Skip to contentSkip to conversation

mlx5 fixes 2022-05-03

GIT PULL Applied Saeed Mahameed 2 participants 17 messages May 4, 2022

Subsystems (4)
lore
Share & tools
lore

Saeed Mahameed

author maintainer
#0 lore raw

This series provides bug fixes to mlx5 driver. Please pull and let me know if there is any problem.

Thanks,
Saeed.

The following changes since commit 0a806ecc4023fcf393fcc1093f4a532813ca4124:

Merge branch 'bnxt_en-bug-fixes' (2022-05-03 17:41:35 -0700)

are available in the Git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-fixes-2022-05-03

for you to fetch changes up to a042d7f5bb68c47f6e0e546ca367d14e1e4b25ba:

net/mlx5: Fix matching on inner TTC (2022-05-04 00:00:07 -0700)

----------------------------------------------------------------
mlx5-fixes-2022-05-03

----------------------------------------------------------------
Ariel Levkovich (4):
net/mlx5e: Fix wrong source vport matching on tunnel rule net/mlx5e: TC, Fix ct_clear overwriting ct action metadata net/mlx5e: TC, fix decap fallback to uplink when int port not supported net/mlx5e: Avoid checking offload capability in post_parse action

Aya Levin (1):
net/mlx5: Fix slab-out-of-bounds while reading resource dump menu

Mark Bloch (1):
net/mlx5: Fix matching on inner TTC

Mark Zhang (1):
net/mlx5e: Fix the calling of update_buffer_lossy() API

Moshe Shemesh (2):
net/mlx5: Fix deadlock in sync reset flow net/mlx5: Avoid double clear or set of sync reset requested

Moshe Tal (1):
net/mlx5e: Fix trust state reset in reload

Paul Blakey (1):
net/mlx5e: CT: Fix queued up restore put() executing after relevant ft release

Vlad Buslov (4):
net/mlx5e: Don't match double-vlan packets if cvlan is not set net/mlx5e: Lag, Fix use-after-free in fib event handler net/mlx5e: Lag, Fix fib_info pointer assignment net/mlx5e: Lag, Don't skip fib events on current dst

 .../ethernet/mellanox/mlx5/core/diag/rsc_dump.c    | 31 ++++++++---
 .../ethernet/mellanox/mlx5/core/en/port_buffer.c   |  4 +-
 .../ethernet/mellanox/mlx5/core/en/tc/act/act.c    |  3 +-
 .../net/ethernet/mellanox/mlx5/core/en/tc/act/ct.c | 34 +++++++++++-
 drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c | 24 ++++-----
 drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.h | 11 ++++
 .../net/ethernet/mellanox/mlx5/core/en/tc_tun.c    |  3 +-
 drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c | 10 ++++
 drivers/net/ethernet/mellanox/mlx5/core/en_tc.c    | 11 ++++
 .../ethernet/mellanox/mlx5/core/eswitch_offloads.c |  2 +-
 drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c | 60 +++++++++++++---------
 drivers/net/ethernet/mellanox/mlx5/core/lag/mp.c   | 38 +++++++++-----
 drivers/net/ethernet/mellanox/mlx5/core/lag/mp.h   |  7 ++-
 .../net/ethernet/mellanox/mlx5/core/lag/port_sel.c |  2 +-
 .../net/ethernet/mellanox/mlx5/core/lib/fs_ttc.c   |  2 +
 15 files changed, 173 insertions(+), 69 deletions(-)
author maintainer
#1 lore raw

PATCH 1/15 net/mlx5e: Fix wrong source vport matching on tunnel rule

From: Ariel Levkovich <[email protected]>

When OVS internal port is the vtep device, the first decap rule is matching on the internal port's vport metadata value and then changes the metadata to be the uplink's value.

Therefore, following rules on the tunnel, in chain > 0, should avoid matching on internal port metadata and use the uplink vport metadata instead.

Select the uplink's metadata value for the source vport match in case the rule is in chain greater than zero, even if the tunnel route device is internal port.

Fixes 166f431ec6be ("net/mlx5e: Add indirect tc offload of ovs internal port")
---
 drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +1 -1
index 3f63df127091..3b151332e2f8 100644
@@ -139,7 +139,7 @@ mlx5_eswitch_set_rule_source_port(struct mlx5_eswitch *esw,
 		if (mlx5_esw_indir_table_decap_vport(attr))
 			vport = mlx5_esw_indir_table_decap_vport(attr);
 
-		if (esw_attr->int_port)
+		if (attr && !attr->chain && esw_attr->int_port)
 			metadata =
 				mlx5e_tc_int_port_get_metadata_for_match(esw_attr->int_port);
 		else
Signature · 1 line
2.35.1
#2 lore raw

Re: [net 01/15] net/mlx5e: Fix wrong source vport matching on tunnel rule

This series was applied to netdev/net.git (master) by Saeed Mahameed <[email protected]>:

Hello:

This series was applied to netdev/net.git (master) by Saeed Mahameed <[email protected]>:

Quoted 11 lines · On Wed, 4 May 2022 00:02:42 -0700 you wrote:

From: Ariel Levkovich <[email protected]>

When OVS internal port is the vtep device, the first decap rule is matching on the internal port's vport metadata value and then changes the metadata to be the uplink's value.

Therefore, following rules on the tunnel, in chain > 0, should avoid matching on internal port metadata and use the uplink vport metadata instead.

[...]

Here is the summary with links:
- [net,01/15] net/mlx5e: Fix wrong source vport matching on tunnel rule

    https://git.kernel.org/netdev/net/c/cb0d54cbf948
  - [net,02/15] net/mlx5: Fix slab-out-of-bounds while reading resource dump menu
    https://git.kernel.org/netdev/net/c/7ba2d9d8de96
  - [net,03/15] net/mlx5e: Don't match double-vlan packets if cvlan is not set
    https://git.kernel.org/netdev/net/c/ada09af92e62
  - [net,04/15] net/mlx5e: Fix the calling of update_buffer_lossy() API
    https://git.kernel.org/netdev/net/c/c4d963a588a6
  - [net,05/15] net/mlx5e: Lag, Fix use-after-free in fib event handler
    https://git.kernel.org/netdev/net/c/27b0420fd959
  - [net,06/15] net/mlx5e: Lag, Fix fib_info pointer assignment
    https://git.kernel.org/netdev/net/c/a6589155ec98
  - [net,07/15] net/mlx5e: Lag, Don't skip fib events on current dst
    https://git.kernel.org/netdev/net/c/4a2a664ed879
  - [net,08/15] net/mlx5e: TC, Fix ct_clear overwriting ct action metadata
    https://git.kernel.org/netdev/net/c/087032ee7021
  - [net,09/15] net/mlx5e: TC, fix decap fallback to uplink when int port not supported
    https://git.kernel.org/netdev/net/c/e3fdc71bcb6f
  - [net,10/15] net/mlx5e: CT: Fix queued up restore put() executing after relevant ft release
    https://git.kernel.org/netdev/net/c/b069e14fff46
  - [net,11/15] net/mlx5e: Avoid checking offload capability in post_parse action
    https://git.kernel.org/netdev/net/c/0e322efd64d2
  - [net,12/15] net/mlx5e: Fix trust state reset in reload
    https://git.kernel.org/netdev/net/c/b781bff882d1
  - [net,13/15] net/mlx5: Fix deadlock in sync reset flow
    https://git.kernel.org/netdev/net/c/cb7786a76ea3
  - [net,14/15] net/mlx5: Avoid double clear or set of sync reset requested
    https://git.kernel.org/netdev/net/c/fc3d3db07b35
  - [net,15/15] net/mlx5: Fix matching on inner TTC
    https://git.kernel.org/netdev/net/c/a042d7f5bb68

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
author maintainer
#3 lore raw

PATCH 2/15 net/mlx5: Fix slab-out-of-bounds while reading resource dump menu

In reply to Saeed Mahameed

From: Aya Levin <[email protected]>

Resource dump menu may span over more than a single page, support it. Otherwise, menu read may result in a memory access violation: reading outside of the allocated page.
Note that page format of the first menu page contains menu headers while the proceeding menu pages contain only records.

The KASAN logs are as follows:

BUG: KASAN: slab-out-of-bounds in strcmp+0x9b/0xb0
Read of size 1 at addr ffff88812b2e1fd0 by task systemd-udevd/496

CPU: 5 PID: 496 Comm: systemd-udevd Tainted: G    B  5.16.0_for_upstream_debug_2022_01_10_23_12 #1
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014
Call Trace:
 <TASK>
 dump_stack_lvl+0x57/0x7d
 print_address_description.constprop.0+0x1f/0x140
 ? strcmp+0x9b/0xb0
 ? strcmp+0x9b/0xb0
 kasan_report.cold+0x83/0xdf
 ? strcmp+0x9b/0xb0
 strcmp+0x9b/0xb0
 mlx5_rsc_dump_init+0x4ab/0x780 [mlx5_core]
 ? mlx5_rsc_dump_destroy+0x80/0x80 [mlx5_core]
 ? lockdep_hardirqs_on_prepare+0x286/0x400
 ? raw_spin_unlock_irqrestore+0x47/0x50
 ? aomic_notifier_chain_register+0x32/0x40
 mlx5_load+0x104/0x2e0 [mlx5_core]
 mlx5_init_one+0x41b/0x610 [mlx5_core]
 ....
The buggy address belongs to the object at ffff88812b2e0000
 which belongs to the cache kmalloc-4k of size 4096
The buggy address is located 4048 bytes to the right of
 4096-byte region [ffff88812b2e0000, ffff88812b2e1000)
The buggy address belongs to the page:
page:000000009d69807a refcount:1 mapcount:0 mapping:0000000000000000 index:0xffff88812b2e6000 pfn:0x12b2e0
head:000000009d69807a order:3 compound_mapcount:0 compound_pincount:0
flags: 0x8000000000010200(slab|head|zone=2)
raw: 8000000000010200 0000000000000000 dead000000000001 ffff888100043040
raw: ffff88812b2e6000 0000000080040000 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected

Memory state around the buggy address:
 ffff88812b2e1e80: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
 ffff88812b2e1f00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>ffff88812b2e1f80: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
                                                 ^
 ffff88812b2e2000: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
 ffff88812b2e2080: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
==================================================================
Fixes 12206b17235a ("net/mlx5: Add support for resource dump")
---
 .../mellanox/mlx5/core/diag/rsc_dump.c        | 31 +++++++++++++++----
 1 file changed, 25 insertions(+), 6 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/diag/rsc_dump.c +25 -6
index 538adab6878b..c5b560a8b026 100644
@@ -31,6 +31,7 @@ static const char *const mlx5_rsc_sgmt_name[] = {
 struct mlx5_rsc_dump {
 	u32 pdn;
 	u32 mkey;
+	u32 number_of_menu_items;
 	u16 fw_segment_type[MLX5_SGMT_TYPE_NUM];
 };
 
@@ -50,21 +51,37 @@ static int mlx5_rsc_dump_sgmt_get_by_name(char *name)
 	return -EINVAL;
 }
 
-static void mlx5_rsc_dump_read_menu_sgmt(struct mlx5_rsc_dump *rsc_dump, struct page *page)
+#define MLX5_RSC_DUMP_MENU_HEADER_SIZE (MLX5_ST_SZ_BYTES(resource_dump_info_segment) + \
+					MLX5_ST_SZ_BYTES(resource_dump_command_segment) + \
+					MLX5_ST_SZ_BYTES(resource_dump_menu_segment))
+
+static int mlx5_rsc_dump_read_menu_sgmt(struct mlx5_rsc_dump *rsc_dump, struct page *page,
+					int read_size, int start_idx)
 {
 	void *data = page_address(page);
 	enum mlx5_sgmt_type sgmt_idx;
 	int num_of_items;
 	char *sgmt_name;
 	void *member;
+	int size = 0;
 	void *menu;
 	int i;
 
-	menu = MLX5_ADDR_OF(menu_resource_dump_response, data, menu);
-	num_of_items = MLX5_GET(resource_dump_menu_segment, menu, num_of_records);
+	if (!start_idx) {
+		menu = MLX5_ADDR_OF(menu_resource_dump_response, data, menu);
+		rsc_dump->number_of_menu_items = MLX5_GET(resource_dump_menu_segment, menu,
+							  num_of_records);
+		size = MLX5_RSC_DUMP_MENU_HEADER_SIZE;
+		data += size;
+	}
+	num_of_items = rsc_dump->number_of_menu_items;
+
+	for (i = 0; start_idx + i < num_of_items; i++) {
+		size += MLX5_ST_SZ_BYTES(resource_dump_menu_record);
+		if (size >= read_size)
+			return start_idx + i;
 
-	for (i = 0; i < num_of_items; i++) {
-		member = MLX5_ADDR_OF(resource_dump_menu_segment, menu, record[i]);
+		member = data + MLX5_ST_SZ_BYTES(resource_dump_menu_record) * i;
 		sgmt_name =  MLX5_ADDR_OF(resource_dump_menu_record, member, segment_name);
 		sgmt_idx = mlx5_rsc_dump_sgmt_get_by_name(sgmt_name);
 		if (sgmt_idx == -EINVAL)
@@ -72,6 +89,7 @@ static void mlx5_rsc_dump_read_menu_sgmt(struct mlx5_rsc_dump *rsc_dump, struct
 		rsc_dump->fw_segment_type[sgmt_idx] = MLX5_GET(resource_dump_menu_record,
 							       member, segment_type);
 	}
+	return 0;
 }
 
 static int mlx5_rsc_dump_trigger(struct mlx5_core_dev *dev, struct mlx5_rsc_dump_cmd *cmd,
@@ -168,6 +186,7 @@ static int mlx5_rsc_dump_menu(struct mlx5_core_dev *dev)
 	struct mlx5_rsc_dump_cmd *cmd = NULL;
 	struct mlx5_rsc_key key = {};
 	struct page *page;
+	int start_idx = 0;
 	int size;
 	int err;
 
@@ -189,7 +208,7 @@ static int mlx5_rsc_dump_menu(struct mlx5_core_dev *dev)
 		if (err < 0)
 			goto destroy_cmd;
 
-		mlx5_rsc_dump_read_menu_sgmt(dev->rsc_dump, page);
+		start_idx = mlx5_rsc_dump_read_menu_sgmt(dev->rsc_dump, page, size, start_idx);
 
 	} while (err > 0);
 
Signature · 1 line
2.35.1
author maintainer
#4 lore raw

PATCH 3/15 net/mlx5e: Don't match double-vlan packets if cvlan is not set

In reply to Saeed Mahameed

From: Vlad Buslov <[email protected]>

Currently, match VLAN rule also matches packets that have multiple VLAN headers. This behavior is similar to buggy flower classifier behavior that has recently been fixed. Fix the issue by matching on outer_second_cvlan_tag with value 0 which will cause the HW to verify the packet doesn't contain second vlan header.

Fixes 699e96ddf47f ("net/mlx5e: Support offloading tc double vlan headers match")
---
 drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c +11 -0
index e3fc15ae7bb1..ac0f73074f7a 100644
@@ -2459,6 +2459,17 @@ static int __parse_cls_flower(struct mlx5e_priv *priv,
 				 match.key->vlan_priority);
 
 			*match_level = MLX5_MATCH_L2;
+
+			if (!flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_CVLAN) &&
+			    match.mask->vlan_eth_type &&
+			    MLX5_CAP_FLOWTABLE_TYPE(priv->mdev,
+						    ft_field_support.outer_second_vid,
+						    fs_type)) {
+				MLX5_SET(fte_match_set_misc, misc_c,
+					 outer_second_cvlan_tag, 1);
+				spec->match_criteria_enable |=
+					MLX5_MATCH_MISC_PARAMETERS;
+			}
 		}
 	} else if (*match_level != MLX5_MATCH_NONE) {
 		/* cvlan_tag enabled in match criteria and
Signature · 1 line
2.35.1
author maintainer
#5 lore raw

PATCH 4/15 net/mlx5e: Fix the calling of update_buffer_lossy() API

In reply to Saeed Mahameed

From: Mark Zhang <[email protected]>

The arguments of update_buffer_lossy() is in a wrong order. Fix it.

Fixes 88b3d5c90e96 ("net/mlx5e: Fix port buffers cell size value")
---
 drivers/net/ethernet/mellanox/mlx5/core/en/port_buffer.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/port_buffer.c +2 -2
index 673f1c82d381..c9d5d8d93994 100644
@@ -309,8 +309,8 @@ int mlx5e_port_manual_buffer_config(struct mlx5e_priv *priv,
 		if (err)
 			return err;
 
-		err = update_buffer_lossy(max_mtu, curr_pfc_en, prio2buffer, port_buff_cell_sz,
-					  xoff, &port_buffer, &update_buffer);
+		err = update_buffer_lossy(max_mtu, curr_pfc_en, prio2buffer, xoff,
+					  port_buff_cell_sz, &port_buffer, &update_buffer);
 		if (err)
 			return err;
 	}
Signature · 1 line
2.35.1
author maintainer
#6 lore raw

PATCH 5/15 net/mlx5e: Lag, Fix use-after-free in fib event handler

In reply to Saeed Mahameed

From: Vlad Buslov <[email protected]>

Recent commit that modified fib route event handler to handle events according to their priority introduced use-after-free[0] in mp->mfi pointer usage. The pointer now is not just cached in order to be compared to following fib_info instances, but is also dereferenced to obtain fib_priority. However, since mlx5 lag code doesn't hold the reference to fin_info during whole mp->mfi lifetime, it could be used after fib_info instance has already been freed be kernel infrastructure code.

Don't ever dereference mp->mfi pointer. Refactor it to be 'const void*' type and cache fib_info priority in dedicated integer. Group fib_info-related data into dedicated 'fib' structure that will be further extended by following patches in the series.

[0]:

[  203.588029] ==================================================================
[  203.590161] BUG: KASAN: use-after-free in mlx5_lag_fib_update+0xabd/0xd60 [mlx5_core]
[  203.592386] Read of size 4 at addr ffff888144df2050 by task kworker/u20:4/138
[  203.594766] CPU: 3 PID: 138 Comm: kworker/u20:4 Tainted: G    B             5.17.0-rc7+ #6
[  203.596751] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014
[  203.598813] Workqueue: mlx5_lag_mp mlx5_lag_fib_update [mlx5_core]
[  203.600053] Call Trace:
[  203.600608]  <TASK>
[  203.601110]  dump_stack_lvl+0x48/0x5e
[  203.601860]  print_address_description.constprop.0+0x1f/0x160
[  203.602950]  ? mlx5_lag_fib_update+0xabd/0xd60 [mlx5_core]
[  203.604073]  ? mlx5_lag_fib_update+0xabd/0xd60 [mlx5_core]
[  203.605177]  kasan_report.cold+0x83/0xdf
[  203.605969]  ? mlx5_lag_fib_update+0xabd/0xd60 [mlx5_core]
[  203.607102]  mlx5_lag_fib_update+0xabd/0xd60 [mlx5_core]
[  203.608199]  ? mlx5_lag_init_fib_work+0x1c0/0x1c0 [mlx5_core]
[  203.609382]  ? read_word_at_a_time+0xe/0x20
[  203.610463]  ? strscpy+0xa0/0x2a0
[  203.611463]  process_one_work+0x722/0x1270
[  203.612344]  worker_thread+0x540/0x11e0
[  203.613136]  ? rescuer_thread+0xd50/0xd50
[  203.613949]  kthread+0x26e/0x300
[  203.614627]  ? kthread_complete_and_exit+0x20/0x20
[  203.615542]  ret_from_fork+0x1f/0x30
[  203.616273]  </TASK>
[  203.617174] Allocated by task 3746:
[  203.617874]  kasan_save_stack+0x1e/0x40
[  203.618644]  __kasan_kmalloc+0x81/0xa0
[  203.619394]  fib_create_info+0xb41/0x3c50
[  203.620213]  fib_table_insert+0x190/0x1ff0
[  203.621020]  fib_magic.isra.0+0x246/0x2e0
[  203.621803]  fib_add_ifaddr+0x19f/0x670
[  203.622563]  fib_inetaddr_event+0x13f/0x270
[  203.623377]  blocking_notifier_call_chain+0xd4/0x130
[  203.624355]  __inet_insert_ifa+0x641/0xb20
[  203.625185]  inet_rtm_newaddr+0xc3d/0x16a0
[  203.626009]  rtnetlink_rcv_msg+0x309/0x880
[  203.626826]  netlink_rcv_skb+0x11d/0x340
[  203.627626]  netlink_unicast+0x4cc/0x790
[  203.628430]  netlink_sendmsg+0x762/0xc00
[  203.629230]  sock_sendmsg+0xb2/0xe0
[  203.629955]  ____sys_sendmsg+0x58a/0x770
[  203.630756]  ___sys_sendmsg+0xd8/0x160
[  203.631523]  __sys_sendmsg+0xb7/0x140
[  203.632294]  do_syscall_64+0x35/0x80
[  203.633045]  entry_SYSCALL_64_after_hwframe+0x44/0xae
[  203.634427] Freed by task 0:
[  203.635063]  kasan_save_stack+0x1e/0x40
[  203.635844]  kasan_set_track+0x21/0x30
[  203.636618]  kasan_set_free_info+0x20/0x30
[  203.637450]  __kasan_slab_free+0xfc/0x140
[  203.638271]  kfree+0x94/0x3b0
[  203.638903]  rcu_core+0x5e4/0x1990
[  203.639640]  __do_softirq+0x1ba/0x5d3
[  203.640828] Last potentially related work creation:
[  203.641785]  kasan_save_stack+0x1e/0x40
[  203.642571]  __kasan_record_aux_stack+0x9f/0xb0
[  203.643478]  call_rcu+0x88/0x9c0
[  203.644178]  fib_release_info+0x539/0x750
[  203.644997]  fib_table_delete+0x659/0xb80
[  203.645809]  fib_magic.isra.0+0x1a3/0x2e0
[  203.646617]  fib_del_ifaddr+0x93f/0x1300
[  203.647415]  fib_inetaddr_event+0x9f/0x270
[  203.648251]  blocking_notifier_call_chain+0xd4/0x130
[  203.649225]  __inet_del_ifa+0x474/0xc10
[  203.650016]  devinet_ioctl+0x781/0x17f0
[  203.650788]  inet_ioctl+0x1ad/0x290
[  203.651533]  sock_do_ioctl+0xce/0x1c0
[  203.652315]  sock_ioctl+0x27b/0x4f0
[  203.653058]  __x64_sys_ioctl+0x124/0x190
[  203.653850]  do_syscall_64+0x35/0x80
[  203.654608]  entry_SYSCALL_64_after_hwframe+0x44/0xae
[  203.666952] The buggy address belongs to the object at ffff888144df2000
                which belongs to the cache kmalloc-256 of size 256
[  203.669250] The buggy address is located 80 bytes inside of
                256-byte region [ffff888144df2000, ffff888144df2100)
[  203.671332] The buggy address belongs to the page:
[  203.672273] page:00000000bf6c9314 refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x144df0
[  203.674009] head:00000000bf6c9314 order:2 compound_mapcount:0 compound_pincount:0
[  203.675422] flags: 0x2ffff800010200(slab|head|node=0|zone=2|lastcpupid=0x1ffff)
[  203.676819] raw: 002ffff800010200 0000000000000000 dead000000000122 ffff888100042b40
[  203.678384] raw: 0000000000000000 0000000080200020 00000001ffffffff 0000000000000000
[  203.679928] page dumped because: kasan: bad access detected
[  203.681455] Memory state around the buggy address:
[  203.682421]  ffff888144df1f00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[  203.683863]  ffff888144df1f80: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[  203.685310] >ffff888144df2000: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[  203.686701]                                                  ^
[  203.687820]  ffff888144df2080: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[  203.689226]  ffff888144df2100: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[  203.690620] ==================================================================
Fixes ad11c4f1d8fd ("net/mlx5e: Lag, Only handle events from highest priority multipath entry")
---
 .../net/ethernet/mellanox/mlx5/core/lag/mp.c  | 26 ++++++++++++-------
 .../net/ethernet/mellanox/mlx5/core/lag/mp.h  |  5 +++-
 2 files changed, 20 insertions(+), 11 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/lag/mp.c +16 -10
index 4a6ec15ef046..bc77aba97ac1 100644
@@ -100,6 +100,12 @@ static void mlx5_lag_fib_event_flush(struct notifier_block *nb)
 	flush_workqueue(mp->wq);
 }
 
+static void mlx5_lag_fib_set(struct lag_mp *mp, struct fib_info *fi)
+{
+	mp->fib.mfi = fi;
+	mp->fib.priority = fi->fib_priority;
+}
+
 struct mlx5_fib_event_work {
 	struct work_struct work;
 	struct mlx5_lag *ldev;
@@ -121,13 +127,13 @@ static void mlx5_lag_fib_route_event(struct mlx5_lag *ldev,
 	/* Handle delete event */
 	if (event == FIB_EVENT_ENTRY_DEL) {
 		/* stop track */
-		if (mp->mfi == fi)
-			mp->mfi = NULL;
+		if (mp->fib.mfi == fi)
+			mp->fib.mfi = NULL;
 		return;
 	}
 
 	/* Handle multipath entry with lower priority value */
-	if (mp->mfi && mp->mfi != fi && fi->fib_priority >= mp->mfi->fib_priority)
+	if (mp->fib.mfi && mp->fib.mfi != fi && fi->fib_priority >= mp->fib.priority)
 		return;
 
 	/* Handle add/replace event */
@@ -145,7 +151,7 @@ static void mlx5_lag_fib_route_event(struct mlx5_lag *ldev,
 			mlx5_lag_set_port_affinity(ldev, i);
 		}
 
-		mp->mfi = fi;
+		mlx5_lag_fib_set(mp, fi);
 		return;
 	}
 
@@ -165,7 +171,7 @@ static void mlx5_lag_fib_route_event(struct mlx5_lag *ldev,
 	}
 
 	/* First time we see multipath route */
-	if (!mp->mfi && !__mlx5_lag_is_active(ldev)) {
+	if (!mp->fib.mfi && !__mlx5_lag_is_active(ldev)) {
 		struct lag_tracker tracker;
 
 		tracker = ldev->tracker;
@@ -173,7 +179,7 @@ static void mlx5_lag_fib_route_event(struct mlx5_lag *ldev,
 	}
 
 	mlx5_lag_set_port_affinity(ldev, MLX5_LAG_NORMAL_AFFINITY);
-	mp->mfi = fi;
+	mlx5_lag_fib_set(mp, fi);
 }
 
 static void mlx5_lag_fib_nexthop_event(struct mlx5_lag *ldev,
@@ -184,7 +190,7 @@ static void mlx5_lag_fib_nexthop_event(struct mlx5_lag *ldev,
 	struct lag_mp *mp = &ldev->lag_mp;
 
 	/* Check the nh event is related to the route */
-	if (!mp->mfi || mp->mfi != fi)
+	if (!mp->fib.mfi || mp->fib.mfi != fi)
 		return;
 
 	/* nh added/removed */
@@ -313,7 +319,7 @@ void mlx5_lag_mp_reset(struct mlx5_lag *ldev)
 	/* Clear mfi, as it might become stale when a route delete event
 	 * has been missed, see mlx5_lag_fib_route_event().
 	 */
-	ldev->lag_mp.mfi = NULL;
+	ldev->lag_mp.fib.mfi = NULL;
 }
 
 int mlx5_lag_mp_init(struct mlx5_lag *ldev)
@@ -324,7 +330,7 @@ int mlx5_lag_mp_init(struct mlx5_lag *ldev)
 	/* always clear mfi, as it might become stale when a route delete event
 	 * has been missed
 	 */
-	mp->mfi = NULL;
+	mp->fib.mfi = NULL;
 
 	if (mp->fib_nb.notifier_call)
 		return 0;
@@ -354,5 +360,5 @@ void mlx5_lag_mp_cleanup(struct mlx5_lag *ldev)
 	unregister_fib_notifier(&init_net, &mp->fib_nb);
 	destroy_workqueue(mp->wq);
 	mp->fib_nb.notifier_call = NULL;
-	mp->mfi = NULL;
+	mp->fib.mfi = NULL;
 }
drivers/net/ethernet/mellanox/mlx5/core/lag/mp.h +4 -1
index 57af962cad29..143226753c3a 100644
@@ -15,7 +15,10 @@ enum mlx5_lag_port_affinity {
 
 struct lag_mp {
 	struct notifier_block     fib_nb;
-	struct fib_info           *mfi; /* used in tracking fib events */
+	struct {
+		const void        *mfi; /* used in tracking fib events */
+		u32               priority;
+	} fib;
 	struct workqueue_struct   *wq;
 };
 
Signature · 1 line
2.35.1
author maintainer
#7 lore raw

PATCH 6/15 net/mlx5e: Lag, Fix fib_info pointer assignment

In reply to Saeed Mahameed

From: Vlad Buslov <[email protected]>

Referenced change incorrectly sets single path fib_info even when LAG is not active. Fix it by moving call to mlx5_lag_fib_set() into conditional that verifies LAG state.

Fixes ad11c4f1d8fd ("net/mlx5e: Lag, Only handle events from highest priority multipath entry")
---
 drivers/net/ethernet/mellanox/mlx5/core/lag/mp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/mellanox/mlx5/core/lag/mp.c +1 -1
index bc77aba97ac1..9a5884e8a8bf 100644
@@ -149,9 +149,9 @@ static void mlx5_lag_fib_route_event(struct mlx5_lag *ldev,
 
 			i++;
 			mlx5_lag_set_port_affinity(ldev, i);
+			mlx5_lag_fib_set(mp, fi);
 		}
 
-		mlx5_lag_fib_set(mp, fi);
 		return;
 	}
 
Signature · 1 line
2.35.1
author maintainer
#8 lore raw

PATCH 7/15 net/mlx5e: Lag, Don't skip fib events on current dst

In reply to Saeed Mahameed

From: Vlad Buslov <[email protected]>

Referenced change added check to skip updating fib when new fib instance has same or lower priority. However, new fib instance can be an update on same dst address as existing one even though the structure is another instance that has different address. Ignoring events on such instances causes multipath LAG state to not be correctly updated.

Track 'dst' and 'dst_len' fields of fib event fib_entry_notifier_info structure and don't skip events that have the same value of that fields.

Fixes ad11c4f1d8fd ("net/mlx5e: Lag, Only handle events from highest priority multipath entry")
---
 .../net/ethernet/mellanox/mlx5/core/lag/mp.c  | 20 +++++++++++--------
 .../net/ethernet/mellanox/mlx5/core/lag/mp.h  |  2 ++
 2 files changed, 14 insertions(+), 8 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/lag/mp.c +12 -8
index 9a5884e8a8bf..d6c3e6dfd71f 100644
@@ -100,10 +100,12 @@ static void mlx5_lag_fib_event_flush(struct notifier_block *nb)
 	flush_workqueue(mp->wq);
 }
 
-static void mlx5_lag_fib_set(struct lag_mp *mp, struct fib_info *fi)
+static void mlx5_lag_fib_set(struct lag_mp *mp, struct fib_info *fi, u32 dst, int dst_len)
 {
 	mp->fib.mfi = fi;
 	mp->fib.priority = fi->fib_priority;
+	mp->fib.dst = dst;
+	mp->fib.dst_len = dst_len;
 }
 
 struct mlx5_fib_event_work {
@@ -116,10 +118,10 @@ struct mlx5_fib_event_work {
 	};
 };
 
-static void mlx5_lag_fib_route_event(struct mlx5_lag *ldev,
-				     unsigned long event,
-				     struct fib_info *fi)
+static void mlx5_lag_fib_route_event(struct mlx5_lag *ldev, unsigned long event,
+				     struct fib_entry_notifier_info *fen_info)
 {
+	struct fib_info *fi = fen_info->fi;
 	struct lag_mp *mp = &ldev->lag_mp;
 	struct fib_nh *fib_nh0, *fib_nh1;
 	unsigned int nhs;
@@ -133,7 +135,9 @@ static void mlx5_lag_fib_route_event(struct mlx5_lag *ldev,
 	}
 
 	/* Handle multipath entry with lower priority value */
-	if (mp->fib.mfi && mp->fib.mfi != fi && fi->fib_priority >= mp->fib.priority)
+	if (mp->fib.mfi && mp->fib.mfi != fi &&
+	    (mp->fib.dst != fen_info->dst || mp->fib.dst_len != fen_info->dst_len) &&
+	    fi->fib_priority >= mp->fib.priority)
 		return;
 
 	/* Handle add/replace event */
@@ -149,7 +153,7 @@ static void mlx5_lag_fib_route_event(struct mlx5_lag *ldev,
 
 			i++;
 			mlx5_lag_set_port_affinity(ldev, i);
-			mlx5_lag_fib_set(mp, fi);
+			mlx5_lag_fib_set(mp, fi, fen_info->dst, fen_info->dst_len);
 		}
 
 		return;
@@ -179,7 +183,7 @@ static void mlx5_lag_fib_route_event(struct mlx5_lag *ldev,
 	}
 
 	mlx5_lag_set_port_affinity(ldev, MLX5_LAG_NORMAL_AFFINITY);
-	mlx5_lag_fib_set(mp, fi);
+	mlx5_lag_fib_set(mp, fi, fen_info->dst, fen_info->dst_len);
 }
 
 static void mlx5_lag_fib_nexthop_event(struct mlx5_lag *ldev,
@@ -220,7 +224,7 @@ static void mlx5_lag_fib_update(struct work_struct *work)
 	case FIB_EVENT_ENTRY_REPLACE:
 	case FIB_EVENT_ENTRY_DEL:
 		mlx5_lag_fib_route_event(ldev, fib_work->event,
-					 fib_work->fen_info.fi);
+					 &fib_work->fen_info);
 		fib_info_put(fib_work->fen_info.fi);
 		break;
 	case FIB_EVENT_NH_ADD:
drivers/net/ethernet/mellanox/mlx5/core/lag/mp.h +2 -0
index 143226753c3a..056a066da604 100644
@@ -18,6 +18,8 @@ struct lag_mp {
 	struct {
 		const void        *mfi; /* used in tracking fib events */
 		u32               priority;
+		u32               dst;
+		int               dst_len;
 	} fib;
 	struct workqueue_struct   *wq;
 };
Signature · 1 line
2.35.1
author maintainer
#9 lore raw

PATCH 8/15 net/mlx5e: TC, Fix ct_clear overwriting ct action metadata

In reply to Saeed Mahameed

From: Ariel Levkovich <[email protected]>

ct_clear action is translated to clearing reg_c metadata which holds ct state and zone information using mod header actions.
These actions are allocated during the actions parsing, as part of the flow attributes main mod header action list.

If ct action exists in the rule, the flow's main mod header is used only in the post action table rule, after the ct tables which set the ct info in the reg_c as part of the ct actions.

Therefore, if the original rule has a ct_clear action followed by a ct action, the ct action reg_c setting will be done first and will be followed by the ct_clear resetting reg_c and overwriting the ct info.

Fix this by moving the ct_clear mod header actions allocation from the ct action parsing stage to the ct action post parsing stage where it is already known if ct_clear is followed by a ct action. In such case, we skip the mod header actions allocation for the ct clear since the ct action will write to reg_c anyway after clearing it.

Fixes 806401c20a0f ("net/mlx5e: CT, Fix multiple allocations and memleak of mod acts")
---
 .../mellanox/mlx5/core/en/tc/act/ct.c         | 34 +++++++++++++++++--
 .../ethernet/mellanox/mlx5/core/en/tc_ct.c    | 20 ++++-------
 .../ethernet/mellanox/mlx5/core/en/tc_ct.h    | 11 ++++++
 3 files changed, 49 insertions(+), 16 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/ct.c +32 -2
index b9d38fe807df..a829c94289c1 100644
@@ -45,12 +45,41 @@ tc_act_parse_ct(struct mlx5e_tc_act_parse_state *parse_state,
 	if (mlx5e_is_eswitch_flow(parse_state->flow))
 		attr->esw_attr->split_count = attr->esw_attr->out_count;
 
-	if (!clear_action) {
+	if (clear_action) {
+		parse_state->ct_clear = true;
+	} else {
 		attr->flags |= MLX5_ATTR_FLAG_CT;
 		flow_flag_set(parse_state->flow, CT);
 		parse_state->ct = true;
 	}
-	parse_state->ct_clear = clear_action;
+
+	return 0;
+}
+
+static int
+tc_act_post_parse_ct(struct mlx5e_tc_act_parse_state *parse_state,
+		     struct mlx5e_priv *priv,
+		     struct mlx5_flow_attr *attr)
+{
+	struct mlx5e_tc_mod_hdr_acts *mod_acts = &attr->parse_attr->mod_hdr_acts;
+	int err;
+
+	/* If ct action exist, we can ignore previous ct_clear actions */
+	if (parse_state->ct)
+		return 0;
+
+	if (parse_state->ct_clear) {
+		err = mlx5_tc_ct_set_ct_clear_regs(parse_state->ct_priv, mod_acts);
+		if (err) {
+			NL_SET_ERR_MSG_MOD(parse_state->extack,
+					   "Failed to set registers for ct clear");
+			return err;
+		}
+		attr->action |= MLX5_FLOW_CONTEXT_ACTION_MOD_HDR;
+
+		/* Prevent handling of additional, redundant clear actions */
+		parse_state->ct_clear = false;
+	}
 
 	return 0;
 }
@@ -70,5 +99,6 @@ struct mlx5e_tc_act mlx5e_tc_act_ct = {
 	.can_offload = tc_act_can_offload_ct,
 	.parse_action = tc_act_parse_ct,
 	.is_multi_table_act = tc_act_is_multi_table_act_ct,
+	.post_parse = tc_act_post_parse_ct,
 };
 
drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c +6 -14
index e49f51124c74..73a1e0a4818d 100644
@@ -582,6 +582,12 @@ mlx5_tc_ct_entry_set_registers(struct mlx5_tc_ct_priv *ct_priv,
 	return 0;
 }
 
+int mlx5_tc_ct_set_ct_clear_regs(struct mlx5_tc_ct_priv *priv,
+				 struct mlx5e_tc_mod_hdr_acts *mod_acts)
+{
+		return mlx5_tc_ct_entry_set_registers(priv, mod_acts, 0, 0, 0, 0);
+}
+
 static int
 mlx5_tc_ct_parse_mangle_to_mod_act(struct flow_action_entry *act,
 				   char *modact)
@@ -1410,9 +1416,6 @@ mlx5_tc_ct_parse_action(struct mlx5_tc_ct_priv *priv,
 			const struct flow_action_entry *act,
 			struct netlink_ext_ack *extack)
 {
-	bool clear_action = act->ct.action & TCA_CT_ACT_CLEAR;
-	int err;
-
 	if (!priv) {
 		NL_SET_ERR_MSG_MOD(extack,
 				   "offload of ct action isn't available");
@@ -1423,17 +1426,6 @@ mlx5_tc_ct_parse_action(struct mlx5_tc_ct_priv *priv,
 	attr->ct_attr.ct_action = act->ct.action;
 	attr->ct_attr.nf_ft = act->ct.flow_table;
 
-	if (!clear_action)
-		goto out;
-
-	err = mlx5_tc_ct_entry_set_registers(priv, mod_acts, 0, 0, 0, 0);
-	if (err) {
-		NL_SET_ERR_MSG_MOD(extack, "Failed to set registers for ct clear");
-		return err;
-	}
-	attr->action |= MLX5_FLOW_CONTEXT_ACTION_MOD_HDR;
-
-out:
 	return 0;
 }
 
drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.h +11 -0
index 36d3652bf829..00a3ba862afb 100644
@@ -129,6 +129,10 @@ bool
 mlx5e_tc_ct_restore_flow(struct mlx5_tc_ct_priv *ct_priv,
 			 struct sk_buff *skb, u8 zone_restore_id);
 
+int
+mlx5_tc_ct_set_ct_clear_regs(struct mlx5_tc_ct_priv *priv,
+			     struct mlx5e_tc_mod_hdr_acts *mod_acts);
+
 #else /* CONFIG_MLX5_TC_CT */
 
 static inline struct mlx5_tc_ct_priv *
@@ -170,6 +174,13 @@ mlx5_tc_ct_add_no_trk_match(struct mlx5_flow_spec *spec)
 	return 0;
 }
 
+static inline int
+mlx5_tc_ct_set_ct_clear_regs(struct mlx5_tc_ct_priv *priv,
+			     struct mlx5e_tc_mod_hdr_acts *mod_acts)
+{
+	return -EOPNOTSUPP;
+}
+
 static inline int
 mlx5_tc_ct_parse_action(struct mlx5_tc_ct_priv *priv,
 			struct mlx5_flow_attr *attr,
Signature · 1 line
2.35.1
author maintainer
#10 lore raw

PATCH 9/15 net/mlx5e: TC, fix decap fallback to uplink when int port not supported

In reply to Saeed Mahameed

From: Ariel Levkovich <[email protected]>

When resolving the decap route device for a tunnel decap rule, the result may be an OVS internal port device.

Prior to adding the support for internal port offload, such case would result in using the uplink as the default decap route device which allowed devices that can't support internal port offload to offload this decap rule.

This behavior got broken by adding the internal port offload which will fail in case the device can't support internal port offload.

To restore the old behavior, use the uplink device as the decap route as before when internal port offload is not supported.

Fixes b16eb3c81fe2 ("net/mlx5: Support internal port as decap route device")
---
 drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun.c +2 -1
index 378fc8e3bd97..d87bbb0be7c8 100644
@@ -713,6 +713,7 @@ int mlx5e_tc_tun_route_lookup(struct mlx5e_priv *priv,
 			      struct net_device *filter_dev)
 {
 	struct mlx5_esw_flow_attr *esw_attr = flow_attr->esw_attr;
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
 	struct mlx5e_tc_int_port *int_port;
 	TC_TUN_ROUTE_ATTR_INIT(attr);
 	u16 vport_num;
@@ -747,7 +748,7 @@ int mlx5e_tc_tun_route_lookup(struct mlx5e_priv *priv,
 		esw_attr->rx_tun_attr->vni = MLX5_GET(fte_match_param, spec->match_value,
 						      misc_parameters.vxlan_vni);
 		esw_attr->rx_tun_attr->decap_vport = vport_num;
-	} else if (netif_is_ovs_master(attr.route_dev)) {
+	} else if (netif_is_ovs_master(attr.route_dev) && mlx5e_tc_int_port_supported(esw)) {
 		int_port = mlx5e_tc_int_port_get(mlx5e_get_int_port_priv(priv),
 						 attr.route_dev->ifindex,
 						 MLX5E_TC_INT_PORT_INGRESS);
Signature · 1 line
2.35.1
author maintainer
#11 lore raw

PATCH 10/15 net/mlx5e: CT: Fix queued up restore put() executing after relevant ft release

In reply to Saeed Mahameed

From: Paul Blakey <[email protected]>

__mlx5_tc_ct_entry_put() queues release of tuple related to some ct FT, if that is the last reference to that tuple, the actual deletion of the tuple can happen after the FT is already destroyed and freed.

Flush the used workqueue before destroying the ct FT.

Fixes a2173131526d ("net/mlx5e: CT: manage the lifetime of the ct entry object")
---
 drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c +4 -0
index 73a1e0a4818d..ab4b0f3ee2a0 100644
@@ -1741,6 +1741,8 @@ mlx5_tc_ct_flush_ft_entry(void *ptr, void *arg)
 static void
 mlx5_tc_ct_del_ft_cb(struct mlx5_tc_ct_priv *ct_priv, struct mlx5_ct_ft *ft)
 {
+	struct mlx5e_priv *priv;
+
 	if (!refcount_dec_and_test(&ft->refcount))
 		return;
 
@@ -1750,6 +1752,8 @@ mlx5_tc_ct_del_ft_cb(struct mlx5_tc_ct_priv *ct_priv, struct mlx5_ct_ft *ft)
 	rhashtable_free_and_destroy(&ft->ct_entries_ht,
 				    mlx5_tc_ct_flush_ft_entry,
 				    ct_priv);
+	priv = netdev_priv(ct_priv->netdev);
+	flush_workqueue(priv->wq);
 	mlx5_tc_ct_free_pre_ct_tables(ft);
 	mapping_remove(ct_priv->zone_mapping, ft->zone_restore_id);
 	kfree(ft);
Signature · 1 line
2.35.1
author maintainer
#12 lore raw

PATCH 11/15 net/mlx5e: Avoid checking offload capability in post_parse action

In reply to Saeed Mahameed

From: Ariel Levkovich <[email protected]>

During TC action parsing, the can_offload callback is called before calling the action's main parsing callback.

Later on, the can_offload callback is called again before handling the action's post_parse callback if exists.

Since the main parsing callback might have changed and set parsing params for the rule, following can_offload checks might fail because some parsing params were already set.

Specifically, the ct action main parsing sets the ct param in the parsing status structure and when the second can_offload for ct action is called, before handling the ct post parsing, it will return an error since it checks this ct param to indicate multiple ct actions which are not supported.

Therefore, the can_offload call is removed from the post parsing handling to prevent such cases.
This is allowed since the first can_offload call will ensure that the action can be offloaded and the fact the code reached the post parsing handling already means that the action can be offloaded.

Fixes 8300f225268b ("net/mlx5e: Create new flow attr for multi table actions")
---
 drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/act.c +1 -2
index af37a8d247a1..2755c25ba324 100644
@@ -145,8 +145,7 @@ mlx5e_tc_act_post_parse(struct mlx5e_tc_act_parse_state *parse_state,
 
 	flow_action_for_each(i, act, flow_action) {
 		tc_act = mlx5e_tc_act_get(act->id, ns_type);
-		if (!tc_act || !tc_act->post_parse ||
-		    !tc_act->can_offload(parse_state, act, i, attr))
+		if (!tc_act || !tc_act->post_parse)
 			continue;
 
 		err = tc_act->post_parse(parse_state, priv, attr);
Signature · 1 line
2.35.1
author maintainer
#13 lore raw

PATCH 12/15 net/mlx5e: Fix trust state reset in reload

In reply to Saeed Mahameed

From: Moshe Tal <[email protected]>

Setting dscp2prio during the driver reload can cause dcb ieee app list to be not empty after the reload finish and as a result to a conflict between the priority trust state reported by the app and the state in the device register.

Reset the dcb ieee app list on initialization in case this is conflicting with the register status.

Fixes 2a5e7a1344f4 ("net/mlx5e: Add dcbnl dscp to priority support")
---
 drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c +10 -0
index d659fe07d464..8ead2c82a52a 100644
@@ -1200,6 +1200,16 @@ static int mlx5e_trust_initialize(struct mlx5e_priv *priv)
 		return err;
 	WRITE_ONCE(priv->dcbx_dp.trust_state, trust_state);
 
+	if (priv->dcbx_dp.trust_state == MLX5_QPTS_TRUST_PCP && priv->dcbx.dscp_app_cnt) {
+		/*
+		 * Align the driver state with the register state.
+		 * Temporary state change is required to enable the app list reset.
+		 */
+		priv->dcbx_dp.trust_state = MLX5_QPTS_TRUST_DSCP;
+		mlx5e_dcbnl_delete_app(priv);
+		priv->dcbx_dp.trust_state = MLX5_QPTS_TRUST_PCP;
+	}
+
 	mlx5e_params_calc_trust_tx_min_inline_mode(priv->mdev, &priv->channels.params,
 						   priv->dcbx_dp.trust_state);
 
Signature · 1 line
2.35.1
author maintainer
#14 lore raw

PATCH 13/15 net/mlx5: Fix deadlock in sync reset flow

In reply to Saeed Mahameed

From: Moshe Shemesh <[email protected]>

The sync reset flow can lead to the following deadlock when poll_sync_reset() is called by timer softirq and waiting on del_timer_sync() for the same timer. Fix that by moving the part of the flow that waits for the timer to reset_reload_work.

It fixes the following kernel Trace:

RIP: 0010:del_timer_sync+0x32/0x40
...
Call Trace:
 <IRQ>
 mlx5_sync_reset_clear_reset_requested+0x26/0x50 [mlx5_core]
 poll_sync_reset.cold+0x36/0x52 [mlx5_core]
 call_timer_fn+0x32/0x130
 __run_timers.part.0+0x180/0x280
 ? tick_sched_handle+0x33/0x60
 ? tick_sched_timer+0x3d/0x80
 ? ktime_get+0x3e/0xa0
 run_timer_softirq+0x2a/0x50
 __do_softirq+0xe1/0x2d6
 ? hrtimer_interrupt+0x136/0x220
 irq_exit+0xae/0xb0
 smp_apic_timer_interrupt+0x7b/0x140
 apic_timer_interrupt+0xf/0x20
 </IRQ>
Fixes 3c5193a87b0f ("net/mlx5: Use del_timer_sync in fw reset flow of halting poll")
---
 .../ethernet/mellanox/mlx5/core/fw_reset.c    | 34 +++++++++----------
 1 file changed, 17 insertions(+), 17 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c +17 -17
index 4aa22dce9b77..ec18d4ccbc11 100644
@@ -155,22 +155,6 @@ static void mlx5_fw_reset_complete_reload(struct mlx5_core_dev *dev)
 	}
 }
 
-static void mlx5_sync_reset_reload_work(struct work_struct *work)
-{
-	struct mlx5_fw_reset *fw_reset = container_of(work, struct mlx5_fw_reset,
-						      reset_reload_work);
-	struct mlx5_core_dev *dev = fw_reset->dev;
-	int err;
-
-	mlx5_enter_error_state(dev, true);
-	mlx5_unload_one(dev);
-	err = mlx5_health_wait_pci_up(dev);
-	if (err)
-		mlx5_core_err(dev, "reset reload flow aborted, PCI reads still not working\n");
-	fw_reset->ret = err;
-	mlx5_fw_reset_complete_reload(dev);
-}
-
 static void mlx5_stop_sync_reset_poll(struct mlx5_core_dev *dev)
 {
 	struct mlx5_fw_reset *fw_reset = dev->priv.fw_reset;
@@ -188,6 +172,23 @@ static void mlx5_sync_reset_clear_reset_requested(struct mlx5_core_dev *dev, boo
 		mlx5_start_health_poll(dev);
 }
 
+static void mlx5_sync_reset_reload_work(struct work_struct *work)
+{
+	struct mlx5_fw_reset *fw_reset = container_of(work, struct mlx5_fw_reset,
+						      reset_reload_work);
+	struct mlx5_core_dev *dev = fw_reset->dev;
+	int err;
+
+	mlx5_sync_reset_clear_reset_requested(dev, false);
+	mlx5_enter_error_state(dev, true);
+	mlx5_unload_one(dev);
+	err = mlx5_health_wait_pci_up(dev);
+	if (err)
+		mlx5_core_err(dev, "reset reload flow aborted, PCI reads still not working\n");
+	fw_reset->ret = err;
+	mlx5_fw_reset_complete_reload(dev);
+}
+
 #define MLX5_RESET_POLL_INTERVAL	(HZ / 10)
 static void poll_sync_reset(struct timer_list *t)
 {
@@ -202,7 +203,6 @@ static void poll_sync_reset(struct timer_list *t)
 
 	if (fatal_error) {
 		mlx5_core_warn(dev, "Got Device Reset\n");
-		mlx5_sync_reset_clear_reset_requested(dev, false);
 		queue_work(fw_reset->wq, &fw_reset->reset_reload_work);
 		return;
 	}
Signature · 1 line
2.35.1
author maintainer
#15 lore raw

PATCH 14/15 net/mlx5: Avoid double clear or set of sync reset requested

In reply to Saeed Mahameed

From: Moshe Shemesh <[email protected]>

Double clear of reset requested state can lead to NULL pointer as it will try to delete the timer twice. This can happen for example on a race between abort from FW and pci error or reset. Avoid such case using test_and_clear_bit() to verify only one time reset requested state clear flow. Similarly use test_and_set_bit() to verify only one time reset requested state set flow.

Fixes 7dd6df329d4c ("net/mlx5: Handle sync reset abort event")
---
 .../ethernet/mellanox/mlx5/core/fw_reset.c    | 28 +++++++++++++------
 1 file changed, 19 insertions(+), 9 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c +19 -9
index ec18d4ccbc11..ca1aba845dd6 100644
@@ -162,14 +162,19 @@ static void mlx5_stop_sync_reset_poll(struct mlx5_core_dev *dev)
 	del_timer_sync(&fw_reset->timer);
 }
 
-static void mlx5_sync_reset_clear_reset_requested(struct mlx5_core_dev *dev, bool poll_health)
+static int mlx5_sync_reset_clear_reset_requested(struct mlx5_core_dev *dev, bool poll_health)
 {
 	struct mlx5_fw_reset *fw_reset = dev->priv.fw_reset;
 
+	if (!test_and_clear_bit(MLX5_FW_RESET_FLAGS_RESET_REQUESTED, &fw_reset->reset_flags)) {
+		mlx5_core_warn(dev, "Reset request was already cleared\n");
+		return -EALREADY;
+	}
+
 	mlx5_stop_sync_reset_poll(dev);
-	clear_bit(MLX5_FW_RESET_FLAGS_RESET_REQUESTED, &fw_reset->reset_flags);
 	if (poll_health)
 		mlx5_start_health_poll(dev);
+	return 0;
 }
 
 static void mlx5_sync_reset_reload_work(struct work_struct *work)
@@ -229,13 +234,17 @@ static int mlx5_fw_reset_set_reset_sync_nack(struct mlx5_core_dev *dev)
 	return mlx5_reg_mfrl_set(dev, MLX5_MFRL_REG_RESET_LEVEL3, 0, 2, false);
 }
 
-static void mlx5_sync_reset_set_reset_requested(struct mlx5_core_dev *dev)
+static int mlx5_sync_reset_set_reset_requested(struct mlx5_core_dev *dev)
 {
 	struct mlx5_fw_reset *fw_reset = dev->priv.fw_reset;
 
+	if (test_and_set_bit(MLX5_FW_RESET_FLAGS_RESET_REQUESTED, &fw_reset->reset_flags)) {
+		mlx5_core_warn(dev, "Reset request was already set\n");
+		return -EALREADY;
+	}
 	mlx5_stop_health_poll(dev, true);
-	set_bit(MLX5_FW_RESET_FLAGS_RESET_REQUESTED, &fw_reset->reset_flags);
 	mlx5_start_sync_reset_poll(dev);
+	return 0;
 }
 
 static void mlx5_fw_live_patch_event(struct work_struct *work)
@@ -264,7 +273,9 @@ static void mlx5_sync_reset_request_event(struct work_struct *work)
 			       err ? "Failed" : "Sent");
 		return;
 	}
-	mlx5_sync_reset_set_reset_requested(dev);
+	if (mlx5_sync_reset_set_reset_requested(dev))
+		return;
+
 	err = mlx5_fw_reset_set_reset_sync_ack(dev);
 	if (err)
 		mlx5_core_warn(dev, "PCI Sync FW Update Reset Ack Failed. Error code: %d\n", err);
@@ -362,7 +373,8 @@ static void mlx5_sync_reset_now_event(struct work_struct *work)
 	struct mlx5_core_dev *dev = fw_reset->dev;
 	int err;
 
-	mlx5_sync_reset_clear_reset_requested(dev, false);
+	if (mlx5_sync_reset_clear_reset_requested(dev, false))
+		return;
 
 	mlx5_core_warn(dev, "Sync Reset now. Device is going to reset.\n");
 
@@ -391,10 +403,8 @@ static void mlx5_sync_reset_abort_event(struct work_struct *work)
 						      reset_abort_work);
 	struct mlx5_core_dev *dev = fw_reset->dev;
 
-	if (!test_bit(MLX5_FW_RESET_FLAGS_RESET_REQUESTED, &fw_reset->reset_flags))
+	if (mlx5_sync_reset_clear_reset_requested(dev, true))
 		return;
-
-	mlx5_sync_reset_clear_reset_requested(dev, true);
 	mlx5_core_warn(dev, "PCI Sync FW Update Reset Aborted.\n");
 }
 
Signature · 1 line
2.35.1
author maintainer
#16 lore raw

PATCH 15/15 net/mlx5: Fix matching on inner TTC

In reply to Saeed Mahameed

From: Mark Bloch <[email protected]>

The cited commits didn't use proper matching on inner TTC as a result distribution of encapsulated packets wasn't symmetric between the physical ports.

Fixes 4c71ce50d2fe ("net/mlx5: Support partial TTC rules")
Fixes 8e25a2bc6687 ("net/mlx5: Lag, add support to create TTC tables for LAG port selection")
---
 drivers/net/ethernet/mellanox/mlx5/core/lag/port_sel.c | 2 +-
 drivers/net/ethernet/mellanox/mlx5/core/lib/fs_ttc.c   | 2 ++
 2 files changed, 3 insertions(+), 1 deletion(-)
drivers/net/ethernet/mellanox/mlx5/core/lag/port_sel.c +1 -1
index a6592f9c3c05..5be322528279 100644
@@ -505,7 +505,7 @@ static int mlx5_lag_create_inner_ttc_table(struct mlx5_lag *ldev)
 	struct ttc_params ttc_params = {};
 
 	mlx5_lag_set_inner_ttc_params(ldev, &ttc_params);
-	port_sel->inner.ttc = mlx5_create_ttc_table(dev, &ttc_params);
+	port_sel->inner.ttc = mlx5_create_inner_ttc_table(dev, &ttc_params);
 	if (IS_ERR(port_sel->inner.ttc))
 		return PTR_ERR(port_sel->inner.ttc);
 
drivers/net/ethernet/mellanox/mlx5/core/lib/fs_ttc.c +2 -0
index b63dec24747a..b78f2ba25c19 100644
@@ -408,6 +408,8 @@ static int mlx5_generate_inner_ttc_table_rules(struct mlx5_core_dev *dev,
 	for (tt = 0; tt < MLX5_NUM_TT; tt++) {
 		struct mlx5_ttc_rule *rule = &rules[tt];
 
+		if (test_bit(tt, params->ignore_dests))
+			continue;
 		rule->rule = mlx5_generate_inner_ttc_rule(dev, ft,
 							  &params->dests[tt],
 							  ttc_rules[tt].etype,
Signature · 1 line
2.35.1