3.0.46-stable review

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 arch/mn10300/Makefile |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

arch/mn10300/Makefile +1 -1

@@ -26,7 +26,7 @@ CHECKFLAGS	+=
 PROCESSOR	:= unset
 UNIT		:= unset
 
-KBUILD_CFLAGS	+= -mam33 -mmem-funcs -DCPU=AM33
+KBUILD_CFLAGS	+= -mam33 -DCPU=AM33 $(call cc-option,-mmem-funcs,)
 KBUILD_AFLAGS	+= -mam33 -DCPU=AM33
 
 ifeq ($(CONFIG_MN10300_CURRENT_IN_E2),y)

Greg Kroah-Hartman

author

2012-10-11 02:12 UTC #2 lore raw

PATCH 2/84 kbuild: make: fix if_changed when command contains backslashes

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

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

From: Sascha Hauer <[email protected]>

commit c353acba28fb3fa1fd05fd6b85a9fc7938330f9c upstream.

The call if_changed mechanism does not work when the command contains backslashes. This basically is an issue with lzo and bzip2 compressed kernels. The compressed binaries do not contain the uncompressed image size, so these use size_append to append the size. This results in backslashes in the executed command. With this if_changed always detects a change in the command and rebuilds the compressed image even if nothing has changed.

Fix this by escaping backslashes in make-cmd

Signed-off-by Sascha Hauer <[email protected]>

Signed-off-by Jan Luebbe <[email protected]>

Cc Sam Ravnborg <[email protected]>

Cc Bernhard Walle <[email protected]>

Cc Michal Marek <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 scripts/Kbuild.include |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

scripts/Kbuild.include +1 -1

@@ -205,7 +205,7 @@ endif
 # >$< substitution to preserve $ when reloading .cmd file
 # note: when using inline perl scripts [perl -e '...$$t=1;...']
 # in $(cmd_xxx) double $$ your perl vars
-make-cmd = $(subst \#,\\\#,$(subst $$,$$$$,$(call escsq,$(cmd_$(1)))))
+make-cmd = $(subst \\,\\\\,$(subst \#,\\\#,$(subst $$,$$$$,$(call escsq,$(cmd_$(1))))))
 
 # Find any prerequisites that is newer than target or that does not exist.
 # PHONY targets skipped in both cases.

Greg Kroah-Hartman

author

2012-10-11 02:11 UTC #3 lore raw

PATCH 3/84 media: rc: ite-cir: Initialise ite_dev::rdev earlier

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

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

From: Ben Hutchings <[email protected]>

commit 4b961180ef275035b1538317ffd0e21e80e63e77 upstream.

ite_dev::rdev is currently initialised in ite_probe() after rc_register_device() returns. If a newly registered device is opened quickly enough, we may enable interrupts and try to use ite_dev::rdev before it has been initialised. Move it up to the earliest point we can, right after calling rc_allocate_device().

Reported-and-tested-by YunQiang Su <[email protected]>

Signed-off-by Ben Hutchings <[email protected]>

Signed-off-by Mauro Carvalho Chehab <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/media/rc/ite-cir.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/media/rc/ite-cir.c +1 -1

@@ -1477,6 +1477,7 @@ static int ite_probe(struct pnp_dev *pde
 	rdev = rc_allocate_device();
 	if (!rdev)
 		goto failure;
+	itdev->rdev = rdev;
 
 	ret = -ENODEV;

@@ -1608,7 +1609,6 @@ static int ite_probe(struct pnp_dev *pde
 	if (ret)
 		goto failure;
 
-	itdev->rdev = rdev;
 	ite_pr(KERN_NOTICE, "driver has been successfully loaded\n");
 
 	return 0;

Greg Kroah-Hartman

author

2012-10-11 02:11 UTC #4 lore raw

PATCH 4/84 ACPI: run _OSC after ACPI_FULL_INITIALIZATION

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

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

From: Lin Ming <[email protected]>

commit fc54ab72959edbf229b65ac74b2f122d799ca002 upstream.

The _OSC method may exist in module level code, so it must be called after ACPI_FULL_INITIALIZATION

On some new platforms with Zero-Power-Optical-Disk-Drive (ZPODD) support, this fix is necessary to save power.

Signed-off-by Lin Ming <[email protected]>

Tested-by Aaron Lu <[email protected]>

Signed-off-by Len Brown <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/acpi/bus.c |    8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

drivers/acpi/bus.c +6 -2

@@ -944,8 +944,6 @@ static int __init acpi_bus_init(void)
 	status = acpi_ec_ecdt_probe();
 	/* Ignore result. Not having an ECDT is not fatal. */
 
-	acpi_bus_osc_support();
-
 	status = acpi_initialize_objects(ACPI_FULL_INITIALIZATION);
 	if (ACPI_FAILURE(status)) {
 		printk(KERN_ERR PREFIX "Unable to initialize ACPI objects\n");

@@ -953,6 +951,12 @@ static int __init acpi_bus_init(void)
 	}
 
 	/*
+	 * _OSC method may exist in module level code,
+	 * so it must be run after ACPI_FULL_INITIALIZATION
+	 */
+	acpi_bus_osc_support();
+
+	/*
 	 * _PDC control method may load dynamic SSDT tables,
 	 * and we need to install the table handler before that.
 	 */

Greg Kroah-Hartman

author

2012-10-11 02:11 UTC #5 lore raw

PATCH 5/84 PCI: acpiphp: check whether _ADR evaluation succeeded

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

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

From: Bjorn Helgaas <[email protected]>

commit dfb117b3e50c52c7b3416db4a4569224b8db80bb upstream.

Check whether we evaluated _ADR successfully. Previously we ignored failure, so we would have used garbage data from the stack as the device and function number.

We return AE_OK so that we ignore only this slot and continue looking for other slots.

Found by Coverity (CID 113981).

Signed-off-by Bjorn Helgaas <[email protected]>

[bwh: Backported to 2.6.32/3.0: adjust context]

Signed-off-by Ben Hutchings <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/pci/hotplug/acpiphp_glue.c |    7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

drivers/pci/hotplug/acpiphp_glue.c +6 -1

@@ -131,7 +131,12 @@ register_slot(acpi_handle handle, u32 lv
 	if (!acpi_pci_check_ejectable(pbus, handle) && !is_dock_device(handle))
 		return AE_OK;
 
-	acpi_evaluate_integer(handle, "_ADR", NULL, &adr);
+	status = acpi_evaluate_integer(handle, "_ADR", NULL, &adr);
+	if (ACPI_FAILURE(status)) {
+		warn("can't evaluate _ADR (%#x)\n", status);
+		return AE_OK;
+	}
+
 	device = (adr >> 16) & 0xffff;
 	function = adr & 0xffff;

Greg Kroah-Hartman

author

2012-10-11 02:11 UTC #6 lore raw

PATCH 6/84 lib/gcd.c: prevent possible div by 0

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

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

From: Davidlohr Bueso <[email protected]>

commit e96875677fb2b7cb739c5d7769824dff7260d31d upstream.

Account for all properties when a and/or b are 0:
gcd(0, 0) = 0
gcd(a, 0) = a
gcd(0, b) = b

Fixes no known problems in current kernels.

Signed-off-by Davidlohr Bueso <[email protected]>

Cc Eric Dumazet <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 lib/gcd.c |    3 +++
 1 file changed, 3 insertions(+)

lib/gcd.c +3 -0

@@ -9,6 +9,9 @@ unsigned long gcd(unsigned long a, unsig
 
 	if (a < b)
 		swap(a, b);
+
+	if (!b)
+		return a;
 	while ((r = a % b) != 0) {
 		a = b;
 		b = r;

Greg Kroah-Hartman

author

2012-10-11 02:31 UTC #7 lore raw

PATCH 7/84 kernel/sys.c: call disable_nonboot_cpus() in kernel_restart()

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

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

From: Shawn Guo <[email protected]>

commit f96972f2dc6365421cf2366ebd61ee4cf060c8d5 upstream.

As kernel_power_off() calls disable_nonboot_cpus(), we may also want to have kernel_restart() call disable_nonboot_cpus(). Doing so can help machines that require boot cpu be the last alive cpu during reboot to survive with kernel restart.

This fixes one reboot issue seen on imx6q (Cortex-A9 Quad). The machine requires that the restart routine be run on the primary cpu rather than secondary ones. Otherwise, the secondary core running the restart routine will fail to come to online after reboot.

Signed-off-by Shawn Guo <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 kernel/sys.c |    1 +
 1 file changed, 1 insertion(+)

kernel/sys.c +1 -0

@@ -334,6 +334,7 @@ void kernel_restart_prepare(char *cmd)
 void kernel_restart(char *cmd)
 {
 	kernel_restart_prepare(cmd);
+	disable_nonboot_cpus();
 	if (!cmd)
 		printk(KERN_EMERG "Restarting system.\n");
 	else

Greg Kroah-Hartman

author

2012-10-11 02:12 UTC #8 lore raw

PATCH 8/84 drivers/scsi/atp870u.c: fix bad use of udelay

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

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

From: Martin Michlmayr <[email protected]>

commit 0f6d93aa9d96cc9022b51bd10d462b03296be146 upstream.

The ACARD driver calls udelay() with a value > 2000, which leads to to the following compilation error on ARM:

ERROR: "__bad_udelay" [drivers/scsi/atp870u.ko] undefined! make[1]: *** [__modpost] Error 1

This is because udelay is defined on ARM, roughly speaking, as

	#define udelay(n) ((n) > 2000 ? __bad_udelay() : \
		__const_udelay((n) * ((2199023U*HZ)>>11)))

The argument to __const_udelay is the number of jiffies to wait divided by 4, but this does not work unless the multiplication does not overflow, and that is what the build error is designed to prevent. The intended behavior can be achieved by using mdelay to call udelay multiple times in a loop.

[[email protected]: adding context]

Signed-off-by Martin Michlmayr <[email protected]>

Signed-off-by Jonathan Nieder <[email protected]>

Cc James Bottomley <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/scsi/atp870u.c |   11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

drivers/scsi/atp870u.c +10 -1

@@ -1174,7 +1174,16 @@ wait_io1:
 	outw(val, tmport);
 	outb(2, 0x80);
 TCM_SYNC:
-	udelay(0x800);
+	/*
+	 * The funny division into multiple delays is to accomodate
+	 * arches like ARM where udelay() multiplies its argument by
+	 * a large number to initialize a loop counter.  To avoid
+	 * overflow, the maximum supported udelay is 2000 microseconds.
+	 *
+	 * XXX it would be more polite to find a way to use msleep()
+	 */
+	mdelay(2);
+	udelay(48);
 	if ((inb(tmport) & 0x80) == 0x00) {	/* bsy ? */
 		outw(0, tmport--);
 		outb(0, tmport);

Greg Kroah-Hartman

author

2012-10-11 02:31 UTC #9 lore raw

PATCH 9/84 workqueue: add missing smp_wmb() in process_one_work()

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

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

From: Tejun Heo <[email protected]>

commit 959d1af8cffc8fd38ed53e8be1cf4ab8782f9c00 upstream.

WORK_STRUCT_PENDING is used to claim ownership of a work item and process_one_work() releases it before starting execution. When someone else grabs PENDING, all pre-release updates to the work item should be visible and all updates made by the new owner should happen afterwards.

Grabbing PENDING uses test_and_set_bit() and thus has a full barrier; however, clearing doesn't have a matching wmb. Given the preceding spin_unlock and use of clear_bit, I don't believe this can be a problem on an actual machine and there hasn't been any related report but it still is theretically possible for clear_pending to permeate upwards and happen before work->entry update.

Add an explicit smp_wmb() before work_clear_pending().

Signed-off-by Tejun Heo <[email protected]>

Cc Oleg Nesterov <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 kernel/workqueue.c |    2 ++
 1 file changed, 2 insertions(+)

kernel/workqueue.c +2 -0

@@ -1868,7 +1868,9 @@ __acquires(&gcwq->lock)
 
 	spin_unlock_irq(&gcwq->lock);
 
+	smp_wmb();	/* paired with test_and_set_bit(PENDING) */
 	work_clear_pending(work);
+
 	lock_map_acquire_read(&cwq->wq->lockdep_map);
 	lock_map_acquire(&lockdep_map);
 	trace_workqueue_execute_start(work);

Greg Kroah-Hartman

author

2012-10-11 02:12 UTC #10 lore raw

PATCH 10/84 xfrm: Workaround incompatibility of ESN and async crypto

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

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

From: Steffen Klassert <[email protected]>

[ Upstream commit 3b59df46a449ec9975146d71318c4777ad086744 ]

ESN for esp is defined in RFC 4303. This RFC assumes that the sequence number counters are always up to date. However, this is not true if an async crypto algorithm is employed.

If the sequence number counters are not up to date on sequence number check, we may incorrectly update the upper 32 bit of the sequence number. This leads to a DOS.

We workaround this by comparing the upper sequence number, (used for authentication) with the upper sequence number computed after the async processing. We drop the packet if these numbers are different.

To do this, we introduce a recheck function that does this check in the ESN case.

Signed-off-by Steffen Klassert <[email protected]>

Acked-by Herbert Xu <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 include/net/xfrm.h     |    3 +++
 net/xfrm/xfrm_input.c  |    2 +-
 net/xfrm/xfrm_replay.c |   15 +++++++++++++++
 3 files changed, 19 insertions(+), 1 deletion(-)

include/net/xfrm.h +3 -0

@@ -269,6 +269,9 @@ struct xfrm_replay {
 	int	(*check)(struct xfrm_state *x,
 			 struct sk_buff *skb,
 			 __be32 net_seq);
+	int	(*recheck)(struct xfrm_state *x,
+			   struct sk_buff *skb,
+			   __be32 net_seq);
 	void	(*notify)(struct xfrm_state *x, int event);
 	int	(*overflow)(struct xfrm_state *x, struct sk_buff *skb);
 };

net/xfrm/xfrm_input.c +1 -1

@@ -212,7 +212,7 @@ resume:
 		/* only the first xfrm gets the encap type */
 		encap_type = 0;
 
-		if (async && x->repl->check(x, skb, seq)) {
+		if (async && x->repl->recheck(x, skb, seq)) {
 			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATESEQERROR);
 			goto drop_unlock;
 		}

net/xfrm/xfrm_replay.c +15 -0

@@ -437,6 +437,18 @@ err:
 	return -EINVAL;
 }
 
+static int xfrm_replay_recheck_esn(struct xfrm_state *x,
+				   struct sk_buff *skb, __be32 net_seq)
+{
+	if (unlikely(XFRM_SKB_CB(skb)->seq.input.hi !=
+		     htonl(xfrm_replay_seqhi(x, net_seq)))) {
+			x->stats.replay_window++;
+			return -EINVAL;
+	}
+
+	return xfrm_replay_check_esn(x, skb, net_seq);
+}
+
 static void xfrm_replay_advance_esn(struct xfrm_state *x, __be32 net_seq)
 {
 	unsigned int bitnr, nr, i;

@@ -508,6 +520,7 @@ static void xfrm_replay_advance_esn(stru
 static struct xfrm_replay xfrm_replay_legacy = {
 	.advance	= xfrm_replay_advance,
 	.check		= xfrm_replay_check,
+	.recheck	= xfrm_replay_check,
 	.notify		= xfrm_replay_notify,
 	.overflow	= xfrm_replay_overflow,
 };

@@ -515,6 +528,7 @@ static struct xfrm_replay xfrm_replay_le
 static struct xfrm_replay xfrm_replay_bmp = {
 	.advance	= xfrm_replay_advance_bmp,
 	.check		= xfrm_replay_check_bmp,
+	.recheck	= xfrm_replay_check_bmp,
 	.notify		= xfrm_replay_notify_bmp,
 	.overflow	= xfrm_replay_overflow_bmp,
 };

@@ -522,6 +536,7 @@ static struct xfrm_replay xfrm_replay_bm
 static struct xfrm_replay xfrm_replay_esn = {
 	.advance	= xfrm_replay_advance_esn,
 	.check		= xfrm_replay_check_esn,
+	.recheck	= xfrm_replay_recheck_esn,
 	.notify		= xfrm_replay_notify_bmp,
 	.overflow	= xfrm_replay_overflow_esn,
 };

Greg Kroah-Hartman

author

2012-10-11 02:30 UTC #11 lore raw

PATCH 11/84 xfrm_user: return error pointer instead of NULL

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

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

From: Mathias Krause <[email protected]>

[ Upstream commit 864745d291b5ba80ea0bd0edcbe67273de368836 ]

When dump_one_state() returns an error, e.g. because of a too small buffer to dump the whole xfrm state, xfrm_state_netlink() returns NULL instead of an error pointer. But its callers expect an error pointer and therefore continue to operate on a NULL skbuff.

This could lead to a privilege escalation (execution of user code in kernel context) if the attacker has CAP_NET_ADMIN and is able to map address 0.

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/xfrm/xfrm_user.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

net/xfrm/xfrm_user.c +4 -2

@@ -862,6 +862,7 @@ static struct sk_buff *xfrm_state_netlin
 {
 	struct xfrm_dump_info info;
 	struct sk_buff *skb;
+	int err;
 
 	skb = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_ATOMIC);
 	if (!skb)

@@ -872,9 +873,10 @@ static struct sk_buff *xfrm_state_netlin
 	info.nlmsg_seq = seq;
 	info.nlmsg_flags = 0;
 
-	if (dump_one_state(x, 0, &info)) {
+	err = dump_one_state(x, 0, &info);
+	if (err) {
 		kfree_skb(skb);
-		return NULL;
+		return ERR_PTR(err);
 	}
 
 	return skb;

Greg Kroah-Hartman

author

2012-10-11 02:11 UTC #12 lore raw

PATCH 12/84 xfrm_user: return error pointer instead of NULL #2

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

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

From: Mathias Krause <[email protected]>

[ Upstream commit c25463722509fef0ed630b271576a8c9a70236f3 ]

When dump_one_policy() returns an error, e.g. because of a too small buffer to dump the whole xfrm policy, xfrm_policy_netlink() returns NULL instead of an error pointer. But its caller expects an error pointer and therefore continues to operate on a NULL skbuff.

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/xfrm/xfrm_user.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

net/xfrm/xfrm_user.c +4 -2

@@ -1531,6 +1531,7 @@ static struct sk_buff *xfrm_policy_netli
 {
 	struct xfrm_dump_info info;
 	struct sk_buff *skb;
+	int err;
 
 	skb = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
 	if (!skb)

@@ -1541,9 +1542,10 @@ static struct sk_buff *xfrm_policy_netli
 	info.nlmsg_seq = seq;
 	info.nlmsg_flags = 0;
 
-	if (dump_one_policy(xp, dir, 0, &info) < 0) {
+	err = dump_one_policy(xp, dir, 0, &info);
+	if (err) {
 		kfree_skb(skb);
-		return NULL;
+		return ERR_PTR(err);
 	}
 
 	return skb;

Greg Kroah-Hartman

author

2012-10-11 02:11 UTC #13 lore raw

PATCH 13/84 xfrm: fix a read lock imbalance in make_blackhole

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

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

From: Li RongQing <[email protected]>

[ Upstream commit 433a19548061bb5457b6ab77ed7ea58ca6e43ddb ]

if xfrm_policy_get_afinfo returns 0, it has already released the read lock, xfrm_policy_put_afinfo should not be called again.

Signed-off-by Li RongQing <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/xfrm/xfrm_policy.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

net/xfrm/xfrm_policy.c +1 -1

@@ -1759,7 +1759,7 @@ static struct dst_entry *make_blackhole(
 
 	if (!afinfo) {
 		dst_release(dst_orig);
-		ret = ERR_PTR(-EINVAL);
+		return ERR_PTR(-EINVAL);
 	} else {
 		ret = afinfo->blackhole_route(net, dst_orig);
 	}

Greg Kroah-Hartman

author

2012-10-11 02:32 UTC #14 lore raw

PATCH 14/84 xfrm_user: fix info leak in copy_to_user_auth()

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

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

From: Mathias Krause <[email protected]>

[ Upstream commit 4c87308bdea31a7b4828a51f6156e6f721a1fcc9 ]

copy_to_user_auth() fails to initialize the remainder of alg_name and therefore discloses up to 54 bytes of heap memory via netlink to userland.

Use strncpy() instead of strcpy() to fill the trailing bytes of alg_name with null bytes.

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/xfrm/xfrm_user.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

net/xfrm/xfrm_user.c +1 -1

@@ -742,7 +742,7 @@ static int copy_to_user_auth(struct xfrm
 		return -EMSGSIZE;
 
 	algo = nla_data(nla);
-	strcpy(algo->alg_name, auth->alg_name);
+	strncpy(algo->alg_name, auth->alg_name, sizeof(algo->alg_name));
 	memcpy(algo->alg_key, auth->alg_key, (auth->alg_key_len + 7) / 8);
 	algo->alg_key_len = auth->alg_key_len;

Greg Kroah-Hartman

author

2012-10-11 02:24 UTC #15 lore raw

PATCH 15/84 xfrm_user: fix info leak in copy_to_user_state()

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

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

From: Mathias Krause <[email protected]>

[ Upstream commit f778a636713a435d3a922c60b1622a91136560c1 ]

The memory reserved to dump the xfrm state includes the padding bytes of struct xfrm_usersa_info added by the compiler for alignment (7 for amd64, 3 for i386). Add an explicit memset(0) before filling the buffer to avoid the info leak.

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/xfrm/xfrm_user.c |    1 +
 1 file changed, 1 insertion(+)

net/xfrm/xfrm_user.c +1 -0

@@ -689,6 +689,7 @@ out:
 
 static void copy_to_user_state(struct xfrm_state *x, struct xfrm_usersa_info *p)
 {
+	memset(p, 0, sizeof(*p));
 	memcpy(&p->id, &x->id, sizeof(p->id));
 	memcpy(&p->sel, &x->sel, sizeof(p->sel));
 	memcpy(&p->lft, &x->lft, sizeof(p->lft));

Greg Kroah-Hartman

author

2012-10-11 02:13 UTC #16 lore raw

PATCH 16/84 xfrm_user: fix info leak in copy_to_user_policy()

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

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

From: Mathias Krause <[email protected]>

[ Upstream commit 7b789836f434c87168eab067cfbed1ec4783dffd ]

The memory reserved to dump the xfrm policy includes multiple padding bytes added by the compiler for alignment (padding bytes in struct xfrm_selector and struct xfrm_userpolicy_info). Add an explicit memset(0) before filling the buffer to avoid the heap info leak.

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/xfrm/xfrm_user.c |    1 +
 1 file changed, 1 insertion(+)

net/xfrm/xfrm_user.c +1 -0

@@ -1300,6 +1300,7 @@ static void copy_from_user_policy(struct
 
 static void copy_to_user_policy(struct xfrm_policy *xp, struct xfrm_userpolicy_info *p, int dir)
 {
+	memset(p, 0, sizeof(*p));
 	memcpy(&p->sel, &xp->selector, sizeof(p->sel));
 	memcpy(&p->lft, &xp->lft, sizeof(p->lft));
 	memcpy(&p->curlft, &xp->curlft, sizeof(p->curlft));

Greg Kroah-Hartman

author

2012-10-11 02:13 UTC #17 lore raw

PATCH 17/84 xfrm_user: fix info leak in copy_to_user_tmpl()

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

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

From: Mathias Krause <[email protected]>

[ Upstream commit 1f86840f897717f86d523a13e99a447e6a5d2fa5 ]

The memory used for the template copy is a local stack variable. As struct xfrm_user_tmpl contains multiple holes added by the compiler for alignment, not initializing the memory will lead to leaking stack bytes to userland. Add an explicit memset(0) to avoid the info leak.

Initial version of the patch by Brad Spengler.

Cc Brad Spengler <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/xfrm/xfrm_user.c |    1 +
 1 file changed, 1 insertion(+)

net/xfrm/xfrm_user.c +1 -0

@@ -1405,6 +1405,7 @@ static int copy_to_user_tmpl(struct xfrm
 		struct xfrm_user_tmpl *up = &vec[i];
 		struct xfrm_tmpl *kp = &xp->xfrm_vec[i];
 
+		memset(up, 0, sizeof(*up));
 		memcpy(&up->id, &kp->id, sizeof(up->id));
 		up->family = kp->encap_family;
 		memcpy(&up->saddr, &kp->saddr, sizeof(up->saddr));

Greg Kroah-Hartman

author

2012-10-11 02:25 UTC #18 lore raw

PATCH 18/84 xfrm_user: dont copy esn replay window twice for new states

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

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

From: Mathias Krause <[email protected]>

[ Upstream commit e3ac104d41a97b42316915020ba228c505447d21 ]

The ESN replay window was already fully initialized in xfrm_alloc_replay_state_esn(). No need to copy it again.

Cc Steffen Klassert <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/xfrm/xfrm_user.c |    9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

net/xfrm/xfrm_user.c +5 -4

@@ -442,10 +442,11 @@ static void copy_from_user_state(struct
  * somehow made shareable and move it to xfrm_state.c - JHS
  *
 */
-static void xfrm_update_ae_params(struct xfrm_state *x, struct nlattr **attrs)
+static void xfrm_update_ae_params(struct xfrm_state *x, struct nlattr **attrs,
+				  int update_esn)
 {
 	struct nlattr *rp = attrs[XFRMA_REPLAY_VAL];
-	struct nlattr *re = attrs[XFRMA_REPLAY_ESN_VAL];
+	struct nlattr *re = update_esn ? attrs[XFRMA_REPLAY_ESN_VAL] : NULL;
 	struct nlattr *lt = attrs[XFRMA_LTIME_VAL];
 	struct nlattr *et = attrs[XFRMA_ETIMER_THRESH];
 	struct nlattr *rt = attrs[XFRMA_REPLAY_THRESH];

@@ -555,7 +556,7 @@ static struct xfrm_state *xfrm_state_con
 		goto error;
 
 	/* override default values from above */
-	xfrm_update_ae_params(x, attrs);
+	xfrm_update_ae_params(x, attrs, 0);
 
 	return x;

@@ -1801,7 +1802,7 @@ static int xfrm_new_ae(struct sk_buff *s
 		goto out;
 
 	spin_lock_bh(&x->lock);
-	xfrm_update_ae_params(x, attrs);
+	xfrm_update_ae_params(x, attrs, 1);
 	spin_unlock_bh(&x->lock);
 
 	c.event = nlh->nlmsg_type;

Greg Kroah-Hartman

author

2012-10-11 02:13 UTC #19 lore raw

PATCH 19/84 xfrm_user: ensure user supplied esn replay window is valid

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

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

From: Mathias Krause <[email protected]>

[ Upstream commit ecd7918745234e423dd87fcc0c077da557909720 ]

The current code fails to ensure that the netlink message actually contains as many bytes as the header indicates. If a user creates a new state or updates an existing one but does not supply the bytes for the whole ESN replay window, the kernel copies random heap bytes into the replay bitmap, the ones happen to follow the XFRMA_REPLAY_ESN_VAL netlink attribute. This leads to following issues:

1. The replay window has random bits set confusing the replay handling code later on.

2. A malicious user could use this flaw to leak up to ~3.5kB of heap memory when she has access to the XFRM netlink interface (requires CAP_NET_ADMIN).

Known users of the ESN replay window are strongSwan and Steffen's iproute2 patch (<http://patchwork.ozlabs.org/patch/85962/>). The latter uses the interface with a bitmap supplied while the former does not. strongSwan is therefore prone to run into issue 1.

To fix both issues without breaking existing userland allow using the XFRMA_REPLAY_ESN_VAL netlink attribute with either an empty bitmap or a fully specified one. For the former case we initialize the in-kernel bitmap with zero, for the latter we copy the user supplied bitmap. For state updates the full bitmap must be supplied.

To prevent overflows in the bitmap length calculation the maximum size of bmp_len is limited to 128 by this patch -- resulting in a maximum replay window of 4096 packets. This should be sufficient for all real life scenarios (RFC 4303 recommends a default replay window size of 64).

Cc Steffen Klassert <[email protected]>

Cc Martin Willi <[email protected]>

Cc Ben Hutchings <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 include/linux/xfrm.h |    2 ++
 net/xfrm/xfrm_user.c |   31 +++++++++++++++++++++++++------
 2 files changed, 27 insertions(+), 6 deletions(-)

include/linux/xfrm.h +2 -0

@@ -84,6 +84,8 @@ struct xfrm_replay_state {
 	__u32	bitmap;
 };
 
+#define XFRMA_REPLAY_ESN_MAX	4096
+
 struct xfrm_replay_state_esn {
 	unsigned int	bmp_len;
 	__u32		oseq;

net/xfrm/xfrm_user.c +25 -6

@@ -123,9 +123,21 @@ static inline int verify_replay(struct x
 				struct nlattr **attrs)
 {
 	struct nlattr *rt = attrs[XFRMA_REPLAY_ESN_VAL];
+	struct xfrm_replay_state_esn *rs;
 
-	if ((p->flags & XFRM_STATE_ESN) && !rt)
-		return -EINVAL;
+	if (p->flags & XFRM_STATE_ESN) {
+		if (!rt)
+			return -EINVAL;
+
+		rs = nla_data(rt);
+
+		if (rs->bmp_len > XFRMA_REPLAY_ESN_MAX / sizeof(rs->bmp[0]) / 8)
+			return -EINVAL;
+
+		if (nla_len(rt) < xfrm_replay_state_esn_len(rs) &&
+		    nla_len(rt) != sizeof(*rs))
+			return -EINVAL;
+	}
 
 	if (!rt)
 		return 0;

@@ -370,14 +382,15 @@ static inline int xfrm_replay_verify_len
 					 struct nlattr *rp)
 {
 	struct xfrm_replay_state_esn *up;
+	int ulen;
 
 	if (!replay_esn || !rp)
 		return 0;
 
 	up = nla_data(rp);
+	ulen = xfrm_replay_state_esn_len(up);
 
-	if (xfrm_replay_state_esn_len(replay_esn) !=
-			xfrm_replay_state_esn_len(up))
+	if (nla_len(rp) < ulen || xfrm_replay_state_esn_len(replay_esn) != ulen)
 		return -EINVAL;
 
 	return 0;

@@ -388,22 +401,28 @@ static int xfrm_alloc_replay_state_esn(s
 				       struct nlattr *rta)
 {
 	struct xfrm_replay_state_esn *p, *pp, *up;
+	int klen, ulen;
 
 	if (!rta)
 		return 0;
 
 	up = nla_data(rta);
+	klen = xfrm_replay_state_esn_len(up);
+	ulen = nla_len(rta) >= klen ? klen : sizeof(*up);
 
-	p = kmemdup(up, xfrm_replay_state_esn_len(up), GFP_KERNEL);
+	p = kzalloc(klen, GFP_KERNEL);
 	if (!p)
 		return -ENOMEM;
 
-	pp = kmemdup(up, xfrm_replay_state_esn_len(up), GFP_KERNEL);
+	pp = kzalloc(klen, GFP_KERNEL);
 	if (!pp) {
 		kfree(p);
 		return -ENOMEM;
 	}
 
+	memcpy(p, up, ulen);
+	memcpy(pp, up, ulen);
+
 	*replay_esn = p;
 	*preplay_esn = pp;

Greg Kroah-Hartman

author

2012-10-11 02:25 UTC #20 lore raw

PATCH 20/84 net: ethernet: davinci_cpdma: decrease the desc count when cleaning up the remaining packets

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

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

From: htbegin <[email protected]>

[ Upstream commit ffb5ba90017505a19e238e986e6d33f09e4df765 ]

chan->count is used by rx channel. If the desc count is not updated by the clean up loop in cpdma_chan_stop, the value written to the rxfree register in cpdma_chan_start will be incorrect.

Signed-off-by Tao Hou <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/davinci_cpdma.c |    1 +
 1 file changed, 1 insertion(+)

drivers/net/davinci_cpdma.c +1 -0

@@ -849,6 +849,7 @@ int cpdma_chan_stop(struct cpdma_chan *c
 
 		next_dma = desc_read(desc, hw_next);
 		chan->head = desc_from_phys(pool, next_dma);
+		chan->count--;
 		chan->stats.teardown_dequeue++;
 
 		/* issue callback without locks held */

Greg Kroah-Hartman

author

2012-10-11 02:13 UTC #21 lore raw

PATCH 21/84 ixp4xx_hss: fix build failure due to missing linux/module.h inclusion

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

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

From: Florian Fainelli <[email protected]>

[ Upstream commit 0b836ddde177bdd5790ade83772860940bd481ea ]

Commit 36a1211970193ce215de50ed1e4e1272bc814df1 (netprio_cgroup.h:
dont include module.h from other includes) made the following build error on ixp4xx_hss pop up:

CC [M] drivers/net/wan/ixp4xx_hss.o
drivers/net/wan/ixp4xx_hss.c:1412:20: error: expected ';', ',' or ')' before string constant drivers/net/wan/ixp4xx_hss.c:1413:25: error: expected ';', ',' or ')' before string constant drivers/net/wan/ixp4xx_hss.c:1414:21: error: expected ';', ',' or ')' before string constant drivers/net/wan/ixp4xx_hss.c:1415:19: error: expected ';', ',' or ')' before string constant
make[8]: *** [drivers/net/wan/ixp4xx_hss.o] Error 1

This was previously hidden because ixp4xx_hss includes linux/hdlc.h which includes linux/netdevice.h which includes linux/netprio_cgroup.h which used to include linux/module.h. The real issue was actually present since the initial commit that added this driver since it uses macros from linux/module.h without including this file.

Signed-off-by Florian Fainelli <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/wan/ixp4xx_hss.c |    1 +
 1 file changed, 1 insertion(+)

drivers/net/wan/ixp4xx_hss.c +1 -0

@@ -8,6 +8,7 @@
  * as published by the Free Software Foundation.
  */
 
+#include <linux/module.h>
 #include <linux/bitops.h>
 #include <linux/cdev.h>
 #include <linux/dma-mapping.h>

Greg Kroah-Hartman

author

2012-10-11 02:24 UTC #22 lore raw

PATCH 22/84 netxen: check for root bus in netxen_mask_aer_correctable

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

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

From: Nikolay Aleksandrov <[email protected]>

[ Upstream commit e4d1aa40e363ed3e0486aeeeb0d173f7f822737e ]

Add a check if pdev->bus->self == NULL (root bus). When attaching a netxen NIC to a VM it can be on the root bus and the guest would crash in netxen_mask_aer_correctable() because of a NULL pointer dereference if CONFIG_PCIEAER is present.

Signed-off-by Nikolay Aleksandrov <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/netxen/netxen_nic_main.c |    4 ++++
 1 file changed, 4 insertions(+)

drivers/net/netxen/netxen_nic_main.c +4 -0

@@ -1288,6 +1288,10 @@ static void netxen_mask_aer_correctable(
 	struct pci_dev *root = pdev->bus->self;
 	u32 aer_pos;
 
+	/* root bus? */
+	if (!root)
+		return;
+
 	if (adapter->ahw.board_type != NETXEN_BRDTYPE_P3_4_GB_MM &&
 		adapter->ahw.board_type != NETXEN_BRDTYPE_P3_10G_TP)
 		return;

Greg Kroah-Hartman

author

2012-10-11 02:13 UTC #23 lore raw

PATCH 23/84 net-sched: sch_cbq: avoid infinite loop

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

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

From: Eric Dumazet <[email protected]>

[ Upstream commit bdfc87f7d1e253e0a61e2fc6a75ea9d76f7fc03a ]

Its possible to setup a bad cbq configuration leading to an infinite loop in cbq_classify()

DEV_OUT=eth0
ICMP="match ip protocol 1 0xff"
U32="protocol ip u32"
DST="match ip dst"
tc qdisc add dev $DEV_OUT root handle 1: cbq avpkt 1000 \
bandwidth 100mbit
tc class add dev $DEV_OUT parent 1: classid 1:1 cbq \
rate 512kbit allot 1500 prio 5 bounded isolated

tc filter add dev $DEV_OUT parent 1: prio 3 $U32 \
	$ICMP $DST 192.168.3.234 flowid 1:

Reported-by Denys Fedoryschenko <[email protected]>

Tested-by Denys Fedoryschenko <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/sched/sch_cbq.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

net/sched/sch_cbq.c +3 -2

@@ -250,10 +250,11 @@ cbq_classify(struct sk_buff *skb, struct
 			else if ((cl = defmap[res.classid & TC_PRIO_MAX]) == NULL)
 				cl = defmap[TC_PRIO_BESTEFFORT];
 
-			if (cl == NULL || cl->level >= head->level)
+			if (cl == NULL)
 				goto fallback;
 		}
-
+		if (cl->level >= head->level)
+			goto fallback;
 #ifdef CONFIG_NET_CLS_ACT
 		switch (result) {
 		case TC_ACT_QUEUED:

Greg Kroah-Hartman

author

2012-10-11 02:13 UTC #24 lore raw

PATCH 24/84 pkt_sched: fix virtual-start-time update in QFQ

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

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

From: Paolo Valente <[email protected]>

[ Upstream commit 71261956973ba9e0637848a5adb4a5819b4bae83 ]

If the old timestamps of a class, say cl, are stale when the class becomes active, then QFQ may assign to cl a much higher start time than the maximum value allowed. This may happen when QFQ assigns to the start time of cl the finish time of a group whose classes are characterized by a higher value of the ratio max_class_pkt/weight_of_the_class with respect to that of cl. Inserting a class with a too high start time into the bucket list corrupts the data structure and may eventually lead to crashes. This patch limits the maximum start time assigned to a class.

Signed-off-by Paolo Valente <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/sched/sch_qfq.c |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

net/sched/sch_qfq.c +4 -1

@@ -829,7 +829,10 @@ static void qfq_update_start(struct qfq_
 		if (mask) {
 			struct qfq_group *next = qfq_ffs(q, mask);
 			if (qfq_gt(roundedF, next->F)) {
-				cl->S = next->F;
+				if (qfq_gt(limit, next->F))
+					cl->S = next->F;
+				else /* preserve timestamp correctness */
+					cl->S = limit;
 				return;
 			}
 		}

Greg Kroah-Hartman

author

2012-10-11 02:12 UTC #25 lore raw

PATCH 25/84 sierra_net: Endianess bug fix.

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

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

From: Lennart Sorensen <[email protected]>

[ Upstream commit 2120c52da6fe741454a60644018ad2a6abd957ac ]

I discovered I couldn't get sierra_net to work on a powerpc. Turns out the firmware attribute check assumes the system is little endian and hence fails because the attributes is a 16 bit value.

Signed-off-by Len Sorensen <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/usb/sierra_net.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/net/usb/sierra_net.c +1 -1

@@ -678,7 +678,7 @@ static int sierra_net_get_fw_attr(struct
 		return -EIO;
 	}
 
-	*datap = *attrdata;
+	*datap = le16_to_cpu(*attrdata);
 
 	kfree(attrdata);
 	return result;

Greg Kroah-Hartman

author

2012-10-11 02:30 UTC #26 lore raw

PATCH 26/84 8021q: fix mac_len recomputation in vlan_untag()

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

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

From: Antonio Quartulli <[email protected]>

[ Upstream commit 5316cf9a5197eb80b2800e1acadde287924ca975 ]

skb_reset_mac_len() relies on the value of the skb->network_header pointer, therefore we must wait for such pointer to be recalculated before computing the new mac_len value.

Signed-off-by Antonio Quartulli <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/8021q/vlan_core.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

net/8021q/vlan_core.c +2 -1

@@ -106,7 +106,6 @@ static struct sk_buff *vlan_reorder_head
 		return NULL;
 	memmove(skb->data - ETH_HLEN, skb->data - VLAN_ETH_HLEN, 2 * ETH_ALEN);
 	skb->mac_header += VLAN_HLEN;
-	skb_reset_mac_len(skb);
 	return skb;
 }

@@ -173,6 +172,8 @@ struct sk_buff *vlan_untag(struct sk_buf
 
 	skb_reset_network_header(skb);
 	skb_reset_transport_header(skb);
+	skb_reset_mac_len(skb);
+
 	return skb;
 
 err_free:

Greg Kroah-Hartman

author

2012-10-11 02:12 UTC #27 lore raw

PATCH 27/84 ipv6: release reference of ip6_null_entrys dst entry in __ip6_del_rt

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

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

From: Gao feng <[email protected]>

[ Upstream commit 6825a26c2dc21eb4f8df9c06d3786ddec97cf53b ]

as we hold dst_entry before we call __ip6_del_rt, so we should alse call dst_release not only return -ENOENT when the rt6_info is ip6_null_entry.

and we already hold the dst entry, so I think it's safe to call dst_release out of the write-read lock.

Signed-off-by Gao feng <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/ipv6/route.c |   11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

net/ipv6/route.c +6 -5

@@ -1399,17 +1399,18 @@ static int __ip6_del_rt(struct rt6_info
 	struct fib6_table *table;
 	struct net *net = dev_net(rt->rt6i_dev);
 
-	if (rt == net->ipv6.ip6_null_entry)
-		return -ENOENT;
+	if (rt == net->ipv6.ip6_null_entry) {
+		err = -ENOENT;
+		goto out;
+	}
 
 	table = rt->rt6i_table;
 	write_lock_bh(&table->tb6_lock);
-
 	err = fib6_del(rt, info);
-	dst_release(&rt->dst);
-
 	write_unlock_bh(&table->tb6_lock);
 
+out:
+	dst_release(&rt->dst);
 	return err;
 }

Greg Kroah-Hartman

author

2012-10-11 02:12 UTC #28 lore raw

PATCH 28/84 tcp: flush DMA queue before sk_wait_data if rcv_wnd is zero

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

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

From: =?UTF-8?q?Michal=20Kube=C4=8Dek?= <[email protected]>

[ Upstream commit 15c041759bfcd9ab0a4e43f1c16e2644977d0467 ]

If recv() syscall is called for a TCP socket so that
- IOAT DMA is used
- MSG_WAITALL flag is used
- requested length is bigger than sk_rcvbuf
- enough data has already arrived to bring rcv_wnd to zero
then when tcp_recvmsg() gets to calling sk_wait_data(), receive window can be still zero while sk_async_wait_queue exhausts enough space to keep it zero. As this queue isn't cleaned until the tcp_service_net_dma() call, sk_wait_data() cannot receive any data and blocks forever.

If zero receive window and non-empty sk_async_wait_queue is detected before calling sk_wait_data(), process the queue first.

Signed-off-by Michal Kubecek <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/ipv4/tcp.c |   10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

net/ipv4/tcp.c +8 -2

@@ -1592,8 +1592,14 @@ int tcp_recvmsg(struct kiocb *iocb, stru
 		}
 
 #ifdef CONFIG_NET_DMA
-		if (tp->ucopy.dma_chan)
-			dma_async_memcpy_issue_pending(tp->ucopy.dma_chan);
+		if (tp->ucopy.dma_chan) {
+			if (tp->rcv_wnd == 0 &&
+			    !skb_queue_empty(&sk->sk_async_wait_queue)) {
+				tcp_service_net_dma(sk, true);
+				tcp_cleanup_rbuf(sk, copied);
+			} else
+				dma_async_memcpy_issue_pending(tp->ucopy.dma_chan);
+		}
 #endif
 		if (copied >= target) {
 			/* Do not sleep, just process backlog. */

Greg Kroah-Hartman

author

2012-10-11 02:12 UTC #29 lore raw

PATCH 29/84 sctp: Dont charge for data in sndbuf again when transmitting packet

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

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

From: Thomas Graf <[email protected]>

[ Upstream commit 4c3a5bdae293f75cdf729c6c00124e8489af2276 ]

SCTP charges wmem_alloc via sctp_set_owner_w() in sctp_sendmsg() and via skb_set_owner_w() in sctp_packet_transmit(). If a sender runs out of sndbuf it will sleep in sctp_wait_for_sndbuf() and expects to be waken up by __sctp_write_space().

Buffer space charged via sctp_set_owner_w() is released in sctp_wfree() which calls __sctp_write_space() directly.

Buffer space charged via skb_set_owner_w() is released via sock_wfree() which calls sk->sk_write_space() _if_ SOCK_USE_WRITE_QUEUE is not set. sctp_endpoint_init() sets SOCK_USE_WRITE_QUEUE on all sockets.

Therefore if sctp_packet_transmit() manages to queue up more than sndbuf bytes, sctp_wait_for_sndbuf() will never be woken up again unless it is interrupted by a signal.

This could be fixed by clearing the SOCK_USE_WRITE_QUEUE flag but ...

Charging for the data twice does not make sense in the first place, it leads to overcharging sndbuf by a factor 2. Therefore this patch only charges a single byte in wmem_alloc when transmitting an SCTP packet to ensure that the socket stays alive until the packet has been released.

This means that control chunks are no longer accounted for in wmem_alloc which I believe is not a problem as skb->truesize will typically lead to overcharging anyway and thus compensates for any control overhead.

Signed-off-by Thomas Graf <[email protected]>

CC Vlad Yasevich <[email protected]>

CC Neil Horman <[email protected]>

CC David Miller <[email protected]>

Acked-by Vlad Yasevich <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/sctp/output.c |   21 ++++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)

net/sctp/output.c +20 -1

@@ -334,6 +334,25 @@ finish:
 	return retval;
 }
 
+static void sctp_packet_release_owner(struct sk_buff *skb)
+{
+	sk_free(skb->sk);
+}
+
+static void sctp_packet_set_owner_w(struct sk_buff *skb, struct sock *sk)
+{
+	skb_orphan(skb);
+	skb->sk = sk;
+	skb->destructor = sctp_packet_release_owner;
+
+	/*
+	 * The data chunks have already been accounted for in sctp_sendmsg(),
+	 * therefore only reserve a single byte to keep socket around until
+	 * the packet has been transmitted.
+	 */
+	atomic_inc(&sk->sk_wmem_alloc);
+}
+
 /* All packets are sent to the network through this function from
  * sctp_outq_tail().
  *

@@ -375,7 +394,7 @@ int sctp_packet_transmit(struct sctp_pac
 	/* Set the owning socket so that we know where to get the
 	 * destination IP address.
 	 */
-	skb_set_owner_w(nskb, sk);
+	sctp_packet_set_owner_w(nskb, sk);
 
 	if (!sctp_transport_dst_check(tp)) {
 		sctp_transport_route(tp, NULL, sctp_sk(sk));

Greg Kroah-Hartman

author

2012-10-11 02:12 UTC #30 lore raw

PATCH 30/84 pppoe: drop PPPOX_ZOMBIEs in pppoe_release

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

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

From: Xiaodong Xu <[email protected]>

[ Upstream commit 2b018d57ff18e5405823e5cb59651a5b4d946d7b ]

When PPPOE is running over a virtual ethernet interface (e.g., a bonding interface) and the user tries to delete the interface in case the PPPOE state is ZOMBIE, the kernel will loop forever while unregistering net_device for the reference count is not decreased to zero which should have been done with dev_put().

Signed-off-by Xiaodong Xu <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/pppoe.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/net/pppoe.c +1 -1

@@ -576,7 +576,7 @@ static int pppoe_release(struct socket *
 
 	po = pppox_sk(sk);
 
-	if (sk->sk_state & (PPPOX_CONNECTED | PPPOX_BOUND)) {
+	if (sk->sk_state & (PPPOX_CONNECTED | PPPOX_BOUND | PPPOX_ZOMBIE)) {
 		dev_put(po->pppoe_dev);
 		po->pppoe_dev = NULL;
 	}

Greg Kroah-Hartman

author

2012-10-11 02:26 UTC #31 lore raw

PATCH 31/84 net: small bug on rxhash calculation

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

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

From: Chema Gonzalez <[email protected]>

[ Upstream commit 6862234238e84648c305526af2edd98badcad1e0 ]

In the current rxhash calculation function, while the sorting of the ports/addrs is coherent (you get the same rxhash for packets sharing the same 4-tuple, in both directions), ports and addrs are sorted independently. This implies packets from a connection between the same addresses but crossed ports hash to the same rxhash.

For example, traffic between A=S:l and B=L:s is hashed (in both directions) from {L, S, {s, l}}. The same rxhash is obtained for packets between C=S:s and D=L:l.

This patch ensures that you either swap both addrs and ports, or you swap none. Traffic between A and B, and traffic between C and D, get their rxhash from different sources ({L, S, {l, s}} for A<->B, and {L, S, {s, l}} for C<->D)

The patch is co-written with Eric Dumazet <[email protected]>

Signed-off-by Chema Gonzalez <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/core/dev.c |   11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

net/core/dev.c +6 -5

@@ -2559,16 +2559,17 @@ __u32 __skb_get_rxhash(struct sk_buff *s
 	poff = proto_ports_offset(ip_proto);
 	if (poff >= 0) {
 		nhoff += ihl * 4 + poff;
-		if (pskb_may_pull(skb, nhoff + 4)) {
+		if (pskb_may_pull(skb, nhoff + 4))
 			ports.v32 = * (__force u32 *) (skb->data + nhoff);
-			if (ports.v16[1] < ports.v16[0])
-				swap(ports.v16[0], ports.v16[1]);
-		}
 	}
 
 	/* get a consistent hash (same value on both flow directions) */
-	if (addr2 < addr1)
+	if (addr2 < addr1 ||
+	    (addr2 == addr1 &&
+	     ports.v16[1] < ports.v16[0])) {
 		swap(addr1, addr2);
+		swap(ports.v16[0], ports.v16[1]);
+	}
 
 	hash = jhash_3words(addr1, addr2, ports.v32, hashrnd);
 	if (!hash)

Greg Kroah-Hartman

author

2012-10-11 02:12 UTC #32 lore raw

PATCH 32/84 net: guard tcp_set_keepalive() to tcp sockets

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

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

From: Eric Dumazet <[email protected]>

[ Upstream commit 3e10986d1d698140747fcfc2761ec9cb64c1d582 ]

Its possible to use RAW sockets to get a crash in tcp_set_keepalive() / sk_reset_timer()

Fix is to make sure socket is a SOCK_STREAM one.

Reported-by Dave Jones <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/core/sock.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

net/core/sock.c +2 -1

@@ -594,7 +594,8 @@ set_rcvbuf:
 
 	case SO_KEEPALIVE:
 #ifdef CONFIG_INET
-		if (sk->sk_protocol == IPPROTO_TCP)
+		if (sk->sk_protocol == IPPROTO_TCP &&
+		    sk->sk_type == SOCK_STREAM)
 			tcp_set_keepalive(sk, valbool);
 #endif
 		sock_valbool_flag(sk, SOCK_KEEPOPEN, valbool);

Greg Kroah-Hartman

author

2012-10-11 02:15 UTC #33 lore raw

PATCH 33/84 ipv4: raw: fix icmp_filter()

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

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

From: Eric Dumazet <[email protected]>

[ Upstream commit ab43ed8b7490cb387782423ecf74aeee7237e591 ]

icmp_filter() should not modify its input, or else its caller would need to recompute ip_hdr() if skb->head is reallocated.

Use skb_header_pointer() instead of pskb_may_pull() and change the prototype to make clear both sk and skb are const.

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/ipv4/raw.c |   14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

net/ipv4/raw.c +8 -6

@@ -130,18 +130,20 @@ found:
  *	0 - deliver
  *	1 - block
  */
-static __inline__ int icmp_filter(struct sock *sk, struct sk_buff *skb)
+static int icmp_filter(const struct sock *sk, const struct sk_buff *skb)
 {
-	int type;
+	struct icmphdr _hdr;
+	const struct icmphdr *hdr;
 
-	if (!pskb_may_pull(skb, sizeof(struct icmphdr)))
+	hdr = skb_header_pointer(skb, skb_transport_offset(skb),
+				 sizeof(_hdr), &_hdr);
+	if (!hdr)
 		return 1;
 
-	type = icmp_hdr(skb)->type;
-	if (type < 32) {
+	if (hdr->type < 32) {
 		__u32 data = raw_sk(sk)->filter.data;
 
-		return ((1 << type) & data) != 0;
+		return ((1U << hdr->type) & data) != 0;
 	}
 
 	/* Do not block unknown ICMP types */

Greg Kroah-Hartman

author

2012-10-11 02:14 UTC #34 lore raw

PATCH 34/84 ipv6: raw: fix icmpv6_filter()

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

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

From: Eric Dumazet <[email protected]>

[ Upstream commit 1b05c4b50edbddbdde715c4a7350629819f6655e ]

icmpv6_filter() should not modify its input, or else its caller would need to recompute ipv6_hdr() if skb->head is reallocated.

Use skb_header_pointer() instead of pskb_may_pull() and change the prototype to make clear both sk and skb are const.

Also, if icmpv6 header cannot be found, do not deliver the packet, as we do in IPv4.

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/ipv6/raw.c |   21 ++++++++++-----------
 1 file changed, 10 insertions(+), 11 deletions(-)

net/ipv6/raw.c +10 ~~-11~~

@@ -106,21 +106,20 @@ found:
  *	0 - deliver
  *	1 - block
  */
-static __inline__ int icmpv6_filter(struct sock *sk, struct sk_buff *skb)
+static int icmpv6_filter(const struct sock *sk, const struct sk_buff *skb)
 {
-	struct icmp6hdr *icmph;
-	struct raw6_sock *rp = raw6_sk(sk);
+	struct icmp6hdr *_hdr;
+	const struct icmp6hdr *hdr;
 
-	if (pskb_may_pull(skb, sizeof(struct icmp6hdr))) {
-		__u32 *data = &rp->filter.data[0];
-		int bit_nr;
+	hdr = skb_header_pointer(skb, skb_transport_offset(skb),
+				 sizeof(_hdr), &_hdr);
+	if (hdr) {
+		const __u32 *data = &raw6_sk(sk)->filter.data[0];
+		unsigned int type = hdr->icmp6_type;
 
-		icmph = (struct icmp6hdr *) skb->data;
-		bit_nr = icmph->icmp6_type;
-
-		return (data[bit_nr >> 5] & (1 << (bit_nr & 31))) != 0;
+		return (data[type >> 5] & (1U << (type & 31))) != 0;
 	}
-	return 0;
+	return 1;
 }
 
 #if defined(CONFIG_IPV6_MIP6) || defined(CONFIG_IPV6_MIP6_MODULE)

Greg Kroah-Hartman

author

2012-10-11 02:14 UTC #35 lore raw

PATCH 35/84 ipv6: mip6: fix mip6_mh_filter()

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

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

From: Eric Dumazet <[email protected]>

[ Upstream commit 96af69ea2a83d292238bdba20e4508ee967cf8cb ]

mip6_mh_filter() should not modify its input, or else its caller would need to recompute ipv6_hdr() if skb->head is reallocated.

Use skb_header_pointer() instead of pskb_may_pull()

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/ipv6/mip6.c |   20 +++++++++++---------
 1 file changed, 11 insertions(+), 9 deletions(-)

net/ipv6/mip6.c +11 -9

@@ -84,28 +84,30 @@ static int mip6_mh_len(int type)
 
 static int mip6_mh_filter(struct sock *sk, struct sk_buff *skb)
 {
-	struct ip6_mh *mh;
+	struct ip6_mh _hdr;
+	const struct ip6_mh *mh;
 
-	if (!pskb_may_pull(skb, (skb_transport_offset(skb)) + 8) ||
-	    !pskb_may_pull(skb, (skb_transport_offset(skb) +
-				 ((skb_transport_header(skb)[1] + 1) << 3))))
+	mh = skb_header_pointer(skb, skb_transport_offset(skb),
+				sizeof(_hdr), &_hdr);
+	if (!mh)
 		return -1;
 
-	mh = (struct ip6_mh *)skb_transport_header(skb);
+	if (((mh->ip6mh_hdrlen + 1) << 3) > skb->len)
+		return -1;
 
 	if (mh->ip6mh_hdrlen < mip6_mh_len(mh->ip6mh_type)) {
 		LIMIT_NETDEBUG(KERN_DEBUG "mip6: MH message too short: %d vs >=%d\n",
 			       mh->ip6mh_hdrlen, mip6_mh_len(mh->ip6mh_type));
-		mip6_param_prob(skb, 0, ((&mh->ip6mh_hdrlen) -
-					 skb_network_header(skb)));
+		mip6_param_prob(skb, 0, offsetof(struct ip6_mh, ip6mh_hdrlen) +
+				skb_network_header_len(skb));
 		return -1;
 	}
 
 	if (mh->ip6mh_proto != IPPROTO_NONE) {
 		LIMIT_NETDEBUG(KERN_DEBUG "mip6: MH invalid payload proto = %d\n",
 			       mh->ip6mh_proto);
-		mip6_param_prob(skb, 0, ((&mh->ip6mh_proto) -
-					 skb_network_header(skb)));
+		mip6_param_prob(skb, 0, offsetof(struct ip6_mh, ip6mh_proto) +
+				skb_network_header_len(skb));
 		return -1;
 	}

Greg Kroah-Hartman

author

2012-10-11 02:14 UTC #36 lore raw

PATCH 36/84 l2tp: fix a typo in l2tp_eth_dev_recv()

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

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

From: Eric Dumazet <[email protected]>

[ Upstream commit c0cc88a7627c333de50b07b7c60b1d49d9d2e6cc ]

While investigating l2tp bug, I hit a bug in eth_type_trans(), because not enough bytes were pulled in skb head.

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/l2tp/l2tp_eth.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

net/l2tp/l2tp_eth.c +1 -1

@@ -132,7 +132,7 @@ static void l2tp_eth_dev_recv(struct l2t
 		printk("\n");
 	}
 
-	if (!pskb_may_pull(skb, sizeof(ETH_HLEN)))
+	if (!pskb_may_pull(skb, ETH_HLEN))
 		goto error;
 
 	secpath_reset(skb);

Greg Kroah-Hartman

author

2012-10-11 02:22 UTC #37 lore raw

PATCH 37/84 netrom: copy_datagram_iovec can fail

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

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

From: Alan Cox <[email protected]>

[ Upstream commit 6cf5c951175abcec4da470c50565cc0afe6cd11d ]

Check for an error from this and if so bail properly.

Signed-off-by Alan Cox <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/netrom/af_netrom.c |    7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

net/netrom/af_netrom.c +6 -1

@@ -1170,7 +1170,12 @@ static int nr_recvmsg(struct kiocb *iocb
 		msg->msg_flags |= MSG_TRUNC;
 	}
 
-	skb_copy_datagram_iovec(skb, 0, msg->msg_iov, copied);
+	er = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, copied);
+	if (er < 0) {
+		skb_free_datagram(sk, skb);
+		release_sock(sk);
+		return er;
+	}
 
 	if (sax != NULL) {
 		sax->sax25_family = AF_NETROM;

Greg Kroah-Hartman

author

2012-10-11 02:14 UTC #38 lore raw

PATCH 38/84 net: do not disable sg for packets requiring no checksum

http://www.spinics.net/lists/linux-mm/msg15184.html

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

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

From: Ed Cashin <[email protected]>

[ Upstream commit c0d680e577ff171e7b37dbdb1b1bf5451e851f04 ]

A change in a series of VLAN-related changes appears to have inadvertently disabled the use of the scatter gather feature of network cards for transmission of non-IP ethernet protocols like ATA over Ethernet (AoE). Below is a reference to the commit that introduces a "harmonize_features" function that turns off scatter gather when the NIC does not support hardware checksumming for the ethernet protocol of an sk buff.

commit f01a5236bd4b140198fbcc550f085e8361fd73fa
Author: Jesse Gross <[email protected]>
Date: Sun Jan 9 06:23:31 2011 +0000

net offloading: Generalize netif_get_vlan_features().

The can_checksum_protocol function is not equipped to consider a protocol that does not require checksumming. Calling it for a protocol that requires no checksum is inappropriate.

The patch below has harmonize_features call can_checksum_protocol when the protocol needs a checksum, so that the network layer is not forced to perform unnecessary skb linearization on the transmission of AoE packets. Unnecessary linearization results in decreased performance and increased memory pressure, as reported here:

The problem has probably not been widely experienced yet, because only recently has the kernel.org-distributed aoe driver acquired the ability to use payloads of over a page in size, with the patchset recently included in the mm tree:

https://lkml.org/lkml/2012/8/28/140

The coraid.com-distributed aoe driver already could use payloads of greater than a page in size, but its users generally do not use the newest kernels.

Signed-off-by Ed Cashin <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 net/core/dev.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

net/core/dev.c +2 -1

@@ -2038,7 +2038,8 @@ static bool can_checksum_protocol(unsign
 
 static u32 harmonize_features(struct sk_buff *skb, __be16 protocol, u32 features)
 {
-	if (!can_checksum_protocol(features, protocol)) {
+	if (skb->ip_summed != CHECKSUM_NONE &&
+	    !can_checksum_protocol(features, protocol)) {
 		features &= ~NETIF_F_ALL_CSUM;
 		features &= ~NETIF_F_SG;
 	} else if (illegal_highdma(skb->dev, skb)) {

Greg Kroah-Hartman

author

2012-10-11 02:22 UTC #39 lore raw

PATCH 39/84 aoe: assert AoE packets marked as requiring no checksum

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

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

From: Ed Cashin <[email protected]>

[ Upstream commit 8babe8cc6570ed896b7b596337eb8fe730c3ff45 ]

In order for the network layer to see that AoE requires no checksumming in a generic way, the packets must be marked as requiring no checksum, so we make this requirement explicit with the assertion.

Signed-off-by Ed Cashin <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/block/aoe/aoecmd.c |    1 +
 1 file changed, 1 insertion(+)

drivers/block/aoe/aoecmd.c +1 -0

@@ -35,6 +35,7 @@ new_skb(ulong len)
 		skb_reset_mac_header(skb);
 		skb_reset_network_header(skb);
 		skb->protocol = __constant_htons(ETH_P_AOE);
+		skb_checksum_none_assert(skb);
 	}
 	return skb;
 }

Greg Kroah-Hartman

author

2012-10-11 02:14 UTC #40 lore raw

PATCH 40/84 tg3: Fix TSO CAP for 5704 devs w / ASF enabled

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

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

From: Matt Carlson <[email protected]>

[ Upstream commit cf9ecf4b631f649a964fa611f1a5e8874f2a76db ]

On the earliest TSO capable devices, TSO was accomplished through firmware. The TSO cannot coexist with ASF management firmware though. The tg3 driver determines whether or not ASF is enabled by calling tg3_get_eeprom_hw_cfg(), which checks a particular bit of NIC memory. Commit dabc5c670d3f86d15ee4f42ab38ec5bd2682487d, entitled "tg3: Move TSO_CAPABLE assignment", accidentally moved the code that determines TSO capabilities earlier than the call to tg3_get_eeprom_hw_cfg(). As a consequence, the driver was attempting to determine TSO capabilities before it had all the data it needed to make the decision.

This patch fixes the problem by revisiting and reevaluating the decision after tg3_get_eeprom_hw_cfg() is called.

Signed-off-by Matt Carlson <[email protected]>

Signed-off-by Michael Chan <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/tg3.c |   14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

drivers/net/tg3.c +12 -2

@@ -13647,9 +13647,13 @@ static int __devinit tg3_get_invariants(
 	if (tg3_flag(tp, HW_TSO_1) ||
 	    tg3_flag(tp, HW_TSO_2) ||
 	    tg3_flag(tp, HW_TSO_3) ||
-	    (tp->fw_needed && !tg3_flag(tp, ENABLE_ASF)))
+	    tp->fw_needed) {
+		/* For firmware TSO, assume ASF is disabled.
+		 * We'll disable TSO later if we discover ASF
+		 * is enabled in tg3_get_eeprom_hw_cfg().
+		 */
 		tg3_flag_set(tp, TSO_CAPABLE);
-	else {
+	} else {
 		tg3_flag_clear(tp, TSO_CAPABLE);
 		tg3_flag_clear(tp, TSO_BUG);
 		tp->fw_needed = NULL;

@@ -13887,6 +13891,12 @@ static int __devinit tg3_get_invariants(
 	 */
 	tg3_get_eeprom_hw_cfg(tp);
 
+	if (tp->fw_needed && tg3_flag(tp, ENABLE_ASF)) {
+		tg3_flag_clear(tp, TSO_CAPABLE);
+		tg3_flag_clear(tp, TSO_BUG);
+		tp->fw_needed = NULL;
+	}
+
 	if (tg3_flag(tp, ENABLE_APE)) {
 		/* Allow reads and writes to the
 		 * APE register and memory space.

Greg Kroah-Hartman

author

2012-10-11 02:21 UTC #41 lore raw

PATCH 41/84 SCSI: zfcp: Make trace record tags unique

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

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

From: Steffen Maier <[email protected]>

commit 0100998dbfe6dfcd90a6e912ca7ed6f255d48f25 upstream.

Duplicate fssrh_2 from a54ca0f62f953898b05549391ac2a8a4dad6482b "[SCSI] zfcp: Redesign of the debug tracing for HBA records." complicates distinction of generic status read response from local link up.
Duplicate fsscth1 from 2c55b750a884b86dea8b4cc5f15e1484cc47a25c "[SCSI] zfcp: Redesign of the debug tracing for SAN records." complicates distinction of good common transport response from invalid port handle.

Reviewed-by Martin Peschke <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/s390/scsi/zfcp_fsf.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

drivers/s390/scsi/zfcp_fsf.c +2 -2

@@ -219,7 +219,7 @@ static void zfcp_fsf_status_read_handler
 		return;
 	}
 
-	zfcp_dbf_hba_fsf_uss("fssrh_2", req);
+	zfcp_dbf_hba_fsf_uss("fssrh_4", req);
 
 	switch (sr_buf->status_type) {
 	case FSF_STATUS_READ_PORT_CLOSED:

@@ -885,7 +885,7 @@ static void zfcp_fsf_send_ct_handler(str
 
 	switch (header->fsf_status) {
         case FSF_GOOD:
-		zfcp_dbf_san_res("fsscth1", req);
+		zfcp_dbf_san_res("fsscth2", req);
 		ct->status = 0;
 		break;
         case FSF_SERVICE_CLASS_NOT_SUPPORTED:

Greg Kroah-Hartman

author

2012-10-11 02:21 UTC #42 lore raw

PATCH 42/84 SCSI: zfcp: Do not wakeup while suspended

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

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

From: Steffen Maier <[email protected]>

commit cb45214960bc989af8b911ebd77da541c797717d upstream.

If the mapping of FCP device bus ID and corresponding subchannel is modified while the Linux image is suspended, the resume of FCP devices can fail. During resume, zfcp gets callbacks from cio regarding the modified subchannels but they can be arbitrarily mixed with the restore/resume callback. Since the cio callbacks would trigger adapter recovery, zfcp could wakeup before the resume callback. Therefore, ignore the cio callbacks regarding subchannels while being suspended. We can safely do so, since zfcp does not deal itself with subchannels. For problem determination purposes, we still trace the ignored callback events.

The following kernel messages could be seen on resume:

kernel: <WWPN>: parent <FCP device bus ID> should not be sleeping

As part of adapter reopen recovery, zfcp performs auto port scanning which can erroneously try to register new remote ports with scsi_transport_fc and the device core code complains about the parent (adapter) still sleeping.

kernel: zfcp.3dff9c: <FCP device bus ID>:\
Setting up the QDIO connection to the FCP adapter failed
<last kernel message repeated 3 more times>
kernel: zfcp.574d43: <FCP device bus ID>:\
ERP cannot recover an error on the FCP device

In such cases, the adapter gave up recovery and remained blocked along with its child objects: remote ports and LUNs/scsi devices. Even the adapter shutdown as part of giving up recovery failed because the ccw device state remained disconnected. Later, the corresponding remote ports ran into dev_loss_tmo. As a result, the LUNs were erroneously not available again after resume.

Even a manually triggered adapter recovery (e.g. sysfs attribute failed, or device offline/online via sysfs) could not recover the adapter due to the remaining disconnected state of the corresponding ccw device.

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/s390/scsi/zfcp_ccw.c |   73 +++++++++++++++++++++++++++++++++++++------
 drivers/s390/scsi/zfcp_dbf.c |   20 +++++++++++
 drivers/s390/scsi/zfcp_dbf.h |    1 
 drivers/s390/scsi/zfcp_def.h |    1 
 drivers/s390/scsi/zfcp_ext.h |    1 
 5 files changed, 86 insertions(+), 10 deletions(-)

drivers/s390/scsi/zfcp_ccw.c +63 ~~-10~~

@@ -38,17 +38,23 @@ void zfcp_ccw_adapter_put(struct zfcp_ad
 	spin_unlock_irqrestore(&zfcp_ccw_adapter_ref_lock, flags);
 }
 
-static int zfcp_ccw_activate(struct ccw_device *cdev)
-
+/**
+ * zfcp_ccw_activate - activate adapter and wait for it to finish
+ * @cdev: pointer to belonging ccw device
+ * @clear: Status flags to clear.
+ * @tag: s390dbf trace record tag
+ */
+static int zfcp_ccw_activate(struct ccw_device *cdev, int clear, char *tag)
 {
 	struct zfcp_adapter *adapter = zfcp_ccw_adapter_by_cdev(cdev);
 
 	if (!adapter)
 		return 0;
 
+	zfcp_erp_clear_adapter_status(adapter, clear);
 	zfcp_erp_set_adapter_status(adapter, ZFCP_STATUS_COMMON_RUNNING);
 	zfcp_erp_adapter_reopen(adapter, ZFCP_STATUS_COMMON_ERP_FAILED,
-				"ccresu2");
+				tag);
 	zfcp_erp_wait(adapter);
 	flush_work(&adapter->scan_work);

@@ -163,26 +169,29 @@ static int zfcp_ccw_set_online(struct cc
 	BUG_ON(!zfcp_reqlist_isempty(adapter->req_list));
 	adapter->req_no = 0;
 
-	zfcp_ccw_activate(cdev);
+	zfcp_ccw_activate(cdev, 0, "ccsonl1");
 	zfcp_ccw_adapter_put(adapter);
 	return 0;
 }
 
 /**
- * zfcp_ccw_set_offline - set_offline function of zfcp driver
+ * zfcp_ccw_offline_sync - shut down adapter and wait for it to finish
  * @cdev: pointer to belonging ccw device
+ * @set: Status flags to set.
+ * @tag: s390dbf trace record tag
  *
  * This function gets called by the common i/o layer and sets an adapter
  * into state offline.
  */
-static int zfcp_ccw_set_offline(struct ccw_device *cdev)
+static int zfcp_ccw_offline_sync(struct ccw_device *cdev, int set, char *tag)
 {
 	struct zfcp_adapter *adapter = zfcp_ccw_adapter_by_cdev(cdev);
 
 	if (!adapter)
 		return 0;
 
-	zfcp_erp_adapter_shutdown(adapter, 0, "ccsoff1");
+	zfcp_erp_set_adapter_status(adapter, set);
+	zfcp_erp_adapter_shutdown(adapter, 0, tag);
 	zfcp_erp_wait(adapter);
 
 	zfcp_ccw_adapter_put(adapter);

@@ -190,6 +199,18 @@ static int zfcp_ccw_set_offline(struct c
 }
 
 /**
+ * zfcp_ccw_set_offline - set_offline function of zfcp driver
+ * @cdev: pointer to belonging ccw device
+ *
+ * This function gets called by the common i/o layer and sets an adapter
+ * into state offline.
+ */
+static int zfcp_ccw_set_offline(struct ccw_device *cdev)
+{
+	return zfcp_ccw_offline_sync(cdev, 0, "ccsoff1");
+}
+
+/**
  * zfcp_ccw_notify - ccw notify function
  * @cdev: pointer to belonging ccw device
  * @event: indicates if adapter was detached or attached

@@ -206,6 +227,11 @@ static int zfcp_ccw_notify(struct ccw_de
 
 	switch (event) {
 	case CIO_GONE:
+		if (atomic_read(&adapter->status) &
+		    ZFCP_STATUS_ADAPTER_SUSPENDED) { /* notification ignore */
+			zfcp_dbf_hba_basic("ccnigo1", adapter);
+			break;
+		}
 		dev_warn(&cdev->dev, "The FCP device has been detached\n");
 		zfcp_erp_adapter_shutdown(adapter, 0, "ccnoti1");
 		break;

@@ -215,6 +241,11 @@ static int zfcp_ccw_notify(struct ccw_de
 		zfcp_erp_adapter_shutdown(adapter, 0, "ccnoti2");
 		break;
 	case CIO_OPER:
+		if (atomic_read(&adapter->status) &
+		    ZFCP_STATUS_ADAPTER_SUSPENDED) { /* notification ignore */
+			zfcp_dbf_hba_basic("ccniop1", adapter);
+			break;
+		}
 		dev_info(&cdev->dev, "The FCP device is operational again\n");
 		zfcp_erp_set_adapter_status(adapter,
 					    ZFCP_STATUS_COMMON_RUNNING);

@@ -250,6 +281,28 @@ static void zfcp_ccw_shutdown(struct ccw
 	zfcp_ccw_adapter_put(adapter);
 }
 
+static int zfcp_ccw_suspend(struct ccw_device *cdev)
+{
+	zfcp_ccw_offline_sync(cdev, ZFCP_STATUS_ADAPTER_SUSPENDED, "ccsusp1");
+	return 0;
+}
+
+static int zfcp_ccw_thaw(struct ccw_device *cdev)
+{
+	/* trace records for thaw and final shutdown during suspend
+	   can only be found in system dump until the end of suspend
+	   but not after resume because it's based on the memory image
+	   right after the very first suspend (freeze) callback */
+	zfcp_ccw_activate(cdev, 0, "ccthaw1");
+	return 0;
+}
+
+static int zfcp_ccw_resume(struct ccw_device *cdev)
+{
+	zfcp_ccw_activate(cdev, ZFCP_STATUS_ADAPTER_SUSPENDED, "ccresu1");
+	return 0;
+}
+
 struct ccw_driver zfcp_ccw_driver = {
 	.driver = {
 		.owner	= THIS_MODULE,

@@ -262,7 +315,7 @@ struct ccw_driver zfcp_ccw_driver = {
 	.set_offline = zfcp_ccw_set_offline,
 	.notify      = zfcp_ccw_notify,
 	.shutdown    = zfcp_ccw_shutdown,
-	.freeze      = zfcp_ccw_set_offline,
-	.thaw	     = zfcp_ccw_activate,
-	.restore     = zfcp_ccw_activate,
+	.freeze      = zfcp_ccw_suspend,
+	.thaw	     = zfcp_ccw_thaw,
+	.restore     = zfcp_ccw_resume,
 };

drivers/s390/scsi/zfcp_dbf.c +20 -0

@@ -163,6 +163,26 @@ void zfcp_dbf_hba_bit_err(char *tag, str
 	spin_unlock_irqrestore(&dbf->hba_lock, flags);
 }
 
+/**
+ * zfcp_dbf_hba_basic - trace event for basic adapter events
+ * @adapter: pointer to struct zfcp_adapter
+ */
+void zfcp_dbf_hba_basic(char *tag, struct zfcp_adapter *adapter)
+{
+	struct zfcp_dbf *dbf = adapter->dbf;
+	struct zfcp_dbf_hba *rec = &dbf->hba_buf;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dbf->hba_lock, flags);
+	memset(rec, 0, sizeof(*rec));
+
+	memcpy(rec->tag, tag, ZFCP_DBF_TAG_LEN);
+	rec->id = ZFCP_DBF_HBA_BASIC;
+
+	debug_event(dbf->hba, 1, rec, sizeof(*rec));
+	spin_unlock_irqrestore(&dbf->hba_lock, flags);
+}
+
 static void zfcp_dbf_set_common(struct zfcp_dbf_rec *rec,
 				struct zfcp_adapter *adapter,
 				struct zfcp_port *port,

drivers/s390/scsi/zfcp_dbf.h +1 -0

@@ -154,6 +154,7 @@ enum zfcp_dbf_hba_id {
 	ZFCP_DBF_HBA_RES	= 1,
 	ZFCP_DBF_HBA_USS	= 2,
 	ZFCP_DBF_HBA_BIT	= 3,
+	ZFCP_DBF_HBA_BASIC	= 4,
 };
 
 /**

drivers/s390/scsi/zfcp_def.h +1 -0

@@ -76,6 +76,7 @@ struct zfcp_reqlist;
 #define ZFCP_STATUS_ADAPTER_SIOSL_ISSUED	0x00000004
 #define ZFCP_STATUS_ADAPTER_XCONFIG_OK		0x00000008
 #define ZFCP_STATUS_ADAPTER_HOST_CON_INIT	0x00000010
+#define ZFCP_STATUS_ADAPTER_SUSPENDED		0x00000040
 #define ZFCP_STATUS_ADAPTER_ERP_PENDING		0x00000100
 #define ZFCP_STATUS_ADAPTER_LINK_UNPLUGGED	0x00000200
 #define ZFCP_STATUS_ADAPTER_DATA_DIV_ENABLED	0x00000400

drivers/s390/scsi/zfcp_ext.h +1 -0

@@ -53,6 +53,7 @@ extern void zfcp_dbf_hba_fsf_uss(char *,
 extern void zfcp_dbf_hba_fsf_res(char *, struct zfcp_fsf_req *);
 extern void zfcp_dbf_hba_bit_err(char *, struct zfcp_fsf_req *);
 extern void zfcp_dbf_hba_berr(struct zfcp_dbf *, struct zfcp_fsf_req *);
+extern void zfcp_dbf_hba_basic(char *, struct zfcp_adapter *);
 extern void zfcp_dbf_san_req(char *, struct zfcp_fsf_req *, u32);
 extern void zfcp_dbf_san_res(char *, struct zfcp_fsf_req *);
 extern void zfcp_dbf_san_in_els(char *, struct zfcp_fsf_req *);

Greg Kroah-Hartman

author

2012-10-11 02:13 UTC #43 lore raw

PATCH 43/84 SCSI: zfcp: remove invalid reference to list iterator variable

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

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

From: Julia Lawall <[email protected]>

commit ca579c9f136af4274ccfd1bcaee7f38a29a0e2e9 upstream.

If list_for_each_entry, etc complete a traversal of the list, the iterator variable ends up pointing to an address at an offset from the list head, and not a meaningful structure. Thus this value should not be used after the end of the iterator. Replace port->adapter->scsi_host by adapter->scsi_host.

This problem was found using Coccinelle (http://coccinelle.lip6.fr/).

Oversight in upsteam commit of v2.6.37 a1ca48319a9aa1c5b57ce142f538e76050bb8972
"[SCSI] zfcp: Move ACL/CFDC code to zfcp_cfdc.c" which merged the content of zfcp_erp_port_access_changed().

Signed-off-by Julia Lawall <[email protected]>

Reviewed-by Martin Peschke <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/s390/scsi/zfcp_cfdc.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/s390/scsi/zfcp_cfdc.c +1 -1

@@ -293,7 +293,7 @@ void zfcp_cfdc_adapter_access_changed(st
 	}
 	read_unlock_irqrestore(&adapter->port_list_lock, flags);
 
-	shost_for_each_device(sdev, port->adapter->scsi_host) {
+	shost_for_each_device(sdev, adapter->scsi_host) {
 		zfcp_sdev = sdev_to_zfcp(sdev);
 		status = atomic_read(&zfcp_sdev->status);
 		if ((status & ZFCP_STATUS_COMMON_ACCESS_DENIED) ||

Greg Kroah-Hartman

author

2012-10-11 02:24 UTC #44 lore raw

PATCH 44/84 SCSI: zfcp: restore refcount check on port_remove

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

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

From: Steffen Maier <[email protected]>

commit d99b601b63386f3395dc26a699ae703a273d9982 upstream.

Upstream commit f3450c7b917201bb49d67032e9f60d5125675d6a "[SCSI] zfcp: Replace local reference counting with common kref" accidentally dropped a reference count check before tearing down zfcp_ports that are potentially in use by zfcp_units. Even remote ports in use can be removed causing unreachable garbage objects zfcp_ports with zfcp_units. Thus units won't come back even after a manual port_rescan. The kref of zfcp_port->dev.kobj is already used by the driver core. We cannot re-use it to track the number of zfcp_units. Re-introduce our own counter for units per port
and check on port_remove.

Reviewed-by Heiko Carstens <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/s390/scsi/zfcp_aux.c   |    1 +
 drivers/s390/scsi/zfcp_def.h   |    1 +
 drivers/s390/scsi/zfcp_ext.h   |    1 +
 drivers/s390/scsi/zfcp_sysfs.c |   18 ++++++++++++++++--
 drivers/s390/scsi/zfcp_unit.c  |   36 ++++++++++++++++++++++++++----------
 5 files changed, 45 insertions(+), 12 deletions(-)

drivers/s390/scsi/zfcp_aux.c +1 -0

@@ -518,6 +518,7 @@ struct zfcp_port *zfcp_port_enqueue(stru
 
 	rwlock_init(&port->unit_list_lock);
 	INIT_LIST_HEAD(&port->unit_list);
+	atomic_set(&port->units, 0);
 
 	INIT_WORK(&port->gid_pn_work, zfcp_fc_port_did_lookup);
 	INIT_WORK(&port->test_link_work, zfcp_fc_link_test_work);

drivers/s390/scsi/zfcp_def.h +1 -0

@@ -204,6 +204,7 @@ struct zfcp_port {
 	struct zfcp_adapter    *adapter;       /* adapter used to access port */
 	struct list_head	unit_list;	/* head of logical unit list */
 	rwlock_t		unit_list_lock; /* unit list lock */
+	atomic_t		units;	       /* zfcp_unit count */
 	atomic_t	       status;	       /* status of this remote port */
 	u64		       wwnn;	       /* WWNN if known */
 	u64		       wwpn;	       /* WWPN */

drivers/s390/scsi/zfcp_ext.h +1 -0

@@ -158,6 +158,7 @@ extern void zfcp_scsi_dif_sense_error(st
 extern struct attribute_group zfcp_sysfs_unit_attrs;
 extern struct attribute_group zfcp_sysfs_adapter_attrs;
 extern struct attribute_group zfcp_sysfs_port_attrs;
+extern struct mutex zfcp_sysfs_port_units_mutex;
 extern struct device_attribute *zfcp_sysfs_sdev_attrs[];
 extern struct device_attribute *zfcp_sysfs_shost_attrs[];

drivers/s390/scsi/zfcp_sysfs.c +16 -2

@@ -227,6 +227,8 @@ static ssize_t zfcp_sysfs_port_rescan_st
 static ZFCP_DEV_ATTR(adapter, port_rescan, S_IWUSR, NULL,
 		     zfcp_sysfs_port_rescan_store);
 
+DEFINE_MUTEX(zfcp_sysfs_port_units_mutex);
+
 static ssize_t zfcp_sysfs_port_remove_store(struct device *dev,
 					    struct device_attribute *attr,
 					    const char *buf, size_t count)

@@ -249,6 +251,16 @@ static ssize_t zfcp_sysfs_port_remove_st
 	else
 		retval = 0;
 
+	mutex_lock(&zfcp_sysfs_port_units_mutex);
+	if (atomic_read(&port->units) > 0) {
+		retval = -EBUSY;
+		mutex_unlock(&zfcp_sysfs_port_units_mutex);
+		goto out;
+	}
+	/* port is about to be removed, so no more unit_add */
+	atomic_set(&port->units, -1);
+	mutex_unlock(&zfcp_sysfs_port_units_mutex);
+
 	write_lock_irq(&adapter->port_list_lock);
 	list_del(&port->list);
 	write_unlock_irq(&adapter->port_list_lock);

@@ -289,12 +301,14 @@ static ssize_t zfcp_sysfs_unit_add_store
 {
 	struct zfcp_port *port = container_of(dev, struct zfcp_port, dev);
 	u64 fcp_lun;
+	int retval;
 
 	if (strict_strtoull(buf, 0, (unsigned long long *) &fcp_lun))
 		return -EINVAL;
 
-	if (zfcp_unit_add(port, fcp_lun))
-		return -EINVAL;
+	retval = zfcp_unit_add(port, fcp_lun);
+	if (retval)
+		return retval;
 
 	return count;
 }

drivers/s390/scsi/zfcp_unit.c +26 ~~-10~~

@@ -104,7 +104,7 @@ static void zfcp_unit_release(struct dev
 {
 	struct zfcp_unit *unit = container_of(dev, struct zfcp_unit, dev);
 
-	put_device(&unit->port->dev);
+	atomic_dec(&unit->port->units);
 	kfree(unit);
 }

@@ -119,16 +119,27 @@ static void zfcp_unit_release(struct dev
 int zfcp_unit_add(struct zfcp_port *port, u64 fcp_lun)
 {
 	struct zfcp_unit *unit;
+	int retval = 0;
+
+	mutex_lock(&zfcp_sysfs_port_units_mutex);
+	if (atomic_read(&port->units) == -1) {
+		/* port is already gone */
+		retval = -ENODEV;
+		goto out;
+	}
 
 	unit = zfcp_unit_find(port, fcp_lun);
 	if (unit) {
 		put_device(&unit->dev);
-		return -EEXIST;
+		retval = -EEXIST;
+		goto out;
 	}
 
 	unit = kzalloc(sizeof(struct zfcp_unit), GFP_KERNEL);
-	if (!unit)
-		return -ENOMEM;
+	if (!unit) {
+		retval = -ENOMEM;
+		goto out;
+	}
 
 	unit->port = port;
 	unit->fcp_lun = fcp_lun;

@@ -139,28 +150,33 @@ int zfcp_unit_add(struct zfcp_port *port
 	if (dev_set_name(&unit->dev, "0x%016llx",
 			 (unsigned long long) fcp_lun)) {
 		kfree(unit);
-		return -ENOMEM;
+		retval = -ENOMEM;
+		goto out;
 	}
 
-	get_device(&port->dev);
-
 	if (device_register(&unit->dev)) {
 		put_device(&unit->dev);
-		return -ENOMEM;
+		retval = -ENOMEM;
+		goto out;
 	}
 
 	if (sysfs_create_group(&unit->dev.kobj, &zfcp_sysfs_unit_attrs)) {
 		device_unregister(&unit->dev);
-		return -EINVAL;
+		retval = -EINVAL;
+		goto out;
 	}
 
+	atomic_inc(&port->units); /* under zfcp_sysfs_port_units_mutex ! */
+
 	write_lock_irq(&port->unit_list_lock);
 	list_add_tail(&unit->list, &port->unit_list);
 	write_unlock_irq(&port->unit_list_lock);
 
 	zfcp_unit_scsi_scan(unit);
 
-	return 0;
+out:
+	mutex_unlock(&zfcp_sysfs_port_units_mutex);
+	return retval;
 }
 
 /**

Greg Kroah-Hartman

author

2012-10-11 02:14 UTC #45 lore raw

PATCH 45/84 SCSI: zfcp: only access zfcp_scsi_dev for valid scsi_device

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

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

From: Martin Peschke <[email protected]>

commit d436de8ce25f53a8a880a931886821f632247943 upstream.

__scsi_remove_device (e.g. due to dev_loss_tmo) calls zfcp_scsi_slave_destroy which in turn sends a close LUN FSF request to the adapter. After 30 seconds without response, zfcp_erp_timeout_handler kicks the ERP thread failing the close LUN ERP action. zfcp_erp_wait in zfcp_erp_lun_shutdown_wait and thus zfcp_scsi_slave_destroy returns and then scsi_device is no longer valid. Sometime later the response to the close LUN FSF request may finally come in. However, commit b62a8d9b45b971a67a0f8413338c230e3117dff5
"[SCSI] zfcp: Use SCSI device data zfcp_scsi_dev instead of zfcp_unit" introduced a number of attempts to unconditionally access struct zfcp_scsi_dev through struct scsi_device causing a use-after-free. This leads to an Oops due to kernel page fault in one of:
zfcp_fsf_abort_fcp_command_handler, zfcp_fsf_open_lun_handler, zfcp_fsf_close_lun_handler, zfcp_fsf_req_trace, zfcp_fsf_fcp_handler_common.
Move dereferencing of zfcp private data zfcp_scsi_dev allocated in scsi_device via scsi_transport_reserve_device after the check for potentially aborted FSF request and thus no longer valid scsi_device. Only then assign sdev_to_zfcp(sdev) to the local auto variable struct zfcp_scsi_dev *zfcp_sdev.

Signed-off-by Martin Peschke <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/s390/scsi/zfcp_fsf.c |   19 ++++++++++++++-----
 1 file changed, 14 insertions(+), 5 deletions(-)

drivers/s390/scsi/zfcp_fsf.c +14 -5

@@ -771,12 +771,14 @@ out:
 static void zfcp_fsf_abort_fcp_command_handler(struct zfcp_fsf_req *req)
 {
 	struct scsi_device *sdev = req->data;
-	struct zfcp_scsi_dev *zfcp_sdev = sdev_to_zfcp(sdev);
+	struct zfcp_scsi_dev *zfcp_sdev;
 	union fsf_status_qual *fsq = &req->qtcb->header.fsf_status_qual;
 
 	if (req->status & ZFCP_STATUS_FSFREQ_ERROR)
 		return;
 
+	zfcp_sdev = sdev_to_zfcp(sdev);
+
 	switch (req->qtcb->header.fsf_status) {
 	case FSF_PORT_HANDLE_NOT_VALID:
 		if (fsq->word[0] == fsq->word[1]) {

@@ -1730,13 +1732,15 @@ static void zfcp_fsf_open_lun_handler(st
 {
 	struct zfcp_adapter *adapter = req->adapter;
 	struct scsi_device *sdev = req->data;
-	struct zfcp_scsi_dev *zfcp_sdev = sdev_to_zfcp(sdev);
+	struct zfcp_scsi_dev *zfcp_sdev;
 	struct fsf_qtcb_header *header = &req->qtcb->header;
 	struct fsf_qtcb_bottom_support *bottom = &req->qtcb->bottom.support;
 
 	if (req->status & ZFCP_STATUS_FSFREQ_ERROR)
 		return;
 
+	zfcp_sdev = sdev_to_zfcp(sdev);
+
 	atomic_clear_mask(ZFCP_STATUS_COMMON_ACCESS_DENIED |
 			  ZFCP_STATUS_COMMON_ACCESS_BOXED |
 			  ZFCP_STATUS_LUN_SHARED |

@@ -1847,11 +1851,13 @@ out:
 static void zfcp_fsf_close_lun_handler(struct zfcp_fsf_req *req)
 {
 	struct scsi_device *sdev = req->data;
-	struct zfcp_scsi_dev *zfcp_sdev = sdev_to_zfcp(sdev);
+	struct zfcp_scsi_dev *zfcp_sdev;
 
 	if (req->status & ZFCP_STATUS_FSFREQ_ERROR)
 		return;
 
+	zfcp_sdev = sdev_to_zfcp(sdev);
+
 	switch (req->qtcb->header.fsf_status) {
 	case FSF_PORT_HANDLE_NOT_VALID:
 		zfcp_erp_adapter_reopen(zfcp_sdev->port->adapter, 0, "fscuh_1");

@@ -1941,7 +1947,7 @@ static void zfcp_fsf_req_trace(struct zf
 {
 	struct fsf_qual_latency_info *lat_in;
 	struct latency_cont *lat = NULL;
-	struct zfcp_scsi_dev *zfcp_sdev = sdev_to_zfcp(scsi->device);
+	struct zfcp_scsi_dev *zfcp_sdev;
 	struct zfcp_blk_drv_data blktrc;
 	int ticks = req->adapter->timer_ticks;

@@ -1956,6 +1962,7 @@ static void zfcp_fsf_req_trace(struct zf
 
 	if (req->adapter->adapter_features & FSF_FEATURE_MEASUREMENT_DATA &&
 	    !(req->status & ZFCP_STATUS_FSFREQ_ERROR)) {
+		zfcp_sdev = sdev_to_zfcp(scsi->device);
 		blktrc.flags |= ZFCP_BLK_LAT_VALID;
 		blktrc.channel_lat = lat_in->channel_lat * ticks;
 		blktrc.fabric_lat = lat_in->fabric_lat * ticks;

@@ -1993,12 +2000,14 @@ static void zfcp_fsf_fcp_handler_common(
 {
 	struct scsi_cmnd *scmnd = req->data;
 	struct scsi_device *sdev = scmnd->device;
-	struct zfcp_scsi_dev *zfcp_sdev = sdev_to_zfcp(sdev);
+	struct zfcp_scsi_dev *zfcp_sdev;
 	struct fsf_qtcb_header *header = &req->qtcb->header;
 
 	if (unlikely(req->status & ZFCP_STATUS_FSFREQ_ERROR))
 		return;
 
+	zfcp_sdev = sdev_to_zfcp(sdev);
+
 	switch (header->fsf_status) {
 	case FSF_HANDLE_MISMATCH:
 	case FSF_PORT_HANDLE_NOT_VALID:

Greg Kroah-Hartman

author

2012-10-11 02:14 UTC #46 lore raw

PATCH 46/84 PCI: Check P2P bridge for invalid secondary/subordinate range

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

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

From: Yinghai Lu <[email protected]>

commit 1965f66e7db08d1ebccd24a59043eba826cc1ce8 upstream.

For bridges with "secondary > subordinate", i.e., invalid bus number apertures, we don't enumerate anything behind the bridge unless the user specified "pci=assign-busses".

This patch makes us automatically try to reassign the downstream bus numbers in this case (just for that bridge, not for all bridges as "pci=assign-busses" does).

We don't discover all the devices on the Intel DP43BF motherboard without this change (or "pci=assign-busses") because its BIOS configures a bridge as:

    pci 0000:00:1e.0: PCI bridge to [bus 20-08] (subtractive decode)

[bhelgaas: changelog, change message to dev_info]
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=18412
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=625754

Reported-by Brian C. Huffman <[email protected]>

Reported-by VL <[email protected]>

Tested-by VL <[email protected]>

Signed-off-by Yinghai Lu <[email protected]>

Signed-off-by Bjorn Helgaas <[email protected]>

---
 drivers/pci/probe.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

drivers/pci/probe.c +4 -2

@@ -664,8 +664,10 @@ int __devinit pci_scan_bridge(struct pci
 
 	/* Check if setup is sensible at all */
 	if (!pass &&
-	    (primary != bus->number || secondary <= bus->number)) {
-		dev_dbg(&dev->dev, "bus configuration invalid, reconfiguring\n");
+	    (primary != bus->number || secondary <= bus->number ||
+	     secondary > subordinate)) {
+		dev_info(&dev->dev, "bridge configuration invalid ([bus %02x-%02x]), reconfiguring\n",
+			 secondary, subordinate);
 		broken = 1;
 	}

Greg Kroah-Hartman

author

2012-10-11 02:23 UTC #47 lore raw

PATCH 47/84 ext4: online defrag is not supported for journaled files

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

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

From: Dmitry Monakhov <[email protected]>

commit f066055a3449f0e5b0ae4f3ceab4445bead47638 upstream.

Proper block swap for inodes with full journaling enabled is truly non obvious task. In order to be on a safe side let's explicitly disable it for now.

Signed-off-by Dmitry Monakhov <[email protected]>

Signed-off-by "Theodore Ts'o" <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 fs/ext4/move_extent.c |    7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

fs/ext4/move_extent.c +6 -1

@@ -1209,7 +1209,12 @@ ext4_move_extents(struct file *o_filp, s
 			orig_inode->i_ino, donor_inode->i_ino);
 		return -EINVAL;
 	}
-
+	/* TODO: This is non obvious task to swap blocks for inodes with full
+	   jornaling enabled */
+	if (ext4_should_journal_data(orig_inode) ||
+	    ext4_should_journal_data(donor_inode)) {
+		return -EINVAL;
+	}
 	/* Protect orig and donor inodes against a truncate */
 	ret1 = mext_inode_double_lock(orig_inode, donor_inode);
 	if (ret1 < 0)

Greg Kroah-Hartman

author

2012-10-11 02:14 UTC #48 lore raw

PATCH 48/84 ext4: always set i_op in ext4_mknod()

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

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

From: Bernd Schubert <[email protected]>

commit 6a08f447facb4f9e29fcc30fb68060bb5a0d21c2 upstream.

ext4_special_inode_operations have their own ifdef CONFIG_EXT4_FS_XATTR to mask those methods. And ext4_iget also always sets it, so there is an inconsistency.

Signed-off-by Bernd Schubert <[email protected]>

Signed-off-by "Theodore Ts'o" <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 fs/ext4/namei.c |    2 --
 1 file changed, 2 deletions(-)

fs/ext4/namei.c +0 -2

@@ -1799,9 +1799,7 @@ retry:
 	err = PTR_ERR(inode);
 	if (!IS_ERR(inode)) {
 		init_special_inode(inode, inode->i_mode, rdev);
-#ifdef CONFIG_EXT4_FS_XATTR
 		inode->i_op = &ext4_special_inode_operations;
-#endif
 		err = ext4_add_nondir(handle, dentry, inode);
 	}
 	ext4_journal_stop(handle);

Greg Kroah-Hartman

author

2012-10-11 02:14 UTC #49 lore raw

PATCH 49/84 ext4: fix fdatasync() for files with only i_size changes

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

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

From: Jan Kara <[email protected]>

commit b71fc079b5d8f42b2a52743c8d2f1d35d655b1c5 upstream.

Code tracking when transaction needs to be committed on fdatasync(2) forgets to handle a situation when only inode's i_size is changed. Thus in such situations fdatasync(2) doesn't force transaction with new i_size to disk and that can result in wrong i_size after a crash.

Fix the issue by updating inode's i_datasync_tid whenever its size is updated.

Reported-by Kristian Nielsen <[email protected]>

Signed-off-by Jan Kara <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 fs/ext4/inode.c |    8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

fs/ext4/inode.c +6 -2

@@ -5151,6 +5151,7 @@ static int ext4_do_update_inode(handle_t
 	struct ext4_inode_info *ei = EXT4_I(inode);
 	struct buffer_head *bh = iloc->bh;
 	int err = 0, rc, block;
+	int need_datasync = 0;
 
 	/* For fields not not tracking in the in-memory inode,
 	 * initialise them to zero for new inodes. */

@@ -5199,7 +5200,10 @@ static int ext4_do_update_inode(handle_t
 		raw_inode->i_file_acl_high =
 			cpu_to_le16(ei->i_file_acl >> 32);
 	raw_inode->i_file_acl_lo = cpu_to_le32(ei->i_file_acl);
-	ext4_isize_set(raw_inode, ei->i_disksize);
+	if (ei->i_disksize != ext4_isize(raw_inode)) {
+		ext4_isize_set(raw_inode, ei->i_disksize);
+		need_datasync = 1;
+	}
 	if (ei->i_disksize > 0x7fffffffULL) {
 		struct super_block *sb = inode->i_sb;
 		if (!EXT4_HAS_RO_COMPAT_FEATURE(sb,

@@ -5252,7 +5256,7 @@ static int ext4_do_update_inode(handle_t
 		err = rc;
 	ext4_clear_inode_state(inode, EXT4_STATE_NEW);
 
-	ext4_update_inode_fsync_trans(handle, inode, 0);
+	ext4_update_inode_fsync_trans(handle, inode, need_datasync);
 out_brelse:
 	brelse(bh);
 	ext4_std_error(inode->i_sb, err);

Greg Kroah-Hartman

author

2012-10-11 02:23 UTC #50 lore raw

PATCH 50/84 ASoC: wm9712: Fix name of Capture Switch

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

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

From: Mark Brown <[email protected]>

commit 689185b78ba6fbe0042f662a468b5565909dff7a upstream.

Help UIs associate it with the matching gain control.

Signed-off-by Mark Brown <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 sound/soc/codecs/wm9712.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

sound/soc/codecs/wm9712.c +1 -1

@@ -144,7 +144,7 @@ SOC_SINGLE("Playback Attenuate (-6dB) Sw
 SOC_SINGLE("Bass Volume", AC97_MASTER_TONE, 8, 15, 1),
 SOC_SINGLE("Treble Volume", AC97_MASTER_TONE, 0, 15, 1),
 
-SOC_SINGLE("Capture ADC Switch", AC97_REC_GAIN, 15, 1, 1),
+SOC_SINGLE("Capture Switch", AC97_REC_GAIN, 15, 1, 1),
 SOC_ENUM("Capture Volume Steps", wm9712_enum[6]),
 SOC_DOUBLE("Capture Volume", AC97_REC_GAIN, 8, 0, 63, 1),
 SOC_SINGLE("Capture ZC Switch", AC97_REC_GAIN, 7, 1, 0),

Greg Kroah-Hartman

author

2012-10-11 02:16 UTC #51 lore raw

PATCH 51/84 mm: fix invalidate_complete_page2() lock ordering

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

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

From: Hugh Dickins <[email protected]>

commit ec4d9f626d5908b6052c2973f37992f1db52e967 upstream.

In fuzzing with trinity, lockdep protested "possible irq lock inversion dependency detected" when isolate_lru_page() reenabled interrupts while still holding the supposedly irq-safe tree_lock:

invalidate_inode_pages2
invalidate_complete_page2

    spin_lock_irq(&mapping->tree_lock)
    clear_page_mlock
      isolate_lru_page
        spin_unlock_irq(&zone->lru_lock)

isolate_lru_page() is correct to enable interrupts unconditionally:
invalidate_complete_page2() is incorrect to call clear_page_mlock() while holding tree_lock, which is supposed to nest inside lru_lock.

Both truncate_complete_page() and invalidate_complete_page() call clear_page_mlock() before taking tree_lock to remove page from radix_tree.
I guess invalidate_complete_page2() preferred to test PageDirty (again)
under tree_lock before committing to the munlock; but since the page has already been unmapped, its state is already somewhat inconsistent, and no worse if clear_page_mlock() moved up.

Reported-by Sasha Levin <[email protected]>

Deciphered-by Andrew Morton <[email protected]>

Signed-off-by Hugh Dickins <[email protected]>

Acked-by Mel Gorman <[email protected]>

Cc Rik van Riel <[email protected]>

Cc Johannes Weiner <[email protected]>

Cc Michel Lespinasse <[email protected]>

Cc Ying Han <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 mm/truncate.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

mm/truncate.c +2 -1

@@ -398,11 +398,12 @@ invalidate_complete_page2(struct address
 	if (page_has_private(page) && !try_to_release_page(page, GFP_KERNEL))
 		return 0;
 
+	clear_page_mlock(page);
+
 	spin_lock_irq(&mapping->tree_lock);
 	if (PageDirty(page))
 		goto failed;
 
-	clear_page_mlock(page);
 	BUG_ON(page_has_private(page));
 	__delete_from_page_cache(page);
 	spin_unlock_irq(&mapping->tree_lock);

Greg Kroah-Hartman

author

2012-10-11 02:19 UTC #52 lore raw

PATCH 52/84 mm: thp: fix pmd_present for split_huge_page and PROT_NONE with THP

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

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

From: Andrea Arcangeli <[email protected]>

commit 027ef6c87853b0a9df53175063028edb4950d476 upstream.

In many places !pmd_present has been converted to pmd_none. For pmds that's equivalent and pmd_none is quicker so using pmd_none is better.

However (unless we delete pmd_present) we should provide an accurate pmd_present too. This will avoid the risk of code thinking the pmd is non present because it's under __split_huge_page_map, see the pmd_mknotpresent there and the comment above it.

If the page has been mprotected as PROT_NONE, it would also lead to a pmd_present false negative in the same way as the race with split_huge_page.

Because the PSE bit stays on at all times (both during split_huge_page and when the _PAGE_PROTNONE bit get set), we could only check for the PSE bit, but checking the PROTNONE bit too is still good to remember pmd_present must always keep PROT_NONE into account.

This explains a not reproducible BUG_ON that was seldom reported on the lists.

The same issue is in pmd_large, it would go wrong with both PROT_NONE and if it races with split_huge_page.

Signed-off-by Andrea Arcangeli <[email protected]>

Acked-by Rik van Riel <[email protected]>

Cc Johannes Weiner <[email protected]>

Cc Hugh Dickins <[email protected]>

Cc Mel Gorman <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 arch/x86/include/asm/pgtable.h |   11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

arch/x86/include/asm/pgtable.h +8 -3

@@ -146,8 +146,7 @@ static inline unsigned long pmd_pfn(pmd_
 
 static inline int pmd_large(pmd_t pte)
 {
-	return (pmd_flags(pte) & (_PAGE_PSE | _PAGE_PRESENT)) ==
-		(_PAGE_PSE | _PAGE_PRESENT);
+	return pmd_flags(pte) & _PAGE_PSE;
 }
 
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE

@@ -415,7 +414,13 @@ static inline int pte_hidden(pte_t pte)
 
 static inline int pmd_present(pmd_t pmd)
 {
-	return pmd_flags(pmd) & _PAGE_PRESENT;
+	/*
+	 * Checking for _PAGE_PSE is needed too because
+	 * split_huge_page will temporarily clear the present bit (but
+	 * the _PAGE_PSE flag will remain set at all times while the
+	 * _PAGE_PRESENT bit is clear).
+	 */
+	return pmd_flags(pmd) & (_PAGE_PRESENT | _PAGE_PROTNONE | _PAGE_PSE);
 }
 
 static inline int pmd_none(pmd_t pmd)

Greg Kroah-Hartman

author

2012-10-11 02:15 UTC #53 lore raw

PATCH 53/84 ALSA: aloop - add locking to timer access

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

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

From: Omair Mohammed Abdullah <[email protected]>

commit d4f1e48bd11e3df6a26811f7a1f06c4225d92f7d upstream.

When the loopback timer handler is running, calling del_timer() (for STOP trigger) will not wait for the handler to complete before deactivating the timer. The timer gets rescheduled in the handler as usual. Then a subsequent START trigger will try to start the timer using add_timer() with a timer pending leading to a kernel panic.

Serialize the calls to add_timer() and del_timer() using a spin lock to avoid this.

Signed-off-by Omair Mohammed Abdullah <[email protected]>

Signed-off-by Vinod Koul <[email protected]>

Signed-off-by Takashi Iwai <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 sound/drivers/aloop.c |    6 ++++++
 1 file changed, 6 insertions(+)

sound/drivers/aloop.c +6 -0

@@ -119,6 +119,7 @@ struct loopback_pcm {
 	unsigned int period_size_frac;
 	unsigned long last_jiffies;
 	struct timer_list timer;
+	spinlock_t timer_lock;
 };
 
 static struct platform_device *devices[SNDRV_CARDS];

@@ -169,6 +170,7 @@ static void loopback_timer_start(struct
 	unsigned long tick;
 	unsigned int rate_shift = get_rate_shift(dpcm);
 
+	spin_lock(&dpcm->timer_lock);
 	if (rate_shift != dpcm->pcm_rate_shift) {
 		dpcm->pcm_rate_shift = rate_shift;
 		dpcm->period_size_frac = frac_pos(dpcm, dpcm->pcm_period_size);

@@ -181,12 +183,15 @@ static void loopback_timer_start(struct
 	tick = (tick + dpcm->pcm_bps - 1) / dpcm->pcm_bps;
 	dpcm->timer.expires = jiffies + tick;
 	add_timer(&dpcm->timer);
+	spin_unlock(&dpcm->timer_lock);
 }
 
 static inline void loopback_timer_stop(struct loopback_pcm *dpcm)
 {
+	spin_lock(&dpcm->timer_lock);
 	del_timer(&dpcm->timer);
 	dpcm->timer.expires = 0;
+	spin_unlock(&dpcm->timer_lock);
 }
 
 #define CABLE_VALID_PLAYBACK	(1 << SNDRV_PCM_STREAM_PLAYBACK)

@@ -658,6 +663,7 @@ static int loopback_open(struct snd_pcm_
 	dpcm->substream = substream;
 	setup_timer(&dpcm->timer, loopback_timer_function,
 		    (unsigned long)dpcm);
+	spin_lock_init(&dpcm->timer_lock);
 
 	cable = loopback->cables[substream->number][dev];
 	if (!cable) {

Greg Kroah-Hartman

author

2012-10-11 02:15 UTC #54 lore raw

PATCH 54/84 ALSA: usb - disable broken hw volume for Tenx TP6911

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

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

From: David Henningsson <[email protected]>

commit c10514394ef9e8de93a4ad8c8904d71dcd82c122 upstream.

While going through Ubuntu bugs, I discovered this patch being posted and a confirmation that the patch works as expected.

Finding out how the hw volume really works would be preferrable to just disabling the broken one, but this would be better than nothing.

Credit: sndfnsdfin (qawsnews)

BugLink https://bugs.launchpad.net/bugs/559939

Signed-off-by David Henningsson <[email protected]>

Signed-off-by Takashi Iwai <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 sound/usb/mixer.c |    7 +++++++
 1 file changed, 7 insertions(+)

sound/usb/mixer.c +7 -0

@@ -1246,6 +1246,13 @@ static int parse_audio_feature_unit(stru
 		/* disable non-functional volume control */
 		master_bits &= ~UAC_CONTROL_BIT(UAC_FU_VOLUME);
 		break;
+	case USB_ID(0x1130, 0xf211):
+		snd_printk(KERN_INFO
+			   "usbmixer: volume control quirk for Tenx TP6911 Audio Headset\n");
+		/* disable non-functional volume control */
+		channels = 0;
+		break;
+
 	}
 	if (channels > 0)
 		first_ch_bits = snd_usb_combine_bytes(bmaControls + csize, csize);

Greg Kroah-Hartman

author

2012-10-11 02:15 UTC #55 lore raw

PATCH 55/84 ALSA: USB: Support for (original) Xbox Communicator

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

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

From: Marko Friedemann <[email protected]>

commit c05fce586d4da2dfe0309bef3795a8586e967bc3 upstream.

Added support for Xbox Communicator to USB quirks.

Signed-off-by Marko Friedemann <[email protected]>

Signed-off-by Takashi Iwai <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 sound/usb/quirks-table.h |   53 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 53 insertions(+)

sound/usb/quirks-table.h +53 -0

@@ -2581,6 +2581,59 @@ YAMAHA_DEVICE(0x7010, "UB99"),
 	}
 },
 
+/* Microsoft XboxLive Headset/Xbox Communicator */
+{
+	USB_DEVICE(0x045e, 0x0283),
+	.bInterfaceClass = USB_CLASS_PER_INTERFACE,
+	.driver_info = (unsigned long) &(const struct snd_usb_audio_quirk) {
+		.vendor_name = "Microsoft",
+		.product_name = "XboxLive Headset/Xbox Communicator",
+		.ifnum = QUIRK_ANY_INTERFACE,
+		.type = QUIRK_COMPOSITE,
+		.data = &(const struct snd_usb_audio_quirk[]) {
+			{
+				/* playback */
+				.ifnum = 0,
+				.type = QUIRK_AUDIO_FIXED_ENDPOINT,
+				.data = &(const struct audioformat) {
+					.formats = SNDRV_PCM_FMTBIT_S16_LE,
+					.channels = 1,
+					.iface = 0,
+					.altsetting = 0,
+					.altset_idx = 0,
+					.attributes = 0,
+					.endpoint = 0x04,
+					.ep_attr = 0x05,
+					.rates = SNDRV_PCM_RATE_CONTINUOUS,
+					.rate_min = 22050,
+					.rate_max = 22050
+				}
+			},
+			{
+				/* capture */
+				.ifnum = 1,
+				.type = QUIRK_AUDIO_FIXED_ENDPOINT,
+				.data = &(const struct audioformat) {
+					.formats = SNDRV_PCM_FMTBIT_S16_LE,
+					.channels = 1,
+					.iface = 1,
+					.altsetting = 0,
+					.altset_idx = 0,
+					.attributes = 0,
+					.endpoint = 0x85,
+					.ep_attr = 0x05,
+					.rates = SNDRV_PCM_RATE_CONTINUOUS,
+					.rate_min = 16000,
+					.rate_max = 16000
+				}
+			},
+			{
+				.ifnum = -1
+			}
+		}
+	}
+},
+
 {
 	/*
 	 * Some USB MIDI devices don't have an audio control interface,

Greg Kroah-Hartman

author

2012-10-11 02:18 UTC #56 lore raw

PATCH 56/84 drm/radeon: only adjust default clocks on NI GPUs

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

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

From: Alex Deucher <[email protected]>

commit 2e3b3b105ab3bb5b6a37198da4f193cd13781d13 upstream.

SI asics store voltage information differently so we don't have a way to deal with it properly yet.

Signed-off-by Alex Deucher <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/gpu/drm/radeon/radeon_pm.c |    8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

drivers/gpu/drm/radeon/radeon_pm.c +6 -2

@@ -535,7 +535,9 @@ void radeon_pm_suspend(struct radeon_dev
 void radeon_pm_resume(struct radeon_device *rdev)
 {
 	/* set up the default clocks if the MC ucode is loaded */
-	if (ASIC_IS_DCE5(rdev) && rdev->mc_fw) {
+	if ((rdev->family >= CHIP_BARTS) &&
+	    (rdev->family <= CHIP_CAYMAN) &&
+	    rdev->mc_fw) {
 		if (rdev->pm.default_vddc)
 			radeon_atom_set_voltage(rdev, rdev->pm.default_vddc,
 						SET_VOLTAGE_TYPE_ASIC_VDDC);

@@ -590,7 +592,9 @@ int radeon_pm_init(struct radeon_device
 		radeon_pm_print_states(rdev);
 		radeon_pm_init_profile(rdev);
 		/* set up the default clocks if the MC ucode is loaded */
-		if (ASIC_IS_DCE5(rdev) && rdev->mc_fw) {
+		if ((rdev->family >= CHIP_BARTS) &&
+		    (rdev->family <= CHIP_CAYMAN) &&
+		    rdev->mc_fw) {
 			if (rdev->pm.default_vddc)
 				radeon_atom_set_voltage(rdev, rdev->pm.default_vddc,
 							SET_VOLTAGE_TYPE_ASIC_VDDC);

Greg Kroah-Hartman

author

2012-10-11 02:18 UTC #57 lore raw

PATCH 57/84 drm/radeon: Add MSI quirk for gateway RS690

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

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

From: Alex Deucher <[email protected]>

commit 3a6d59df80897cc87812b6826d70085905bed013 upstream.

Fixes another system on:
https://bugs.freedesktop.org/show_bug.cgi?id=37679

Signed-off-by Alex Deucher <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/gpu/drm/radeon/radeon_irq_kms.c |    6 ++++++
 1 file changed, 6 insertions(+)

drivers/gpu/drm/radeon/radeon_irq_kms.c +6 -0

@@ -143,6 +143,12 @@ static bool radeon_msi_ok(struct radeon_
 	    (rdev->pdev->subsystem_device == 0x01fd))
 		return true;
 
+	/* Gateway RS690 only seems to work with MSIs. */
+	if ((rdev->pdev->device == 0x791f) &&
+	    (rdev->pdev->subsystem_vendor == 0x107b) &&
+	    (rdev->pdev->subsystem_device == 0x0185))
+		return true;
+
 	/* RV515 seems to have MSI issues where it loses
 	 * MSI rearms occasionally. This leads to lockups and freezes.
 	 * disable it by default.

Greg Kroah-Hartman

author

2012-10-11 02:16 UTC #58 lore raw

PATCH 58/84 drm/radeon: force MSIs on RS690 asics

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

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

From: Alex Deucher <[email protected]>

commit fb6ca6d154cdcd53e7f27f8dbba513830372699b upstream.

There are so many quirks, lets just try and force
this for all RS690s. See:
https://bugs.freedesktop.org/show_bug.cgi?id=37679

Signed-off-by Alex Deucher <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/gpu/drm/radeon/radeon_irq_kms.c |    4 ++++
 1 file changed, 4 insertions(+)

drivers/gpu/drm/radeon/radeon_irq_kms.c +4 -0

@@ -149,6 +149,10 @@ static bool radeon_msi_ok(struct radeon_
 	    (rdev->pdev->subsystem_device == 0x0185))
 		return true;
 
+	/* try and enable MSIs by default on all RS690s */
+	if (rdev->family == CHIP_RS690)
+		return true;
+
 	/* RV515 seems to have MSI issues where it loses
 	 * MSI rearms occasionally. This leads to lockups and freezes.
 	 * disable it by default.

Greg Kroah-Hartman

author

2012-10-11 02:17 UTC #59 lore raw

PATCH 59/84 rcu: Fix day-one dyntick-idle stall-warning bug

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

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

From: "Paul E. McKenney" <[email protected]>

commit a10d206ef1a83121ab7430cb196e0376a7145b22 upstream.

Each grace period is supposed to have at least one callback waiting for that grace period to complete. However, if CONFIG_NO_HZ=n, an extra callback-free grace period is no big problem -- it will chew up a tiny bit of CPU time, but it will complete normally. In contrast, CONFIG_NO_HZ=y kernels have the potential for all the CPUs to go to sleep indefinitely, in turn indefinitely delaying completion of the callback-free grace period. Given that nothing is waiting on this grace period, this is also not a problem.

That is, unless RCU CPU stall warnings are also enabled, as they are in recent kernels. In this case, if a CPU wakes up after at least one minute of inactivity, an RCU CPU stall warning will result. The reason that no one noticed until quite recently is that most systems have enough OS noise that they will never remain absolutely idle for a full minute. But there are some embedded systems with cut-down userspace configurations that consistently get into this situation.

All this begs the question of exactly how a callback-free grace period gets started in the first place. This can happen due to the fact that CPUs do not necessarily agree on which grace period is in progress. If a CPU still believes that the grace period that just completed is still ongoing, it will believe that it has callbacks that need to wait for another grace period, never mind the fact that the grace period that they were waiting for just completed. This CPU can therefore erroneously decide to start a new grace period. Note that this can happen in TREE_RCU and TREE_PREEMPT_RCU even on a single-CPU system: Deadlock considerations mean that the CPU that detected the end of the grace period is not necessarily officially informed of this fact for some time.

Once this CPU notices that the earlier grace period completed, it will invoke its callbacks. It then won't have any callbacks left. If no other CPU has any callbacks, we now have a callback-free grace period.

This commit therefore makes CPUs check more carefully before starting a new grace period. This new check relies on an array of tail pointers into each CPU's list of callbacks. If the CPU is up to date on which grace periods have completed, it checks to see if any callbacks follow the RCU_DONE_TAIL segment, otherwise it checks to see if any callbacks follow the RCU_WAIT_TAIL segment. The reason that this works is that the RCU_WAIT_TAIL segment will be promoted to the RCU_DONE_TAIL segment as soon as the CPU is officially notified that the old grace period has ended.

This change is to cpu_needs_another_gp(), which is called in a number of places. The only one that really matters is in rcu_start_gp(), where the root rcu_node structure's ->lock is held, which prevents any other CPU from starting or completing a grace period, so that the comparison that determines whether the CPU is missing the completion of a grace period is stable.

Reported-by Becky Bruce <[email protected]>

Reported-by Subodh Nijsure <[email protected]>

Reported-by Paul Walmsley <[email protected]>

Signed-off-by Paul E. McKenney <[email protected]>

Tested-by Paul Walmsley <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 kernel/rcutree.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

kernel/rcutree.c +3 -1

@@ -283,7 +283,9 @@ cpu_has_callbacks_ready_to_invoke(struct
 static int
 cpu_needs_another_gp(struct rcu_state *rsp, struct rcu_data *rdp)
 {
-	return *rdp->nxttail[RCU_DONE_TAIL] && !rcu_gp_in_progress(rsp);
+	return *rdp->nxttail[RCU_DONE_TAIL +
+			     ACCESS_ONCE(rsp->completed) != rdp->completed] &&
+	       !rcu_gp_in_progress(rsp);
 }
 
 /*

Greg Kroah-Hartman

author

2012-10-11 02:16 UTC #60 lore raw

PATCH 60/84 r8169: fix wake on lan setting for non-8111E.

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

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

From: Hayes Wang <[email protected]>

commit d4ed95d796e5126bba51466dc07e287cebc8bd19 upstream.

Only 8111E needs enable RxConfig bit 0 ~ 3 when suspending or shutdowning for wake on lan.

Signed-off-by Hayes Wang <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

drivers/net/r8169.c +4 -2

@@ -3126,8 +3126,10 @@ static void r8168_pll_power_down(struct
 		rtl_writephy(tp, 0x1f, 0x0000);
 		rtl_writephy(tp, MII_BMCR, 0x0000);
 
-		RTL_W32(RxConfig, RTL_R32(RxConfig) |
-			AcceptBroadcast | AcceptMulticast | AcceptMyPhys);
+		if (tp->mac_version == RTL_GIGA_MAC_VER_32 ||
+		    tp->mac_version == RTL_GIGA_MAC_VER_33)
+			RTL_W32(RxConfig, RTL_R32(RxConfig) | AcceptBroadcast |
+				AcceptMulticast | AcceptMyPhys);
 		return;
 	}

Jonathan Nieder

2012-10-11 07:16 UTC #61 lore raw

Re: [ 60/84] r8169: fix wake on lan setting for non-8111E.

Hi Greg,

Quoted 4 lines · Greg Kroah-Hartman

commit d4ed95d796e5126bba51466dc07e287cebc8bd19 upstream.
Only 8111E needs enable RxConfig bit 0 ~ 3 when suspending or shutdowning for wake on lan.

Sounds sensible to me, but I haven't reviewed this --- e.g., I don't know what symptoms this addresses or what model they were experienced on.

Here's a patch to remove the reviewed-bys that seem to have been added by mistake.

Thanks,
Jonathan

-- >8 --
Subject: queue-3.0: remove reviewed-by on patches I haven't reviewed

---
 ...ater-require-bit-0x20-to-be-set-in-config2-for-pme-signaling.patch |    1 -
 .../r8169-call-netif_napi_del-at-errpaths-and-at-driver-unload.patch  |    1 -
 queue-3.0/r8169-config1-is-read-only-on-8168c-and-later.patch         |    1 -
 queue-3.0/r8169-don-t-enable-rx-when-shutdown.patch                   |    1 -
 queue-3.0/r8169-expand-received-packet-length-indication.patch        |    1 -
 queue-3.0/r8169-fix-config2-msienable-bit-setting.patch               |    1 -
 queue-3.0/r8169-fix-unsigned-int-wraparound-with-tso.patch            |    1 -
 queue-3.0/r8169-fix-wake-on-lan-setting-for-non-8111e.patch           |    1 -
 .../r8169-increase-the-delay-parameter-of-pm_schedule_suspend.patch   |    1 -
 queue-3.0/r8169-jumbo-fixes.patch                                     |    1 -
 queue-3.0/r8169-missing-barriers.patch                                |    1 -
 queue-3.0/r8169-remove-erroneous-processing-of-always-set-bit.patch   |    1 -
 queue-3.0/r8169-runtime-resume-before-shutdown.patch                  |    1 -
 13 files changed, 13 deletions(-)

queue-3.0/r8169-8168c-and-later-require-bit-0x20-to-be-set-in-config2-for-pme-signaling.patch +0 -1

index b4d1b9d4..9dd53b4e 100644
@@ -12,7 +12,6 @@ The new 84xx stopped flying below the radars.
 
 Signed-off-by: Francois Romieu <[email protected]>
 Cc: Hayes Wang <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Acked-by: David S. Miller <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---

queue-3.0/r8169-call-netif_napi_del-at-errpaths-and-at-driver-unload.patch +0 -1

index c2600140..f9148f64 100644
@@ -14,7 +14,6 @@ delete the NAPI before unregistering netdevice using unregister_netdev.
 
 Signed-off-by: Devendra Naga <[email protected]>
 Signed-off-by: David S. Miller <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---
  drivers/net/r8169.c |    3 +++

queue-3.0/r8169-config1-is-read-only-on-8168c-and-later.patch +0 -1

index dd62d63a..d8690d0d 100644
@@ -12,7 +12,6 @@ Suggested by Hayes.
 
 Signed-off-by: Francois Romieu <[email protected]>
 Cc: Hayes Wang <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Acked-by: David S. Miller <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---

queue-3.0/r8169-don-t-enable-rx-when-shutdown.patch +0 -1

index ca876bb5..a511d5ec 100644
@@ -12,7 +12,6 @@ Only 8111b needs to enable rx when shutdowning with WoL.
 
 Signed-off-by: Hayes Wang <[email protected]>
 Acked-by: Francois Romieu <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Acked-by: David S. Miller <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---

queue-3.0/r8169-expand-received-packet-length-indication.patch +0 -1

index 341ddee4..bbe8a00d 100644
@@ -18,7 +18,6 @@ Realtek's drivers use the same mask for the 816x and the fast ethernet
 only 810x.
 
 Signed-off-by: Francois Romieu <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Acked-by: David S. Miller <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---

queue-3.0/r8169-fix-config2-msienable-bit-setting.patch +0 -1

index 027e5658..92b29990 100644
@@ -17,7 +17,6 @@ Reported-by: Su Kang Yin <[email protected]>
 Signed-off-by: Francois Romieu <[email protected]>
 Cc: Hayes Wang <[email protected]>
 Signed-off-by: David S. Miller <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---
  drivers/net/r8169.c |   14 ++++++++------

queue-3.0/r8169-fix-unsigned-int-wraparound-with-tso.patch +0 -1

index 665d18a0..82c37928 100644
@@ -20,7 +20,6 @@ Zotac 890GXITX.
 Signed-off-by: Julien Ducourthial <[email protected]>
 Acked-by: Francois Romieu <[email protected]>
 Signed-off-by: David S. Miller <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---
  drivers/net/r8169.c |   16 ++++++++++------

queue-3.0/r8169-fix-wake-on-lan-setting-for-non-8111e.patch +0 -1

index 4fabe266..fda216e1 100644
@@ -13,7 +13,6 @@ shutdowning for wake on lan.
 
 Signed-off-by: Hayes Wang <[email protected]>
 Acked-by: Francois Romieu <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Acked-by: David S. Miller <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---

queue-3.0/r8169-increase-the-delay-parameter-of-pm_schedule_suspend.patch +0 -1

index b157a375..1cfd3eaa 100644
@@ -17,7 +17,6 @@ would find the driver keep in a loop of runtime_suspend and rumtime_resume.
 Signed-off-by: Hayes Wang <[email protected]>
 Acked-by: Francois Romieu <[email protected]>
 Signed-off-by: David S. Miller <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---
  drivers/net/r8169.c |    2 +-

queue-3.0/r8169-jumbo-fixes.patch +0 -1

index a34f3739..eb33accd 100644
@@ -19,7 +19,6 @@ commit d58d46b5d85139d18eb939aa7279c160bab70484 upstream.
 - rtl_tx_performance_tweak is moved forward. No change.
 
 Signed-off-by: Francois Romieu <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Acked-by: David S. Miller <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---

queue-3.0/r8169-missing-barriers.patch +0 -1

index 07f8691b..714f73ab 100644
@@ -9,7 +9,6 @@ commit 1e874e041fc7c222cbd85b20c4406070be1f687a upstream.
 
 Signed-off-by: Francois Romieu <[email protected]>
 Cc: Hayes Wang <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Acked-by: David S. Miller <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---

queue-3.0/r8169-remove-erroneous-processing-of-always-set-bit.patch +0 -1

index 6f79703a..c0002754 100644
@@ -11,7 +11,6 @@ When set, RxFOVF (resp. RxBOVF) is always 1 (resp. 0).
 
 Signed-off-by: Francois Romieu <[email protected]>
 Cc: Hayes <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Acked-by: David S. Miller <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---

queue-3.0/r8169-runtime-resume-before-shutdown.patch +0 -1

index 5b2757a7..2dec927f 100644
@@ -27,7 +27,6 @@ Cc: Hayes Wang <[email protected]>
 Cc: Alan Stern <[email protected]>
 Acked-by: Rafael J. Wysocki <[email protected]>
 Signed-off-by: David S. Miller <[email protected]>
-Reviewed-by: Jonathan Nieder <[email protected]>
 Signed-off-by: Greg Kroah-Hartman <[email protected]>
 ---
  drivers/net/r8169.c |    5 +++++

Greg Kroah-Hartman

author

2012-10-11 10:59 UTC #62 lore raw

Re: [ 60/84] r8169: fix wake on lan setting for non-8111E.

Quoted 9 lines · Jonathan Nieder

Hi Greg,
Quoted 4 lines · Greg Kroah-Hartman
commit d4ed95d796e5126bba51466dc07e287cebc8bd19 upstream.
Only 8111E needs enable RxConfig bit 0 ~ 3 when suspending or shutdowning for wake on lan.

Quoted 6 lines

Sounds sensible to me, but I haven't reviewed this --- e.g., I don't know what symptoms this addresses or what model they were experienced on.
Here's a patch to remove the reviewed-bys that seem to have been added by mistake.

Sorry, yes, I added your reviewed-by to the whole series, my fault. I've now applied your patch removing them.

greg k-h

Greg Kroah-Hartman

author

2012-10-11 02:15 UTC #63 lore raw

PATCH 61/84 r8169: dont enable rx when shutdown.

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

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

From: Hayes Wang <[email protected]>

commit aaa89c08d9ffa3739c93d65d98b73ec2aa2e93a5 upstream.

Only 8111b needs to enable rx when shutdowning with WoL.

Signed-off-by Hayes Wang <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |    7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

drivers/net/r8169.c +5 -2

@@ -5391,8 +5391,11 @@ static void rtl_shutdown(struct pci_dev
 	spin_unlock_irq(&tp->lock);
 
 	if (system_state == SYSTEM_POWER_OFF) {
-		/* WoL fails with some 8168 when the receiver is disabled. */
-		if (tp->features & RTL_FEATURE_WOL) {
+		/* WoL fails with 8168b when the receiver is disabled. */
+		if ((tp->mac_version == RTL_GIGA_MAC_VER_11 ||
+		     tp->mac_version == RTL_GIGA_MAC_VER_12 ||
+		     tp->mac_version == RTL_GIGA_MAC_VER_17) &&
+		    (tp->features & RTL_FEATURE_WOL)) {
 			pci_clear_master(pdev);
 
 			RTL_W8(ChipCmd, CmdRxEnb);

Greg Kroah-Hartman

author

2012-10-11 02:20 UTC #64 lore raw

PATCH 62/84 r8169: remove erroneous processing of always set bit.

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

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

From: Francois Romieu <[email protected]>

commit e03f33af79f0772156e1a1a1e36bdddf8012b2e4 upstream.

When set, RxFOVF (resp. RxBOVF) is always 1 (resp. 0).

Cc Hayes <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |    7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

drivers/net/r8169.c +6 -1

@@ -388,6 +388,7 @@ enum rtl_register_content {
 	RxOK		= 0x0001,
 
 	/* RxStatusDesc */
+	RxBOVF	= (1 << 24),
 	RxFOVF	= (1 << 23),
 	RxRWT	= (1 << 22),
 	RxRES	= (1 << 21),

@@ -666,6 +667,7 @@ struct rtl8169_private {
 	struct mii_if_info mii;
 	struct rtl8169_counters counters;
 	u32 saved_wolopts;
+	u32 opts1_mask;
 
 	const struct firmware *fw;
 #define RTL_FIRMWARE_UNKNOWN	ERR_PTR(-EAGAIN);

@@ -3442,6 +3444,9 @@ rtl8169_init_one(struct pci_dev *pdev, c
 	tp->intr_event = cfg->intr_event;
 	tp->napi_event = cfg->napi_event;
 
+	tp->opts1_mask = (tp->mac_version != RTL_GIGA_MAC_VER_01) ?
+		~(RxBOVF | RxFOVF) : ~0;
+
 	init_timer(&tp->timer);
 	tp->timer.data = (unsigned long) dev;
 	tp->timer.function = rtl8169_phy_timer;

@@ -4920,7 +4925,7 @@ static int rtl8169_rx_interrupt(struct n
 		u32 status;
 
 		rmb();
-		status = le32_to_cpu(desc->opts1);
+		status = le32_to_cpu(desc->opts1) & tp->opts1_mask;
 
 		if (status & DescOwn)
 			break;

Greg Kroah-Hartman

author

2012-10-11 02:15 UTC #65 lore raw

PATCH 63/84 r8169: jumbo fixes.

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

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

From: Francois Romieu <[email protected]>

commit d58d46b5d85139d18eb939aa7279c160bab70484 upstream.

- fix features : jumbo frames and checksumming can not be used at the same time.

- introduce hw_jumbo_{enable / disable} helpers. Their content has been creatively extracted from Realtek's own drivers. As an illustration, it would be nice to know how/if the MaxTxPacketSize register operates when the device can work with a 9k jumbo frame as its documentation (8168c) can not be applied beyond ~7k.

- rtl_tx_performance_tweak is moved forward. No change.

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |  302 ++++++++++++++++++++++++++++++++++++++++++----------
 1 file changed, 247 insertions(+), 55 deletions(-)

drivers/net/r8169.c +247 ~~-55~~

@@ -140,82 +140,101 @@ enum rtl_tx_desc_version {
 	RTL_TD_1	= 1,
 };
 
-#define _R(NAME,TD,FW) \
-	{ .name = NAME, .txd_version = TD, .fw_name = FW }
+#define JUMBO_1K	ETH_DATA_LEN
+#define JUMBO_4K	(4*1024 - ETH_HLEN - 2)
+#define JUMBO_6K	(6*1024 - ETH_HLEN - 2)
+#define JUMBO_7K	(7*1024 - ETH_HLEN - 2)
+#define JUMBO_9K	(9*1024 - ETH_HLEN - 2)
+
+#define _R(NAME,TD,FW,SZ,B) {	\
+	.name = NAME,		\
+	.txd_version = TD,	\
+	.fw_name = FW,		\
+	.jumbo_max = SZ,	\
+	.jumbo_tx_csum = B	\
+}
 
 static const struct {
 	const char *name;
 	enum rtl_tx_desc_version txd_version;
 	const char *fw_name;
+	u16 jumbo_max;
+	bool jumbo_tx_csum;
 } rtl_chip_infos[] = {
 	/* PCI devices. */
 	[RTL_GIGA_MAC_VER_01] =
-		_R("RTL8169",		RTL_TD_0, NULL),
+		_R("RTL8169",		RTL_TD_0, NULL, JUMBO_7K, true),
 	[RTL_GIGA_MAC_VER_02] =
-		_R("RTL8169s",		RTL_TD_0, NULL),
+		_R("RTL8169s",		RTL_TD_0, NULL, JUMBO_7K, true),
 	[RTL_GIGA_MAC_VER_03] =
-		_R("RTL8110s",		RTL_TD_0, NULL),
+		_R("RTL8110s",		RTL_TD_0, NULL, JUMBO_7K, true),
 	[RTL_GIGA_MAC_VER_04] =
-		_R("RTL8169sb/8110sb",	RTL_TD_0, NULL),
+		_R("RTL8169sb/8110sb",	RTL_TD_0, NULL, JUMBO_7K, true),
 	[RTL_GIGA_MAC_VER_05] =
-		_R("RTL8169sc/8110sc",	RTL_TD_0, NULL),
+		_R("RTL8169sc/8110sc",	RTL_TD_0, NULL, JUMBO_7K, true),
 	[RTL_GIGA_MAC_VER_06] =
-		_R("RTL8169sc/8110sc",	RTL_TD_0, NULL),
+		_R("RTL8169sc/8110sc",	RTL_TD_0, NULL, JUMBO_7K, true),
 	/* PCI-E devices. */
 	[RTL_GIGA_MAC_VER_07] =
-		_R("RTL8102e",		RTL_TD_1, NULL),
+		_R("RTL8102e",		RTL_TD_1, NULL, JUMBO_1K, true),
 	[RTL_GIGA_MAC_VER_08] =
-		_R("RTL8102e",		RTL_TD_1, NULL),
+		_R("RTL8102e",		RTL_TD_1, NULL, JUMBO_1K, true),
 	[RTL_GIGA_MAC_VER_09] =
-		_R("RTL8102e",		RTL_TD_1, NULL),
+		_R("RTL8102e",		RTL_TD_1, NULL, JUMBO_1K, true),
 	[RTL_GIGA_MAC_VER_10] =
-		_R("RTL8101e",		RTL_TD_0, NULL),
+		_R("RTL8101e",		RTL_TD_0, NULL, JUMBO_1K, true),
 	[RTL_GIGA_MAC_VER_11] =
-		_R("RTL8168b/8111b",	RTL_TD_0, NULL),
+		_R("RTL8168b/8111b",	RTL_TD_0, NULL, JUMBO_4K, false),
 	[RTL_GIGA_MAC_VER_12] =
-		_R("RTL8168b/8111b",	RTL_TD_0, NULL),
+		_R("RTL8168b/8111b",	RTL_TD_0, NULL, JUMBO_4K, false),
 	[RTL_GIGA_MAC_VER_13] =
-		_R("RTL8101e",		RTL_TD_0, NULL),
+		_R("RTL8101e",		RTL_TD_0, NULL, JUMBO_1K, true),
 	[RTL_GIGA_MAC_VER_14] =
-		_R("RTL8100e",		RTL_TD_0, NULL),
+		_R("RTL8100e",		RTL_TD_0, NULL, JUMBO_1K, true),
 	[RTL_GIGA_MAC_VER_15] =
-		_R("RTL8100e",		RTL_TD_0, NULL),
+		_R("RTL8100e",		RTL_TD_0, NULL, JUMBO_1K, true),
 	[RTL_GIGA_MAC_VER_16] =
-		_R("RTL8101e",		RTL_TD_0, NULL),
+		_R("RTL8101e",		RTL_TD_0, NULL, JUMBO_1K, true),
 	[RTL_GIGA_MAC_VER_17] =
-		_R("RTL8168b/8111b",	RTL_TD_0, NULL),
+		_R("RTL8168b/8111b",	RTL_TD_1, NULL, JUMBO_4K, false),
 	[RTL_GIGA_MAC_VER_18] =
-		_R("RTL8168cp/8111cp",	RTL_TD_1, NULL),
+		_R("RTL8168cp/8111cp",	RTL_TD_1, NULL, JUMBO_6K, false),
 	[RTL_GIGA_MAC_VER_19] =
-		_R("RTL8168c/8111c",	RTL_TD_1, NULL),
+		_R("RTL8168c/8111c",	RTL_TD_1, NULL, JUMBO_6K, false),
 	[RTL_GIGA_MAC_VER_20] =
-		_R("RTL8168c/8111c",	RTL_TD_1, NULL),
+		_R("RTL8168c/8111c",	RTL_TD_1, NULL, JUMBO_6K, false),
 	[RTL_GIGA_MAC_VER_21] =
-		_R("RTL8168c/8111c",	RTL_TD_1, NULL),
+		_R("RTL8168c/8111c",	RTL_TD_1, NULL, JUMBO_6K, false),
 	[RTL_GIGA_MAC_VER_22] =
-		_R("RTL8168c/8111c",	RTL_TD_1, NULL),
+		_R("RTL8168c/8111c",	RTL_TD_1, NULL, JUMBO_6K, false),
 	[RTL_GIGA_MAC_VER_23] =
-		_R("RTL8168cp/8111cp",	RTL_TD_1, NULL),
+		_R("RTL8168cp/8111cp",	RTL_TD_1, NULL, JUMBO_6K, false),
 	[RTL_GIGA_MAC_VER_24] =
-		_R("RTL8168cp/8111cp",	RTL_TD_1, NULL),
+		_R("RTL8168cp/8111cp",	RTL_TD_1, NULL, JUMBO_6K, false),
 	[RTL_GIGA_MAC_VER_25] =
-		_R("RTL8168d/8111d",	RTL_TD_1, FIRMWARE_8168D_1),
+		_R("RTL8168d/8111d",	RTL_TD_1, FIRMWARE_8168D_1,
+							JUMBO_9K, false),
 	[RTL_GIGA_MAC_VER_26] =
-		_R("RTL8168d/8111d",	RTL_TD_1, FIRMWARE_8168D_2),
+		_R("RTL8168d/8111d",	RTL_TD_1, FIRMWARE_8168D_2,
+							JUMBO_9K, false),
 	[RTL_GIGA_MAC_VER_27] =
-		_R("RTL8168dp/8111dp",	RTL_TD_1, NULL),
+		_R("RTL8168dp/8111dp",	RTL_TD_1, NULL, JUMBO_9K, false),
 	[RTL_GIGA_MAC_VER_28] =
-		_R("RTL8168dp/8111dp",	RTL_TD_1, NULL),
+		_R("RTL8168dp/8111dp",	RTL_TD_1, NULL, JUMBO_9K, false),
 	[RTL_GIGA_MAC_VER_29] =
-		_R("RTL8105e",		RTL_TD_1, FIRMWARE_8105E_1),
+		_R("RTL8105e",		RTL_TD_1, FIRMWARE_8105E_1,
+							JUMBO_1K, true),
 	[RTL_GIGA_MAC_VER_30] =
-		_R("RTL8105e",		RTL_TD_1, FIRMWARE_8105E_1),
+		_R("RTL8105e",		RTL_TD_1, FIRMWARE_8105E_1,
+							JUMBO_1K, true),
 	[RTL_GIGA_MAC_VER_31] =
-		_R("RTL8168dp/8111dp",	RTL_TD_1, NULL),
+		_R("RTL8168dp/8111dp",	RTL_TD_1, NULL, JUMBO_9K, false),
 	[RTL_GIGA_MAC_VER_32] =
-		_R("RTL8168e/8111e",	RTL_TD_1, FIRMWARE_8168E_1),
+		_R("RTL8168e/8111e",	RTL_TD_1, FIRMWARE_8168E_1,
+							JUMBO_9K, false),
 	[RTL_GIGA_MAC_VER_33] =
-		_R("RTL8168e/8111e",	RTL_TD_1, FIRMWARE_8168E_2)
+		_R("RTL8168e/8111e",	RTL_TD_1, FIRMWARE_8168E_2,
+							JUMBO_9K, false)
 };
 #undef _R

@@ -443,8 +462,12 @@ enum rtl_register_content {
 	/* Config3 register p.25 */
 	MagicPacket	= (1 << 5),	/* Wake up when receives a Magic Packet */
 	LinkUp		= (1 << 4),	/* Wake up when the cable connection is re-established */
+	Jumbo_En0	= (1 << 2),	/* 8168 only. Reserved in the 8168b */
 	Beacon_en	= (1 << 0),	/* 8168 only. Reserved in the 8168b */
 
+	/* Config4 register */
+	Jumbo_En1	= (1 << 1),	/* 8168 only. Reserved in the 8168b */
+
 	/* Config5 register p.27 */
 	BWF		= (1 << 6),	/* Accept Broadcast wakeup frame */
 	MWF		= (1 << 5),	/* Accept Multicast wakeup frame */

@@ -653,6 +676,11 @@ struct rtl8169_private {
 		void (*up)(struct rtl8169_private *);
 	} pll_power_ops;
 
+	struct jumbo_ops {
+		void (*enable)(struct rtl8169_private *);
+		void (*disable)(struct rtl8169_private *);
+	} jumbo_ops;
+
 	int (*set_speed)(struct net_device *, u8 aneg, u16 sp, u8 dpx, u32 adv);
 	int (*get_settings)(struct net_device *, struct ethtool_cmd *);
 	void (*phy_reset_enable)(struct rtl8169_private *tp);

@@ -707,6 +735,21 @@ static int rtl8169_poll(struct napi_stru
 static const unsigned int rtl8169_rx_config =
 	(RX_FIFO_THRESH << RxCfgFIFOShift) | (RX_DMA_BURST << RxCfgDMAShift);
 
+static void rtl_tx_performance_tweak(struct pci_dev *pdev, u16 force)
+{
+	struct net_device *dev = pci_get_drvdata(pdev);
+	struct rtl8169_private *tp = netdev_priv(dev);
+	int cap = tp->pcie_cap;
+
+	if (cap) {
+		u16 ctl;
+
+		pci_read_config_word(pdev, cap + PCI_EXP_DEVCTL, &ctl);
+		ctl = (ctl & ~PCI_EXP_DEVCTL_READRQ) | force;
+		pci_write_config_word(pdev, cap + PCI_EXP_DEVCTL, ctl);
+	}
+}
+
 static u32 ocp_read(struct rtl8169_private *tp, u8 mask, u16 reg)
 {
 	void __iomem *ioaddr = tp->mmio_addr;

@@ -1375,9 +1418,15 @@ static int rtl8169_set_settings(struct n
 
 static u32 rtl8169_fix_features(struct net_device *dev, u32 features)
 {
+	struct rtl8169_private *tp = netdev_priv(dev);
+
 	if (dev->mtu > TD_MSS_MAX)
 		features &= ~NETIF_F_ALL_TSO;
 
+	if (dev->mtu > JUMBO_1K &&
+	    !rtl_chip_infos[tp->mac_version].jumbo_tx_csum)
+		features &= ~NETIF_F_IP_CSUM;
+
 	return features;
 }

@@ -3176,8 +3225,8 @@ static void r8168_pll_power_up(struct rt
 	r8168_phy_power_up(tp);
 }
 
-static void rtl_pll_power_op(struct rtl8169_private *tp,
-			     void (*op)(struct rtl8169_private *))
+static void rtl_generic_op(struct rtl8169_private *tp,
+			   void (*op)(struct rtl8169_private *))
 {
 	if (op)
 		op(tp);

@@ -3185,12 +3234,12 @@ static void rtl_pll_power_op(struct rtl8
 
 static void rtl_pll_power_down(struct rtl8169_private *tp)
 {
-	rtl_pll_power_op(tp, tp->pll_power_ops.down);
+	rtl_generic_op(tp, tp->pll_power_ops.down);
 }
 
 static void rtl_pll_power_up(struct rtl8169_private *tp)
 {
-	rtl_pll_power_op(tp, tp->pll_power_ops.up);
+	rtl_generic_op(tp, tp->pll_power_ops.up);
 }
 
 static void __devinit rtl_init_pll_power_ops(struct rtl8169_private *tp)

@@ -3237,6 +3286,149 @@ static void __devinit rtl_init_pll_power
 	}
 }
 
+static void rtl_hw_jumbo_enable(struct rtl8169_private *tp)
+{
+	rtl_generic_op(tp, tp->jumbo_ops.enable);
+}
+
+static void rtl_hw_jumbo_disable(struct rtl8169_private *tp)
+{
+	rtl_generic_op(tp, tp->jumbo_ops.disable);
+}
+
+static void r8168c_hw_jumbo_enable(struct rtl8169_private *tp)
+{
+	void __iomem *ioaddr = tp->mmio_addr;
+
+	RTL_W8(Config3, RTL_R8(Config3) | Jumbo_En0);
+	RTL_W8(Config4, RTL_R8(Config4) | Jumbo_En1);
+	rtl_tx_performance_tweak(tp->pci_dev, 0x2 << MAX_READ_REQUEST_SHIFT);
+}
+
+static void r8168c_hw_jumbo_disable(struct rtl8169_private *tp)
+{
+	void __iomem *ioaddr = tp->mmio_addr;
+
+	RTL_W8(Config3, RTL_R8(Config3) & ~Jumbo_En0);
+	RTL_W8(Config4, RTL_R8(Config4) & ~Jumbo_En1);
+	rtl_tx_performance_tweak(tp->pci_dev, 0x5 << MAX_READ_REQUEST_SHIFT);
+}
+
+static void r8168dp_hw_jumbo_enable(struct rtl8169_private *tp)
+{
+	void __iomem *ioaddr = tp->mmio_addr;
+
+	RTL_W8(Config3, RTL_R8(Config3) | Jumbo_En0);
+}
+
+static void r8168dp_hw_jumbo_disable(struct rtl8169_private *tp)
+{
+	void __iomem *ioaddr = tp->mmio_addr;
+
+	RTL_W8(Config3, RTL_R8(Config3) & ~Jumbo_En0);
+}
+
+static void r8168e_hw_jumbo_enable(struct rtl8169_private *tp)
+{
+	void __iomem *ioaddr = tp->mmio_addr;
+	struct pci_dev *pdev = tp->pci_dev;
+
+	RTL_W8(MaxTxPacketSize, 0x3f);
+	RTL_W8(Config3, RTL_R8(Config3) | Jumbo_En0);
+	RTL_W8(Config4, RTL_R8(Config4) | 0x01);
+	pci_write_config_byte(pdev, 0x79, 0x20);
+}
+
+static void r8168e_hw_jumbo_disable(struct rtl8169_private *tp)
+{
+	void __iomem *ioaddr = tp->mmio_addr;
+	struct pci_dev *pdev = tp->pci_dev;
+
+	RTL_W8(MaxTxPacketSize, 0x0c);
+	RTL_W8(Config3, RTL_R8(Config3) & ~Jumbo_En0);
+	RTL_W8(Config4, RTL_R8(Config4) & ~0x01);
+	pci_write_config_byte(pdev, 0x79, 0x50);
+}
+
+static void r8168b_0_hw_jumbo_enable(struct rtl8169_private *tp)
+{
+	rtl_tx_performance_tweak(tp->pci_dev,
+		(0x2 << MAX_READ_REQUEST_SHIFT) | PCI_EXP_DEVCTL_NOSNOOP_EN);
+}
+
+static void r8168b_0_hw_jumbo_disable(struct rtl8169_private *tp)
+{
+	rtl_tx_performance_tweak(tp->pci_dev,
+		(0x5 << MAX_READ_REQUEST_SHIFT) | PCI_EXP_DEVCTL_NOSNOOP_EN);
+}
+
+static void r8168b_1_hw_jumbo_enable(struct rtl8169_private *tp)
+{
+	void __iomem *ioaddr = tp->mmio_addr;
+
+	r8168b_0_hw_jumbo_enable(tp);
+
+	RTL_W8(Config4, RTL_R8(Config4) | (1 << 0));
+}
+
+static void r8168b_1_hw_jumbo_disable(struct rtl8169_private *tp)
+{
+	void __iomem *ioaddr = tp->mmio_addr;
+
+	r8168b_0_hw_jumbo_disable(tp);
+
+	RTL_W8(Config4, RTL_R8(Config4) & ~(1 << 0));
+}
+
+static void __devinit rtl_init_jumbo_ops(struct rtl8169_private *tp)
+{
+	struct jumbo_ops *ops = &tp->jumbo_ops;
+
+	switch (tp->mac_version) {
+	case RTL_GIGA_MAC_VER_11:
+		ops->disable	= r8168b_0_hw_jumbo_disable;
+		ops->enable	= r8168b_0_hw_jumbo_enable;
+		break;
+	case RTL_GIGA_MAC_VER_12:
+	case RTL_GIGA_MAC_VER_17:
+		ops->disable	= r8168b_1_hw_jumbo_disable;
+		ops->enable	= r8168b_1_hw_jumbo_enable;
+		break;
+	case RTL_GIGA_MAC_VER_18: /* Wild guess. Needs info from Realtek. */
+	case RTL_GIGA_MAC_VER_19:
+	case RTL_GIGA_MAC_VER_20:
+	case RTL_GIGA_MAC_VER_21: /* Wild guess. Needs info from Realtek. */
+	case RTL_GIGA_MAC_VER_22:
+	case RTL_GIGA_MAC_VER_23:
+	case RTL_GIGA_MAC_VER_24:
+	case RTL_GIGA_MAC_VER_25:
+	case RTL_GIGA_MAC_VER_26:
+		ops->disable	= r8168c_hw_jumbo_disable;
+		ops->enable	= r8168c_hw_jumbo_enable;
+		break;
+	case RTL_GIGA_MAC_VER_27:
+	case RTL_GIGA_MAC_VER_28:
+		ops->disable	= r8168dp_hw_jumbo_disable;
+		ops->enable	= r8168dp_hw_jumbo_enable;
+		break;
+	case RTL_GIGA_MAC_VER_31: /* Wild guess. Needs info from Realtek. */
+	case RTL_GIGA_MAC_VER_32:
+	case RTL_GIGA_MAC_VER_33:
+		ops->disable	= r8168e_hw_jumbo_disable;
+		ops->enable	= r8168e_hw_jumbo_enable;
+		break;
+
+	/*
+	 * No action needed for jumbo frames with 8169.
+	 * No jumbo for 810x at all.
+	 */
+	default:
+		ops->disable	= NULL;
+		ops->enable	= NULL;
+		break;
+	}
+}
+
 static void rtl_hw_reset(struct rtl8169_private *tp)
 {
 	void __iomem *ioaddr = tp->mmio_addr;

@@ -3378,6 +3570,7 @@ rtl8169_init_one(struct pci_dev *pdev, c
 
 	rtl_init_mdio_ops(tp);
 	rtl_init_pll_power_ops(tp);
+	rtl_init_jumbo_ops(tp);
 
 	rtl8169_print_mac_version(tp);

@@ -3462,6 +3655,12 @@ rtl8169_init_one(struct pci_dev *pdev, c
 	netif_info(tp, probe, dev, "%s at 0x%lx, %pM, XID %08x IRQ %d\n",
 		   rtl_chip_infos[chipset].name, dev->base_addr, dev->dev_addr,
 		   (u32)(RTL_R32(TxConfig) & 0x9cf0f8ff), dev->irq);
+	if (rtl_chip_infos[chipset].jumbo_max != JUMBO_1K) {
+		netif_info(tp, probe, dev, "jumbo features [frames: %d bytes, "
+			   "tx checksumming: %s]\n",
+			   rtl_chip_infos[chipset].jumbo_max,
+			   rtl_chip_infos[chipset].jumbo_tx_csum ? "ok" : "ko");
+	}
 
 	if (tp->mac_version == RTL_GIGA_MAC_VER_27 ||
 	    tp->mac_version == RTL_GIGA_MAC_VER_28 ||

@@ -3786,21 +3985,6 @@ static void rtl_hw_start_8169(struct net
 	RTL_W16(IntrMask, tp->intr_event);
 }
 
-static void rtl_tx_performance_tweak(struct pci_dev *pdev, u16 force)
-{
-	struct net_device *dev = pci_get_drvdata(pdev);
-	struct rtl8169_private *tp = netdev_priv(dev);
-	int cap = tp->pcie_cap;
-
-	if (cap) {
-		u16 ctl;
-
-		pci_read_config_word(pdev, cap + PCI_EXP_DEVCTL, &ctl);
-		ctl = (ctl & ~PCI_EXP_DEVCTL_READRQ) | force;
-		pci_write_config_word(pdev, cap + PCI_EXP_DEVCTL, ctl);
-	}
-}
-
 static void rtl_csi_access_enable(void __iomem *ioaddr, u32 bits)
 {
 	u32 csi;

@@ -4343,9 +4527,17 @@ static void rtl_hw_start_8101(struct net
 
 static int rtl8169_change_mtu(struct net_device *dev, int new_mtu)
 {
-	if (new_mtu < ETH_ZLEN || new_mtu > SafeMtu)
+	struct rtl8169_private *tp = netdev_priv(dev);
+
+	if (new_mtu < ETH_ZLEN ||
+	    new_mtu > rtl_chip_infos[tp->mac_version].jumbo_max)
 		return -EINVAL;
 
+	if (new_mtu > ETH_DATA_LEN)
+		rtl_hw_jumbo_enable(tp);
+	else
+		rtl_hw_jumbo_disable(tp);
+
 	dev->mtu = new_mtu;
 	netdev_update_features(dev);

Greg Kroah-Hartman

author

2012-10-11 02:20 UTC #66 lore raw

PATCH 64/84 r8169: expand received packet length indication.

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

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

From: Francois Romieu <[email protected]>

commit deb9d93c89d311714a60809b28160e538e1cbb43 upstream.

8168d and above allow jumbo frames beyond 8k. Bump the received packet length check before enabling jumbo frames on these chipsets.

Frame length indication covers bits 0..13 of the first Rx descriptor 32 bits for the 8169 and 8168. I only have authoritative documentation for the allowed use of the extra (13) bit with the 8169 and 8168c. Realtek's drivers use the same mask for the 816x and the fast ethernet only 810x.

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/net/r8169.c +1 -1

@@ -5137,7 +5137,7 @@ static int rtl8169_rx_interrupt(struct n
 		} else {
 			struct sk_buff *skb;
 			dma_addr_t addr = le64_to_cpu(desc->addr);
-			int pkt_size = (status & 0x00001FFF) - 4;
+			int pkt_size = (status & 0x00003fff) - 4;
 
 			/*
 			 * The driver does not support incoming fragmented

Greg Kroah-Hartman

author

2012-10-11 02:15 UTC #67 lore raw

PATCH 65/84 r8169: increase the delay parameter of pm_schedule_suspend

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

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

From: hayeswang <[email protected]>

commit 10953db8e1a278742ef7e64a3d1491802bcfa98b upstream
The link down would occur when reseting PHY. And it would take about 2 ~ 5 seconds from link down to link up. If the delay of pm_schedule_suspend is not long enough, the device would enter runtime_suspend before link up. After link up, the device would wake up and reset PHY again. Then, you would find the driver keep in a loop of runtime_suspend and rumtime_resume.

Signed-off-by Hayes Wang <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/net/r8169.c +1 -1

@@ -1157,7 +1157,7 @@ static void __rtl8169_check_link_status(
 		netif_carrier_off(dev);
 		netif_info(tp, ifdown, dev, "link down\n");
 		if (pm)
-			pm_schedule_suspend(&tp->pci_dev->dev, 100);
+			pm_schedule_suspend(&tp->pci_dev->dev, 5000);
 	}
 	spin_unlock_irqrestore(&tp->lock, flags);
 }

Greg Kroah-Hartman

author

2012-10-11 02:20 UTC #68 lore raw

PATCH 66/84 r8169: Rx FIFO overflow fixes.

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

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

From: Francois Romieu <[email protected]>

commit 811fd3010cf512f2e23e6c4c912aad54516dc706 upstream.

Realtek has specified that the post 8168c gigabit chips and the post 8105e fast ethernet chips recover automatically from a Rx FIFO overflow. The driver does not need to clear the RxFIFOOver bit of IntrStatus and it should rather avoid messing it.

The implementation deserves some explanation:
1. events outside of the intr_event bit mask are now ignored. It enforces a no-processing policy for the events that either should not be there or should be ignored.

2. RxFIFOOver was already ignored in rtl_cfg_infos[RTL_CFG_1] for the whole 8168 line of chips with two exceptions:
- RTL_GIGA_MAC_VER_22 since b5ba6d12bdac21bc0620a5089e0f24e362645efd ("use RxFIFO overflow workaround for 8168c chipset."). This one should now be correctly handled.
- RTL_GIGA_MAC_VER_11 (8168b) which requires a different Rx FIFO overflow processing.

Though it does not conform to Realtek suggestion above, the updated driver includes no change for RTL_GIGA_MAC_VER_12 and RTL_GIGA_MAC_VER_17. Both are 8168b. RTL_GIGA_MAC_VER_12 is common and a bit old so I'd rather wait for experimental evidence that the change suggested by Realtek really helps or does not hurt in unexpected ways.

Removed case statements in rtl8169_interrupt are only 8168 relevant.

3. RxFIFOOver is masked for post 8105e 810x chips, namely the sole 8105e (RTL_GIGA_MAC_VER_30) itself.

Cc hayeswang <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |   54 +++++++++++++++++++++++-----------------------------
 1 file changed, 24 insertions(+), 30 deletions(-)

drivers/net/r8169.c +24 ~~-30~~

@@ -1088,17 +1088,21 @@ static u8 rtl8168d_efuse_read(void __iom
 	return value;
 }
 
-static void rtl8169_irq_mask_and_ack(void __iomem *ioaddr)
+static void rtl8169_irq_mask_and_ack(struct rtl8169_private *tp)
 {
-	RTL_W16(IntrMask, 0x0000);
+	void __iomem *ioaddr = tp->mmio_addr;
 
-	RTL_W16(IntrStatus, 0xffff);
+	RTL_W16(IntrMask, 0x0000);
+	RTL_W16(IntrStatus, tp->intr_event);
+	RTL_R8(ChipCmd);
 }
 
-static void rtl8169_asic_down(void __iomem *ioaddr)
+static void rtl8169_asic_down(struct rtl8169_private *tp)
 {
+	void __iomem *ioaddr = tp->mmio_addr;
+
 	RTL_W8(ChipCmd, 0x00);
-	rtl8169_irq_mask_and_ack(ioaddr);
+	rtl8169_irq_mask_and_ack(tp);
 	RTL_R16(CPlusCmd);
 }

@@ -3817,7 +3821,7 @@ static void rtl8169_hw_reset(struct rtl8
 	void __iomem *ioaddr = tp->mmio_addr;
 
 	/* Disable interrupts */
-	rtl8169_irq_mask_and_ack(ioaddr);
+	rtl8169_irq_mask_and_ack(tp);
 
 	if (tp->mac_version == RTL_GIGA_MAC_VER_27 ||
 	    tp->mac_version == RTL_GIGA_MAC_VER_28 ||

@@ -4284,8 +4288,7 @@ static void rtl_hw_start_8168(struct net
 	RTL_W16(IntrMitigate, 0x5151);
 
 	/* Work around for RxFIFO overflow. */
-	if (tp->mac_version == RTL_GIGA_MAC_VER_11 ||
-	    tp->mac_version == RTL_GIGA_MAC_VER_22) {
+	if (tp->mac_version == RTL_GIGA_MAC_VER_11) {
 		tp->intr_event |= RxFIFOOver | PCSTimeout;
 		tp->intr_event &= ~RxOverflow;
 	}

@@ -4467,6 +4470,11 @@ static void rtl_hw_start_8101(struct net
 	void __iomem *ioaddr = tp->mmio_addr;
 	struct pci_dev *pdev = tp->pci_dev;
 
+	if (tp->mac_version >= RTL_GIGA_MAC_VER_30) {
+		tp->intr_event &= ~RxFIFOOver;
+		tp->napi_event &= ~RxFIFOOver;
+	}
+
 	if (tp->mac_version == RTL_GIGA_MAC_VER_13 ||
 	    tp->mac_version == RTL_GIGA_MAC_VER_16) {
 		int cap = tp->pcie_cap;

@@ -4738,7 +4746,7 @@ static void rtl8169_wait_for_quiescence(
 	/* Wait for any pending NAPI task to complete */
 	napi_disable(&tp->napi);
 
-	rtl8169_irq_mask_and_ack(ioaddr);
+	rtl8169_irq_mask_and_ack(tp);
 
 	tp->intr_mask = 0xffff;
 	RTL_W16(IntrMask, tp->intr_event);

@@ -5200,13 +5208,17 @@ static irqreturn_t rtl8169_interrupt(int
 	 */
 	status = RTL_R16(IntrStatus);
 	while (status && status != 0xffff) {
+		status &= tp->intr_event;
+		if (!status)
+			break;
+
 		handled = 1;
 
 		/* Handle all of the error cases first. These will reset
 		 * the chip, so just exit the loop.
 		 */
 		if (unlikely(!netif_running(dev))) {
-			rtl8169_asic_down(ioaddr);
+			rtl8169_asic_down(tp);
 			break;
 		}

@@ -5214,27 +5226,9 @@ static irqreturn_t rtl8169_interrupt(int
 			switch (tp->mac_version) {
 			/* Work around for rx fifo overflow */
 			case RTL_GIGA_MAC_VER_11:
-			case RTL_GIGA_MAC_VER_22:
-			case RTL_GIGA_MAC_VER_26:
 				netif_stop_queue(dev);
 				rtl8169_tx_timeout(dev);
 				goto done;
-			/* Testers needed. */
-			case RTL_GIGA_MAC_VER_17:
-			case RTL_GIGA_MAC_VER_19:
-			case RTL_GIGA_MAC_VER_20:
-			case RTL_GIGA_MAC_VER_21:
-			case RTL_GIGA_MAC_VER_23:
-			case RTL_GIGA_MAC_VER_24:
-			case RTL_GIGA_MAC_VER_27:
-			case RTL_GIGA_MAC_VER_28:
-			case RTL_GIGA_MAC_VER_31:
-			/* Experimental science. Pktgen proof. */
-			case RTL_GIGA_MAC_VER_12:
-			case RTL_GIGA_MAC_VER_25:
-				if (status == RxFIFOOver)
-					goto done;
-				break;
 			default:
 				break;
 			}

@@ -5329,7 +5323,7 @@ static void rtl8169_down(struct net_devi
 
 	spin_lock_irq(&tp->lock);
 
-	rtl8169_asic_down(ioaddr);
+	rtl8169_asic_down(tp);
 	/*
 	 * At this point device interrupts can not be enabled in any function,
 	 * as netif_running is not true (rtl8169_interrupt, rtl8169_reset_task,

@@ -5583,7 +5577,7 @@ static void rtl_shutdown(struct pci_dev
 
 	spin_lock_irq(&tp->lock);
 
-	rtl8169_asic_down(ioaddr);
+	rtl8169_asic_down(tp);
 
 	spin_unlock_irq(&tp->lock);

Greg Kroah-Hartman

author

2012-10-11 02:15 UTC #69 lore raw

PATCH 67/84 r8169: fix Config2 MSIEnable bit setting.

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

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

From: françois romieu <[email protected]>

commit 2ca6cf06d988fea21e812a86be79353352677c9c upstream.

The MSIEnable bit is only available for the 8169.

Avoid Config2 writes for the post-8169 8168 and 810x.

Reported-by Su Kang Yin <[email protected]>

Cc Hayes Wang <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |   14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

drivers/net/r8169.c +8 -6

@@ -448,7 +448,6 @@ enum rtl_register_content {
 	/* Config1 register p.24 */
 	LEDS1		= (1 << 7),
 	LEDS0		= (1 << 6),
-	MSIEnable	= (1 << 5),	/* Enable Message Signaled Interrupt */
 	Speed_down	= (1 << 4),
 	MEMMAP		= (1 << 3),
 	IOMAP		= (1 << 2),

@@ -456,6 +455,7 @@ enum rtl_register_content {
 	PMEnable	= (1 << 0),	/* Power Management Enable */
 
 	/* Config2 register p. 25 */
+	MSIEnable	= (1 << 5),	/* 8169 only. Reserved in the 8168. */
 	PCI_Clock_66MHz = 0x01,
 	PCI_Clock_33MHz = 0x00,

@@ -3003,22 +3003,24 @@ static const struct rtl_cfg_info {
 };
 
 /* Cfg9346_Unlock assumed. */
-static unsigned rtl_try_msi(struct pci_dev *pdev, void __iomem *ioaddr,
+static unsigned rtl_try_msi(struct rtl8169_private *tp,
 			    const struct rtl_cfg_info *cfg)
 {
+	void __iomem *ioaddr = tp->mmio_addr;
 	unsigned msi = 0;
 	u8 cfg2;
 
 	cfg2 = RTL_R8(Config2) & ~MSIEnable;
 	if (cfg->features & RTL_FEATURE_MSI) {
-		if (pci_enable_msi(pdev)) {
-			dev_info(&pdev->dev, "no MSI. Back to INTx.\n");
+		if (pci_enable_msi(tp->pci_dev)) {
+			netif_info(tp, hw, tp->dev, "no MSI. Back to INTx.\n");
 		} else {
 			cfg2 |= MSIEnable;
 			msi = RTL_FEATURE_MSI;
 		}
 	}
-	RTL_W8(Config2, cfg2);
+	if (tp->mac_version <= RTL_GIGA_MAC_VER_06)
+		RTL_W8(Config2, cfg2);
 	return msi;
 }

@@ -3588,7 +3590,7 @@ rtl8169_init_one(struct pci_dev *pdev, c
 		tp->features |= RTL_FEATURE_WOL;
 	if ((RTL_R8(Config5) & (UWF | BWF | MWF)) != 0)
 		tp->features |= RTL_FEATURE_WOL;
-	tp->features |= rtl_try_msi(pdev, ioaddr, cfg);
+	tp->features |= rtl_try_msi(tp, cfg);
 	RTL_W8(Cfg9346, Cfg9346_Lock);
 
 	if ((tp->mac_version <= RTL_GIGA_MAC_VER_06) &&

Greg Kroah-Hartman

author

2012-10-11 02:27 UTC #70 lore raw

PATCH 68/84 r8169: missing barriers.

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

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

From: Francois Romieu <[email protected]>

commit 1e874e041fc7c222cbd85b20c4406070be1f687a upstream.

Cc Hayes Wang <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |    5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

drivers/net/r8169.c +2 -3

@@ -4957,7 +4957,7 @@ static netdev_tx_t rtl8169_start_xmit(st
 
 	if (TX_BUFFS_AVAIL(tp) < MAX_SKB_FRAGS) {
 		netif_stop_queue(dev);
-		smp_rmb();
+		smp_mb();
 		if (TX_BUFFS_AVAIL(tp) >= MAX_SKB_FRAGS)
 			netif_wake_queue(dev);
 	}

@@ -5058,7 +5058,7 @@ static void rtl8169_tx_interrupt(struct
 
 	if (tp->dirty_tx != dirty_tx) {
 		tp->dirty_tx = dirty_tx;
-		smp_wmb();
+		smp_mb();
 		if (netif_queue_stopped(dev) &&
 		    (TX_BUFFS_AVAIL(tp) >= MAX_SKB_FRAGS)) {
 			netif_wake_queue(dev);

@@ -5069,7 +5069,6 @@ static void rtl8169_tx_interrupt(struct
 		 * of start_xmit activity is detected (if it is not detected,
 		 * it is slow enough). -- FR
 		 */
-		smp_rmb();
 		if (tp->cur_tx != dirty_tx)
 			RTL_W8(TxPoll, NPQ);
 	}

Greg Kroah-Hartman

author

2012-10-11 02:26 UTC #71 lore raw

PATCH 69/84 r8169: runtime resume before shutdown.

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

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

From: françois romieu <[email protected]>

commit 2a15cd2ff488a9fdb55e5e34060f499853b27c77 upstream.

With runtime PM, if the ethernet cable is disconnected, the device is transitioned to D3 state to conserve energy. If the system is shutdown in this state, any register accesses in rtl_shutdown are dropped on the floor. As the device was programmed by .runtime_suspend() to wake on link changes, it is thus brought back up as soon as the link recovers.

Resuming every suspended device through the driver core would slow things down and it is not clear how many devices really need it now.

Original report and D0 transition patch by Sameer Nanda. Patch has been changed to comply with advices by Rafael J. Wysocki and the PM folks.

Reported-by Sameer Nanda <[email protected]>

Cc Rafael J. Wysocki <[email protected]>

Cc Hayes Wang <[email protected]>

Cc Alan Stern <[email protected]>

Acked-by Rafael J. Wysocki <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |    5 +++++
 1 file changed, 5 insertions(+)

drivers/net/r8169.c +5 -0

@@ -5570,6 +5570,9 @@ static void rtl_shutdown(struct pci_dev
 	struct net_device *dev = pci_get_drvdata(pdev);
 	struct rtl8169_private *tp = netdev_priv(dev);
 	void __iomem *ioaddr = tp->mmio_addr;
+	struct device *d = &pdev->dev;
+
+	pm_runtime_get_sync(d);
 
 	rtl8169_net_suspend(dev);

@@ -5598,6 +5601,8 @@ static void rtl_shutdown(struct pci_dev
 		pci_wake_from_d3(pdev, true);
 		pci_set_power_state(pdev, PCI_D3hot);
 	}
+
+	pm_runtime_put_noidle(d);
 }
 
 static struct pci_driver rtl8169_pci_driver = {

Greg Kroah-Hartman

author

2012-10-11 02:26 UTC #72 lore raw

PATCH 70/84 r8169: Config1 is read-only on 8168c and later.

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

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

From: Francois Romieu <[email protected]>

commit 851e60221926a53344b4227879858bef841b0477 upstream.

Suggested by Hayes.

Cc Hayes Wang <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |   15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)

drivers/net/r8169.c +13 -2

@@ -1223,7 +1223,6 @@ static void __rtl8169_set_wol(struct rtl
 		u16 reg;
 		u8  mask;
 	} cfg[] = {
-		{ WAKE_ANY,   Config1, PMEnable },
 		{ WAKE_PHY,   Config3, LinkUp },
 		{ WAKE_MAGIC, Config3, MagicPacket },
 		{ WAKE_UCAST, Config5, UWF },

@@ -1231,16 +1230,28 @@ static void __rtl8169_set_wol(struct rtl
 		{ WAKE_MCAST, Config5, MWF },
 		{ WAKE_ANY,   Config5, LanWake }
 	};
+	u8 options;
 
 	RTL_W8(Cfg9346, Cfg9346_Unlock);
 
 	for (i = 0; i < ARRAY_SIZE(cfg); i++) {
-		u8 options = RTL_R8(cfg[i].reg) & ~cfg[i].mask;
+		options = RTL_R8(cfg[i].reg) & ~cfg[i].mask;
 		if (wolopts & cfg[i].opt)
 			options |= cfg[i].mask;
 		RTL_W8(cfg[i].reg, options);
 	}
 
+	switch (tp->mac_version) {
+	case RTL_GIGA_MAC_VER_01 ... RTL_GIGA_MAC_VER_17:
+		options = RTL_R8(Config1) & ~PMEnable;
+		if (wolopts)
+			options |= PMEnable;
+		RTL_W8(Config1, options);
+		break;
+	default:
+		break;
+	}
+
 	RTL_W8(Cfg9346, Cfg9346_Lock);
 }

Greg Kroah-Hartman

author

2012-10-11 02:28 UTC #73 lore raw

PATCH 71/84 r8169: 8168c and later require bit 0x20 to be set in Config2 for PME signaling.

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

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

From: Francois Romieu <[email protected]>

commit d387b427c973974dd619a33549c070ac5d0e089f upstream.

The new 84xx stopped flying below the radars.

Cc Hayes Wang <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |    6 ++++++
 1 file changed, 6 insertions(+)

drivers/net/r8169.c +6 -0

@@ -299,6 +299,8 @@ enum rtl_registers {
 	Config0		= 0x51,
 	Config1		= 0x52,
 	Config2		= 0x53,
+#define PME_SIGNAL			(1 << 5)	/* 8168c and later */
+
 	Config3		= 0x54,
 	Config4		= 0x55,
 	Config5		= 0x56,

@@ -1249,6 +1251,10 @@ static void __rtl8169_set_wol(struct rtl
 		RTL_W8(Config1, options);
 		break;
 	default:
+		options = RTL_R8(Config2) & ~PME_SIGNAL;
+		if (wolopts)
+			options |= PME_SIGNAL;
+		RTL_W8(Config2, options);
 		break;
 	}

Greg Kroah-Hartman

author

2012-10-11 02:28 UTC #74 lore raw

PATCH 72/84 r8169: fix unsigned int wraparound with TSO

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

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

From: Julien Ducourthial <[email protected]>

commit 477206a018f902895bfcd069dd820bfe94c187b1 upstream.

The r8169 may get stuck or show bad behaviour after activating TSO :
the net_device is not stopped when it has no more TX descriptors. This problem comes from TX_BUFS_AVAIL which may reach -1 when all transmit descriptors are in use. The patch simply tries to keep positive values.

Tested with 8111d(onboard) on a D510MO, and with 8111e(onboard) on a Zotac 890GXITX.

Signed-off-by Julien Ducourthial <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |   16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

drivers/net/r8169.c +10 -6

@@ -58,8 +58,12 @@
 #define R8169_MSG_DEFAULT \
 	(NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_IFUP | NETIF_MSG_IFDOWN)
 
-#define TX_BUFFS_AVAIL(tp) \
-	(tp->dirty_tx + NUM_TX_DESC - tp->cur_tx - 1)
+#define TX_SLOTS_AVAIL(tp) \
+	(tp->dirty_tx + NUM_TX_DESC - tp->cur_tx)
+
+/* A skbuff with nr_frags needs nr_frags+1 entries in the tx queue */
+#define TX_FRAGS_READY_FOR(tp,nr_frags) \
+	(TX_SLOTS_AVAIL(tp) >= (nr_frags + 1))
 
 /* Maximum number of multicast addresses to filter (vs. Rx-all-multicast).
    The RTL chips use a 64 element hash table based on the Ethernet CRC. */

@@ -4924,7 +4928,7 @@ static netdev_tx_t rtl8169_start_xmit(st
 	u32 opts[2];
 	int frags;
 
-	if (unlikely(TX_BUFFS_AVAIL(tp) < skb_shinfo(skb)->nr_frags)) {
+	if (unlikely(!TX_FRAGS_READY_FOR(tp, skb_shinfo(skb)->nr_frags))) {
 		netif_err(tp, drv, dev, "BUG! Tx Ring full when queue awake!\n");
 		goto err_stop_0;
 	}

@@ -4972,10 +4976,10 @@ static netdev_tx_t rtl8169_start_xmit(st
 
 	RTL_W8(TxPoll, NPQ);
 
-	if (TX_BUFFS_AVAIL(tp) < MAX_SKB_FRAGS) {
+	if (!TX_FRAGS_READY_FOR(tp, MAX_SKB_FRAGS)) {
 		netif_stop_queue(dev);
 		smp_mb();
-		if (TX_BUFFS_AVAIL(tp) >= MAX_SKB_FRAGS)
+		if (TX_FRAGS_READY_FOR(tp, MAX_SKB_FRAGS))
 			netif_wake_queue(dev);
 	}

@@ -5077,7 +5081,7 @@ static void rtl8169_tx_interrupt(struct
 		tp->dirty_tx = dirty_tx;
 		smp_mb();
 		if (netif_queue_stopped(dev) &&
-		    (TX_BUFFS_AVAIL(tp) >= MAX_SKB_FRAGS)) {
+		    TX_FRAGS_READY_FOR(tp, MAX_SKB_FRAGS)) {
 			netif_wake_queue(dev);
 		}
 		/*

Greg Kroah-Hartman

author

2012-10-11 02:26 UTC #75 lore raw

PATCH 73/84 r8169: call netif_napi_del at errpaths and at driver unload

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

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

From: Devendra Naga <[email protected]>

commit ad1be8d345416a794dea39761a374032aa471a76 upstream.

When register_netdev fails, the init'ed NAPIs by netif_napi_add must be deleted with netif_napi_del, and also when driver unloads, it should delete the NAPI before unregistering netdevice using unregister_netdev.

Signed-off-by Devendra Naga <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/net/r8169.c |    3 +++
 1 file changed, 3 insertions(+)

drivers/net/r8169.c +3 -0

@@ -3706,6 +3706,7 @@ out:
 	return rc;
 
 err_out_msi_4:
+	netif_napi_del(&tp->napi);
 	rtl_disable_msi(pdev, tp);
 	iounmap(ioaddr);
 err_out_free_res_3:

@@ -3731,6 +3732,8 @@ static void __devexit rtl8169_remove_one
 
 	cancel_delayed_work_sync(&tp->task);
 
+	netif_napi_del(&tp->napi);
+
 	unregister_netdev(dev);
 
 	rtl_release_firmware(tp);

Greg Kroah-Hartman

author

2012-10-11 02:28 UTC #76 lore raw

PATCH 74/84 revert "mm: mempolicy: Let vma_merge and vma_split handle vma->vm_policy linkages"

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

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

From: KOSAKI Motohiro <[email protected]>

commit 8d34694c1abf29df1f3c7317936b7e3e2e308d9b upstream.

Commit 05f144a0d5c2 ("mm: mempolicy: Let vma_merge and vma_split handle vma->vm_policy linkages") removed vma->vm_policy updates code but it is the purpose of mbind_range(). Now, mbind_range() is virtually a no-op and while it does not allow memory corruption it is not the right fix. This patch is a revert.

[[email protected]: Edited changelog]

Signed-off-by Mel Gorman <[email protected]>

Cc Christoph Lameter <[email protected]>

Cc Josh Boyer <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 mm/mempolicy.c |   41 ++++++++++++++++++++++++-----------------
 1 file changed, 24 insertions(+), 17 deletions(-)

mm/mempolicy.c +24 ~~-17~~

@@ -606,6 +606,27 @@ check_range(struct mm_struct *mm, unsign
 	return first;
 }
 
+/* Apply policy to a single VMA */
+static int policy_vma(struct vm_area_struct *vma, struct mempolicy *new)
+{
+	int err = 0;
+	struct mempolicy *old = vma->vm_policy;
+
+	pr_debug("vma %lx-%lx/%lx vm_ops %p vm_file %p set_policy %p\n",
+		 vma->vm_start, vma->vm_end, vma->vm_pgoff,
+		 vma->vm_ops, vma->vm_file,
+		 vma->vm_ops ? vma->vm_ops->set_policy : NULL);
+
+	if (vma->vm_ops && vma->vm_ops->set_policy)
+		err = vma->vm_ops->set_policy(vma, new);
+	if (!err) {
+		mpol_get(new);
+		vma->vm_policy = new;
+		mpol_put(old);
+	}
+	return err;
+}
+
 /* Step 2: apply policy to a range and do splits. */
 static int mbind_range(struct mm_struct *mm, unsigned long start,
 		       unsigned long end, struct mempolicy *new_pol)

@@ -645,23 +666,9 @@ static int mbind_range(struct mm_struct
 			if (err)
 				goto out;
 		}
-
-		/*
-		 * Apply policy to a single VMA. The reference counting of
-		 * policy for vma_policy linkages has already been handled by
-		 * vma_merge and split_vma as necessary. If this is a shared
-		 * policy then ->set_policy will increment the reference count
-		 * for an sp node.
-		 */
-		pr_debug("vma %lx-%lx/%lx vm_ops %p vm_file %p set_policy %p\n",
-			vma->vm_start, vma->vm_end, vma->vm_pgoff,
-			vma->vm_ops, vma->vm_file,
-			vma->vm_ops ? vma->vm_ops->set_policy : NULL);
-		if (vma->vm_ops && vma->vm_ops->set_policy) {
-			err = vma->vm_ops->set_policy(vma, new_pol);
-			if (err)
-				goto out;
-		}
+		err = policy_vma(vma, new_pol);
+		if (err)
+			goto out;
 	}
 
  out:

Greg Kroah-Hartman

author

2012-10-11 02:27 UTC #77 lore raw

PATCH 75/84 mempolicy: remove mempolicy sharing

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

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

From: KOSAKI Motohiro <[email protected]>

commit 869833f2c5c6e4dd09a5378cfc665ffb4615e5d2 upstream.

Dave Jones' system call fuzz testing tool "trinity" triggered the following bug error with slab debugging enabled

    =============================================================================
    BUG numa_policy (Not tainted): Poison overwritten
    -----------------------------------------------------------------------------

    INFO: 0xffff880146498250-0xffff880146498250. First byte 0x6a instead of 0x6b
    INFO: Allocated in mpol_new+0xa3/0x140 age=46310 cpu=6 pid=32154
     __slab_alloc+0x3d3/0x445
     kmem_cache_alloc+0x29d/0x2b0
     mpol_new+0xa3/0x140
     sys_mbind+0x142/0x620
     system_call_fastpath+0x16/0x1b

    INFO: Freed in __mpol_put+0x27/0x30 age=46268 cpu=6 pid=32154
     __slab_free+0x2e/0x1de
     kmem_cache_free+0x25a/0x260
     __mpol_put+0x27/0x30
     remove_vma+0x68/0x90
     exit_mmap+0x118/0x140
     mmput+0x73/0x110
     exit_mm+0x108/0x130
     do_exit+0x162/0xb90
     do_group_exit+0x4f/0xc0
     sys_exit_group+0x17/0x20
     system_call_fastpath+0x16/0x1b

    INFO: Slab 0xffffea0005192600 objects=27 used=27 fp=0x          (null) flags=0x20000000004080
    INFO: Object 0xffff880146498250 @offset=592 fp=0xffff88014649b9d0

The problem is that the structure is being prematurely freed due to a reference count imbalance. In the following case mbind(addr, len) should replace the memory policies of both vma1 and vma2 and thus they will become to share the same mempolicy and the new mempolicy will have the MPOL_F_SHARED flag.

  +-------------------+-------------------+
  |     vma1          |     vma2(shmem)   |
  +-------------------+-------------------+
  |                                       |
 addr                                 addr+len

alloc_pages_vma() uses get_vma_policy() and mpol_cond_put() pair for maintaining the mempolicy reference count. The current rule is that get_vma_policy() only increments refcount for shmem VMA and mpol_conf_put() only decrements refcount if the policy has MPOL_F_SHARED.

In above case, vma1 is not shmem vma and vma->policy has MPOL_F_SHARED! The reference count will be decreased even though was not increased whenever alloc_page_vma() is called. This has been broken since commit [52cd3b07: mempolicy: rework mempolicy Reference Counting] in 2008.

There is another serious bug with the sharing of memory policies. Currently, mempolicy rebind logic (it is called from cpuset rebinding) ignores a refcount of mempolicy and override it forcibly. Thus, any mempolicy sharing may cause mempolicy corruption. The bug was introduced by commit [68860ec1: cpusets: automatic numa mempolicy rebinding].

Ideally, the shared policy handling would be rewritten to either properly handle COW of the policy structures or at least reference count MPOL_F_SHARED based exclusively on information within the policy. However, this patch takes the easier approach of disabling any policy sharing between VMAs. Each new range allocated with sp_alloc will allocate a new policy, set the reference count to 1 and drop the reference count of the old policy. This increases the memory footprint but is not expected to be a major problem as mbind() is unlikely to be used for fine-grained ranges. It is also inefficient because it means we allocate a new policy even in cases where mbind_range() could use the new_policy passed to it. However, it is more straight-forward and the change should be invisible to the user.

[[email protected]: Edited changelog]

Reported-by Dave Jones <[email protected]>

Cc Christoph Lameter <[email protected]>

Signed-off-by Mel Gorman <[email protected]>

Cc Josh Boyer <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 mm/mempolicy.c |   52 ++++++++++++++++++++++++++++++++++++++--------------
 1 file changed, 38 insertions(+), 14 deletions(-)

mm/mempolicy.c +38 ~~-14~~

@@ -606,24 +606,39 @@ check_range(struct mm_struct *mm, unsign
 	return first;
 }
 
-/* Apply policy to a single VMA */
-static int policy_vma(struct vm_area_struct *vma, struct mempolicy *new)
+/*
+ * Apply policy to a single VMA
+ * This must be called with the mmap_sem held for writing.
+ */
+static int vma_replace_policy(struct vm_area_struct *vma,
+						struct mempolicy *pol)
 {
-	int err = 0;
-	struct mempolicy *old = vma->vm_policy;
+	int err;
+	struct mempolicy *old;
+	struct mempolicy *new;
 
 	pr_debug("vma %lx-%lx/%lx vm_ops %p vm_file %p set_policy %p\n",
 		 vma->vm_start, vma->vm_end, vma->vm_pgoff,
 		 vma->vm_ops, vma->vm_file,
 		 vma->vm_ops ? vma->vm_ops->set_policy : NULL);
 
-	if (vma->vm_ops && vma->vm_ops->set_policy)
+	new = mpol_dup(pol);
+	if (IS_ERR(new))
+		return PTR_ERR(new);
+
+	if (vma->vm_ops && vma->vm_ops->set_policy) {
 		err = vma->vm_ops->set_policy(vma, new);
-	if (!err) {
-		mpol_get(new);
-		vma->vm_policy = new;
-		mpol_put(old);
+		if (err)
+			goto err_out;
 	}
+
+	old = vma->vm_policy;
+	vma->vm_policy = new; /* protected by mmap_sem */
+	mpol_put(old);
+
+	return 0;
+ err_out:
+	mpol_put(new);
 	return err;
 }

@@ -666,7 +681,7 @@ static int mbind_range(struct mm_struct
 			if (err)
 				goto out;
 		}
-		err = policy_vma(vma, new_pol);
+		err = vma_replace_policy(vma, new_pol);
 		if (err)
 			goto out;
 	}

@@ -2091,15 +2106,24 @@ static void sp_delete(struct shared_poli
 static struct sp_node *sp_alloc(unsigned long start, unsigned long end,
 				struct mempolicy *pol)
 {
-	struct sp_node *n = kmem_cache_alloc(sn_cache, GFP_KERNEL);
+	struct sp_node *n;
+	struct mempolicy *newpol;
 
+	n = kmem_cache_alloc(sn_cache, GFP_KERNEL);
 	if (!n)
 		return NULL;
+
+	newpol = mpol_dup(pol);
+	if (IS_ERR(newpol)) {
+		kmem_cache_free(sn_cache, n);
+		return NULL;
+	}
+	newpol->flags |= MPOL_F_SHARED;
+
 	n->start = start;
 	n->end = end;
-	mpol_get(pol);
-	pol->flags |= MPOL_F_SHARED;	/* for unref */
-	n->policy = pol;
+	n->policy = newpol;
+
 	return n;
 }

Greg Kroah-Hartman

author

2012-10-11 02:27 UTC #78 lore raw

PATCH 76/84 mempolicy: fix a race in shared_policy_replace()

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

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

From: Mel Gorman <[email protected]>

commit b22d127a39ddd10d93deee3d96e643657ad53a49 upstream.

shared_policy_replace() use of sp_alloc() is unsafe. 1) sp_node cannot be dereferenced if sp->lock is not held and 2) another thread can modify sp_node between spin_unlock for allocating a new sp node and next spin_lock. The bug was introduced before 2.6.12-rc2.

Kosaki's original patch for this problem was to allocate an sp node and policy within shared_policy_replace and initialise it when the lock is reacquired. I was not keen on this approach because it partially duplicates sp_alloc(). As the paths were sp->lock is taken are not that performance critical this patch converts sp->lock to sp->mutex so it can sleep when calling sp_alloc().

[[email protected]: Original patch]

Signed-off-by Mel Gorman <[email protected]>

Acked-by KOSAKI Motohiro <[email protected]>

Cc Josh Boyer <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 include/linux/mempolicy.h |    2 +-
 mm/mempolicy.c            |   37 ++++++++++++++++---------------------
 2 files changed, 17 insertions(+), 22 deletions(-)

include/linux/mempolicy.h +1 -1

@@ -188,7 +188,7 @@ struct sp_node {
 
 struct shared_policy {
 	struct rb_root root;
-	spinlock_t lock;
+	struct mutex mutex;
 };
 
 void mpol_shared_policy_init(struct shared_policy *sp, struct mempolicy *mpol);

mm/mempolicy.c +16 ~~-21~~

@@ -2021,7 +2021,7 @@ int __mpol_equal(struct mempolicy *a, st
  */
 
 /* lookup first element intersecting start-end */
-/* Caller holds sp->lock */
+/* Caller holds sp->mutex */
 static struct sp_node *
 sp_lookup(struct shared_policy *sp, unsigned long start, unsigned long end)
 {

@@ -2085,13 +2085,13 @@ mpol_shared_policy_lookup(struct shared_
 
 	if (!sp->root.rb_node)
 		return NULL;
-	spin_lock(&sp->lock);
+	mutex_lock(&sp->mutex);
 	sn = sp_lookup(sp, idx, idx+1);
 	if (sn) {
 		mpol_get(sn->policy);
 		pol = sn->policy;
 	}
-	spin_unlock(&sp->lock);
+	mutex_unlock(&sp->mutex);
 	return pol;
 }

@@ -2131,10 +2131,10 @@ static struct sp_node *sp_alloc(unsigned
 static int shared_policy_replace(struct shared_policy *sp, unsigned long start,
 				 unsigned long end, struct sp_node *new)
 {
-	struct sp_node *n, *new2 = NULL;
+	struct sp_node *n;
+	int ret = 0;
 
-restart:
-	spin_lock(&sp->lock);
+	mutex_lock(&sp->mutex);
 	n = sp_lookup(sp, start, end);
 	/* Take care of old policies in the same range. */
 	while (n && n->start < end) {

@@ -2147,16 +2147,14 @@ restart:
 		} else {
 			/* Old policy spanning whole new range. */
 			if (n->end > end) {
+				struct sp_node *new2;
+				new2 = sp_alloc(end, n->end, n->policy);
 				if (!new2) {
-					spin_unlock(&sp->lock);
-					new2 = sp_alloc(end, n->end, n->policy);
-					if (!new2)
-						return -ENOMEM;
-					goto restart;
+					ret = -ENOMEM;
+					goto out;
 				}
 				n->end = start;
 				sp_insert(sp, new2);
-				new2 = NULL;
 				break;
 			} else
 				n->end = start;

@@ -2167,12 +2165,9 @@ restart:
 	}
 	if (new)
 		sp_insert(sp, new);
-	spin_unlock(&sp->lock);
-	if (new2) {
-		mpol_put(new2->policy);
-		kmem_cache_free(sn_cache, new2);
-	}
-	return 0;
+out:
+	mutex_unlock(&sp->mutex);
+	return ret;
 }
 
 /**

@@ -2190,7 +2185,7 @@ void mpol_shared_policy_init(struct shar
 	int ret;
 
 	sp->root = RB_ROOT;		/* empty tree == default mempolicy */
-	spin_lock_init(&sp->lock);
+	mutex_init(&sp->mutex);
 
 	if (mpol) {
 		struct vm_area_struct pvma;

@@ -2256,7 +2251,7 @@ void mpol_free_shared_policy(struct shar
 
 	if (!p->root.rb_node)
 		return;
-	spin_lock(&p->lock);
+	mutex_lock(&p->mutex);
 	next = rb_first(&p->root);
 	while (next) {
 		n = rb_entry(next, struct sp_node, nd);

@@ -2265,7 +2260,7 @@ void mpol_free_shared_policy(struct shar
 		mpol_put(n->policy);
 		kmem_cache_free(sn_cache, n);
 	}
-	spin_unlock(&p->lock);
+	mutex_unlock(&p->mutex);
 }
 
 /* assumes fs == KERNEL_DS */

Greg Kroah-Hartman

author

2012-10-11 02:27 UTC #79 lore raw

PATCH 77/84 mempolicy: fix refcount leak in mpol_set_shared_policy()

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

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

From: KOSAKI Motohiro <[email protected]>

commit 63f74ca21f1fad36d075e063f06dcc6d39fe86b2 upstream.

When shared_policy_replace() fails to allocate new->policy is not freed correctly by mpol_set_shared_policy(). The problem is that shared mempolicy code directly call kmem_cache_free() in multiple places where it is easy to make a mistake.

This patch creates an sp_free wrapper function and uses it. The bug was introduced pre-git age (IOW, before 2.6.12-rc2).

[[email protected]: Editted changelog]

Signed-off-by Mel Gorman <[email protected]>

Cc Josh Boyer <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 mm/mempolicy.c |   15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)

mm/mempolicy.c +9 -6

@@ -2095,12 +2095,17 @@ mpol_shared_policy_lookup(struct shared_
 	return pol;
 }
 
+static void sp_free(struct sp_node *n)
+{
+	mpol_put(n->policy);
+	kmem_cache_free(sn_cache, n);
+}
+
 static void sp_delete(struct shared_policy *sp, struct sp_node *n)
 {
 	pr_debug("deleting %lx-l%lx\n", n->start, n->end);
 	rb_erase(&n->nd, &sp->root);
-	mpol_put(n->policy);
-	kmem_cache_free(sn_cache, n);
+	sp_free(n);
 }
 
 static struct sp_node *sp_alloc(unsigned long start, unsigned long end,

@@ -2239,7 +2244,7 @@ int mpol_set_shared_policy(struct shared
 	}
 	err = shared_policy_replace(info, vma->vm_pgoff, vma->vm_pgoff+sz, new);
 	if (err && new)
-		kmem_cache_free(sn_cache, new);
+		sp_free(new);
 	return err;
 }

@@ -2256,9 +2261,7 @@ void mpol_free_shared_policy(struct shar
 	while (next) {
 		n = rb_entry(next, struct sp_node, nd);
 		next = rb_next(&n->nd);
-		rb_erase(&n->nd, &p->root);
-		mpol_put(n->policy);
-		kmem_cache_free(sn_cache, n);
+		sp_delete(p, n);
 	}
 	mutex_unlock(&p->mutex);
 }

Greg Kroah-Hartman

author

2012-10-11 02:16 UTC #80 lore raw

PATCH 78/84 mempolicy: fix a memory corruption by refcount imbalance in alloc_pages_vma()

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

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

From: Mel Gorman <[email protected]>

commit 00442ad04a5eac08a98255697c510e708f6082e2 upstream.

Commit cc9a6c877661 ("cpuset: mm: reduce large amounts of memory barrier related damage v3") introduced a potential memory corruption. shmem_alloc_page() uses a pseudo vma and it has one significant unique combination, vma->vm_ops=NULL and vma->policy->flags & MPOL_F_SHARED.

get_vma_policy() does NOT increase a policy ref when vma->vm_ops=NULL and mpol_cond_put() DOES decrease a policy ref when a policy has MPOL_F_SHARED. Therefore, when a cpuset update race occurs, alloc_pages_vma() falls in 'goto retry_cpuset' path, decrements the reference count and frees the policy prematurely.

Signed-off-by Mel Gorman <[email protected]>

Cc Josh Boyer <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 mm/mempolicy.c |   12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

mm/mempolicy.c +11 -1

@@ -1511,8 +1511,18 @@ struct mempolicy *get_vma_policy(struct
 									addr);
 			if (vpol)
 				pol = vpol;
-		} else if (vma->vm_policy)
+		} else if (vma->vm_policy) {
 			pol = vma->vm_policy;
+
+			/*
+			 * shmem_alloc_page() passes MPOL_F_SHARED policy with
+			 * a pseudo vma whose vma->vm_ops=NULL. Take a reference
+			 * count on these policies which will be dropped by
+			 * mpol_cond_put() later
+			 */
+			if (mpol_needs_cond_ref(pol))
+				mpol_get(pol);
+		}
 	}
 	if (!pol)
 		pol = &default_policy;

Greg Kroah-Hartman

author

2012-10-11 02:17 UTC #81 lore raw

PATCH 79/84 CPU hotplug, cpusets, suspend: Dont modify cpusets during suspend/resume

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

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

From: "Srivatsa S. Bhat" <[email protected]>

commit d35be8bab9b0ce44bed4b9453f86ebf64062721e upstream.

In the event of CPU hotplug, the kernel modifies the cpusets' cpus_allowed masks as and when necessary to ensure that the tasks belonging to the cpusets have some place (online CPUs) to run on. And regular CPU hotplug is destructive in the sense that the kernel doesn't remember the original cpuset configurations set by the user, across hotplug operations.

However, suspend/resume (which uses CPU hotplug) is a special case in which the kernel has the responsibility to restore the system (during resume), to exactly the same state it was in before suspend.

In order to achieve that, do the following:

1. Don't modify cpusets during suspend/resume. At all.
In particular, don't move the tasks from one cpuset to another, and don't modify any cpuset's cpus_allowed mask. So, simply ignore cpusets during the CPU hotplug operations that are carried out in the suspend/resume path.

2. However, cpusets and sched domains are related. We just want to avoid altering cpusets alone. So, to keep the sched domains updated, build a single sched domain (containing all active cpus) during each of the CPU hotplug operations carried out in s/r path, effectively ignoring the cpusets' cpus_allowed masks.

(Since userspace is frozen while doing all this, it will go unnoticed.)

3. During the last CPU online operation during resume, build the sched domains by looking up the (unaltered) cpusets' cpus_allowed masks. That will bring back the system to the same original state as it was in before suspend.

Ultimately, this will not only solve the cpuset problem related to suspend resume (ie., restores the cpusets to exactly what it was before suspend, by not touching it at all) but also speeds up suspend/resume because we avoid running cpuset update code for every CPU being offlined/onlined.

Signed-off-by Srivatsa S. Bhat <[email protected]>

Signed-off-by Peter Zijlstra <[email protected]>

Cc Linus Torvalds <[email protected]>

Cc Andrew Morton <[email protected]>

Cc Thomas Gleixner <[email protected]>

Link http://lkml.kernel.org/r/[email protected] open here

Signed-off-by Ingo Molnar <[email protected]>

Signed-off-by Preeti U Murthy <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 kernel/cpuset.c |    3 +++
 kernel/sched.c  |   40 ++++++++++++++++++++++++++++++++++++----
 2 files changed, 39 insertions(+), 4 deletions(-)

kernel/cpuset.c +3 -0

@@ -2080,6 +2080,9 @@ static void scan_for_empty_cpusets(struc
  * (of no affect) on systems that are actively using CPU hotplug
  * but making no active use of cpusets.
  *
+ * The only exception to this is suspend/resume, where we don't
+ * modify cpusets at all.
+ *
  * This routine ensures that top_cpuset.cpus_allowed tracks
  * cpu_active_mask on each CPU hotplug (cpuhp) event.
  *

kernel/sched.c +36 -4

@@ -7777,34 +7777,66 @@ int __init sched_create_sysfs_power_savi
 }
 #endif /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */
 
+static int num_cpus_frozen;	/* used to mark begin/end of suspend/resume */
+
 /*
  * Update cpusets according to cpu_active mask.  If cpusets are
  * disabled, cpuset_update_active_cpus() becomes a simple wrapper
  * around partition_sched_domains().
+ *
+ * If we come here as part of a suspend/resume, don't touch cpusets because we
+ * want to restore it back to its original state upon resume anyway.
  */
 static int cpuset_cpu_active(struct notifier_block *nfb, unsigned long action,
 			     void *hcpu)
 {
-	switch (action & ~CPU_TASKS_FROZEN) {
+	switch (action) {
+	case CPU_ONLINE_FROZEN:
+	case CPU_DOWN_FAILED_FROZEN:
+
+		/*
+		 * num_cpus_frozen tracks how many CPUs are involved in suspend
+		 * resume sequence. As long as this is not the last online
+		 * operation in the resume sequence, just build a single sched
+		 * domain, ignoring cpusets.
+		 */
+		num_cpus_frozen--;
+		if (likely(num_cpus_frozen)) {
+			partition_sched_domains(1, NULL, NULL);
+			break;
+		}
+
+		/*
+		 * This is the last CPU online operation. So fall through and
+		 * restore the original sched domains by considering the
+		 * cpuset configurations.
+		 */
+
 	case CPU_ONLINE:
 	case CPU_DOWN_FAILED:
 		cpuset_update_active_cpus();
-		return NOTIFY_OK;
+		break;
 	default:
 		return NOTIFY_DONE;
 	}
+	return NOTIFY_OK;
 }
 
 static int cpuset_cpu_inactive(struct notifier_block *nfb, unsigned long action,
 			       void *hcpu)
 {
-	switch (action & ~CPU_TASKS_FROZEN) {
+	switch (action) {
 	case CPU_DOWN_PREPARE:
 		cpuset_update_active_cpus();
-		return NOTIFY_OK;
+		break;
+	case CPU_DOWN_PREPARE_FROZEN:
+		num_cpus_frozen++;
+		partition_sched_domains(1, NULL, NULL);
+		break;
 	default:
 		return NOTIFY_DONE;
 	}
+	return NOTIFY_OK;
 }
 
 static int update_runtime(struct notifier_block *nfb,

Greg Kroah-Hartman

author

2012-10-11 02:16 UTC #82 lore raw

PATCH 80/84 mtd: autcpu12-nvram: Fix compile breakage

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

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

From: Alexander Shiyan <[email protected]>

commit d1f55c680e5d021e7066f4461dd678d42af18898 upstream.

Update driver autcpu12-nvram.c so it compiles; map_read32/map_write32 no longer exist in the kernel so the driver is totally broken. Additionally, map_info name passed to simple_map_init is incorrect.

Signed-off-by Alexander Shiyan <[email protected]>

Acked-by Arnd Bergmann <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/mtd/maps/autcpu12-nvram.c |   19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)

drivers/mtd/maps/autcpu12-nvram.c +11 -8

@@ -43,7 +43,8 @@ struct map_info autcpu12_sram_map = {
 
 static int __init init_autcpu12_sram (void)
 {
-	int err, save0, save1;
+	map_word tmp, save0, save1;
+	int err;
 
 	autcpu12_sram_map.virt = ioremap(0x12000000, SZ_128K);
 	if (!autcpu12_sram_map.virt) {

@@ -51,7 +52,7 @@ static int __init init_autcpu12_sram (vo
 		err = -EIO;
 		goto out;
 	}
-	simple_map_init(&autcpu_sram_map);
+	simple_map_init(&autcpu12_sram_map);
 
 	/*
 	 * Check for 32K/128K

@@ -61,20 +62,22 @@ static int __init init_autcpu12_sram (vo
 	 * Read	and check result on ofs 0x0
 	 * Restore contents
 	 */
-	save0 = map_read32(&autcpu12_sram_map,0);
-	save1 = map_read32(&autcpu12_sram_map,0x10000);
-	map_write32(&autcpu12_sram_map,~save0,0x10000);
+	save0 = map_read(&autcpu12_sram_map, 0);
+	save1 = map_read(&autcpu12_sram_map, 0x10000);
+	tmp.x[0] = ~save0.x[0];
+	map_write(&autcpu12_sram_map, tmp, 0x10000);
 	/* if we find this pattern on 0x0, we have 32K size
 	 * restore contents and exit
 	 */
-	if ( map_read32(&autcpu12_sram_map,0) != save0) {
-		map_write32(&autcpu12_sram_map,save0,0x0);
+	tmp = map_read(&autcpu12_sram_map, 0);
+	if (!map_word_equal(&autcpu12_sram_map, tmp, save0)) {
+		map_write(&autcpu12_sram_map, save0, 0x0);
 		goto map;
 	}
 	/* We have a 128K found, restore 0x10000 and set size
 	 * to 128K
 	 */
-	map_write32(&autcpu12_sram_map,save1,0x10000);
+	map_write(&autcpu12_sram_map, save1, 0x10000);
 	autcpu12_sram_map.size = SZ_128K;
 
 map:

Greg Kroah-Hartman

author

2012-10-11 02:16 UTC #83 lore raw

PATCH 81/84 mtd: nandsim: bugfix: fail if overridesize is too big

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

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

From: Richard Genoud <[email protected]>

commit bb0a13a13411c4ce24c48c8ff3cdf7b48d237240 upstream.

If override size is too big, the module was actually loaded instead of failing, because retval was not set.

This lead to memory corruption with the use of the freed structs nandsim and nand_chip.

Signed-off-by Richard Genoud <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/mtd/nand/nandsim.c |    1 +
 1 file changed, 1 insertion(+)

drivers/mtd/nand/nandsim.c +1 -0

@@ -2355,6 +2355,7 @@ static int __init ns_init_module(void)
 		uint64_t new_size = (uint64_t)nsmtd->erasesize << overridesize;
 		if (new_size >> overridesize != nsmtd->erasesize) {
 			NS_ERR("overridesize is too big\n");
+			retval = -EINVAL;
 			goto err_exit;
 		}
 		/* N.B. This relies on nand_scan not doing anything with the size before we change it */

Greg Kroah-Hartman

author

2012-10-11 02:17 UTC #84 lore raw

PATCH 82/84 mtd: nand: Use the mirror BBT descriptor when reading its version

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

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

From: Shmulik Ladkani <[email protected]>

commit 7bb9c75436212813b38700c34df4bbb6eb82debe upstream.

The code responsible for reading the version of the mirror bbt was incorrectly using the descriptor of the main bbt.

Pass the mirror bbt descriptor to 'scan_read_raw' when reading the version of the mirror bbt.

Signed-off-by Shmulik Ladkani <[email protected]>

Acked-by Sebastian Andrzej Siewior <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/mtd/nand/nand_bbt.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/mtd/nand/nand_bbt.c +1 -1

@@ -429,7 +429,7 @@ static int read_abs_bbts(struct mtd_info
 	/* Read the mirror version, if available */
 	if (md && (md->options & NAND_BBT_VERSION)) {
 		scan_read_raw(mtd, buf, (loff_t)md->pages[0] << this->page_shift,
-			      mtd->writesize, td);
+			      mtd->writesize, md);
 		md->version[0] = buf[bbt_get_ver_offs(mtd, md)];
 		printk(KERN_DEBUG "Bad block table at page %d, version 0x%02X\n",
 		       md->pages[0], md->version[0]);

Greg Kroah-Hartman

author

2012-10-11 02:29 UTC #85 lore raw

PATCH 83/84 mtd: omap2: fix omap_nand_remove segfault

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

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

From: Andreas Bießmann <[email protected]>

commit 7d9b110269253b1d5858cfa57d68dfc7bf50dd77 upstream.

Do not kfree() the mtd_info; it is handled in the mtd subsystem and already freed by nand_release(). Instead kfree() the struct omap_nand_info allocated in omap_nand_probe which was not freed before.

This patch fixes following error when unloading the omap2 module:

---8<---
~ $ rmmod omap2

------------[ cut here ]------------
kernel BUG at mm/slab.c:3126!
Internal error: Oops - BUG: 0 [#1] PREEMPT ARM
Modules linked in: omap2(-)
CPU: 0    Not tainted  (3.6.0-rc3-00230-g155e36d-dirty #3)
PC is at cache_free_debugcheck+0x2d4/0x36c
LR is at kfree+0xc8/0x2ac
pc : [<c01125a0>]    lr : [<c0112efc>]    psr: 200d0193

sp : c521fe08 ip : c0e8ef90 fp : c521fe5c
r10: bf0001fc r9 : c521e000 r8 : c0d99c8c
r7 : c661ebc0 r6 : c065d5a4 r5 : c65c4060 r4 : c78005c0 r3 : 00000000 r2 : 00001000 r1 : c65c4000 r0 : 00000001
Flags: nzCv IRQs off FIQs on Mode SVC_32 ISA ARM Segment user
Control: 10c5387d Table: 86694019 DAC: 00000015 Process rmmod (pid: 549, stack limit = 0xc521e2f0)

Stack: (0xc521fe08 to 0xc5220000)

fe00: c008a874 c00bf44c c515c6d0 200d0193 c65c4860 c515c240
fe20: c521fe3c c521fe30 c008a9c0 c008a854 c521fe5c c65c4860 c78005c0 bf0001fc
fe40: c780ff40 a00d0113 c521e000 00000000 c521fe84 c521fe60 c0112efc c01122d8
fe60: c65c4860 c0673778 c06737ac 00000000 00070013 00000000 c521fe9c c521fe88
fe80: bf0001fc c0112e40 c0673778 bf001ca8 c521feac c521fea0 c02ca11c bf0001ac
fea0: c521fec4 c521feb0 c02c82c4 c02ca100 c0673778 bf001ca8 c521fee4 c521fec8
fec0: c02c8dd8 c02c8250 00000000 bf001ca8 bf001ca8 c0804ee0 c521ff04 c521fee8
fee0: c02c804c c02c8d20 bf001924 00000000 bf001ca8 c521e000 c521ff1c c521ff08
ff00: c02c950c c02c7fbc bf001d48 00000000 c521ff2c c521ff20 c02ca3a4 c02c94b8
ff20: c521ff3c c521ff30 bf001938 c02ca394 c521ffa4 c521ff40 c009beb4 bf001930
ff40: c521ff6c 70616d6f b6fe0032 c0014f84 70616d6f b6fe0032 00000081 60070010
ff60: c521ff84 c521ff70 c008e1f4 c00bf328 0001a004 70616d6f c521ff94 0021ff88
ff80: c008e368 0001a004 70616d6f b6fe0032 00000081 c0015028 00000000 c521ffa8
ffa0: c0014dc0 c009bcd0 0001a004 70616d6f bec2ab38 00000880 bec2ab38 00000880
ffc0: 0001a004 70616d6f b6fe0032 00000081 00000319 00000000 b6fe1000 00000000
ffe0: bec2ab30 bec2ab20 00019f00 b6f539c0 60070010 bec2ab38 aaaaaaaa aaaaaaaa

Backtrace:
[<c01122cc>] (cache_free_debugcheck+0x0/0x36c) from [<c0112efc>] (kfree+0xc8/0x2ac)
[<c0112e34>] (kfree+0x0/0x2ac) from [<bf0001fc>] (omap_nand_remove+0x5c/0x64 [omap2])
[<bf0001a0>] (omap_nand_remove+0x0/0x64 [omap2]) from [<c02ca11c>] (platform_drv_remove+0x28/0x2c)
 r5:bf001ca8 r4:c0673778
[<c02ca0f4>] (platform_drv_remove+0x0/0x2c) from [<c02c82c4>] (__device_release_driver+0x80/0xdc)
[<c02c8244>] (__device_release_driver+0x0/0xdc) from [<c02c8dd8>] (driver_detach+0xc4/0xc8)
 r5:bf001ca8 r4:c0673778
[<c02c8d14>] (driver_detach+0x0/0xc8) from [<c02c804c>] (bus_remove_driver+0x9c/0x104)
 r6:c0804ee0 r5:bf001ca8 r4:bf001ca8 r3:00000000
[<c02c7fb0>] (bus_remove_driver+0x0/0x104) from [<c02c950c>] (driver_unregister+0x60/0x80)
 r6:c521e000 r5:bf001ca8 r4:00000000 r3:bf001924
[<c02c94ac>] (driver_unregister+0x0/0x80) from [<c02ca3a4>] (platform_driver_unregister+0x1c/0x20)
 r5:00000000 r4:bf001d48
[<c02ca388>] (platform_driver_unregister+0x0/0x20) from [<bf001938>] (omap_nand_driver_exit+0x14/0x1c [omap2])
[<bf001924>] (omap_nand_driver_exit+0x0/0x1c [omap2]) from [<c009beb4>] (sys_delete_module+0x1f0/0x2ec)
[<c009bcc4>] (sys_delete_module+0x0/0x2ec) from [<c0014dc0>] (ret_fast_syscall+0x0/0x48)
 r8:c0015028 r7:00000081 r6:b6fe0032 r5:70616d6f r4:0001a004
Code: e1a00005 eb0d9172 e7f001f2 e7f001f2 (e7f001f2)
---[ end trace 6a30b24d8c0cc2ee ]---

Segmentation fault
--->8---

This error was introduced in 67ce04bf2746f8a1f8c2a104b313d20c63f68378 which was the first commit of this driver.

Signed-off-by Andreas Bießmann <[email protected]>

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/mtd/nand/omap2.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

drivers/mtd/nand/omap2.c +1 -1

@@ -1139,7 +1139,7 @@ static int omap_nand_remove(struct platf
 	/* Release NAND device, its internal structures and partitions */
 	nand_release(&info->mtd);
 	iounmap(info->nand.IO_ADDR_R);
-	kfree(&info->mtd);
+	kfree(info);
 	return 0;
 }

Greg Kroah-Hartman

author

2012-10-11 02:26 UTC #86 lore raw

PATCH 84/84 mtd: omap2: fix module loading

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

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

From: Andreas Bießmann <[email protected]>

commit 4d3d688da8e7016f15483e9319b41311e1db9515 upstream.

Unloading the omap2 nand driver missed to release the memory region which will result in not being able to request it again if one want to load the driver later on.

This patch fixes following error when loading omap2 module after unloading:
---8<---
~ $ rmmod omap2
~ $ modprobe omap2

[   37.420928] omap2-nand: probe of omap2-nand.0 failed with error -16

~ $
--->8---

This error was introduced in 67ce04bf2746f8a1f8c2a104b313d20c63f68378 which was the first commit of this driver.

Signed-off-by Andreas Bießmann <[email protected]>