Skip to contentSkip to conversation

virtchnl: fix fake 1-elem arrays

PATCH In review Alexander Lobakin 4 participants 20 messages Jul 28, 2023 – Aug 16, 2023

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

cover virtchnl: fix fake 1-elem arrays

6.5-rc1 started spitting warning splats when composing virtchnl messages, precisely on virtchnl_rss_key and virtchnl_lut:

[   84.167709] memcpy: detected field-spanning write (size 52) of single
field "vrk->key" at drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1095
(size 1)
[   84.169915] WARNING: CPU: 3 PID: 11 at drivers/net/ethernet/intel/
iavf/iavf_virtchnl.c:1095 iavf_set_rss_key+0x123/0x140 [iavf]
...
[   84.191982] Call Trace:
[   84.192439]  <TASK>
[   84.192900]  ? __warn+0xc9/0x1a0
[   84.193353]  ? iavf_set_rss_key+0x123/0x140 [iavf]
[   84.193818]  ? report_bug+0x12c/0x1b0
[   84.194266]  ? handle_bug+0x42/0x70
[   84.194714]  ? exc_invalid_op+0x1a/0x50
[   84.195149]  ? asm_exc_invalid_op+0x1a/0x20
[   84.195592]  ? iavf_set_rss_key+0x123/0x140 [iavf]
[   84.196033]  iavf_watchdog_task+0xb0c/0xe00 [iavf]
...
[   84.225476] memcpy: detected field-spanning write (size 64) of single
field "vrl->lut" at drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1127
(size 1)
[   84.227190] WARNING: CPU: 27 PID: 1044 at drivers/net/ethernet/intel/
iavf/iavf_virtchnl.c:1127 iavf_set_rss_lut+0x123/0x140 [iavf]
...
[   84.246601] Call Trace:
[   84.247228]  <TASK>
[   84.247840]  ? __warn+0xc9/0x1a0
[   84.248263]  ? iavf_set_rss_lut+0x123/0x140 [iavf]
[   84.248698]  ? report_bug+0x12c/0x1b0
[   84.249122]  ? handle_bug+0x42/0x70
[   84.249549]  ? exc_invalid_op+0x1a/0x50
[   84.249970]  ? asm_exc_invalid_op+0x1a/0x20
[   84.250390]  ? iavf_set_rss_lut+0x123/0x140 [iavf]
[   84.250820]  iavf_watchdog_task+0xb16/0xe00 [iavf]

Gustavo already tried to fix those back in 2021[0][1]. Unfortunately, a VM can run a different kernel than the host, meaning that those structures are sorta ABI.
However, it is possible to have proper flex arrays + struct_size() calculations and still send the very same messages with the same sizes. The common rule is:

elem[1] -> elem[]
size = struct_size() + <difference between the old and the new msg size>

The "old" size in the current code is calculated 3 different ways for 10 virtchnl structures total. Each commit addresses one of the ways cumulatively instead of per-structure.

I was planning to send it to -net initially, but given that virtchnl was renamed from i40evf and got some fat style cleanup commits in the past, it's not very straightforward to even pick appropriate SHAs, not speaking of automatic portability. I may send manual backports for a couple of the latest supported kernels later on if anyone needs it at all.

[0] https://lore.kernel.org/all/20210525230912.GA175802@embeddedor
[1] https://lore.kernel.org/all/20210525231851.GA176647@embeddedor

Alexander Lobakin (3):
virtchnl: fix fake 1-elem arrays in structs allocated as nents + 1 -

    1
  virtchnl: fix fake 1-elem arrays in structures allocated as `nents +
    1`

virtchnl: fix fake 1-elem arrays for structures allocated as nents

 .../ethernet/intel/i40e/i40e_virtchnl_pf.c    |   9 +-
 drivers/net/ethernet/intel/iavf/iavf.h        |   6 +-
 drivers/net/ethernet/intel/iavf/iavf_client.c |   4 +-
 drivers/net/ethernet/intel/iavf/iavf_client.h |   2 +-
 .../net/ethernet/intel/iavf/iavf_virtchnl.c   |  75 +++++------
 drivers/net/ethernet/intel/ice/ice_virtchnl.c |   2 +-
 include/linux/avf/virtchnl.h                  | 127 +++++++++++-------
 7 files changed, 124 insertions(+), 101 deletions(-)
Signature · 1 line
2.41.0
#1 lore raw

PATCH 1/3 virtchnl: fix fake 1-elem arrays in structs allocated as `nents + 1` - 1

The two most problematic virtchnl structures are virtchnl_rss_key and virtchnl_rss_lut. Their "flex" arrays have the type of u8, thus, when allocating / checking, the actual size is calculated as sizeof + nents - 1 byte. But their sizeof() is not 1 byte larger than the size of such structure with proper flex array, it's two bytes larger due to the padding. That said, their size is always 1 byte larger unless there are no tail elements -- then it's +2 bytes.
Add virtchnl_struct_size() macro which will handle this case (and later other cases as well). Make its calling conv the same as we call struct_size() to allow it to be drop-in, even though it's unlikely to become possible to switch to generic API. The macro will calculate a proper size of a structure with a flex array at the end, so that it becomes transparent for the compilers, but add the difference from the old values, so that the real size of sorta-ABI-messages doesn't change. Use it on the allocation side in IAVF and the receiving side (defined as static inline in virtchnl.h) for the mentioned two structures.

---
 .../net/ethernet/intel/iavf/iavf_virtchnl.c   |  6 ++--
 include/linux/avf/virtchnl.h                  | 31 ++++++++++++++-----
 2 files changed, 25 insertions(+), 12 deletions(-)
drivers/net/ethernet/intel/iavf/iavf_virtchnl.c +2 -4
index be3c007ce90a..10f03054a603 100644
@@ -1085,8 +1085,7 @@ void iavf_set_rss_key(struct iavf_adapter *adapter)
 			adapter->current_op);
 		return;
 	}
-	len = sizeof(struct virtchnl_rss_key) +
-	      (adapter->rss_key_size * sizeof(u8)) - 1;
+	len = virtchnl_struct_size(vrk, key, adapter->rss_key_size);
 	vrk = kzalloc(len, GFP_KERNEL);
 	if (!vrk)
 		return;
@@ -1117,8 +1116,7 @@ void iavf_set_rss_lut(struct iavf_adapter *adapter)
 			adapter->current_op);
 		return;
 	}
-	len = sizeof(struct virtchnl_rss_lut) +
-	      (adapter->rss_lut_size * sizeof(u8)) - 1;
+	len = virtchnl_struct_size(vrl, lut, adapter->rss_lut_size);
 	vrl = kzalloc(len, GFP_KERNEL);
 	if (!vrl)
 		return;
include/linux/avf/virtchnl.h +23 -8
index c15221dcb75e..3ab207c14809 100644
@@ -866,18 +866,20 @@ VIRTCHNL_CHECK_STRUCT_LEN(4, virtchnl_promisc_info);
 struct virtchnl_rss_key {
 	u16 vsi_id;
 	u16 key_len;
-	u8 key[1];         /* RSS hash key, packed bytes */
+	u8 key[];          /* RSS hash key, packed bytes */
 };
 
-VIRTCHNL_CHECK_STRUCT_LEN(6, virtchnl_rss_key);
+VIRTCHNL_CHECK_STRUCT_LEN(4, virtchnl_rss_key);
+#define virtchnl_rss_key_LEGACY_SIZEOF	6
 
 struct virtchnl_rss_lut {
 	u16 vsi_id;
 	u16 lut_entries;
-	u8 lut[1];        /* RSS lookup table */
+	u8 lut[];         /* RSS lookup table */
 };
 
-VIRTCHNL_CHECK_STRUCT_LEN(6, virtchnl_rss_lut);
+VIRTCHNL_CHECK_STRUCT_LEN(4, virtchnl_rss_lut);
+#define virtchnl_rss_lut_LEGACY_SIZEOF	6
 
 /* VIRTCHNL_OP_GET_RSS_HENA_CAPS
  * VIRTCHNL_OP_SET_RSS_HENA
@@ -1367,6 +1369,17 @@ struct virtchnl_fdir_del {
 
 VIRTCHNL_CHECK_STRUCT_LEN(12, virtchnl_fdir_del);
 
+#define __vss_byone(p, member, count, old)				      \
+	(struct_size(p, member, count) + (old - 1 - struct_size(p, member, 0)))
+
+#define __vss(type, func, p, member, count)		\
+	struct type: func(p, member, count, type##_LEGACY_SIZEOF)
+
+#define virtchnl_struct_size(p, m, c)					      \
+	_Generic(*p,							      \
+		 __vss(virtchnl_rss_key, __vss_byone, p, m, c),		      \
+		 __vss(virtchnl_rss_lut, __vss_byone, p, m, c))
+
 /**
  * virtchnl_vc_validate_vf_msg
  * @ver: Virtchnl version info
@@ -1479,19 +1492,21 @@ virtchnl_vc_validate_vf_msg(struct virtchnl_version_info *ver, u32 v_opcode,
 		}
 		break;
 	case VIRTCHNL_OP_CONFIG_RSS_KEY:
-		valid_len = sizeof(struct virtchnl_rss_key);
+		valid_len = virtchnl_rss_key_LEGACY_SIZEOF;
 		if (msglen >= valid_len) {
 			struct virtchnl_rss_key *vrk =
 				(struct virtchnl_rss_key *)msg;
-			valid_len += vrk->key_len - 1;
+			valid_len = virtchnl_struct_size(vrk, key,
+							 vrk->key_len);
 		}
 		break;
 	case VIRTCHNL_OP_CONFIG_RSS_LUT:
-		valid_len = sizeof(struct virtchnl_rss_lut);
+		valid_len = virtchnl_rss_lut_LEGACY_SIZEOF;
 		if (msglen >= valid_len) {
 			struct virtchnl_rss_lut *vrl =
 				(struct virtchnl_rss_lut *)msg;
-			valid_len += vrl->lut_entries - 1;
+			valid_len = virtchnl_struct_size(vrl, lut,
+							 vrl->lut_entries);
 		}
 		break;
 	case VIRTCHNL_OP_GET_RSS_HENA_CAPS:
Signature · 1 line
2.41.0
#2 lore raw

Re: [PATCH net-next 1/3] virtchnl: fix fake 1-elem arrays in structs allocated as `nents + 1` - 1

Quoted 8 lines · Alexander Lobakin

The two most problematic virtchnl structures are virtchnl_rss_key and virtchnl_rss_lut. Their "flex" arrays have the type of u8, thus, when allocating / checking, the actual size is calculated as sizeof + nents - 1 byte. But their sizeof() is not 1 byte larger than the size of such structure with proper flex array, it's two bytes larger due to the padding. That said, their size is always 1 byte larger unless there are no tail elements -- then it's +2 bytes.
Add virtchnl_struct_size() macro which will handle this case (and later

Quoted 8 lines

other cases as well). Make its calling conv the same as we call struct_size() to allow it to be drop-in, even though it's unlikely to become possible to switch to generic API. The macro will calculate a proper size of a structure with a flex array at the end, so that it becomes transparent for the compilers, but add the difference from the old values, so that the real size of sorta-ABI-messages doesn't change. Use it on the allocation side in IAVF and the receiving side (defined as static inline in virtchnl.h) for the mentioned two structures.

This all looks workable, but it's a unique solution in the kernel. That is fine, of course, but would it be easier to maintain/test if it went with the union style solutions?

struct foo {
	...
	union {
		type legacy_padding;
		DECLARE_FLEX_ARRAY(type, member);
	};
};

Then the size doesn't change and "member" can still be used. (i.e. no collateral changes needed.)

-Kees

Signature · 5 lines
Kees Cook
_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan
#3 lore raw

Re: [PATCH net-next 1/3] virtchnl: fix fake 1-elem arrays in structs allocated as `nents + 1` - 1

From: Kees Cook <[email protected]>
Date: Fri, 28 Jul 2023 15:43:26 -0700

Quoted 30 lines · Alexander Lobakin
Quoted 8 lines

The two most problematic virtchnl structures are virtchnl_rss_key and virtchnl_rss_lut. Their "flex" arrays have the type of u8, thus, when allocating / checking, the actual size is calculated as sizeof + nents - 1 byte. But their sizeof() is not 1 byte larger than the size of such structure with proper flex array, it's two bytes larger due to the padding. That said, their size is always 1 byte larger unless there are no tail elements -- then it's +2 bytes.
Add virtchnl_struct_size() macro which will handle this case (and later

Quoted 8 lines

other cases as well). Make its calling conv the same as we call struct_size() to allow it to be drop-in, even though it's unlikely to become possible to switch to generic API. The macro will calculate a proper size of a structure with a flex array at the end, so that it becomes transparent for the compilers, but add the difference from the old values, so that the real size of sorta-ABI-messages doesn't change. Use it on the allocation side in IAVF and the receiving side (defined as static inline in virtchnl.h) for the mentioned two structures.

This all looks workable, but it's a unique solution in the kernel. That is fine, of course, but would it be easier to maintain/test if it went with the union style solutions?

struct foo {
	...
	union {
		type legacy_padding;
		DECLARE_FLEX_ARRAY(type, member);
	};
};
Quoted 2 lines

Then the size doesn't change and "member" can still be used. (i.e. no collateral changes needed.)

This wouldn't work unfortunately. I mean, you'd still need weird calculations.
Speaking of e.g. virtchnl_rss_lut, it's always struct_size(nents + 1) - 1, you can't just use the pattern above and then struct_size(). Not only legacy padding is needed, but also calculation adjustments.
Or did you mean define the structures as above and leave the calculations as they are? It makes me feel we can miss something that way. With the series, all the broken structures are processed in one place and I thought _this_ would be actually easier to maintain...

Quoted 3 lines

-Kees

Thanks,
Olek

Reviewed-by on v1 1/3
#4 lore raw

Re: [PATCH net-next 1/3] virtchnl: fix fake 1-elem arrays in structs allocated as `nents + 1` - 1

In reply to Alexander Lobakin
Quoted 17 lines · Alexander Lobakin

The two most problematic virtchnl structures are virtchnl_rss_key and virtchnl_rss_lut. Their "flex" arrays have the type of u8, thus, when allocating / checking, the actual size is calculated as sizeof + nents - 1 byte. But their sizeof() is not 1 byte larger than the size of such structure with proper flex array, it's two bytes larger due to the padding. That said, their size is always 1 byte larger unless there are no tail elements -- then it's +2 bytes.
Add virtchnl_struct_size() macro which will handle this case (and later other cases as well). Make its calling conv the same as we call struct_size() to allow it to be drop-in, even though it's unlikely to become possible to switch to generic API. The macro will calculate a proper size of a structure with a flex array at the end, so that it becomes transparent for the compilers, but add the difference from the old values, so that the real size of sorta-ABI-messages doesn't change. Use it on the allocation side in IAVF and the receiving side (defined as static inline in virtchnl.h) for the mentioned two structures.

Quoted 1 line

This is a novel approach to solving the ABI issues for a 1-elem conversion, but I have been convinced it's a workable approach here. :) Thanks for doing this conversion!

Signature · 5 lines
Kees Cook
_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan
#5 lore raw

Re: [PATCH net-next 1/3] virtchnl: fix fake 1-elem arrays in structs allocated as `nents + 1` - 1

From: Kees Cook <[email protected]>
Date: Fri, 4 Aug 2023 01:27:02 -0700

Quoted 18 lines · Alexander Lobakin
Quoted 17 lines

The two most problematic virtchnl structures are virtchnl_rss_key and virtchnl_rss_lut. Their "flex" arrays have the type of u8, thus, when allocating / checking, the actual size is calculated as sizeof + nents - 1 byte. But their sizeof() is not 1 byte larger than the size of such structure with proper flex array, it's two bytes larger due to the padding. That said, their size is always 1 byte larger unless there are no tail elements -- then it's +2 bytes.
Add virtchnl_struct_size() macro which will handle this case (and later other cases as well). Make its calling conv the same as we call struct_size() to allow it to be drop-in, even though it's unlikely to become possible to switch to generic API. The macro will calculate a proper size of a structure with a flex array at the end, so that it becomes transparent for the compilers, but add the difference from the old values, so that the real size of sorta-ABI-messages doesn't change. Use it on the allocation side in IAVF and the receiving side (defined as static inline in virtchnl.h) for the mentioned two structures.

Quoted 8 lines
Quoted 1 line

This is a novel approach to solving the ABI issues for a 1-elem conversion, but I have been convinced it's a workable approach here. :) Thanks for doing this conversion!

Thanks a lot!
You gave Reviewed-by for patches #1 and #3, does it mean the whole series or something is wrong with the patch #2? :D

Thanks,
Olek

#6 lore raw

Re: [PATCH net-next 1/3] virtchnl: fix fake 1-elem arrays in structs allocated as `nents + 1` - 1

Quoted 27 lines · Alexander Lobakin

From: Kees Cook <[email protected]>
Date: Fri, 4 Aug 2023 01:27:02 -0700

Quoted 24 lines · Alexander Lobakin
Quoted 17 lines

The two most problematic virtchnl structures are virtchnl_rss_key and virtchnl_rss_lut. Their "flex" arrays have the type of u8, thus, when allocating / checking, the actual size is calculated as sizeof + nents - 1 byte. But their sizeof() is not 1 byte larger than the size of such structure with proper flex array, it's two bytes larger due to the padding. That said, their size is always 1 byte larger unless there are no tail elements -- then it's +2 bytes.
Add virtchnl_struct_size() macro which will handle this case (and later other cases as well). Make its calling conv the same as we call struct_size() to allow it to be drop-in, even though it's unlikely to become possible to switch to generic API. The macro will calculate a proper size of a structure with a flex array at the end, so that it becomes transparent for the compilers, but add the difference from the old values, so that the real size of sorta-ABI-messages doesn't change. Use it on the allocation side in IAVF and the receiving side (defined as static inline in virtchnl.h) for the mentioned two structures.

Quoted 1 line

This is a novel approach to solving the ABI issues for a 1-elem conversion, but I have been convinced it's a workable approach here. :) Thanks for doing this conversion!

Quoted 6 lines
Quoted 2 lines

Thanks a lot!
You gave Reviewed-by for patches #1 and #3, does it mean the whole series or something is wrong with the patch #2? :D

Hm, maybe delivery was delayed? I see it on lore:
https://lore.kernel.org/all/202308040128.667940394B@keescook/

Signature · 1 line
Kees Cook
#7 lore raw

Re: [PATCH net-next 1/3] virtchnl: fix fake 1-elem arrays in structs allocated as `nents + 1` - 1

From: Kees Cook <[email protected]>
Date: Fri, 4 Aug 2023 10:29:48 -0700

Quoted 31 lines · Alexander Lobakin
Quoted 30 lines

From: Kees Cook <[email protected]>
Date: Fri, 4 Aug 2023 01:27:02 -0700

Quoted 26 lines · Alexander Lobakin
Quoted 17 lines

The two most problematic virtchnl structures are virtchnl_rss_key and virtchnl_rss_lut. Their "flex" arrays have the type of u8, thus, when allocating / checking, the actual size is calculated as sizeof + nents - 1 byte. But their sizeof() is not 1 byte larger than the size of such structure with proper flex array, it's two bytes larger due to the padding. That said, their size is always 1 byte larger unless there are no tail elements -- then it's +2 bytes.
Add virtchnl_struct_size() macro which will handle this case (and later other cases as well). Make its calling conv the same as we call struct_size() to allow it to be drop-in, even though it's unlikely to become possible to switch to generic API. The macro will calculate a proper size of a structure with a flex array at the end, so that it becomes transparent for the compilers, but add the difference from the old values, so that the real size of sorta-ABI-messages doesn't change. Use it on the allocation side in IAVF and the receiving side (defined as static inline in virtchnl.h) for the mentioned two structures.

Quoted 1 line

This is a novel approach to solving the ABI issues for a 1-elem conversion, but I have been convinced it's a workable approach here. :) Thanks for doing this conversion!

Quoted 7 lines
Quoted 3 lines

Thanks a lot!
You gave Reviewed-by for patches #1 and #3, does it mean the whole series or something is wrong with the patch #2? :D

Hm, maybe delivery was delayed? I see it on lore:
https://lore.kernel.org/all/202308040128.667940394B@keescook/

Nevermind, my mail rules did put it in the folder other than the one where the main thread was, sorry :s
Much thanks, I'm now a fan of _Generic() too :D

Olek

#8 lore raw

RE: [Intel-wired-lan] [PATCH net-next 1/3] virtchnl: fix fake 1-elem arrays in structs allocated as `nents + 1` - 1

Quoted 62 lines

-----Original Message-----
From: Intel-wired-lan <[email protected]> On Behalf Of Alexander Lobakin
Sent: piątek, 4 sierpnia 2023 19:34
To: Kees Cook <[email protected]>
Cc: Andy Shevchenko <[email protected]>; Zaremba, Larysa <[email protected]>; [email protected]; Gustavo A. R. Silva <[email protected]>; [email protected]; Eric Dumazet <[email protected]>; [email protected]; linux- [email protected]; Jakub Kicinski <[email protected]>; Paolo Abeni <[email protected]>; David S. Miller <[email protected]>
Subject: Re: [Intel-wired-lan] [PATCH net-next 1/3] virtchnl: fix fake 1-elem arrays in structs allocated as nents + 1 - 1

From: Kees Cook <[email protected]>
Date: Fri, 4 Aug 2023 10:29:48 -0700

Quoted 5 lines · Alexander Lobakin
Quoted 4 lines

From: Kees Cook <[email protected]>
Date: Fri, 4 Aug 2023 01:27:02 -0700

Quoted 1 line · Alexander Lobakin
Quoted 15 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

The two most problematic virtchnl structures are virtchnl_rss_key and virtchnl_rss_lut. Their "flex" arrays have the type of u8, thus, when allocating / checking, the actual size is calculated as sizeof + nents - 1 byte. But their sizeof() is not 1 byte larger than the size of such structure with proper flex array, it's two bytes larger due to the padding. That said, their size is always 1 byte larger unless there are no tail elements -- then it's +2 bytes.

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

Add virtchnl_struct_size() macro which will handle this case (and later other cases as well). Make its calling conv the same as we call
struct_size() to allow it to be drop-in, even though it's unlikely to become possible to switch to generic API. The macro will calculate a proper size of a structure with a flex array at the end, so that it becomes transparent for the compilers, but add the difference from the old values, so that the real size of sorta-ABI-

messages doesn't change.

Quoted 12 lines
Quoted 12 lines
Quoted 11 lines
Quoted 4 lines

Use it on the allocation side in IAVF and the receiving side (defined as static inline in virtchnl.h) for the mentioned two structures.

This is a novel approach to solving the ABI issues for a 1-elem conversion, but I have been convinced it's a workable approach here. :) Thanks for doing this conversion!

Quoted 7 lines
Quoted 3 lines

Thanks a lot!
You gave Reviewed-by for patches #1 and #3, does it mean the whole series or something is wrong with the patch #2? :D

Hm, maybe delivery was delayed? I see it on lore:
https://lore.kernel.org/all/202308040128.667940394B@keescook/

Nevermind, my mail rules did put it in the folder other than the one where the main thread was, sorry :s Much thanks, I'm now a fan of _Generic() too :D

Quoted 5 lines

Olek
_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan

#9 lore raw

PATCH 2/3 virtchnl: fix fake 1-elem arrays in structures allocated as `nents + 1`

In reply to Alexander Lobakin

There are five virtchnl structures, which are allocated and checked in the code as nents + 1, meaning that they always have memory for one excessive element regardless of their actual number. This comes from that their sizeof() includes space for 1 element and then they get allocated via struct_size() or its open-coded equivalents, passing the actual number of elements.
Expand virtchnl_struct_size() to handle such structures and replace those 1-elem arrays with proper flex ones. Also fix several places which open-code %IAVF_VIRTCHNL_VF_RESOURCE_SIZE. Finally, let the virtchnl_ether_addr_list size be computed automatically when there's no enough space for the whole list, otherwise we have to open-code reverse struct_size() logics.

---
 .../ethernet/intel/i40e/i40e_virtchnl_pf.c    |  2 +-
 drivers/net/ethernet/intel/iavf/iavf.h        |  6 +-
 .../net/ethernet/intel/iavf/iavf_virtchnl.c   | 44 +++++++-------
 drivers/net/ethernet/intel/ice/ice_virtchnl.c |  2 +-
 include/linux/avf/virtchnl.h                  | 57 ++++++++++++-------
 5 files changed, 59 insertions(+), 52 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c +1 -1
index 398fb4854cbe..030738409f76 100644
@@ -2103,7 +2103,7 @@ static int i40e_vc_get_vf_resources_msg(struct i40e_vf *vf, u8 *msg)
 		goto err;
 	}
 
-	len = struct_size(vfres, vsi_res, num_vsis);
+	len = virtchnl_struct_size(vfres, vsi_res, num_vsis);
 	vfres = kzalloc(len, GFP_KERNEL);
 	if (!vfres) {
 		aq_ret = I40E_ERR_NO_MEMORY;
drivers/net/ethernet/intel/iavf/iavf.h +3 -3
index 8cbdebc5b698..85fba85fbb23 100644
@@ -92,9 +92,9 @@ struct iavf_vsi {
 #define IAVF_MBPS_DIVISOR	125000 /* divisor to convert to Mbps */
 #define IAVF_MBPS_QUANTA	50
 
-#define IAVF_VIRTCHNL_VF_RESOURCE_SIZE (sizeof(struct virtchnl_vf_resource) + \
-					(IAVF_MAX_VF_VSI * \
-					 sizeof(struct virtchnl_vsi_resource)))
+#define IAVF_VIRTCHNL_VF_RESOURCE_SIZE					\
+	virtchnl_struct_size((struct virtchnl_vf_resource *)NULL,	\
+			     vsi_res, IAVF_MAX_VF_VSI)
 
 /* MAX_MSIX_Q_VECTORS of these are allocated,
  * but we only use one per queue-specific vector.
drivers/net/ethernet/intel/iavf/iavf_virtchnl.c +19 -25

100 lines not on this page · raw message

drivers/net/ethernet/intel/ice/ice_virtchnl.c +1 -1
index 625da88e7965..a5a82833cdc6 100644
@@ -428,7 +428,7 @@ static int ice_vc_get_vf_res_msg(struct ice_vf *vf, u8 *msg)
 		goto err;
 	}
 
-	len = sizeof(struct virtchnl_vf_resource);
+	len = virtchnl_struct_size(vfres, vsi_res, 0);
 
 	vfres = kzalloc(len, GFP_KERNEL);
 	if (!vfres) {
include/linux/avf/virtchnl.h +35 -22

143 lines not on this page · raw message

Signature · 1 line
2.41.0
Reviewed-by on v1 2/3
#10 lore raw

Re: [PATCH net-next 2/3] virtchnl: fix fake 1-elem arrays in structures allocated as `nents + 1`

Quoted 13 lines · Alexander Lobakin

There are five virtchnl structures, which are allocated and checked in the code as nents + 1, meaning that they always have memory for one excessive element regardless of their actual number. This comes from that their sizeof() includes space for 1 element and then they get allocated via struct_size() or its open-coded equivalents, passing the actual number of elements.
Expand virtchnl_struct_size() to handle such structures and replace those 1-elem arrays with proper flex ones. Also fix several places which open-code %IAVF_VIRTCHNL_VF_RESOURCE_SIZE. Finally, let the virtchnl_ether_addr_list size be computed automatically when there's no enough space for the whole list, otherwise we have to open-code reverse struct_size() logics.

Quoted 1 line

I remain a fan of _Generic uses. :)

Signature · 1 line
Kees Cook
#11 lore raw

RE: [Intel-wired-lan] [PATCH net-next 2/3] virtchnl: fix fake 1-elem arrays in structures allocated as `nents + 1`

Quoted 32 lines

-----Original Message-----
From: Intel-wired-lan <[email protected]> On Behalf Of Kees Cook
Sent: piątek, 4 sierpnia 2023 10:29
To: Lobakin, Aleksander <[email protected]>
Cc: Andy Shevchenko <[email protected]>; Zaremba, Larysa <[email protected]>; [email protected]; Gustavo A. R. Silva <[email protected]>; [email protected]; Eric Dumazet <[email protected]>; [email protected]; linux- [email protected]; Jakub Kicinski <[email protected]>; Paolo Abeni <[email protected]>; David S. Miller <[email protected]>
Subject: Re: [Intel-wired-lan] [PATCH net-next 2/3] virtchnl: fix fake 1-elem arrays in structures allocated as nents + 1

Quoted 13 lines · Alexander Lobakin

There are five virtchnl structures, which are allocated and checked in the code as nents + 1, meaning that they always have memory for one excessive element regardless of their actual number. This comes from that their sizeof() includes space for 1 element and then they get allocated via struct_size() or its open-coded equivalents, passing the actual number of elements.
Expand virtchnl_struct_size() to handle such structures and replace those 1-elem arrays with proper flex ones. Also fix several places which open-code %IAVF_VIRTCHNL_VF_RESOURCE_SIZE. Finally, let the virtchnl_ether_addr_list size be computed automatically when there's no enough space for the whole list, otherwise we have to open-code reverse struct_size() logics.

Quoted 1 line

I remain a fan of _Generic uses. :)

Quoted 8 lines

--
Kees Cook
_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan

_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan

#12 lore raw

PATCH 3/3 virtchnl: fix fake 1-elem arrays for structures allocated as `nents`

In reply to Alexander Lobakin

Finally, fix 3 structures which are allocated technically correctly, i.e. the calculated size equals to the one that struct_size() would return, except for sizeof(). For &virtchnl_vlan_filter_list_v2, use the same approach when there are no enough space as taken previously for &virtchnl_vlan_filter_list, i.e. let the maximum size be calculated automatically instead of trying to guestimate it using maths.

---
 .../ethernet/intel/i40e/i40e_virtchnl_pf.c    |  7 ++--
 drivers/net/ethernet/intel/iavf/iavf_client.c |  4 +-
 drivers/net/ethernet/intel/iavf/iavf_client.h |  2 +-
 .../net/ethernet/intel/iavf/iavf_virtchnl.c   | 25 +++++-------
 include/linux/avf/virtchnl.h                  | 39 ++++++++++++-------
 5 files changed, 40 insertions(+), 37 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c +4 -3
index 030738409f76..90a76e927fab 100644
@@ -506,6 +506,7 @@ i40e_config_rdma_qvlist(struct i40e_vf *vf,
 	struct virtchnl_rdma_qv_info *qv_info;
 	u32 v_idx, i, reg_idx, reg;
 	u32 next_q_idx, next_q_type;
+	size_t size;
 	u32 msix_vf;
 	int ret = 0;
 
@@ -521,9 +522,9 @@ i40e_config_rdma_qvlist(struct i40e_vf *vf,
 	}
 
 	kfree(vf->qvlist_info);
-	vf->qvlist_info = kzalloc(struct_size(vf->qvlist_info, qv_info,
-					      qvlist_info->num_vectors - 1),
-				  GFP_KERNEL);
+	size = virtchnl_struct_size(vf->qvlist_info, qv_info,
+				    qvlist_info->num_vectors);
+	vf->qvlist_info = kzalloc(size, GFP_KERNEL);
 	if (!vf->qvlist_info) {
 		ret = -ENOMEM;
 		goto err_out;
drivers/net/ethernet/intel/iavf/iavf_client.c +2 -2
index 93c903c02c64..e6051b6355aa 100644
@@ -469,8 +469,8 @@ static int iavf_client_setup_qvlist(struct iavf_info *ldev,
 	}
 
 	v_qvlist_info = (struct virtchnl_rdma_qvlist_info *)qvlist_info;
-	msg_size = struct_size(v_qvlist_info, qv_info,
-			       v_qvlist_info->num_vectors - 1);
+	msg_size = virtchnl_struct_size(v_qvlist_info, qv_info,
+					v_qvlist_info->num_vectors);
 
 	adapter->client_pending |= BIT(VIRTCHNL_OP_CONFIG_RDMA_IRQ_MAP);
 	err = iavf_aq_send_msg_to_pf(&adapter->hw,
drivers/net/ethernet/intel/iavf/iavf_client.h +1 -1
index c5d51d7dc7cc..500269bc0f5b 100644
@@ -53,7 +53,7 @@ struct iavf_qv_info {
 
 struct iavf_qvlist_info {
 	u32 num_vectors;
-	struct iavf_qv_info qv_info[1];
+	struct iavf_qv_info qv_info[];
 };
 
 #define IAVF_CLIENT_MSIX_ALL 0xFFFFFFFF
drivers/net/ethernet/intel/iavf/iavf_virtchnl.c +9 -16
index 4fdac698eb38..f9727e9c3d63 100644
@@ -727,15 +727,12 @@ void iavf_add_vlans(struct iavf_adapter *adapter)
 			more = true;
 		}
 
-		len = sizeof(*vvfl_v2) + ((count - 1) *
-					  sizeof(struct virtchnl_vlan_filter));
+		len = virtchnl_struct_size(vvfl_v2, filters, count);
 		if (len > IAVF_MAX_AQ_BUF_SIZE) {
 			dev_warn(&adapter->pdev->dev, "Too many add VLAN changes in one request\n");
-			count = (IAVF_MAX_AQ_BUF_SIZE - sizeof(*vvfl_v2)) /
-				sizeof(struct virtchnl_vlan_filter);
-			len = sizeof(*vvfl_v2) +
-				((count - 1) *
-				 sizeof(struct virtchnl_vlan_filter));
+			while (len > IAVF_MAX_AQ_BUF_SIZE)
+				len = virtchnl_struct_size(vvfl_v2, filters,
+							   --count);
 			more = true;
 		}
 
@@ -879,16 +876,12 @@ void iavf_del_vlans(struct iavf_adapter *adapter)
 
 		adapter->current_op = VIRTCHNL_OP_DEL_VLAN_V2;
 
-		len = sizeof(*vvfl_v2) +
-			((count - 1) * sizeof(struct virtchnl_vlan_filter));
+		len = virtchnl_struct_size(vvfl_v2, filters, count);
 		if (len > IAVF_MAX_AQ_BUF_SIZE) {
 			dev_warn(&adapter->pdev->dev, "Too many add VLAN changes in one request\n");
-			count = (IAVF_MAX_AQ_BUF_SIZE -
-				 sizeof(*vvfl_v2)) /
-				sizeof(struct virtchnl_vlan_filter);
-			len = sizeof(*vvfl_v2) +
-				((count - 1) *
-				 sizeof(struct virtchnl_vlan_filter));
+			while (len > IAVF_MAX_AQ_BUF_SIZE)
+				len = virtchnl_struct_size(vvfl_v2, filters,
+							   --count);
 			more = true;
 		}
 
@@ -1492,7 +1485,7 @@ void iavf_enable_channels(struct iavf_adapter *adapter)
 		return;
 	}
 
-	len = struct_size(vti, list, adapter->num_tc - 1);
+	len = virtchnl_struct_size(vti, list, adapter->num_tc);
 	vti = kzalloc(len, GFP_KERNEL);
 	if (!vti)
 		return;
include/linux/avf/virtchnl.h +24 -15
index c1a20b533fc0..d0807ad43f93 100644
@@ -716,10 +716,11 @@ struct virtchnl_vlan_filter_list_v2 {
 	u16 vport_id;
 	u16 num_elements;
 	u8 pad[4];
-	struct virtchnl_vlan_filter filters[1];
+	struct virtchnl_vlan_filter filters[];
 };
 
-VIRTCHNL_CHECK_STRUCT_LEN(40, virtchnl_vlan_filter_list_v2);
+VIRTCHNL_CHECK_STRUCT_LEN(8, virtchnl_vlan_filter_list_v2);
+#define virtchnl_vlan_filter_list_v2_LEGACY_SIZEOF	40
 
 /* VIRTCHNL_OP_ENABLE_VLAN_STRIPPING_V2
  * VIRTCHNL_OP_DISABLE_VLAN_STRIPPING_V2
@@ -918,10 +919,11 @@ VIRTCHNL_CHECK_STRUCT_LEN(16, virtchnl_channel_info);
 struct virtchnl_tc_info {
 	u32	num_tc;
 	u32	pad;
-	struct	virtchnl_channel_info list[1];
+	struct virtchnl_channel_info list[];
 };
 
-VIRTCHNL_CHECK_STRUCT_LEN(24, virtchnl_tc_info);
+VIRTCHNL_CHECK_STRUCT_LEN(8, virtchnl_tc_info);
+#define virtchnl_tc_info_LEGACY_SIZEOF	24
 
 /* VIRTCHNL_ADD_CLOUD_FILTER
  * VIRTCHNL_DEL_CLOUD_FILTER
@@ -1059,10 +1061,11 @@ VIRTCHNL_CHECK_STRUCT_LEN(12, virtchnl_rdma_qv_info);
 
 struct virtchnl_rdma_qvlist_info {
 	u32 num_vectors;
-	struct virtchnl_rdma_qv_info qv_info[1];
+	struct virtchnl_rdma_qv_info qv_info[];
 };
 
-VIRTCHNL_CHECK_STRUCT_LEN(16, virtchnl_rdma_qvlist_info);
+VIRTCHNL_CHECK_STRUCT_LEN(4, virtchnl_rdma_qvlist_info);
+#define virtchnl_rdma_qvlist_info_LEGACY_SIZEOF	16
 
 /* VF reset states - these are written into the RSTAT register:
  * VFGEN_RSTAT on the VF
@@ -1377,6 +1380,9 @@ VIRTCHNL_CHECK_STRUCT_LEN(12, virtchnl_fdir_del);
 #define __vss_byone(p, member, count, old)				      \
 	(struct_size(p, member, count) + (old - 1 - struct_size(p, member, 0)))
 
+#define __vss_byelem(p, member, count, old)				      \
+	(struct_size(p, member, count - 1) + (old - struct_size(p, member, 0)))
+
 #define __vss_full(p, member, count, old)				      \
 	(struct_size(p, member, count) + (old - struct_size(p, member, 0)))
 
@@ -1390,6 +1396,9 @@ VIRTCHNL_CHECK_STRUCT_LEN(12, virtchnl_fdir_del);
 		 __vss(virtchnl_irq_map_info, __vss_full, p, m, c),	      \
 		 __vss(virtchnl_ether_addr_list, __vss_full, p, m, c),	      \
 		 __vss(virtchnl_vlan_filter_list, __vss_full, p, m, c),	      \
+		 __vss(virtchnl_vlan_filter_list_v2, __vss_byelem, p, m, c),  \
+		 __vss(virtchnl_tc_info, __vss_byelem, p, m, c),	      \
+		 __vss(virtchnl_rdma_qvlist_info, __vss_byelem, p, m, c),     \
 		 __vss(virtchnl_rss_key, __vss_byone, p, m, c),		      \
 		 __vss(virtchnl_rss_lut, __vss_byone, p, m, c))
 
@@ -1495,13 +1504,13 @@ virtchnl_vc_validate_vf_msg(struct virtchnl_version_info *ver, u32 v_opcode,
 	case VIRTCHNL_OP_RELEASE_RDMA_IRQ_MAP:
 		break;
 	case VIRTCHNL_OP_CONFIG_RDMA_IRQ_MAP:
-		valid_len = sizeof(struct virtchnl_rdma_qvlist_info);
+		valid_len = virtchnl_rdma_qvlist_info_LEGACY_SIZEOF;
 		if (msglen >= valid_len) {
 			struct virtchnl_rdma_qvlist_info *qv =
 				(struct virtchnl_rdma_qvlist_info *)msg;
 
-			valid_len += ((qv->num_vectors - 1) *
-				sizeof(struct virtchnl_rdma_qv_info));
+			valid_len = virtchnl_struct_size(qv, qv_info,
+							 qv->num_vectors);
 		}
 		break;
 	case VIRTCHNL_OP_CONFIG_RSS_KEY:
@@ -1534,12 +1543,12 @@ virtchnl_vc_validate_vf_msg(struct virtchnl_version_info *ver, u32 v_opcode,
 		valid_len = sizeof(struct virtchnl_vf_res_request);
 		break;
 	case VIRTCHNL_OP_ENABLE_CHANNELS:
-		valid_len = sizeof(struct virtchnl_tc_info);
+		valid_len = virtchnl_tc_info_LEGACY_SIZEOF;
 		if (msglen >= valid_len) {
 			struct virtchnl_tc_info *vti =
 				(struct virtchnl_tc_info *)msg;
-			valid_len += (vti->num_tc - 1) *
-				     sizeof(struct virtchnl_channel_info);
+			valid_len = virtchnl_struct_size(vti, list,
+							 vti->num_tc);
 			if (vti->num_tc == 0)
 				err_msg_format = true;
 		}
@@ -1566,13 +1575,13 @@ virtchnl_vc_validate_vf_msg(struct virtchnl_version_info *ver, u32 v_opcode,
 		break;
 	case VIRTCHNL_OP_ADD_VLAN_V2:
 	case VIRTCHNL_OP_DEL_VLAN_V2:
-		valid_len = sizeof(struct virtchnl_vlan_filter_list_v2);
+		valid_len = virtchnl_vlan_filter_list_v2_LEGACY_SIZEOF;
 		if (msglen >= valid_len) {
 			struct virtchnl_vlan_filter_list_v2 *vfl =
 			    (struct virtchnl_vlan_filter_list_v2 *)msg;
 
-			valid_len += (vfl->num_elements - 1) *
-				sizeof(struct virtchnl_vlan_filter);
+			valid_len = virtchnl_struct_size(vfl, filters,
+							 vfl->num_elements);
 
 			if (vfl->num_elements == 0) {
 				err_msg_format = true;
Signature · 1 line
2.41.0
Reviewed-by on v1 3/3
#13 lore raw

Re: [PATCH net-next 3/3] virtchnl: fix fake 1-elem arrays for structures allocated as `nents`

Quoted 8 lines · Alexander Lobakin

Finally, fix 3 structures which are allocated technically correctly, i.e. the calculated size equals to the one that struct_size() would return, except for sizeof(). For &virtchnl_vlan_filter_list_v2, use the same approach when there are no enough space as taken previously for &virtchnl_vlan_filter_list, i.e. let the maximum size be calculated automatically instead of trying to guestimate it using maths.

Signature · 1 line
Kees Cook
#14 lore raw

RE: [Intel-wired-lan] [PATCH net-next 3/3] virtchnl: fix fake 1-elem arrays for structures allocated as `nents`

Quoted 24 lines

-----Original Message-----
From: Intel-wired-lan <[email protected]> On Behalf Of Kees Cook
Sent: piątek, 4 sierpnia 2023 10:30
To: Lobakin, Aleksander <[email protected]>
Cc: Andy Shevchenko <[email protected]>; Zaremba, Larysa <[email protected]>; [email protected]; Gustavo A. R. Silva <[email protected]>; [email protected]; Eric Dumazet <[email protected]>; [email protected]; linux- [email protected]; Jakub Kicinski <[email protected]>; Paolo Abeni <[email protected]>; David S. Miller <[email protected]>
Subject: Re: [Intel-wired-lan] [PATCH net-next 3/3] virtchnl: fix fake 1-elem arrays for structures allocated as nents

Quoted 8 lines · Alexander Lobakin

Finally, fix 3 structures which are allocated technically correctly, i.e. the calculated size equals to the one that struct_size() would return, except for sizeof(). For &virtchnl_vlan_filter_list_v2, use the same approach when there are no enough space as taken previously for &virtchnl_vlan_filter_list, i.e. let the maximum size be calculated automatically instead of trying to guestimate it using maths.

Quoted 8 lines

--
Kees Cook
_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan

#15 lore raw
In reply to Alexander Lobakin

From: Alexander Lobakin <[email protected]>
Date: Fri, 28 Jul 2023 17:52:04 +0200

Quoted 8 lines

6.5-rc1 started spitting warning splats when composing virtchnl messages, precisely on virtchnl_rss_key and virtchnl_lut:

[   84.167709] memcpy: detected field-spanning write (size 52) of single
field "vrk->key" at drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1095
(size 1)
[   84.169915] WARNING: CPU: 3 PID: 11 at drivers/net/ethernet/intel/

iavf/iavf_virtchnl.c:1095 iavf_set_rss_key+0x123/0x140 [iavf]

(tender ping :D)

Thanks,
Olek

#16 lore raw
In reply to Alexander Lobakin

From: Alexander Lobakin <[email protected]>
Date: Fri, 28 Jul 2023 17:52:04 +0200

Quoted 8 lines

6.5-rc1 started spitting warning splats when composing virtchnl messages, precisely on virtchnl_rss_key and virtchnl_lut:

[   84.167709] memcpy: detected field-spanning write (size 52) of single
field "vrk->key" at drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1095
(size 1)
[   84.169915] WARNING: CPU: 3 PID: 11 at drivers/net/ethernet/intel/

iavf/iavf_virtchnl.c:1095 iavf_set_rss_key+0x123/0x140 [iavf]

[...]

Quoted 9 lines
 .../ethernet/intel/i40e/i40e_virtchnl_pf.c    |   9 +-
 drivers/net/ethernet/intel/iavf/iavf.h        |   6 +-
 drivers/net/ethernet/intel/iavf/iavf_client.c |   4 +-
 drivers/net/ethernet/intel/iavf/iavf_client.h |   2 +-
 .../net/ethernet/intel/iavf/iavf_virtchnl.c   |  75 +++++------
 drivers/net/ethernet/intel/ice/ice_virtchnl.c |   2 +-
 include/linux/avf/virtchnl.h                  | 127 +++++++++++-------
 7 files changed, 124 insertions(+), 101 deletions(-)

Tony, could you please take it via your next tree? I'd like the validation to make sure more different host <-> guest pairs work.

(with Kees' tags, assuming he reviewed and approved the whole series, I
asked about #2 already)

Thanks,
Olek

maintainer
#17 lore raw
Quoted 24 lines · Alexander Lobakin

From: Alexander Lobakin <[email protected]>
Date: Fri, 28 Jul 2023 17:52:04 +0200

Quoted 8 lines

6.5-rc1 started spitting warning splats when composing virtchnl messages, precisely on virtchnl_rss_key and virtchnl_lut:

[   84.167709] memcpy: detected field-spanning write (size 52) of single
field "vrk->key" at drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1095
(size 1)
[   84.169915] WARNING: CPU: 3 PID: 11 at drivers/net/ethernet/intel/

iavf/iavf_virtchnl.c:1095 iavf_set_rss_key+0x123/0x140 [iavf]

[...]

Quoted 9 lines
  .../ethernet/intel/i40e/i40e_virtchnl_pf.c    |   9 +-
  drivers/net/ethernet/intel/iavf/iavf.h        |   6 +-
  drivers/net/ethernet/intel/iavf/iavf_client.c |   4 +-
  drivers/net/ethernet/intel/iavf/iavf_client.h |   2 +-
  .../net/ethernet/intel/iavf/iavf_virtchnl.c   |  75 +++++------
  drivers/net/ethernet/intel/ice/ice_virtchnl.c |   2 +-
  include/linux/avf/virtchnl.h                  | 127 +++++++++++-------
  7 files changed, 124 insertions(+), 101 deletions(-)
Quoted 8 lines

Tony, could you please take it via your next tree? I'd like the validation to make sure more different host <-> guest pairs work.

(with Kees' tags, assuming he reviewed and approved the whole series, I
asked about #2 already)

Thanks,
Olek

Ok, will apply it today. For the future if you want it through IWL, can you tag it with the iwl-* target (and have IWL in the To)? Since this had 'net-next' and was 'To' netdev maintainers, I took it that you wanted it taken through netdev.

Thanks,
Tony
_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan

#18 lore raw

From: Tony Nguyen <[email protected]>
Date: Fri, 4 Aug 2023 11:07:14 -0700

Quoted 28 lines · Alexander Lobakin
Quoted 27 lines

From: Alexander Lobakin <[email protected]>
Date: Fri, 28 Jul 2023 17:52:04 +0200

Quoted 8 lines

6.5-rc1 started spitting warning splats when composing virtchnl messages, precisely on virtchnl_rss_key and virtchnl_lut:

[   84.167709] memcpy: detected field-spanning write (size 52) of single field "vrk->key" at drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1095 (size 1)
[   84.169915] WARNING: CPU: 3 PID: 11 at drivers/net/ethernet/intel/ iavf/iavf_virtchnl.c:1095 iavf_set_rss_key+0x123/0x140 [iavf]

[...]

Quoted 9 lines

  .../ethernet/intel/i40e/i40e_virtchnl_pf.c    |   9 +- drivers/net/ethernet/intel/iavf/iavf.h        |   6 +- drivers/net/ethernet/intel/iavf/iavf_client.c |   4 +- drivers/net/ethernet/intel/iavf/iavf_client.h |   2 +- .../net/ethernet/intel/iavf/iavf_virtchnl.c   |  75 +++++------ drivers/net/ethernet/intel/ice/ice_virtchnl.c |   2 +- include/linux/avf/virtchnl.h                  | 127 +++++++++++-------

  7 files changed, 124 insertions(+), 101 deletions(-)

Tony, could you please take it via your next tree? I'd like the validation to make sure more different host <-> guest pairs work.

Quoted 7 lines
Quoted 5 lines

(with Kees' tags, assuming he reviewed and approved the whole series, I asked about #2 already)

Thanks,
Olek

Ok, will apply it today. For the future if you want it through IWL, can

Great, thanks!

Quoted 3 lines

you tag it with the iwl-* target (and have IWL in the To)? Since this had 'net-next' and was 'To' netdev maintainers, I took it that you wanted it taken through netdev.

Sure, I know, just for some reason targeted this directly to net at first, but then realized it would be better for this to go via IWL :clownface:

Quoted 3 lines

Thanks,
Tony

Thanks,
Olek
_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan

#19 lore raw
Quoted 73 lines

-----Original Message-----
From: Intel-wired-lan <[email protected]> On Behalf Of Alexander Lobakin
Sent: piątek, 4 sierpnia 2023 20:09
To: Nguyen, Anthony L <[email protected]>
Cc: Andy Shevchenko <[email protected]>; Kees Cook <[email protected]>; Zaremba, Larysa <[email protected]>; [email protected]; Gustavo A. R. Silva <[email protected]>; [email protected]; Eric Dumazet <[email protected]>; intel- [email protected]; [email protected]; Jakub Kicinski <[email protected]>; Paolo Abeni <[email protected]>; David S. Miller <[email protected]>
Subject: Re: [Intel-wired-lan] [PATCH net-next 0/3] virtchnl: fix fake 1-elem arrays

From: Tony Nguyen <[email protected]>
Date: Fri, 4 Aug 2023 11:07:14 -0700

Quoted 35 lines · Alexander Lobakin
Quoted 34 lines

From: Alexander Lobakin <[email protected]>
Date: Fri, 28 Jul 2023 17:52:04 +0200

Quoted 10 lines

6.5-rc1 started spitting warning splats when composing virtchnl messages, precisely on virtchnl_rss_key and virtchnl_lut:

[   84.167709] memcpy: detected field-spanning write (size 52) of single field "vrk->key" at drivers/net/ethernet/intel/iavf/iavf_virtchnl.c:1095 (size 1)
[   84.169915] WARNING: CPU: 3 PID: 11 at drivers/net/ethernet/intel/
iavf/iavf_virtchnl.c:1095 iavf_set_rss_key+0x123/0x140 [iavf]

[...]

Quoted 10 lines

  .../ethernet/intel/i40e/i40e_virtchnl_pf.c    |   9 +- drivers/net/ethernet/intel/iavf/iavf.h        |   6 +- drivers/net/ethernet/intel/iavf/iavf_client.c |   4 +- drivers/net/ethernet/intel/iavf/iavf_client.h |   2 +- .../net/ethernet/intel/iavf/iavf_virtchnl.c   |  75 +++++------ drivers/net/ethernet/intel/ice/ice_virtchnl.c |   2 +- include/linux/avf/virtchnl.h                  | 127
+++++++++++-------

  7 files changed, 124 insertions(+), 101 deletions(-)

Tony, could you please take it via your next tree? I'd like the validation to make sure more different host <-> guest pairs work.

(with Kees' tags, assuming he reviewed and approved the whole series, I
  asked about #2 already)

Quoted 5 lines
Quoted 2 lines

Thanks,
Olek

Ok, will apply it today. For the future if you want it through IWL, can

Great, thanks!

Quoted 3 lines

you tag it with the iwl-* target (and have IWL in the To)? Since this had 'net-next' and was 'To' netdev maintainers, I took it that you wanted it taken through netdev.

Sure, I know, just for some reason targeted this directly to net at first, but then realized it would be better for this to go via IWL :clownface:

Quoted 3 lines

Thanks,
Tony

Quoted 6 lines

Thanks,
Olek
_______________________________________________
Intel-wired-lan mailing list
[email protected]
https://lists.osuosl.org/mailman/listinfo/intel-wired-lan