Skip to contentSkip to conversation

Drivers: hv: Miscellaneous fixes

PATCH Superseded [email protected] 7 participants 21 messages Oct 17, 2018 – Oct 21, 2018

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

cover Drivers: hv: Miscellaneous fixes

From: "K. Y. Srinivasan" <[email protected]>

Miscellaneous fixes.

Dexuan Cui (3):
Drivers: hv: kvp: Fix the recent regression caused by incorrect
clean-up
Drivers: hv: kvp: Use %u to print U32
Tools: hv: kvp: Fix a warning of buffer overflow with gcc 8.0.1

Haiyang Zhang (1):
hv_utils: update name in struct hv_driver util_drv

K. Y. Srinivasan (1):
Drivers: hv: vmbus: Get rid of unnecessary state in hv_context

 drivers/hv/hv.c           | 10 +++-------
 drivers/hv/hv_kvp.c       | 28 +++++++++++++++++++++++-----
 drivers/hv/hv_util.c      |  2 +-
 drivers/hv/hyperv_vmbus.h |  2 --
 tools/hv/hv_kvp_daemon.c  |  2 +-
 5 files changed, 28 insertions(+), 16 deletions(-)
Signature · 1 line
2.18.0
#1 lore raw

PATCH 1/5 Drivers: hv: vmbus: Get rid of unnecessary state in hv_context

From: "K. Y. Srinivasan" <[email protected]>

Currently we are replicating state in struct hv_context that is unnecessary - this state can be retrieved from the hypervisor. Furthermore, this is a per-cpu state that is being maintained as a global state in struct hv_context. Get rid of this state in struct hv_context.

Reply-To: [email protected]

---
 drivers/hv/hv.c           | 10 +++-------
 drivers/hv/hyperv_vmbus.h |  2 --
 2 files changed, 3 insertions(+), 9 deletions(-)
drivers/hv/hv.c +3 -7
index 332d7c34be5c..166c2501de17 100644
@@ -33,9 +33,7 @@
 #include "hyperv_vmbus.h"
 
 /* The one and only */
-struct hv_context hv_context = {
-	.synic_initialized	= false,
-};
+struct hv_context hv_context;
 
 /*
  * If false, we're using the old mechanism for stimer0 interrupts
@@ -326,8 +324,6 @@ int hv_synic_init(unsigned int cpu)
 
 	hv_set_synic_state(sctrl.as_uint64);
 
-	hv_context.synic_initialized = true;
-
 	/*
 	 * Register the per-cpu clockevent source.
 	 */
@@ -373,7 +369,8 @@ int hv_synic_cleanup(unsigned int cpu)
 	bool channel_found = false;
 	unsigned long flags;
 
-	if (!hv_context.synic_initialized)
+	hv_get_synic_state(sctrl.as_uint64);
+	if (sctrl.enable != 1)
 		return -EFAULT;
 
 	/*
@@ -435,7 +432,6 @@ int hv_synic_cleanup(unsigned int cpu)
 	hv_set_siefp(siefp.as_uint64);
 
 	/* Disable the global synic bit */
-	hv_get_synic_state(sctrl.as_uint64);
 	sctrl.enable = 0;
 	hv_set_synic_state(sctrl.as_uint64);
 
drivers/hv/hyperv_vmbus.h +0 -2
index 72eaba3d50fc..f17c06a5e74b 100644
@@ -230,8 +230,6 @@ struct hv_context {
 
 	void *tsc_page;
 
-	bool synic_initialized;
-
 	struct hv_per_cpu_context __percpu *cpu_context;
 
 	/*
Signature · 1 line
2.18.0
#2 lore raw

PATCH 2/5 hv_utils: update name in struct hv_driver util_drv

From: Haiyang Zhang <[email protected]>

The correct module name is hv_utils. This patch corrects the name in struct hv_driver util_drv.

---
 drivers/hv/hv_util.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/hv/hv_util.c +1 -1
index 423205077bf6..f10eeb120c8b 100644
@@ -483,7 +483,7 @@ MODULE_DEVICE_TABLE(vmbus, id_table);
 
 /* The one and only one */
 static  struct hv_driver util_drv = {
-	.name = "hv_util",
+	.name = "hv_utils",
 	.id_table = id_table,
 	.probe =  util_probe,
 	.remove =  util_remove,
Signature · 1 line
2.18.0
#3 lore raw

PATCH 3/5 Drivers: hv: kvp: Fix the recent regression caused by incorrect clean-up

In reply to [email protected]

From: Dexuan Cui <[email protected]>

In kvp_send_key(), we do need call process_ib_ipinfo() if message->kvp_hdr.operation is KVP_OP_GET_IP_INFO, because it turns out the userland hv_kvp_daemon needs the info of operation, adapter_id and addr_family. With the incorrect fc62c3b1977d, the host can't get the VM's IP via KVP.

And, fc62c3b1977d added a "break;", but actually forgot to initialize the key_size/value in the case of KVP_OP_SET, so the default key_size of 0 is passed to the kvp daemon, and the pool files /var/lib/hyperv/.kvp_pool_* can't be updated.

This patch effectively rolls back the previous fc62c3b1977d, and correctly fixes the "this statement may fall through" warnings.

This patch is tested on WS 2012 R2 and 2016.

Fixes fc62c3b1977d ("Drivers: hv: kvp: Fix two "this statement may fall through" warnings")
Cc K. Y. Srinivasan <[email protected]>
Cc Haiyang Zhang <[email protected]>
Cc Stephen Hemminger <[email protected]>
---
 drivers/hv/hv_kvp.c | 26 ++++++++++++++++++++++----
 1 file changed, 22 insertions(+), 4 deletions(-)
drivers/hv/hv_kvp.c +22 -4
index a7513a8a8e37..9fbb15c62c6c 100644
@@ -353,6 +353,9 @@ static void process_ib_ipinfo(void *in_msg, void *out_msg, int op)
 
 		out->body.kvp_ip_val.dhcp_enabled = in->kvp_ip_val.dhcp_enabled;
 
+		__attribute__ ((fallthrough));
+
+	case KVP_OP_GET_IP_INFO:
 		utf16s_to_utf8s((wchar_t *)in->kvp_ip_val.adapter_id,
 				MAX_ADAPTER_ID_SIZE,
 				UTF16_LITTLE_ENDIAN,
@@ -405,7 +408,11 @@ kvp_send_key(struct work_struct *dummy)
 		process_ib_ipinfo(in_msg, message, KVP_OP_SET_IP_INFO);
 		break;
 	case KVP_OP_GET_IP_INFO:
-		/* We only need to pass on message->kvp_hdr.operation.  */
+		/*
+		 * We only need to pass on the info of operation, adapter_id
+		 * and addr_family to the userland kvp daemon.
+		 */
+		process_ib_ipinfo(in_msg, message, KVP_OP_GET_IP_INFO);
 		break;
 	case KVP_OP_SET:
 		switch (in_msg->body.kvp_set.data.value_type) {
@@ -446,9 +453,9 @@ kvp_send_key(struct work_struct *dummy)
 
 		}
 
-		break;
-
-	case KVP_OP_GET:
+		/*
+		 * The key is always a string - utf16 encoding.
+		 */
 		message->body.kvp_set.data.key_size =
 			utf16s_to_utf8s(
 			(wchar_t *)in_msg->body.kvp_set.data.key,
@@ -456,6 +463,17 @@ kvp_send_key(struct work_struct *dummy)
 			UTF16_LITTLE_ENDIAN,
 			message->body.kvp_set.data.key,
 			HV_KVP_EXCHANGE_MAX_KEY_SIZE - 1) + 1;
+
+		break;
+
+	case KVP_OP_GET:
+		message->body.kvp_get.data.key_size =
+			utf16s_to_utf8s(
+			(wchar_t *)in_msg->body.kvp_get.data.key,
+			in_msg->body.kvp_get.data.key_size,
+			UTF16_LITTLE_ENDIAN,
+			message->body.kvp_get.data.key,
+			HV_KVP_EXCHANGE_MAX_KEY_SIZE - 1) + 1;
 		break;
 
 	case KVP_OP_DELETE:
Signature · 1 line
2.18.0
maintainer
#4 lore raw

Re: [PATCH 3/5] Drivers: hv: kvp: Fix the recent regression caused by incorrect clean-up

Quoted 34 lines · On Wed, Oct 17, 2018 at 03:14:04AM +0000, [email protected] wrote:

From: Dexuan Cui <[email protected]>

In kvp_send_key(), we do need call process_ib_ipinfo() if message->kvp_hdr.operation is KVP_OP_GET_IP_INFO, because it turns out the userland hv_kvp_daemon needs the info of operation, adapter_id and addr_family. With the incorrect fc62c3b1977d, the host can't get the VM's IP via KVP.

And, fc62c3b1977d added a "break;", but actually forgot to initialize the key_size/value in the case of KVP_OP_SET, so the default key_size of 0 is passed to the kvp daemon, and the pool files /var/lib/hyperv/.kvp_pool_* can't be updated.

This patch effectively rolls back the previous fc62c3b1977d, and correctly fixes the "this statement may fall through" warnings.

This patch is tested on WS 2012 R2 and 2016.

Fixes fc62c3b1977d ("Drivers: hv: kvp: Fix two "this statement may fall through" warnings")
Cc K. Y. Srinivasan <[email protected]>
Cc Haiyang Zhang <[email protected]>
Cc Stephen Hemminger <[email protected]>
---
 drivers/hv/hv_kvp.c | 26 ++++++++++++++++++++++----
 1 file changed, 22 insertions(+), 4 deletions(-)
drivers/hv/hv_kvp.c +0 -0
index a7513a8a8e37..9fbb15c62c6c 100644

drivers/hv/hv_kvp.c static void process_ib_ipinfo(void *in_msg, void *out_msg, int op) · line 356

Quoted 4 lines
drivers/hv/hv_kvp.c +1 -0
@@ -353,6 +353,9 @@ static void process_ib_ipinfo(void *in_msg, void *out_msg, int op)
 
 		out->body.kvp_ip_val.dhcp_enabled = in->kvp_ip_val.dhcp_enabled;
 
+		__attribute__ ((fallthrough));

The comment should be sufficient for this, right? I haven't seen many uses of this attribute before, how common is it?

Quoted 36 lines
drivers/hv/hv_kvp.c +11 -4
+
+	case KVP_OP_GET_IP_INFO:
 		utf16s_to_utf8s((wchar_t *)in->kvp_ip_val.adapter_id,
 				MAX_ADAPTER_ID_SIZE,
 				UTF16_LITTLE_ENDIAN,
@@ -405,7 +408,11 @@ kvp_send_key(struct work_struct *dummy)
 		process_ib_ipinfo(in_msg, message, KVP_OP_SET_IP_INFO);
 		break;
 	case KVP_OP_GET_IP_INFO:
-		/* We only need to pass on message->kvp_hdr.operation.  */
+		/*
+		 * We only need to pass on the info of operation, adapter_id
+		 * and addr_family to the userland kvp daemon.
+		 */
+		process_ib_ipinfo(in_msg, message, KVP_OP_GET_IP_INFO);
 		break;
 	case KVP_OP_SET:
 		switch (in_msg->body.kvp_set.data.value_type) {
@@ -446,9 +453,9 @@ kvp_send_key(struct work_struct *dummy)
 
 		}
 
-		break;
-
-	case KVP_OP_GET:
+		/*
+		 * The key is always a string - utf16 encoding.
+		 */
 		message->body.kvp_set.data.key_size =
 			utf16s_to_utf8s(
 			(wchar_t *)in_msg->body.kvp_set.data.key,
@@ -456,6 +463,17 @@ kvp_send_key(struct work_struct *dummy)
 			UTF16_LITTLE_ENDIAN,
 			message->body.kvp_set.data.key,
 			HV_KVP_EXCHANGE_MAX_KEY_SIZE - 1) + 1;
+

drivers/hv/hv_kvp.c kvp_send_key(struct work_struct *dummy) · line 476

Quoted 10 lines
drivers/hv/hv_kvp.c +10 -0
+		break;
+
+	case KVP_OP_GET:
+		message->body.kvp_get.data.key_size =
+			utf16s_to_utf8s(
+			(wchar_t *)in_msg->body.kvp_get.data.key,
+			in_msg->body.kvp_get.data.key_size,
+			UTF16_LITTLE_ENDIAN,
+			message->body.kvp_get.data.key,
+			HV_KVP_EXCHANGE_MAX_KEY_SIZE - 1) + 1;

Worst indentation ever :(

Yeah, I know it follows the others above it, but you should reconsider it in the future...

thanks,

greg k-h

#5 lore raw

Re: [PATCH 3/5] Drivers: hv: kvp: Fix the recent regression caused by incorrect clean-up

Quoted 36 lines · Greg KH
Quoted 35 lines · On Wed, Oct 17, 2018 at 03:14:04AM +0000, [email protected] wrote:

From: Dexuan Cui <[email protected]>

In kvp_send_key(), we do need call process_ib_ipinfo() if message->kvp_hdr.operation is KVP_OP_GET_IP_INFO, because it turns out the userland hv_kvp_daemon needs the info of operation, adapter_id and addr_family. With the incorrect fc62c3b1977d, the host can't get the VM's IP via KVP.

And, fc62c3b1977d added a "break;", but actually forgot to initialize the key_size/value in the case of KVP_OP_SET, so the default key_size of 0 is passed to the kvp daemon, and the pool files /var/lib/hyperv/.kvp_pool_* can't be updated.

This patch effectively rolls back the previous fc62c3b1977d, and correctly fixes the "this statement may fall through" warnings.

This patch is tested on WS 2012 R2 and 2016.

Fixes fc62c3b1977d ("Drivers: hv: kvp: Fix two "this statement may fall through" warnings")
Cc K. Y. Srinivasan <[email protected]>
Cc Haiyang Zhang <[email protected]>
Cc Stephen Hemminger <[email protected]>
---
 drivers/hv/hv_kvp.c | 26 ++++++++++++++++++++++----
 1 file changed, 22 insertions(+), 4 deletions(-)
drivers/hv/hv_kvp.c +0 -0
index a7513a8a8e37..9fbb15c62c6c 100644
@@ -353,6 +353,9 @@ static void process_ib_ipinfo(void *in_msg, void *out_msg, int op)
 
Quoted 7 lines

drivers/hv/hv_kvp.c static void process_ib_ipinfo(void *in_msg, void *out_msg, int op) · line 356

Quoted 3 lines
drivers/hv/hv_kvp.c +1 -0
 		out->body.kvp_ip_val.dhcp_enabled = in->kvp_ip_val.dhcp_enabled;
 
+		__attribute__ ((fallthrough));

The comment should be sufficient for this, right? I haven't seen many uses of this attribute before, how common is it?

It's not common at all.

Dexuan, please use a comment instead: /* fall through */

Thanks
--
Gustavo

Quoted 57 lines
Quoted 37 lines
drivers/hv/hv_kvp.c +11 -4
+
+	case KVP_OP_GET_IP_INFO:
 		utf16s_to_utf8s((wchar_t *)in->kvp_ip_val.adapter_id,
 				MAX_ADAPTER_ID_SIZE,
 				UTF16_LITTLE_ENDIAN,
@@ -405,7 +408,11 @@ kvp_send_key(struct work_struct *dummy)
 		process_ib_ipinfo(in_msg, message, KVP_OP_SET_IP_INFO);
 		break;
 	case KVP_OP_GET_IP_INFO:
-		/* We only need to pass on message->kvp_hdr.operation.  */
+		/*
+		 * We only need to pass on the info of operation, adapter_id
+		 * and addr_family to the userland kvp daemon.
+		 */
+		process_ib_ipinfo(in_msg, message, KVP_OP_GET_IP_INFO);
 		break;
 	case KVP_OP_SET:
 		switch (in_msg->body.kvp_set.data.value_type) {
@@ -446,9 +453,9 @@ kvp_send_key(struct work_struct *dummy)
 
 		}
 
-		break;
-
-	case KVP_OP_GET:
+		/*
+		 * The key is always a string - utf16 encoding.
+		 */
 		message->body.kvp_set.data.key_size =
 			utf16s_to_utf8s(
 			(wchar_t *)in_msg->body.kvp_set.data.key,
@@ -456,6 +463,17 @@ kvp_send_key(struct work_struct *dummy)
 			UTF16_LITTLE_ENDIAN,
 			message->body.kvp_set.data.key,
 			HV_KVP_EXCHANGE_MAX_KEY_SIZE - 1) + 1;
+

drivers/hv/hv_kvp.c kvp_send_key(struct work_struct *dummy) · line 476

Quoted 10 lines
drivers/hv/hv_kvp.c +10 -0
+		break;
+
+	case KVP_OP_GET:
+		message->body.kvp_get.data.key_size =
+			utf16s_to_utf8s(
+			(wchar_t *)in_msg->body.kvp_get.data.key,
+			in_msg->body.kvp_get.data.key_size,
+			UTF16_LITTLE_ENDIAN,
+			message->body.kvp_get.data.key,
+			HV_KVP_EXCHANGE_MAX_KEY_SIZE - 1) + 1;

Worst indentation ever :(

Yeah, I know it follows the others above it, but you should reconsider it in the future...

thanks,

greg k-h

#6 lore raw

RE: [PATCH 3/5] Drivers: hv: kvp: Fix the recent regression caused by incorrect clean-up

In reply to Greg KH
Quoted 53 lines

-----Original Message-----
From: Greg KH <[email protected]>
Sent: Tuesday, October 16, 2018 10:07 PM
To: KY Srinivasan <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Stephen Hemminger <[email protected]>; Michael Kelley (EOSG) <[email protected]>; vkuznets <[email protected]>; Haiyang Zhang <[email protected]>; [email protected]
Subject: Re: [PATCH 3/5] Drivers: hv: kvp: Fix the recent regression caused by incorrect clean-up

Quoted 4 lines · On Wed, Oct 17, 2018 at 03:14:04AM +0000, [email protected] wrote:

From: Dexuan Cui <[email protected]>

In kvp_send_key(), we do need call process_ib_ipinfo() if message->kvp_hdr.operation is KVP_OP_GET_IP_INFO, because it turns

out

Quoted 9 lines

the userland hv_kvp_daemon needs the info of operation, adapter_id and addr_family. With the incorrect fc62c3b1977d, the host can't get the VM's IP via KVP.

And, fc62c3b1977d added a "break;", but actually forgot to initialize the key_size/value in the case of KVP_OP_SET, so the default key_size of 0 is passed to the kvp daemon, and the pool files /var/lib/hyperv/.kvp_pool_* can't be updated.

Quoted 6 lines

This patch effectively rolls back the previous fc62c3b1977d, and correctly fixes the "this statement may fall through" warnings.

This patch is tested on WS 2012 R2 and 2016.

Fixes fc62c3b1977d ("Drivers: hv: kvp: Fix two "this statement may fall

through" warnings")

drivers/hv/hv_kvp.c static void process_ib_ipinfo(void *in_msg, void

Quoted 15 lines
Cc K. Y. Srinivasan <[email protected]>
Cc Haiyang Zhang <[email protected]>
Cc Stephen Hemminger <[email protected]>
---
 drivers/hv/hv_kvp.c | 26 ++++++++++++++++++++++----
 1 file changed, 22 insertions(+), 4 deletions(-)
drivers/hv/hv_kvp.c +0 -0
index a7513a8a8e37..9fbb15c62c6c 100644
@@ -353,6 +353,9 @@ static void process_ib_ipinfo(void *in_msg, void

*out_msg, int op)

Quoted 1 line
Quoted 7 lines

drivers/hv/hv_kvp.c static void process_ib_ipinfo(void *in_msg, void *out_msg, int op) · line 356

Quoted 4 lines

out->body.kvp_ip_val.dhcp_enabled = in-
kvp_ip_val.dhcp_enabled;

drivers/hv/hv_kvp.c +1 -0
+		__attribute__ ((fallthrough));

The comment should be sufficient for this, right? I haven't seen many uses of this attribute before, how common is it?

Yes; a common should be sufficient.

Quoted 59 lines

drivers/hv/hv_kvp.c kvp_send_key(struct work_struct *dummy) · line 408

Quoted 12 lines
drivers/hv/hv_kvp.c +2 -1
+
+	case KVP_OP_GET_IP_INFO:
 		utf16s_to_utf8s((wchar_t *)in->kvp_ip_val.adapter_id,
 				MAX_ADAPTER_ID_SIZE,
 				UTF16_LITTLE_ENDIAN,
@@ -405,7 +408,11 @@ kvp_send_key(struct work_struct *dummy)
 		process_ib_ipinfo(in_msg, message, KVP_OP_SET_IP_INFO);
 		break;
 	case KVP_OP_GET_IP_INFO:
-		/* We only need to pass on message->kvp_hdr.operation.

*/

drivers/hv/hv_kvp.c kvp_send_key(struct work_struct *dummy)

Quoted 5 lines
drivers/hv/hv_kvp.c +5 -0
+		/*
+		 * We only need to pass on the info of operation, adapter_id
+		 * and addr_family to the userland kvp daemon.
+		 */
+		process_ib_ipinfo(in_msg, message,

KVP_OP_GET_IP_INFO);

Quoted 21 lines
drivers/hv/hv_kvp.c +4 -3
 		break;
 	case KVP_OP_SET:
 		switch (in_msg->body.kvp_set.data.value_type) {
@@ -446,9 +453,9 @@ kvp_send_key(struct work_struct *dummy)
 
 		}
 
-		break;
-
-	case KVP_OP_GET:
+		/*
+		 * The key is always a string - utf16 encoding.
+		 */
 		message->body.kvp_set.data.key_size =
 			utf16s_to_utf8s(
 			(wchar_t *)in_msg->body.kvp_set.data.key,
@@ -456,6 +463,17 @@ kvp_send_key(struct work_struct *dummy)
 			UTF16_LITTLE_ENDIAN,
 			message->body.kvp_set.data.key,
 			HV_KVP_EXCHANGE_MAX_KEY_SIZE - 1) + 1;
+

drivers/hv/hv_kvp.c kvp_send_key(struct work_struct *dummy) · line 476

Quoted 10 lines
drivers/hv/hv_kvp.c +10 -0
+		break;
+
+	case KVP_OP_GET:
+		message->body.kvp_get.data.key_size =
+			utf16s_to_utf8s(
+			(wchar_t *)in_msg->body.kvp_get.data.key,
+			in_msg->body.kvp_get.data.key_size,
+			UTF16_LITTLE_ENDIAN,
+			message->body.kvp_get.data.key,
+			HV_KVP_EXCHANGE_MAX_KEY_SIZE - 1) + 1;

Worst indentation ever :(

Yeah, I know it follows the others above it, but you should reconsider it in the future...

thanks,

greg k-h

maintainer
#7 lore raw

RE: [PATCH 3/5] Drivers: hv: kvp: Fix the recent regression caused by incorrect clean-up

Quoted 11 lines

From: devel <[email protected]> On Behalf Of KY Srinivasan
Sent: Tuesday, October 16, 2018 23:02

Quoted 8 lines

drivers/hv/hv_kvp.c static void process_ib_ipinfo(void *in_msg, void

Quoted 3 lines
drivers/hv/hv_kvp.c +0 -0
@@ -353,6 +353,9 @@ static void process_ib_ipinfo(void *in_msg, void

*out_msg, int op)

Quoted 4 lines
drivers/hv/hv_kvp.c +0 -0
 		out->body.kvp_ip_val.dhcp_enabled = in-
 kvp_ip_val.dhcp_enabled;
 
Quoted 6 lines
Quoted 4 lines

drivers/hv/hv_kvp.c static void process_ib_ipinfo(void *in_msg, void *out_msg, int op) · line 356

Quoted 1 line
drivers/hv/hv_kvp.c +1 -0
+		__attribute__ ((fallthrough));

The comment should be sufficient for this, right? I haven't seen many uses of this attribute before, how common is it?

Yes; a common should be sufficient.

You all are right.
Right now, I realized the gcc warning can also be suppressed by a simple line of comment:

/* fallthrough */

It looks gcc treats this comment specially.

If I add something more in the comment, like
/* add fallthrough */
, the warning can not be suppressed. :-)

I'll do a new version for KY.

Thank you all!

-- Dexuan

#8 lore raw

Re: [PATCH 3/5] Drivers: hv: kvp: Fix the recent regression caused by incorrect clean-up

In reply to Greg KH
Quoted 36 lines · Greg KH
Quoted 35 lines · On Wed, Oct 17, 2018 at 03:14:04AM +0000, [email protected] wrote:

From: Dexuan Cui <[email protected]>

In kvp_send_key(), we do need call process_ib_ipinfo() if message->kvp_hdr.operation is KVP_OP_GET_IP_INFO, because it turns out the userland hv_kvp_daemon needs the info of operation, adapter_id and addr_family. With the incorrect fc62c3b1977d, the host can't get the VM's IP via KVP.

And, fc62c3b1977d added a "break;", but actually forgot to initialize the key_size/value in the case of KVP_OP_SET, so the default key_size of 0 is passed to the kvp daemon, and the pool files /var/lib/hyperv/.kvp_pool_* can't be updated.

This patch effectively rolls back the previous fc62c3b1977d, and correctly fixes the "this statement may fall through" warnings.

This patch is tested on WS 2012 R2 and 2016.

Fixes fc62c3b1977d ("Drivers: hv: kvp: Fix two "this statement may fall through" warnings")
Cc K. Y. Srinivasan <[email protected]>
Cc Haiyang Zhang <[email protected]>
Cc Stephen Hemminger <[email protected]>
---
 drivers/hv/hv_kvp.c | 26 ++++++++++++++++++++++----
 1 file changed, 22 insertions(+), 4 deletions(-)
drivers/hv/hv_kvp.c +0 -0
index a7513a8a8e37..9fbb15c62c6c 100644
@@ -353,6 +353,9 @@ static void process_ib_ipinfo(void *in_msg, void *out_msg, int op)
 
Quoted 7 lines

drivers/hv/hv_kvp.c static void process_ib_ipinfo(void *in_msg, void *out_msg, int op) · line 356

Quoted 3 lines
drivers/hv/hv_kvp.c +1 -0
 		out->body.kvp_ip_val.dhcp_enabled = in->kvp_ip_val.dhcp_enabled;
 
+		__attribute__ ((fallthrough));

The comment should be sufficient for this, right? I haven't seen many uses of this attribute before, how common is it?

It's not common at all. It should be wrapped in a macro and put into compiler.h.

But I hope it does become adopted. It's better than randomly grepping for non-standard comments.

regards,
dan carpenter

#9 lore raw

Re: [PATCH 3/5] Drivers: hv: kvp: Fix the recent regression caused by incorrect clean-up

Quoted 6 lines · Dan Carpenter

It's not common at all. It should be wrapped in a macro and put into compiler.h.

But I hope it does become adopted. It's better than randomly grepping for non-standard comments.

Using an attribute is indeed better whenever possible. In C++17 it is an standard attribute and there have been proposals to include some of them for C as well since 2016 at least, e.g. the latest for fallthrough at:

http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2268.pdf

I have taken a quick look into supporting it (typing it here to save it on the mailing list :-), and we have:

* gcc >= 7.1 supports -Wimplicit-fallthrough with
__attribute__((fallthrough)), the comment syntax and the C++ [[fallthrough]] syntax.
* gcc < 7.1 complains about empty declarations (it does not know
about attributes for null statements) and also -Wdeclaration-after-statement.
* clang 7 supports -Wimplicit-fallthrough (not enabled in
-Wall/extra/pedantic like gcc, though) but *only* in C++ mode and with the C++ syntax [[fallthrough]]. In other words, in C mode, no syntax works and no diagnostics are emitted. It complains about Wmissing-declarations. [IMO they should allow the __attribute__ syntax for fallthrough (and enable it on C mode) to be compatible with gcc. Maybe they are simply waiting for the C2x attributes... :-)]
* icc 19 does not know about -Wimplicit-fallthrough at all (but
seems to allow [[fallthrough]] on C++17 mode to comply with the standard).

Therefore, the only improvement we could do right now is starting to use the attribute for gcc > 7.1, and a comment for everybody else. However, even if that was worth the trouble of changing the 2500+ instances of fall through markings that we have, comments are replaced before the preprocessor stage, so we would need some (probably non-portable) macro magic.

So, I would say, let's revisit this again in a few years. Possibly, when we move the minimum version to gcc 7.1, clang and icc may both support the fallthrough warning in C mode; so that we can always use the __attribute__((fallthrough)) unconditionally under a __fallthrough #define.

Cheers,
Miguel

#10 lore raw

Re: [PATCH 3/5] Drivers: hv: kvp: Fix the recent regression caused by incorrect clean-up

Quoted 18 lines · Miguel Ojeda
Quoted 6 lines · Dan Carpenter

It's not common at all. It should be wrapped in a macro and put into compiler.h.

But I hope it does become adopted. It's better than randomly grepping for non-standard comments.

Using an attribute is indeed better whenever possible. In C++17 it is an standard attribute and there have been proposals to include some of them for C as well since 2016 at least, e.g. the latest for fallthrough at:

http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2268.pdf

I have taken a quick look into supporting it (typing it here to save it on the mailing list :-), and we have:

Quoted 6 lines

* gcc >= 7.1 supports -Wimplicit-fallthrough with
__attribute__((fallthrough)), the comment syntax and the C++ [[fallthrough]] syntax.
* gcc < 7.1 complains about empty declarations (it does not know
about attributes for null statements) and also -Wdeclaration-after-statement.

I'm not sure I understand about empty decalarations. The idea is that we would do:

	case 3:
		frob();
		__fall_through();
	case 4:
		frob();

#if GCC > 7.1
#define __fall_through() __attribute__((fallthrough)) #else
#define __fall_through()
#endif

So long as GCC and static analysis tools understand about the attribute that's enought to catch the bugs. No one cares what clang and icc do. We would just disable the fall through warnings on those until they are updated to support the fallthrough attribute.

We wouldn't update all the fall through comments until later, but going forward people could use the __fall_through() macro if they want.

regards,
dan carpenter

#11 lore raw

Re: [PATCH 3/5] Drivers: hv: kvp: Fix the recent regression caused by incorrect clean-up

Hi Dan,

Quoted 19 lines · Dan Carpenter
Quoted 10 lines · Miguel Ojeda

Using an attribute is indeed better whenever possible. In C++17 it is an standard attribute and there have been proposals to include some of them for C as well since 2016 at least, e.g. the latest for fallthrough at:

http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2268.pdf

I have taken a quick look into supporting it (typing it here to save it on the mailing list :-), and we have:

Quoted 6 lines

* gcc >= 7.1 supports -Wimplicit-fallthrough with
__attribute__((fallthrough)), the comment syntax and the C++ [[fallthrough]] syntax.
* gcc < 7.1 complains about empty declarations (it does not know
about attributes for null statements) and also -Wdeclaration-after-statement.

Quoted 2 lines

I'm not sure I understand about empty decalarations. The idea is that we would do:

That paragraph tried to explain that gcc < 7.1 did not know about __attribute__((fallthrough)); i.e. that everything was introduced in gcc 7.1.

Anyway, the conclusion was a neuron misfiring of mine -- in my mind I was thinking clang supported the comment syntax (when I clearly wrote that it didn't). Never mind --- thanks for pointing it out!

Quoted 7 lines
        case 3:
                frob();
                __fall_through();
        case 4:
                frob();
Quoted 5 lines

#if GCC > 7.1
#define __fall_through() __attribute__((fallthrough)) #else
#define __fall_through()
#endif

Yes, of course, that is what we do for other attributes -- actually in -next we have pending a better way for checking, using __has_attribute:

  #if __has_attribute(fallthrough)
  #define __fallthrough __attribute__((fallthrough))
  #else
  #define __fallthrough
  #endif
Quoted 3 lines

So long as GCC and static analysis tools understand about the attribute that's enought to catch the bugs. No one cares what clang and icc do.

Not so sure about that -- there are quite some people looking forward to building Linux with clang, even if only to have another compiler to check for warnings and to use the clang/llvm-related tools (and to write new ones).

Quoted 2 lines

We would just disable the fall through warnings on those until they are updated to support the fallthrough attribute.

You mean if they start supporting the warning but not the attribute? I don't think that would be likely (i.e. if clang enables the warning on C mode, they will have to introduce a way to suppress it; which should be the attribute (at least), since they typically like to be compatible with gcc and since they already support it in C++ >= 11), but everything is possible.

Quoted 3 lines

We wouldn't update all the fall through comments until later, but going forward people could use the __fall_through() macro if they want.

Agreed. I will introduce it in the compiler-attributes tree -- should be there for -rc1 if no one complains. CC'ing all of you in the patch.

Cheers,
Miguel

maintainer
#12 lore raw

RE: [PATCH 3/5] Drivers: hv: kvp: Fix the recent regression caused by incorrect clean-up

In reply to Greg KH
Quoted 15 lines

From: devel <[email protected]> On Behalf Of Greg KH
Sent: Tuesday, October 16, 2018 22:07

drivers/hv/hv_kvp.c kvp_send_key(struct work_struct *dummy) · line 476

Quoted 9 lines

...

drivers/hv/hv_kvp.c +8 -0
+	case KVP_OP_GET:
+		message->body.kvp_get.data.key_size =
+			utf16s_to_utf8s(
+			(wchar_t *)in_msg->body.kvp_get.data.key,
+			in_msg->body.kvp_get.data.key_size,
+			UTF16_LITTLE_ENDIAN,
+			message->body.kvp_get.data.key,
+			HV_KVP_EXCHANGE_MAX_KEY_SIZE - 1) + 1;

Worst indentation ever :(

Quoted 3 lines

Yeah, I know it follows the others above it, but you should reconsider it in the future...
greg k-h

Agreed. We should consider improving this in future.

-- Dexuan

#13 lore raw

PATCH 4/5 Drivers: hv: kvp: Use %u to print U32

In reply to [email protected]

From: Dexuan Cui <[email protected]>

I didn't find a real issue. Let's just make it consistent with the next "case REG_U64:" where %llu is used.

Cc K. Y. Srinivasan <[email protected]>
Cc Haiyang Zhang <[email protected]>
Cc Stephen Hemminger <[email protected]>
---
 drivers/hv/hv_kvp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/hv/hv_kvp.c +1 -1
index 9fbb15c62c6c..3b8590ff94ba 100644
@@ -437,7 +437,7 @@ kvp_send_key(struct work_struct *dummy)
 			val32 = in_msg->body.kvp_set.data.value_u32;
 			message->body.kvp_set.data.value_size =
 				sprintf(message->body.kvp_set.data.value,
-					"%d", val32) + 1;
+					"%u", val32) + 1;
 			break;
 
 		case REG_U64:
Signature · 1 line
2.18.0
maintainer
#14 lore raw

Re: [PATCH 4/5] Drivers: hv: kvp: Use %u to print U32

Quoted 10 lines · On Wed, Oct 17, 2018 at 03:14:05AM +0000, [email protected] wrote:

From: Dexuan Cui <[email protected]>

I didn't find a real issue. Let's just make it consistent with the next "case REG_U64:" where %llu is used.

Cc K. Y. Srinivasan <[email protected]>
Cc Haiyang Zhang <[email protected]>
Cc Stephen Hemminger <[email protected]>

Why is this a stable patch?

confused,

greg k-h

#15 lore raw

RE: [PATCH 4/5] Drivers: hv: kvp: Use %u to print U32

Quoted 19 lines

-----Original Message-----
From: Greg KH <[email protected]>
Sent: Tuesday, October 16, 2018 10:04 PM
To: KY Srinivasan <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Stephen Hemminger <[email protected]>; Michael Kelley (EOSG) <[email protected]>; vkuznets <[email protected]>; Haiyang Zhang <[email protected]>; [email protected]
Subject: Re: [PATCH 4/5] Drivers: hv: kvp: Use %u to print U32

Quoted 5 lines · On Wed, Oct 17, 2018 at 03:14:05AM +0000, [email protected] wrote:

From: Dexuan Cui <[email protected]>

I didn't find a real issue. Let's just make it consistent with the next "case REG_U64:" where %llu is used.

Quoted 7 lines
Quoted 5 lines
Cc K. Y. Srinivasan <[email protected]>
Cc Haiyang Zhang <[email protected]>
Cc Stephen Hemminger <[email protected]>

Why is this a stable patch?

My mistake; I will resend.

K. Y

Quoted 4 lines

confused,

greg k-h

#16 lore raw

PATCH 5/5 Tools: hv: kvp: Fix a warning of buffer overflow with gcc 8.0.1

In reply to [email protected]

From: Dexuan Cui <[email protected]>

The patch fixes:

hv_kvp_daemon.c: In function 'kvp_set_ip_info':
hv_kvp_daemon.c:1305:2: note: 'snprintf' output between 41 and 4136 bytes into a destination of size 4096

Cc K. Y. Srinivasan <[email protected]>
Cc Haiyang Zhang <[email protected]>
Cc Stephen Hemminger <[email protected]>
---
 tools/hv/hv_kvp_daemon.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/hv/hv_kvp_daemon.c +1 -1
index bbb2a8ef367c..b7c2cf7eaba5 100644
@@ -1176,7 +1176,7 @@ static int kvp_set_ip_info(char *if_name, struct hv_kvp_ipaddr_value *new_val)
 	int error = 0;
 	char if_file[PATH_MAX];
 	FILE *file;
-	char cmd[PATH_MAX];
+	char cmd[PATH_MAX * 2];
 	char *mac_addr;
 
 	/*
Signature · 1 line
2.18.0
maintainer
#17 lore raw

Re: [PATCH 5/5] Tools: hv: kvp: Fix a warning of buffer overflow with gcc 8.0.1

Quoted 23 lines · On Wed, Oct 17, 2018 at 03:14:06AM +0000, [email protected] wrote:

From: Dexuan Cui <[email protected]>

The patch fixes:

hv_kvp_daemon.c: In function 'kvp_set_ip_info':
hv_kvp_daemon.c:1305:2: note: 'snprintf' output between 41 and 4136 bytes into a destination of size 4096

Cc K. Y. Srinivasan <[email protected]>
Cc Haiyang Zhang <[email protected]>
Cc Stephen Hemminger <[email protected]>
---
 tools/hv/hv_kvp_daemon.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/hv/hv_kvp_daemon.c +0 -0
index bbb2a8ef367c..b7c2cf7eaba5 100644

tools/hv/hv_kvp_daemon.c static int kvp_set_ip_info(char *if_name, struct hv_kvp_ipaddr_value *new_val) · line 1179

Quoted 5 lines
tools/hv/hv_kvp_daemon.c +1 -1
@@ -1176,7 +1176,7 @@ static int kvp_set_ip_info(char *if_name, struct hv_kvp_ipaddr_value *new_val)
 	int error = 0;
 	char if_file[PATH_MAX];
 	FILE *file;
-	char cmd[PATH_MAX];
+	char cmd[PATH_MAX * 2];

Overkill?

maintainer
#18 lore raw

RE: [PATCH 5/5] Tools: hv: kvp: Fix a warning of buffer overflow with gcc 8.0.1

Quoted 17 lines

From: devel <[email protected]> On Behalf Of Greg KH
Sent: Tuesday, October 16, 2018 22:07

tools/hv/hv_kvp_daemon.c static int kvp_set_ip_info(char *if_name, struct

Quoted 11 lines · On Wed, Oct 17, 2018 at 03:14:06AM +0000, [email protected] wrote:

From: Dexuan Cui <[email protected]>

The patch fixes:

hv_kvp_daemon.c: In function 'kvp_set_ip_info':
hv_kvp_daemon.c:1305:2: note: 'snprintf' output between 41 and 4136 bytes into a destination of size 4096
...

tools/hv/hv_kvp_daemon.c +0 -0
@@ -1176,7 +1176,7 @@ static int kvp_set_ip_info(char *if_name, struct

hv_kvp_ipaddr_value *new_val)

Quoted 7 lines

tools/hv/hv_kvp_daemon.c static int kvp_set_ip_info(char *if_name, struct hv_kvp_ipaddr_value *new_val) · line 1179

Quoted 5 lines
tools/hv/hv_kvp_daemon.c +1 -1
 	int error = 0;
 	char if_file[PATH_MAX];
 	FILE *file;
-	char cmd[PATH_MAX];
+	char cmd[PATH_MAX * 2];

Overkill?

Well, I found the -Wformat-truncation warning can be suppressed by checking the return value of snprintf, so I'm going to do a new patch for KY to resend like this:

@@ -1178,6 +1178,7 @@ static int kvp_set_ip_info(char *if_name, struct hv_kvp_ipaddr_value *new_val)

        FILE *file;
        char cmd[PATH_MAX];
        char *mac_addr;

+ int str_len;

/*
* Set the configuration for the specified interface with
@@ -1301,8 +1302,19 @@ static int kvp_set_ip_info(char *if_name, struct hv_kvp_ipaddr_value *new_val)
* invoke the external script to do its magic.
*/

- snprintf(cmd, sizeof(cmd), KVP_SCRIPTS_PATH "%s %s",
- "hv_set_ifconfig", if_file);
+ str_len = snprintf(cmd, sizeof(cmd), KVP_SCRIPTS_PATH "%s %s",
+ "hv_set_ifconfig", if_file);
+
+ /*

+        * This is a little overcautious, but it's necessary to suppress some
+        * false warnings from gcc 8.0.1.
+        */

+ if (str_len <= 0 || (unsigned int)str_len >= sizeof(cmd)) {
+ syslog(LOG_ERR, "Cmd '%s' (len=%d) may be too long",
+ cmd, str_len);
+ return HV_E_FAIL;
+ }

Thanks,
-- Dexuan

maintainer
#19 lore raw

Re: [PATCH 1/5] Drivers: hv: vmbus: Get rid of unnecessary state in hv_context

In reply to [email protected]
Quoted 8 lines · On Wed, Oct 17, 2018 at 03:14:02AM +0000, [email protected] wrote:

From: "K. Y. Srinivasan" <[email protected]>

Currently we are replicating state in struct hv_context that is unnecessary - this state can be retrieved from the hypervisor. Furthermore, this is a per-cpu state that is being maintained as a global state in struct hv_context. Get rid of this state in struct hv_context.

Reply-To: [email protected]

Why is this here?

greg k-h

#20 lore raw

RE: [PATCH 1/5] Drivers: hv: vmbus: Get rid of unnecessary state in hv_context

Quoted 18 lines

-----Original Message-----
From: Greg KH <[email protected]>
Sent: Tuesday, October 16, 2018 10:05 PM
To: KY Srinivasan <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Stephen Hemminger <[email protected]>; Michael Kelley (EOSG) <[email protected]>; vkuznets <[email protected]>
Subject: Re: [PATCH 1/5] Drivers: hv: vmbus: Get rid of unnecessary state in hv_context

Quoted 4 lines · On Wed, Oct 17, 2018 at 03:14:02AM +0000, [email protected] wrote:

From: "K. Y. Srinivasan" <[email protected]>

Currently we are replicating state in struct hv_context that is unnecessary - this state can be retrieved from the hypervisor. Furthermore, this is a per-

Quoted 7 lines

cpu

Quoted 4 lines

state that is being maintained as a global state in struct hv_context. Get rid of this state in struct hv_context.

Reply-To: [email protected]

Why is this here?

No sure how this happened; will fix it and resend.

K. Y

Quoted 2 lines

greg k-h