Skip to contentSkip to conversation

iommu/vt-d: Add RPLS to quirk list to skip TE disabling

PATCH In review Sasha Levin 8 participants 172 messages May 30, 2022 – Jun 5, 2022 Page 2 of 2

Subsystems (39)
lore
Share & tools
lore
author maintainer
#160 lore raw

PATCH 149/159 m68k: atari: Make Atari ROM port I/O write macros return void

In reply to Sasha Levin

From: Geert Uytterhoeven <[email protected]>

[ Upstream commit 30b5e6ef4a32ea4985b99200e06d6660a69f9246 ]

The macros implementing Atari ROM port I/O writes do not cast away their output, unlike similar implementations for other I/O buses. When they are combined using conditional expressions in the definitions of outb() and friends, this triggers sparse warnings like:

drivers/net/appletalk/cops.c:382:17: error: incompatible types in conditional expression (different base types):
drivers/net/appletalk/cops.c:382:17: unsigned char
drivers/net/appletalk/cops.c:382:17: void

Fix this by adding casts to "void".

---
 arch/m68k/include/asm/raw_io.h | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
arch/m68k/include/asm/raw_io.h +3 -3
index 80eb2396d01e..3ba40bc1dfaa 100644
@@ -80,14 +80,14 @@
 	({ u16 __v = le16_to_cpu(*(__force volatile u16 *) (addr)); __v; })
 
 #define rom_out_8(addr, b)	\
-	({u8 __maybe_unused __w, __v = (b);  u32 _addr = ((u32) (addr)); \
+	(void)({u8 __maybe_unused __w, __v = (b);  u32 _addr = ((u32) (addr)); \
 	__w = ((*(__force volatile u8 *)  ((_addr | 0x10000) + (__v<<1)))); })
 #define rom_out_be16(addr, w)	\
-	({u16 __maybe_unused __w, __v = (w); u32 _addr = ((u32) (addr)); \
+	(void)({u16 __maybe_unused __w, __v = (w); u32 _addr = ((u32) (addr)); \
 	__w = ((*(__force volatile u16 *) ((_addr & 0xFFFF0000UL) + ((__v & 0xFF)<<1)))); \
 	__w = ((*(__force volatile u16 *) ((_addr | 0x10000) + ((__v >> 8)<<1)))); })
 #define rom_out_le16(addr, w)	\
-	({u16 __maybe_unused __w, __v = (w); u32 _addr = ((u32) (addr)); \
+	(void)({u16 __maybe_unused __w, __v = (w); u32 _addr = ((u32) (addr)); \
 	__w = ((*(__force volatile u16 *) ((_addr & 0xFFFF0000UL) + ((__v >> 8)<<1)))); \
 	__w = ((*(__force volatile u16 *) ((_addr | 0x10000) + ((__v & 0xFF)<<1)))); })
 
Signature · 1 line
2.35.1
author maintainer
#161 lore raw

PATCH 150/159 hwmon: (pmbus) Add get_voltage/set_voltage ops

In reply to Sasha Levin

From: Mårten Lindahl <[email protected]>

[ Upstream commit 28bf22ef93eceb42c7583f0909bc9dedc07770e3 ]

The pmbus core does not have operations for getting or setting voltage. Add functions get/set voltage for the dynamic regulator framework.

---
 drivers/hwmon/pmbus/pmbus_core.c | 67 ++++++++++++++++++++++++++++++++
 1 file changed, 67 insertions(+)
drivers/hwmon/pmbus/pmbus_core.c +67 -0
index d93574d6a1fb..5a1796650f5b 100644
@@ -2548,11 +2548,78 @@ static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsigned
 	return 0;
 }
 
+static int pmbus_regulator_get_voltage(struct regulator_dev *rdev)
+{
+	struct device *dev = rdev_get_dev(rdev);
+	struct i2c_client *client = to_i2c_client(dev->parent);
+	struct pmbus_data *data = i2c_get_clientdata(client);
+	struct pmbus_sensor s = {
+		.page = rdev_get_id(rdev),
+		.class = PSC_VOLTAGE_OUT,
+		.convert = true,
+	};
+
+	s.data = _pmbus_read_word_data(client, s.page, 0xff, PMBUS_READ_VOUT);
+	if (s.data < 0)
+		return s.data;
+
+	return (int)pmbus_reg2data(data, &s) * 1000; /* unit is uV */
+}
+
+static int pmbus_regulator_set_voltage(struct regulator_dev *rdev, int min_uv,
+				       int max_uv, unsigned int *selector)
+{
+	struct device *dev = rdev_get_dev(rdev);
+	struct i2c_client *client = to_i2c_client(dev->parent);
+	struct pmbus_data *data = i2c_get_clientdata(client);
+	struct pmbus_sensor s = {
+		.page = rdev_get_id(rdev),
+		.class = PSC_VOLTAGE_OUT,
+		.convert = true,
+		.data = -1,
+	};
+	int val = DIV_ROUND_CLOSEST(min_uv, 1000); /* convert to mV */
+	int low, high;
+
+	*selector = 0;
+
+	if (pmbus_check_word_register(client, s.page, PMBUS_MFR_VOUT_MIN))
+		s.data = _pmbus_read_word_data(client, s.page, 0xff, PMBUS_MFR_VOUT_MIN);
+	if (s.data < 0) {
+		s.data = _pmbus_read_word_data(client, s.page, 0xff, PMBUS_VOUT_MARGIN_LOW);
+		if (s.data < 0)
+			return s.data;
+	}
+	low = pmbus_reg2data(data, &s);
+
+	s.data = -1;
+	if (pmbus_check_word_register(client, s.page, PMBUS_MFR_VOUT_MAX))
+		s.data = _pmbus_read_word_data(client, s.page, 0xff, PMBUS_MFR_VOUT_MAX);
+	if (s.data < 0) {
+		s.data = _pmbus_read_word_data(client, s.page, 0xff, PMBUS_VOUT_MARGIN_HIGH);
+		if (s.data < 0)
+			return s.data;
+	}
+	high = pmbus_reg2data(data, &s);
+
+	/* Make sure we are within margins */
+	if (low > val)
+		val = low;
+	if (high < val)
+		val = high;
+
+	val = pmbus_data2reg(data, &s, val);
+
+	return _pmbus_write_word_data(client, s.page, PMBUS_VOUT_COMMAND, (u16)val);
+}
+
 const struct regulator_ops pmbus_regulator_ops = {
 	.enable = pmbus_regulator_enable,
 	.disable = pmbus_regulator_disable,
 	.is_enabled = pmbus_regulator_is_enabled,
 	.get_error_flags = pmbus_regulator_get_error_flags,
+	.get_voltage = pmbus_regulator_get_voltage,
+	.set_voltage = pmbus_regulator_set_voltage,
 };
 EXPORT_SYMBOL_NS_GPL(pmbus_regulator_ops, PMBUS);
 
Signature · 1 line
2.35.1
author maintainer
#162 lore raw

PATCH 151/159 hwmon: Make chip parameter for with_info API mandatory

In reply to Sasha Levin

From: Guenter Roeck <[email protected]>

[ Upstream commit ddaefa209c4ac791c1262e97c9b2d0440c8ef1d5 ]

Various attempts were made recently to "convert" the old hwmon_device_register() API to devm_hwmon_device_register_with_info() by just changing the function name without actually converting the driver. Prevent this from happening by making the 'chip' parameter of devm_hwmon_device_register_with_info() mandatory.

---
 Documentation/hwmon/hwmon-kernel-api.rst |  2 +-
 drivers/hwmon/hwmon.c                    | 16 +++++++---------
 2 files changed, 8 insertions(+), 10 deletions(-)
Documentation/hwmon/hwmon-kernel-api.rst +1 -1
index c41eb6108103..23f27fe78e37 100644
@@ -72,7 +72,7 @@ hwmon_device_register_with_info is the most comprehensive and preferred means
 to register a hardware monitoring device. It creates the standard sysfs
 attributes in the hardware monitoring core, letting the driver focus on reading
 from and writing to the chip instead of having to bother with sysfs attributes.
-The parent device parameter cannot be NULL with non-NULL chip info. Its
+The parent device parameter as well as the chip parameter must not be NULL. Its
 parameters are described in more detail below.
 
 devm_hwmon_device_register_with_info is similar to
drivers/hwmon/hwmon.c +7 -9
index 989e2c8496dd..187212988b6c 100644
@@ -886,11 +886,12 @@ EXPORT_SYMBOL_GPL(hwmon_device_register_with_groups);
 
 /**
  * hwmon_device_register_with_info - register w/ hwmon
- * @dev: the parent device
- * @name: hwmon name attribute
- * @drvdata: driver data to attach to created device
- * @chip: pointer to hwmon chip information
+ * @dev: the parent device (mandatory)
+ * @name: hwmon name attribute (mandatory)
+ * @drvdata: driver data to attach to created device (optional)
+ * @chip: pointer to hwmon chip information (mandatory)
  * @extra_groups: pointer to list of additional non-standard attribute groups
+ *	(optional)
  *
  * hwmon_device_unregister() must be called when the device is no
  * longer needed.
@@ -903,13 +904,10 @@ hwmon_device_register_with_info(struct device *dev, const char *name,
 				const struct hwmon_chip_info *chip,
 				const struct attribute_group **extra_groups)
 {
-	if (!name)
-		return ERR_PTR(-EINVAL);
-
-	if (chip && (!chip->ops || !chip->ops->is_visible || !chip->info))
+	if (!dev || !name || !chip)
 		return ERR_PTR(-EINVAL);
 
-	if (chip && !dev)
+	if (!chip->ops || !chip->ops->is_visible || !chip->info)
 		return ERR_PTR(-EINVAL);
 
 	return __hwmon_device_register(dev, name, drvdata, chip, extra_groups);
Signature · 1 line
2.35.1
maintainer
#163 lore raw

Re: [PATCH AUTOSEL 5.18 151/159] hwmon: Make chip parameter for with_info API mandatory

Quoted 4 lines · Sasha Levin

From: Guenter Roeck <[email protected]>

[ Upstream commit ddaefa209c4ac791c1262e97c9b2d0440c8ef1d5 ]

Quoted 8 lines

Various attempts were made recently to "convert" the old hwmon_device_register() API to devm_hwmon_device_register_with_info() by just changing the function name without actually converting the driver. Prevent this from happening by making the 'chip' parameter of devm_hwmon_device_register_with_info() mandatory.

Please drop.

Quoted 55 lines
---
  Documentation/hwmon/hwmon-kernel-api.rst |  2 +-
  drivers/hwmon/hwmon.c                    | 16 +++++++---------
  2 files changed, 8 insertions(+), 10 deletions(-)
Documentation/hwmon/hwmon-kernel-api.rst +1 -1
index c41eb6108103..23f27fe78e37 100644
@@ -72,7 +72,7 @@ hwmon_device_register_with_info is the most comprehensive and preferred means
  to register a hardware monitoring device. It creates the standard sysfs
  attributes in the hardware monitoring core, letting the driver focus on reading
  from and writing to the chip instead of having to bother with sysfs attributes.
-The parent device parameter cannot be NULL with non-NULL chip info. Its
+The parent device parameter as well as the chip parameter must not be NULL. Its
  parameters are described in more detail below.
 
  devm_hwmon_device_register_with_info is similar to
drivers/hwmon/hwmon.c +7 -9
index 989e2c8496dd..187212988b6c 100644
@@ -886,11 +886,12 @@ EXPORT_SYMBOL_GPL(hwmon_device_register_with_groups);
 
  /**
   * hwmon_device_register_with_info - register w/ hwmon
- * @dev: the parent device
- * @name: hwmon name attribute
- * @drvdata: driver data to attach to created device
- * @chip: pointer to hwmon chip information
+ * @dev: the parent device (mandatory)
+ * @name: hwmon name attribute (mandatory)
+ * @drvdata: driver data to attach to created device (optional)
+ * @chip: pointer to hwmon chip information (mandatory)
   * @extra_groups: pointer to list of additional non-standard attribute groups
+ *	(optional)
   *
   * hwmon_device_unregister() must be called when the device is no
   * longer needed.
@@ -903,13 +904,10 @@ hwmon_device_register_with_info(struct device *dev, const char *name,
  				const struct hwmon_chip_info *chip,
  				const struct attribute_group **extra_groups)
  {
-	if (!name)
-		return ERR_PTR(-EINVAL);
-
-	if (chip && (!chip->ops || !chip->ops->is_visible || !chip->info))
+	if (!dev || !name || !chip)
  		return ERR_PTR(-EINVAL);
 
-	if (chip && !dev)
+	if (!chip->ops || !chip->ops->is_visible || !chip->info)
  		return ERR_PTR(-EINVAL);
 
  	return __hwmon_device_register(dev, name, drvdata, chip, extra_groups);
author maintainer
#164 lore raw

PATCH 152/159 rxrpc: Return an error to sendmsg if call failed

In reply to Sasha Levin

From: David Howells <[email protected]>

[ Upstream commit 4ba68c5192554876bd8c3afd904e3064d2915341 ]

If at the end of rxrpc sendmsg() or rxrpc_kernel_send_data() the call that was being given data was aborted remotely or otherwise failed, return an error rather than returning the amount of data buffered for transmission.

The call (presumably) did not complete, so there's not much point continuing with it. AF_RXRPC considers it "complete" and so will be unwilling to do anything else with it - and won't send a notification for it, deeming the return from sendmsg sufficient.

Not returning an error causes afs to incorrectly handle a StoreData operation that gets interrupted by a change of address due to NAT reconfiguration.

This doesn't normally affect most operations since their request parameters tend to fit into a single UDP packet and afs_make_call() returns before the server responds; StoreData is different as it involves transmission of a lot of data.

This can be triggered on a client by doing something like:

	dd if=/dev/zero of=/afs/example.com/foo bs=1M count=512

at one prompt, and then changing the network address at another prompt, e.g.:

	ifconfig enp6s0 inet 192.168.6.2 && route add 192.168.6.1 dev enp6s0

Tracing packets on an Auristor fileserver looks something like:

192.168.6.1 -> 192.168.6.3  RX 107 ACK Idle  Seq: 0  Call: 4  Source Port: 7000  Destination Port: 7001
192.168.6.3 -> 192.168.6.1  AFS (RX) 1482 FS Request: Unknown(64538) (64538)
192.168.6.3 -> 192.168.6.1  AFS (RX) 1482 FS Request: Unknown(64538) (64538)
192.168.6.1 -> 192.168.6.3  RX 107 ACK Idle  Seq: 0  Call: 4  Source Port: 7000  Destination Port: 7001
<ARP exchange for 192.168.6.2>
192.168.6.2 -> 192.168.6.1  AFS (RX) 1482 FS Request: Unknown(0) (0)
192.168.6.2 -> 192.168.6.1  AFS (RX) 1482 FS Request: Unknown(0) (0)
192.168.6.1 -> 192.168.6.2  RX 107 ACK Exceeds Window  Seq: 0  Call: 4  Source Port: 7000  Destination Port: 7001
192.168.6.1 -> 192.168.6.2  RX 74 ABORT  Seq: 0  Call: 4  Source Port: 7000  Destination Port: 7001
192.168.6.1 -> 192.168.6.2  RX 74 ABORT  Seq: 29321  Call: 4  Source Port: 7000  Destination Port: 7001

The Auristor fileserver logs code -453 (RXGEN_SS_UNMARSHAL), but the abort code received by kafs is -5 (RX_PROTOCOL_ERROR) as the rx layer sees the condition and generates an abort first and the unmarshal error is a consequence of that at the application layer.

---
 net/rxrpc/sendmsg.c | 6 ++++++
 1 file changed, 6 insertions(+)
net/rxrpc/sendmsg.c +6 -0
index af8ad6c30b9f..1d38e279e2ef 100644
@@ -444,6 +444,12 @@ static int rxrpc_send_data(struct rxrpc_sock *rx,
 
 success:
 	ret = copied;
+	if (READ_ONCE(call->state) == RXRPC_CALL_COMPLETE) {
+		read_lock_bh(&call->state_lock);
+		if (call->error < 0)
+			ret = call->error;
+		read_unlock_bh(&call->state_lock);
+	}
 out:
 	call->tx_pending = skb;
 	_leave(" = %d", ret);
Signature · 1 line
2.35.1
author maintainer
#165 lore raw

PATCH 153/159 rxrpc, afs: Fix selection of abort codes

In reply to Sasha Levin

From: David Howells <[email protected]>

[ Upstream commit de696c4784f0706884458893c5a6c39b3a3ff65c ]

The RX_USER_ABORT code should really only be used to indicate that the user of the rxrpc service (ie. userspace) implicitly caused a call to be aborted
- for instance if the AF_RXRPC socket is closed whilst the call was in progress. (The user may also explicitly abort a call and specify the abort code to use).

Change some of the points of generation to use other abort codes instead:

(1) Abort the call with RXGEN_SS_UNMARSHAL or RXGEN_CC_UNMARSHAL if we see ENOMEM and EFAULT during received data delivery and abort with RX_CALL_DEAD in the default case.

(2) Abort with RXGEN_SS_MARSHAL if we get ENOMEM whilst trying to send a reply.

(3) Abort with RX_CALL_DEAD if we stop hearing from the peer if we had heard from the peer and abort with RX_CALL_TIMEOUT if we hadn't.

(4) Abort with RX_CALL_DEAD if we try to disconnect a call that's not completed successfully or been aborted.

---
 fs/afs/rxrpc.c          | 8 +++++---
 net/rxrpc/call_event.c  | 4 ++--
 net/rxrpc/conn_object.c | 2 +-
 3 files changed, 8 insertions(+), 6 deletions(-)
fs/afs/rxrpc.c +5 -3
index 23a1a92d64bb..a5434f3e57c6 100644
@@ -537,6 +537,8 @@ static void afs_deliver_to_call(struct afs_call *call)
 		case -ENODATA:
 		case -EBADMSG:
 		case -EMSGSIZE:
+		case -ENOMEM:
+		case -EFAULT:
 			abort_code = RXGEN_CC_UNMARSHAL;
 			if (state != AFS_CALL_CL_AWAIT_REPLY)
 				abort_code = RXGEN_SS_UNMARSHAL;
@@ -544,7 +546,7 @@ static void afs_deliver_to_call(struct afs_call *call)
 						abort_code, ret, "KUM");
 			goto local_abort;
 		default:
-			abort_code = RX_USER_ABORT;
+			abort_code = RX_CALL_DEAD;
 			rxrpc_kernel_abort_call(call->net->socket, call->rxcall,
 						abort_code, ret, "KER");
 			goto local_abort;
@@ -836,7 +838,7 @@ void afs_send_empty_reply(struct afs_call *call)
 	case -ENOMEM:
 		_debug("oom");
 		rxrpc_kernel_abort_call(net->socket, call->rxcall,
-					RX_USER_ABORT, -ENOMEM, "KOO");
+					RXGEN_SS_MARSHAL, -ENOMEM, "KOO");
 		fallthrough;
 	default:
 		_leave(" [error]");
@@ -878,7 +880,7 @@ void afs_send_simple_reply(struct afs_call *call, const void *buf, size_t len)
 	if (n == -ENOMEM) {
 		_debug("oom");
 		rxrpc_kernel_abort_call(net->socket, call->rxcall,
-					RX_USER_ABORT, -ENOMEM, "KOO");
+					RXGEN_SS_MARSHAL, -ENOMEM, "KOO");
 	}
 	_leave(" [error]");
 }
net/rxrpc/call_event.c +2 -2
index 22e05de5d1ca..e426f6831aab 100644
@@ -377,9 +377,9 @@ void rxrpc_process_call(struct work_struct *work)
 		if (test_bit(RXRPC_CALL_RX_HEARD, &call->flags) &&
 		    (int)call->conn->hi_serial - (int)call->rx_serial > 0) {
 			trace_rxrpc_call_reset(call);
-			rxrpc_abort_call("EXP", call, 0, RX_USER_ABORT, -ECONNRESET);
+			rxrpc_abort_call("EXP", call, 0, RX_CALL_DEAD, -ECONNRESET);
 		} else {
-			rxrpc_abort_call("EXP", call, 0, RX_USER_ABORT, -ETIME);
+			rxrpc_abort_call("EXP", call, 0, RX_CALL_TIMEOUT, -ETIME);
 		}
 		set_bit(RXRPC_CALL_EV_ABORT, &call->events);
 		goto recheck_state;
net/rxrpc/conn_object.c +1 -1
index b2159dbf5412..660cd9b1a465 100644
@@ -183,7 +183,7 @@ void __rxrpc_disconnect_call(struct rxrpc_connection *conn,
 			chan->last_type = RXRPC_PACKET_TYPE_ABORT;
 			break;
 		default:
-			chan->last_abort = RX_USER_ABORT;
+			chan->last_abort = RX_CALL_DEAD;
 			chan->last_type = RXRPC_PACKET_TYPE_ABORT;
 			break;
 		}
Signature · 1 line
2.35.1
author maintainer
#166 lore raw

PATCH 154/159 afs: Adjust ACK interpretation to try and cope with NAT

In reply to Sasha Levin

From: David Howells <[email protected]>

[ Upstream commit adc9613ff66c26ebaff9814973181ac178beb90b ]

If a client's address changes, say if it is NAT'd, this can disrupt an in progress operation. For most operations, this is not much of a problem, but StoreData can be different as some servers modify the target file as the data comes in, so if a store request is disrupted, the file can get corrupted on the server.

The problem is that the server doesn't recognise packets that come after the change of address as belonging to the original client and will bounce them, either by sending an OUT_OF_SEQUENCE ACK to the apparent new call if the packet number falls within the initial sequence number window of a call or by sending an EXCEEDS_WINDOW ACK if it falls outside and then aborting it. In both cases, firstPacket will be 1 and previousPacket will be 0 in the ACK information.

Fix this by the following means:

(1) If a client call receives an EXCEEDS_WINDOW ACK with firstPacket as 1 and previousPacket as 0, assume this indicates that the server saw the incoming packets from a different peer and thus as a different call. Fail the call with error -ENETRESET.

(2) Also fail the call if a similar OUT_OF_SEQUENCE ACK occurs if the first packet has been hard-ACK'd. If it hasn't been hard-ACK'd, the ACK packet will cause it to get retransmitted, so the call will just be repeated.

(3) Make afs_select_fileserver() treat -ENETRESET as a straight fail of the operation.

(4) Prioritise the error code over things like -ECONNRESET as the server did actually respond.

(5) Make writeback treat -ENETRESET as a retryable error and make it redirty all the pages involved in a write so that the VM will retry.

Note that there is still a circumstance that I can't easily deal with: if the operation is fully received and processed by the server, but the reply is lost due to address change. There's no way to know if the op happened. We can examine the server, but a conflicting change could have been made by a third party - and we can't tell the difference. In such a case, a message like:

    kAFS: vnode modified {100058:146266} b7->b8 YFS.StoreData64 (op=2646a)

will be logged to dmesg on the next op to touch the file and the client will reset the inode state, including invalidating clean parts of the pagecache.

---
 fs/afs/misc.c     |  5 ++++-
 fs/afs/rotate.c   |  4 ++++
 fs/afs/write.c    |  1 +
 net/rxrpc/input.c | 27 +++++++++++++++++++++++++++
 4 files changed, 36 insertions(+), 1 deletion(-)
fs/afs/misc.c +4 -1
index 1d1a8debe472..933e67fcdab1 100644
@@ -163,8 +163,11 @@ void afs_prioritise_error(struct afs_error *e, int error, u32 abort_code)
 		return;
 
 	case -ECONNABORTED:
+		error = afs_abort_to_error(abort_code);
+		fallthrough;
+	case -ENETRESET: /* Responded, but we seem to have changed address */
 		e->responded = true;
-		e->error = afs_abort_to_error(abort_code);
+		e->error = error;
 		return;
 	}
 }
fs/afs/rotate.c +4 -0
index 79e1a5f6701b..a840c3588ebb 100644
@@ -292,6 +292,10 @@ bool afs_select_fileserver(struct afs_operation *op)
 		op->error = error;
 		goto iterate_address;
 
+	case -ENETRESET:
+		pr_warn("kAFS: Peer reset %s (op=%x)\n",
+			op->type ? op->type->name : "???", op->debug_id);
+		fallthrough;
 	case -ECONNRESET:
 		_debug("call reset");
 		op->error = error;
fs/afs/write.c +1 -0
index 4763132ca57e..c1bc52ac7de1 100644
@@ -636,6 +636,7 @@ static ssize_t afs_write_back_from_locked_folio(struct address_space *mapping,
 	case -EKEYEXPIRED:
 	case -EKEYREJECTED:
 	case -EKEYREVOKED:
+	case -ENETRESET:
 		afs_redirty_pages(wbc, mapping, start, len);
 		mapping_set_error(mapping, ret);
 		break;
net/rxrpc/input.c +27 -0
index dc201363f2c4..67d3eba60dc7 100644
@@ -903,6 +903,33 @@ static void rxrpc_input_ack(struct rxrpc_call *call, struct sk_buff *skb)
 				  rxrpc_propose_ack_respond_to_ack);
 	}
 
+	/* If we get an EXCEEDS_WINDOW ACK from the server, it probably
+	 * indicates that the client address changed due to NAT.  The server
+	 * lost the call because it switched to a different peer.
+	 */
+	if (unlikely(buf.ack.reason == RXRPC_ACK_EXCEEDS_WINDOW) &&
+	    first_soft_ack == 1 &&
+	    prev_pkt == 0 &&
+	    rxrpc_is_client_call(call)) {
+		rxrpc_set_call_completion(call, RXRPC_CALL_REMOTELY_ABORTED,
+					  0, -ENETRESET);
+		return;
+	}
+
+	/* If we get an OUT_OF_SEQUENCE ACK from the server, that can also
+	 * indicate a change of address.  However, we can retransmit the call
+	 * if we still have it buffered to the beginning.
+	 */
+	if (unlikely(buf.ack.reason == RXRPC_ACK_OUT_OF_SEQUENCE) &&
+	    first_soft_ack == 1 &&
+	    prev_pkt == 0 &&
+	    call->tx_hard_ack == 0 &&
+	    rxrpc_is_client_call(call)) {
+		rxrpc_set_call_completion(call, RXRPC_CALL_REMOTELY_ABORTED,
+					  0, -ENETRESET);
+		return;
+	}
+
 	/* Discard any out-of-order or duplicate ACKs (outside lock). */
 	if (!rxrpc_is_ack_valid(call, first_soft_ack, prev_pkt)) {
 		trace_rxrpc_rx_discard_ack(call->debug_id, ack_serial,
Signature · 1 line
2.35.1
author maintainer
#167 lore raw

PATCH 155/159 eth: tg3: silence the GCC 12 array-bounds warning

In reply to Sasha Levin

From: Jakub Kicinski <[email protected]>

[ Upstream commit 9dec850fd7c210a04b4707df8e6c95bfafdd6a4b ]

GCC 12 currently generates a rather inconsistent warning:

drivers/net/ethernet/broadcom/tg3.c:17795:51: warning: array subscript 5 is above array bounds of ‘struct tg3_napi[5]’ [-Warray-bounds]
17795 | struct tg3_napi *tnapi = &tp->napi[i];
| ~~~~~~~~^~~

i is guaranteed < tp->irq_max which in turn is either 1 or 5. There are more loops like this one in the driver, but strangely GCC 12 dislikes only this single one.

Silence this silliness for now.

---
 drivers/net/ethernet/broadcom/Makefile | 5 +++++
 1 file changed, 5 insertions(+)
drivers/net/ethernet/broadcom/Makefile +5 -0
index 0ddfb5b5d53c..2e6c5f258a1f 100644
@@ -17,3 +17,8 @@ obj-$(CONFIG_BGMAC_BCMA) += bgmac-bcma.o bgmac-bcma-mdio.o
 obj-$(CONFIG_BGMAC_PLATFORM) += bgmac-platform.o
 obj-$(CONFIG_SYSTEMPORT) += bcmsysport.o
 obj-$(CONFIG_BNXT) += bnxt/
+
+# FIXME: temporarily silence -Warray-bounds on non W=1+ builds
+ifndef KBUILD_EXTRA_WARN
+CFLAGS_tg3.o += -Wno-array-bounds
+endif
Signature · 1 line
2.35.1
author maintainer
#168 lore raw

PATCH 156/159 char: tpm: cr50_i2c: Suppress duplicated error message in .remove()

In reply to Sasha Levin

From: Uwe Kleine-König <[email protected]>

[ Upstream commit e0687fe958f763f1790f22ed5483025b7624e744 ]

Returning an error value in an i2c remove callback results in an error message being emitted by the i2c core, but otherwise it doesn't make a difference. The device goes away anyhow and the devm cleanups are called.

As tpm_cr50_i2c_remove() emits an error message already and the additional error message by the i2c core doesn't add any useful information, change the return value to zero to suppress this error message.

Note that if i2c_clientdata is NULL, there is something really fishy. Assuming no memory corruption happened (then all bets are lost anyhow), tpm_cr50_i2c_remove() is only called after tpm_cr50_i2c_probe() returned successfully. So there was a tpm chip registered before and after tpm_cr50_i2c_remove() its privdata is freed but the associated character device isn't removed. If after that happened userspace accesses the character device it's likely that the freed memory is accessed. For that reason the warning message is made a bit more frightening.

---
 drivers/char/tpm/tpm_tis_i2c_cr50.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/char/tpm/tpm_tis_i2c_cr50.c +2 -2
index f6c0affbb456..bf608b6af339 100644
@@ -768,8 +768,8 @@ static int tpm_cr50_i2c_remove(struct i2c_client *client)
 	struct device *dev = &client->dev;
 
 	if (!chip) {
-		dev_err(dev, "Could not get client data at remove\n");
-		return -ENODEV;
+		dev_crit(dev, "Could not get client data at remove, memory corruption ahead\n");
+		return 0;
 	}
 
 	tpm_chip_unregister(chip);
Signature · 1 line
2.35.1
author maintainer
#169 lore raw

PATCH 157/159 selftests/bpf: fix btf_dump/btf_dump due to recent clang change

In reply to Sasha Levin

From: Yonghong Song <[email protected]>

[ Upstream commit 4050764cbaa25760aab40857f723393c07898474 ]

Latest llvm-project upstream had a change of behavior related to qualifiers on function return type ([1]). This caused selftests btf_dump/btf_dump failure. The following example shows what changed.

  $ cat t.c
  typedef const char * const (* const (* const fn_ptr_arr2_t[5])())(char * (*)(int));
  struct t {
    int a;
    fn_ptr_arr2_t l;
  };
  int foo(struct t *arg) {
    return arg->a;
  }

Compiled with latest upstream llvm15,

  $ clang -O2 -g -target bpf -S -emit-llvm t.c
The related generated debuginfo IR looks like:
  !16 = !DIDerivedType(tag: DW_TAG_typedef, name: "fn_ptr_arr2_t", file: !1, line: 1, baseType: !17)
  !17 = !DICompositeType(tag: DW_TAG_array_type, baseType: !18, size: 320, elements: !32)
  !18 = !DIDerivedType(tag: DW_TAG_const_type, baseType: !19)
  !19 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !20, size: 64)
  !20 = !DISubroutineType(types: !21)
  !21 = !{!22, null}
  !22 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !23, size: 64)
  !23 = !DISubroutineType(types: !24)
  !24 = !{!25, !28}
  !25 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !26, size: 64)
  !26 = !DIDerivedType(tag: DW_TAG_const_type, baseType: !27)
  !27 = !DIBasicType(name: "char", size: 8, encoding: DW_ATE_signed_char)
You can see two intermediate const qualifier to pointer are dropped in debuginfo IR.

With llvm14, we have following debuginfo IR:
!16 = !DIDerivedType(tag: DW_TAG_typedef, name: "fn_ptr_arr2_t", file: !1, line: 1, baseType: !17) !17 = !DICompositeType(tag: DW_TAG_array_type, baseType: !18, size: 320, elements: !34) !18 = !DIDerivedType(tag: DW_TAG_const_type, baseType: !19)
!19 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !20, size: 64)
!20 = !DISubroutineType(types: !21)
!21 = !{!22, null}
!22 = !DIDerivedType(tag: DW_TAG_const_type, baseType: !23)
!23 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !24, size: 64)
!24 = !DISubroutineType(types: !25)
!25 = !{!26, !30}
!26 = !DIDerivedType(tag: DW_TAG_const_type, baseType: !27)
!27 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !28, size: 64)
!28 = !DIDerivedType(tag: DW_TAG_const_type, baseType: !29)
!29 = !DIBasicType(name: "char", size: 8, encoding: DW_ATE_signed_char)
All const qualifiers are preserved.

To adapt the selftest to both old and new llvm, this patch removed the intermediate const qualifier in const-to-ptr types, to make the test succeed again.

[1] https://reviews.llvm.org/D125919

---
 tools/testing/selftests/bpf/progs/btf_dump_test_case_syntax.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/bpf/progs/btf_dump_test_case_syntax.c +1 -1
index 1c7105fcae3c..4ee4748133fe 100644
@@ -94,7 +94,7 @@ typedef void (* (*signal_t)(int, void (*)(int)))(int);
 
 typedef char * (*fn_ptr_arr1_t[10])(int **);
 
-typedef char * (* const (* const fn_ptr_arr2_t[5])())(char * (*)(int));
+typedef char * (* (* const fn_ptr_arr2_t[5])())(char * (*)(int));
 
 struct struct_w_typedefs {
 	int_t a;
Signature · 1 line
2.35.1
author maintainer
#170 lore raw

PATCH 158/159 gfs2: use i_lock spin_lock for inode qadata

In reply to Sasha Levin

From: Bob Peterson <[email protected]>

[ Upstream commit 5fcff61eea9efd1f4b60e89d2d686b5feaea100f ]

Before this patch, functions gfs2_qa_get and _put used the i_rw_mutex to prevent simultaneous access to its i_qadata. But i_rw_mutex is now used for many other things, including iomap_begin and end, which causes a conflict according to lockdep. We cannot just remove the lock since simultaneous opens (gfs2_open -> gfs2_open_common -> gfs2_qa_get) can then stomp on each others values for i_qadata.

This patch solves the conflict by using the i_lock spin_lock in the inode to prevent simultaneous access.

---
 fs/gfs2/quota.c | 32 ++++++++++++++++++++------------
 1 file changed, 20 insertions(+), 12 deletions(-)
fs/gfs2/quota.c +20 -12
index be0997e24d60..dc77080a82bb 100644
@@ -531,34 +531,42 @@ static void qdsb_put(struct gfs2_quota_data *qd)
  */
 int gfs2_qa_get(struct gfs2_inode *ip)
 {
-	int error = 0;
 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
+	struct inode *inode = &ip->i_inode;
 
 	if (sdp->sd_args.ar_quota == GFS2_QUOTA_OFF)
 		return 0;
 
-	down_write(&ip->i_rw_mutex);
+	spin_lock(&inode->i_lock);
 	if (ip->i_qadata == NULL) {
-		ip->i_qadata = kmem_cache_zalloc(gfs2_qadata_cachep, GFP_NOFS);
-		if (!ip->i_qadata) {
-			error = -ENOMEM;
-			goto out;
-		}
+		struct gfs2_qadata *tmp;
+
+		spin_unlock(&inode->i_lock);
+		tmp = kmem_cache_zalloc(gfs2_qadata_cachep, GFP_NOFS);
+		if (!tmp)
+			return -ENOMEM;
+
+		spin_lock(&inode->i_lock);
+		if (ip->i_qadata == NULL)
+			ip->i_qadata = tmp;
+		else
+			kmem_cache_free(gfs2_qadata_cachep, tmp);
 	}
 	ip->i_qadata->qa_ref++;
-out:
-	up_write(&ip->i_rw_mutex);
-	return error;
+	spin_unlock(&inode->i_lock);
+	return 0;
 }
 
 void gfs2_qa_put(struct gfs2_inode *ip)
 {
-	down_write(&ip->i_rw_mutex);
+	struct inode *inode = &ip->i_inode;
+
+	spin_lock(&inode->i_lock);
 	if (ip->i_qadata && --ip->i_qadata->qa_ref == 0) {
 		kmem_cache_free(gfs2_qadata_cachep, ip->i_qadata);
 		ip->i_qadata = NULL;
 	}
-	up_write(&ip->i_rw_mutex);
+	spin_unlock(&inode->i_lock);
 }
 
 int gfs2_quota_hold(struct gfs2_inode *ip, kuid_t uid, kgid_t gid)
Signature · 1 line
2.35.1
author maintainer
#171 lore raw

PATCH 159/159 linux/types.h: reinstate "__bitwise__" macro for user space use

In reply to Sasha Levin

From: Linus Torvalds <[email protected]>

[ Upstream commit caa28984163cb63ea0be4cb8dbf05defdc7303f9 ]

Commit c724c866bb70 ("linux/types.h: remove unnecessary __bitwise__") was right that there are no users of __bitwise__ in the kernel, but it turns out there are user space users of it that do expect it.

It is, after all, in the uapi directory, so user space usage is to be expected.

Instead of reverting the commit completely, let's just clarify the situation so that it doesn't happen again, and have some in-code explanations for why that "__bitwise__" still exists.

---
 include/uapi/linux/types.h | 3 +++
 1 file changed, 3 insertions(+)
include/uapi/linux/types.h +3 -0
index c4dc597f3dcf..308433be33c2 100644
@@ -26,6 +26,9 @@
 #define __bitwise
 #endif
 
+/* The kernel doesn't use this legacy form, but user space does */
+#define __bitwise__ __bitwise
+
 typedef __u16 __bitwise __le16;
 typedef __u16 __bitwise __be16;
 typedef __u32 __bitwise __le32;
Signature · 1 line
2.35.1