 files changed, 7 insertions(+), 7 deletions(-)
arch/ppc/platforms/4xx/ebony.h renamed from arch/ppc/platforms/4xx/ebony.h~ppc32-compilation-fixes-for-ebony-luan-and-ocotea +2 -2
diff -puN arch/ppc/platforms/4xx/ebony.h~ppc32-compilation-fixes-for-ebony-luan-and-ocotea arch/ppc/platforms/4xx/ebony.h
@@ -61,8 +61,8 @@
  */
 
 /* OpenBIOS defined UART mappings, used before early_serial_setup */
-#define UART0_IO_BASE	(u8 *) 0xE0000200
-#define UART1_IO_BASE	(u8 *) 0xE0000300
+#define UART0_IO_BASE	0xE0000200
+#define UART1_IO_BASE	0xE0000300
 
 /* external Epson SG-615P */
 #define BASE_BAUD	691200
arch/ppc/platforms/4xx/luan.h renamed from arch/ppc/platforms/4xx/luan.h~ppc32-compilation-fixes-for-ebony-luan-and-ocotea +3 -3
diff -puN arch/ppc/platforms/4xx/luan.h~ppc32-compilation-fixes-for-ebony-luan-and-ocotea arch/ppc/platforms/4xx/luan.h
@@ -47,9 +47,9 @@
 #define RS_TABLE_SIZE	3
 
 /* PIBS defined UART mappings, used before early_serial_setup */
-#define UART0_IO_BASE	(u8 *) 0xa0000200
-#define UART1_IO_BASE	(u8 *) 0xa0000300
-#define UART2_IO_BASE	(u8 *) 0xa0000600
+#define UART0_IO_BASE	0xa0000200
+#define UART1_IO_BASE	0xa0000300
+#define UART2_IO_BASE	0xa0000600
 
 #define BASE_BAUD	11059200
 #define STD_UART_OP(num)					\
arch/ppc/platforms/4xx/ocotea.h renamed from arch/ppc/platforms/4xx/ocotea.h~ppc32-compilation-fixes-for-ebony-luan-and-ocotea +2 -2
diff -puN arch/ppc/platforms/4xx/ocotea.h~ppc32-compilation-fixes-for-ebony-luan-and-ocotea arch/ppc/platforms/4xx/ocotea.h
@@ -56,8 +56,8 @@
 #define RS_TABLE_SIZE	2
 
 /* OpenBIOS defined UART mappings, used before early_serial_setup */
-#define UART0_IO_BASE	(u8 *) 0xE0000200
-#define UART1_IO_BASE	(u8 *) 0xE0000300
+#define UART0_IO_BASE	0xE0000200
+#define UART1_IO_BASE	0xE0000300
 
 #define BASE_BAUD	11059200/16
 #define STD_UART_OP(num)					\

_

author
#9 lore raw

PATCH 5/11 Fix i2c messsage flags in video drivers

In reply to Greg KH

-stable review patch. If anyone has any objections, please let us know.

------------------

While working on the saa7110 driver I found a problem with the way various video drivers (found on Zoran-based boards) prepare i2c messages to be used by i2c_transfer. The drivers improperly copy the i2c client flags as the message flags, while both sets are mostly unrelated. The net effect in this case is to trigger an I2C block read instead of the expected I2C block write. The fix is simply not to pass any flag, because none are needed.

I think this patch qualifies hands down as a "critical bug fix" to be included in whatever bug-fix-only trees exist these days. As far as I can see, all Zoran-based boards are broken in 2.6.11 without this patch.

Signed-off-by Jean Delvare <[email protected]>
Signed-off-by Chris Wright <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>
drivers/media/video/adv7170.c renamed from drivers/media/video.orig/adv7170.c +1 -1
diff -u -rN linux-2.6.11-bk3/drivers/media/video.orig/adv7170.c linux-2.6.11-bk3/drivers/media/video/adv7170.c
@@ -130,7 +130,7 @@
 		u8 block_data[32];
 
 		msg.addr = client->addr;
-		msg.flags = client->flags;
+		msg.flags = 0;
 		while (len >= 2) {
 			msg.buf = (char *) block_data;
 			msg.len = 0;
drivers/media/video/adv7175.c renamed from drivers/media/video.orig/adv7175.c +1 -1
diff -u -rN linux-2.6.11-bk3/drivers/media/video.orig/adv7175.c linux-2.6.11-bk3/drivers/media/video/adv7175.c
@@ -126,7 +126,7 @@
 		u8 block_data[32];
 
 		msg.addr = client->addr;
-		msg.flags = client->flags;
+		msg.flags = 0;
 		while (len >= 2) {
 			msg.buf = (char *) block_data;
 			msg.len = 0;
drivers/media/video/bt819.c renamed from drivers/media/video.orig/bt819.c +1 -1
diff -u -rN linux-2.6.11-bk3/drivers/media/video.orig/bt819.c linux-2.6.11-bk3/drivers/media/video/bt819.c
@@ -146,7 +146,7 @@
 		u8 block_data[32];
 
 		msg.addr = client->addr;
-		msg.flags = client->flags;
+		msg.flags = 0;
 		while (len >= 2) {
 			msg.buf = (char *) block_data;
 			msg.len = 0;
drivers/media/video/saa7114.c renamed from drivers/media/video.orig/saa7114.c +1 -1
diff -u -rN linux-2.6.11-bk3/drivers/media/video.orig/saa7114.c linux-2.6.11-bk3/drivers/media/video/saa7114.c
@@ -163,7 +163,7 @@
 		u8 block_data[32];
 
 		msg.addr = client->addr;
-		msg.flags = client->flags;
+		msg.flags = 0;
 		while (len >= 2) {
 			msg.buf = (char *) block_data;
 			msg.len = 0;
drivers/media/video/saa7185.c renamed from drivers/media/video.orig/saa7185.c +1 -1
diff -u -rN linux-2.6.11-bk3/drivers/media/video.orig/saa7185.c linux-2.6.11-bk3/drivers/media/video/saa7185.c
@@ -118,7 +118,7 @@
 		u8 block_data[32];
 
 		msg.addr = client->addr;
-		msg.flags = client->flags;
+		msg.flags = 0;
 		while (len >= 2) {
 			msg.buf = (char *) block_data;
 			msg.len = 0;
Signature · 6 lines
Jean Delvare
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
author
#10 lore raw

PATCH 6/11 : Put back tcp_timer_bug_msg[] symbol export.

In reply to Greg KH

-stable review patch. If anyone has any objections, please let us know.

------------------

This wrecks the ipv6 modular build for a lot of people. In fact, since I always build ipv6 modular I am surprised I never hit this. My best guess is that my compiler is optimizing the reference away, but that can never be depended upon and the symbol export really is needed.

[TCP]: Put back tcp_timer_bug_msg[] symbol export.

It is needed for tcp_reset_xmit_timer(), which is invoked by tcp_prequeue() which is invoked from tcp_ipv6.c

Signed-off-by Hideaki YOSHIFUJI <[email protected]>
Signed-off-by David S. Miller <[email protected]>
Signed-off-by Chris Wright <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>
net/ipv4/tcp_timer.c +1 -0
diff -Nru a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
@@ -38,6 +38,7 @@
 
 #ifdef TCP_DEBUG
 const char tcp_timer_bug_msg[] = KERN_DEBUG "tcpbug: unknown timer value\n";
+EXPORT_SYMBOL(tcp_timer_bug_msg);
 #endif
 
 /*
#11 lore raw

Re: [06/11] [TCP]: Put back tcp_timer_bug_msg[] symbol export.

net/ipv4/tcp_timer.c line 42

Quoted 8 lines
net/ipv4/tcp_timer.c +1 -0
@@ -38,6 +38,7 @@
 
 #ifdef TCP_DEBUG
 const char tcp_timer_bug_msg[] = KERN_DEBUG "tcpbug: unknown timer value\n";
+EXPORT_SYMBOL(tcp_timer_bug_msg);
 #endif

not complaining about putting this into -stable, but why do people have TCP_DEBUG turned on for normal builds?

maintainer
#12 lore raw

Re: [06/11] [TCP]: Put back tcp_timer_bug_msg[] symbol export.

Quoted 11 lines · On Thu, 10 Mar 2005 23:20:14 +0000 Christoph Hellwig <[email protected]> wrote:

net/ipv4/tcp_timer.c line 42

Quoted 8 lines
net/ipv4/tcp_timer.c +1 -0
@@ -38,6 +38,7 @@
 
 #ifdef TCP_DEBUG
 const char tcp_timer_bug_msg[] = KERN_DEBUG "tcpbug: unknown timer value\n";
+EXPORT_SYMBOL(tcp_timer_bug_msg);
 #endif

not complaining about putting this into -stable, but why do people have TCP_DEBUG turned on for normal builds?

It is on in everyone's build unless they edit include/net/tcp.h

author
#13 lore raw

PATCH 7/11 ppc32: trivial fix for e500 oprofile build

In reply to Greg KH

-stable review patch. If anyone has any objections, please let us know.

------------------

Fix for trivial fix for 2.6.11 oprofile compilation on e500 based ppc.

Signed-off-by Andy Fleming <[email protected]>
Signed-off-by Kumar Gala <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>
arch/ppc/oprofile/op_model_fsl_booke.c +1 -2
diff -Nru a/arch/ppc/oprofile/op_model_fsl_booke.c b/arch/ppc/oprofile/op_model_fsl_booke.c
@@ -150,7 +150,6 @@
 	int is_kernel;
 	int val;
 	int i;
-	unsigned int cpu = smp_processor_id();
 
 	/* set the PMM bit (see comment below) */
 	mtmsr(mfmsr() | MSR_PMM);

@@ -162,7 +161,7 @@
 		val = ctr_read(i);
 		if (val < 0) {
 			if (oprofile_running && ctr[i].enabled) {
-				oprofile_add_sample(pc, is_kernel, i, cpu);
+				oprofile_add_pc(pc, is_kernel, i);
 				ctr_write(i, reset_value[i]);
 			} else {
 				ctr_write(i, 0);
author
#14 lore raw

PATCH 8/11 PCI-E: fix hotplug double free

In reply to Greg KH

-stable review patch. If anyone has any objections, please let us know.

------------------

[PATCH] PCI: fix hotplug double free

With the brackets missed out func could be freed twice.

Found by Coverity tool

Signed-off-by Alexander Nyberg <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>
drivers/pci/hotplug/pciehp_ctrl.c +2 -1
diff -Nru a/drivers/pci/hotplug/pciehp_ctrl.c b/drivers/pci/hotplug/pciehp_ctrl.c
@@ -1354,10 +1354,11 @@
 				dbg("PCI Bridge Hot-Remove s:b:d:f(%02x:%02x:%02x:%02x)\n", 
 					ctrl->seg, func->bus, func->device, func->function);
 				bridge_slot_remove(func);
-			} else
+			} else {
 				dbg("PCI Function Hot-Remove s:b:d:f(%02x:%02x:%02x:%02x)\n", 
 					ctrl->seg, func->bus, func->device, func->function);
 				slot_remove(func);
+			}
 
 			func = pciehp_slot_find(ctrl->slot_bus, device, 0);
 		}
author
#15 lore raw

PATCH 9/11 r8169: receive descriptor length fix

In reply to Greg KH

-stable review patch. If anyone has any objections, please let us know.

------------------

The status and received packets indication in the Rx descriptor ring are not correctly reset when a descriptor is recycled.

Signed-off-by Francois Romieu <[email protected]>
Signed-off-by Chris Wright <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>
drivers/net/r8169.c renamed from drivers/net/r8169.c~r8169-490 +10 -7
diff -puN drivers/net/r8169.c~r8169-490 drivers/net/r8169.c
@@ -1683,16 +1683,19 @@ static void rtl8169_free_rx_skb(struct r
 	rtl8169_make_unusable_by_asic(desc);
 }
 
-static inline void rtl8169_return_to_asic(struct RxDesc *desc, int rx_buf_sz)
+static inline void rtl8169_mark_to_asic(struct RxDesc *desc, u32 rx_buf_sz)
 {
-	desc->opts1 |= cpu_to_le32(DescOwn + rx_buf_sz);
+	u32 eor = le32_to_cpu(desc->opts1) & RingEnd;
+
+	desc->opts1 = cpu_to_le32(DescOwn | eor | rx_buf_sz);
 }
 
-static inline void rtl8169_give_to_asic(struct RxDesc *desc, dma_addr_t mapping,
-					int rx_buf_sz)
+static inline void rtl8169_map_to_asic(struct RxDesc *desc, dma_addr_t mapping,
+				       u32 rx_buf_sz)
 {
 	desc->addr = cpu_to_le64(mapping);
-	desc->opts1 |= cpu_to_le32(DescOwn + rx_buf_sz);
+	wmb();
+	rtl8169_mark_to_asic(desc, rx_buf_sz);
 }
 
 static int rtl8169_alloc_rx_skb(struct pci_dev *pdev, struct sk_buff **sk_buff,

@@ -1712,7 +1715,7 @@ static int rtl8169_alloc_rx_skb(struct p
 	mapping = pci_map_single(pdev, skb->tail, rx_buf_sz,
 				 PCI_DMA_FROMDEVICE);
 
-	rtl8169_give_to_asic(desc, mapping, rx_buf_sz);
+	rtl8169_map_to_asic(desc, mapping, rx_buf_sz);
 
 out:
 	return ret;

@@ -2150,7 +2153,7 @@ static inline int rtl8169_try_rx_copy(st
 			skb_reserve(skb, NET_IP_ALIGN);
 			eth_copy_and_sum(skb, sk_buff[0]->tail, pkt_size, 0);
 			*sk_buff = skb;
-			rtl8169_return_to_asic(desc, rx_buf_sz);
+			rtl8169_mark_to_asic(desc, rx_buf_sz);
 			ret = 0;
 		}
 	}

_

Signature · 1 line
Ueimor
author
#16 lore raw

PATCH 10/11 sis900 kernel oops fix

In reply to Greg KH

-stable review patch. If anyone has any objections, please let us know.

------------------

Backport of fix described below.

From: Herbert Xu <[email protected]>

Fix bug #4223.

OK, this happened because we got preempted before sis900_mii_probe finished setting the sis_priv->mii. Theoretically this can happen with SMP as well but I suppose the number of SMP machines with sis900 is fairly small.

Anyway, the fix is to make sure that sis900_mii_probe is done before the device can be opened. This patch does it by moving the setup before register_netdevice.

Since the netdev name is not available before register_netdev, I've changed the relevant printk's to use pci_name instead. Note that one of those printk's may be called after register_netdev as well.

Signed-off-by Chris Wright <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>
drivers/net/sis900.c +14 -16
@@ -236,7 +236,7 @@ static int __devinit sis900_get_mac_addr
 	signature = (u16) read_eeprom(ioaddr, EEPROMSignature);    
 	if (signature == 0xffff || signature == 0x0000) {
 		printk (KERN_INFO "%s: Error EERPOM read %x\n", 
-			net_dev->name, signature);
+			pci_name(pci_dev), signature);
 		return 0;
 	}
 

@@ -268,7 +268,7 @@ static int __devinit sis630e_get_mac_add
 	if (!isa_bridge)
 		isa_bridge = pci_get_device(PCI_VENDOR_ID_SI, 0x0018, isa_bridge);
 	if (!isa_bridge) {
-		printk("%s: Can not find ISA bridge\n", net_dev->name);
+		printk("%s: Can not find ISA bridge\n", pci_name(pci_dev));
 		return 0;
 	}
 	pci_read_config_byte(isa_bridge, 0x48, &reg);

@@ -456,10 +456,6 @@ static int __devinit sis900_probe(struct
 	net_dev->tx_timeout = sis900_tx_timeout;
 	net_dev->watchdog_timeo = TX_TIMEOUT;
 	net_dev->ethtool_ops = &sis900_ethtool_ops;
-	
-	ret = register_netdev(net_dev);
-	if (ret)
-		goto err_unmap_rx;
 		
 	/* Get Mac address according to the chip revision */
 	pci_read_config_byte(pci_dev, PCI_CLASS_REVISION, &revision);

@@ -476,7 +472,7 @@ static int __devinit sis900_probe(struct
 
 	if (ret == 0) {
 		ret = -ENODEV;
-		goto err_out_unregister;
+		goto err_unmap_rx;
 	}
 	
 	/* 630ET : set the mii access mode as software-mode */

@@ -486,7 +482,7 @@ static int __devinit sis900_probe(struct
 	/* probe for mii transceiver */
 	if (sis900_mii_probe(net_dev) == 0) {
 		ret = -ENODEV;
-		goto err_out_unregister;
+		goto err_unmap_rx;
 	}
 
 	/* save our host bridge revision */

@@ -496,6 +492,10 @@ static int __devinit sis900_probe(struct
 		pci_dev_put(dev);
 	}
 
+	ret = register_netdev(net_dev);
+	if (ret)
+		goto err_unmap_rx;
+
 	/* print some information about our NIC */
 	printk(KERN_INFO "%s: %s at %#lx, IRQ %d, ", net_dev->name,
 	       card_name, ioaddr, net_dev->irq);

@@ -505,8 +505,6 @@ static int __devinit sis900_probe(struct
 
 	return 0;
 
- err_out_unregister:
- 	unregister_netdev(net_dev);
  err_unmap_rx:
 	pci_free_consistent(pci_dev, RX_TOTAL_SIZE, sis_priv->rx_ring,
 		sis_priv->rx_ring_dma);

@@ -533,6 +531,7 @@ static int __devinit sis900_probe(struct
 static int __init sis900_mii_probe(struct net_device * net_dev)
 {
 	struct sis900_private * sis_priv = net_dev->priv;
+	const char *dev_name = pci_name(sis_priv->pci_dev);
 	u16 poll_bit = MII_STAT_LINK, status = 0;
 	unsigned long timeout = jiffies + 5 * HZ;
 	int phy_addr;

@@ -582,21 +581,20 @@ static int __init sis900_mii_probe(struc
 					mii_phy->phy_types =
 					    (mii_status & (MII_STAT_CAN_TX_FDX | MII_STAT_CAN_TX)) ? LAN : HOME;
 				printk(KERN_INFO "%s: %s transceiver found at address %d.\n",
-				       net_dev->name, mii_chip_table[i].name,
+				       dev_name, mii_chip_table[i].name,
 				       phy_addr);
 				break;
 			}
 			
 		if( !mii_chip_table[i].phy_id1 ) {
 			printk(KERN_INFO "%s: Unknown PHY transceiver found at address %d.\n",
-			       net_dev->name, phy_addr);
+			       dev_name, phy_addr);
 			mii_phy->phy_types = UNKNOWN;
 		}
 	}
 	
 	if (sis_priv->mii == NULL) {
-		printk(KERN_INFO "%s: No MII transceivers found!\n",
-			net_dev->name);
+		printk(KERN_INFO "%s: No MII transceivers found!\n", dev_name);
 		return 0;
 	}
 

@@ -621,7 +619,7 @@ static int __init sis900_mii_probe(struc
 			poll_bit ^= (mdio_read(net_dev, sis_priv->cur_phy, MII_STATUS) & poll_bit);
 			if (time_after_eq(jiffies, timeout)) {
 				printk(KERN_WARNING "%s: reset phy and link down now\n",
-					net_dev->name);
+				       dev_name);
 				return -ETIME;
 			}
 		}

@@ -691,7 +689,7 @@ static u16 sis900_default_phy(struct net
 		sis_priv->mii = default_phy;
 		sis_priv->cur_phy = default_phy->phy_addr;
 		printk(KERN_INFO "%s: Using transceiver found at address %d as default\n",
-					net_dev->name,sis_priv->cur_phy);
+		       pci_name(sis_priv->pci_dev), sis_priv->cur_phy);
 	}
 	
 	status = mdio_read(net_dev, sis_priv->cur_phy, MII_CONTROL);
author
#19 lore raw

PATCH 11/11 older chips oops on shutdown

In reply to Greg KH

-stable review patch. If anyone has any objections, please let us know.

------------------

Kernel 2.6.11, hardware is a MSI KT333-based board with an XP1800.

I'm oopsing on shutdown on a machine that has a Via Rhine adapter in it:

Unable to handle kernel paging request at virtual address e0803003
  printing eip:
c01f262c
*pde = 014dc067
*pte = 00000000
Oops: 0000 [#1]
Modules linked in: cpufreq_userspace cpufreq_powersave cpufreq_ondemand
CPU:    0
EIP:    0060:[<c01f262c>]    Not tainted VLI
EFLAGS: 00010292   (2.6.11)

EIP is at ioread8+0x2c/0x40
eax: e0803003 ebx: e0803003 ecx: c026b430 edx: e0803003
esi: dff90260 edi: e0802f80 ebp: dd117e74 esp: dd117e74
ds: 007b es: 007b ss: 0068
Process reboot (pid: 5769, threadinfo=dd117000 task=dfafa080)

Stack: dd117e8c c026b490 dff90040 c151ccd4 c044a1a8 b7fdc078 dd117ea4 
c0253ad9
        c151ccd4 00000042 fee1dead 00000001 dd117fbc c012461c c04d72a8 00000001
        00000000 00010800 00000000 dd117ed8 c013b40b dffe7380 00030800 00000000
Call Trace:
  [<c0103d5f>] show_stack+0x7f/0xa0
  [<c0103efa>] show_registers+0x15a/0x1c0
  [<c01040ce>] die+0xce/0x150
  [<c0113406>] do_page_fault+0x356/0x692
  [<c01039ff>] error_code+0x2b/0x30
  [<c026b490>] rhine_shutdown+0x60/0x140
  [<c0253ad9>] device_shutdown+0x89/0x8b
  [<c012461c>] sys_reboot+0xac/0x200
  [<c0102f71>] sysenter_past_esp+0x52/0x75
Code: 3d ff ff 03 00 89 c2 89 e5 77 20 66 31 c0 3d 00 00 01 00 75 0c

81 e2 ff ff 00 00 ec 0f b6 c0 c9 c3 0f 0b 37 00 7b 65 3b c0 eb ea <0f> b6 00 eb ec eb 0d 90 90 90 90 90 90 90 90 90 90 90 90 90 55

Seems like it is the ioread8 in:

         /* Hit power state D3 (sleep) */
         iowrite8(ioread8(ioaddr + StickyHW) | 0x03, ioaddr + StickyHW);

that fails. StickyHW is 0x83. lspci says:

0000:00:07.0 Ethernet controller: VIA Technologies, Inc. VT86C100A [Rhine] (rev 06)

         Flags: bus master, medium devsel, latency 32, IRQ 18
         I/O ports at ec00 [size=128]
         Memory at dfffff80 (32-bit, non-prefetchable) [size=128]

In other words, it's trying to read outside of the I/O range (0x80), which matches the fauling address.

I'm guessing my chip revision doesn't support WOL, it's a crappy noname card.

It does seem as if rhine_power_init checks quirks for rqWOL before touching any registers. Should rhine_shutdown do the same? Proposed patch below, which resolves the problem on my system.

Check to make sure WOL is supported before setting it up in rhine_shutdown.

Signed-off-by Olof Johansson <[email protected]>
Signed-off-by Chris Wright <[email protected]>
Signed-off-by Greg Kroah-Hartman <[email protected]>

--- linux-2.6.11.orig/drivers/net/via-rhine.c 2005-03-02 01:38:32.000000000 -0600
+++ linux-2.6.11/drivers/net/via-rhine.c 2005-03-05 12:25:34.000000000
-0600
@@ -1899,6 +1899,9 @@
struct rhine_private *rp = netdev_priv(dev); void __iomem *ioaddr = rp->base;

+ if (!(rp->quirks & rqWOL))
+ return; /* Nothing to do for non-WOL adapters */
+
rhine_power_init(dev);

/* Make sure we use pattern 0, 1 and not 4, 5 */