bpf: Don't EFAULT for getsockopt with optval=NULL

[ Upstream commit 00e74ae0863827d944e36e56a4ce1e77e50edb91 ]

Some socket options do getsockopt with optval=NULL to estimate the size
of the final buffer (which is returned via optlen). This breaks BPF
getsockopt assumptions about permitted optval buffer size. Let's enforce
these assumptions only when non-NULL optval is provided.

Fixes: 0d01da6afc54 ("bpf: implement getsockopt and setsockopt hooks")
Reported-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/ZD7Js4fj5YyI2oLd@google.com/T/#mb68daf700f87a9244a15d01d00c3f0e5b08f49f7
Link: https://lore.kernel.org/bpf/20230418225343.553806-2-sdf@google.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpf: Fix race between btf_put and btf_idr walk.

[ Upstream commit acf1c3d68e9a31f10d92bc67ad4673cdae5e8d92 ]

Florian and Eduard reported hard dead lock:
[   58.433327]  _raw_spin_lock_irqsave+0x40/0x50
[   58.433334]  btf_put+0x43/0x90
[   58.433338]  bpf_find_btf_id+0x157/0x240
[   58.433353]  btf_parse_fields+0x921/0x11c0

This happens since btf->refcount can be 1 at the time of btf_put() and
btf_put() will call btf_free_id() which will try to grab btf_idr_lock
and will dead lock.
Avoid the issue by doing btf_put() without locking.

Fixes: 3d78417b60fb ("bpf: Add bpf_btf_find_by_name_kind() helper.")
Fixes: 1e89106da253 ("bpf: Add bpf_core_add_cands() and wire it into bpf_core_apply_relo_insn().")
Reported-by: Florian Westphal <fw@strlen.de>
Reported-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/bpf/20230421014901.70908-1-alexei.starovoitov@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: stmmac:fix system hang when setting up tag_8021q VLAN for DSA ports

[ Upstream commit 35226750f7ab9d49140d95bc7d38a2a9b0f4fdfc ]

The system hang because of dsa_tag_8021q_port_setup()->
stmmac_vlan_rx_add_vid().

I found in stmmac_drv_probe() that cailing pm_runtime_put()
disabled the clock.

First, when the kernel is compiled with CONFIG_PM=y,The stmmac's
resume/suspend is active.

Secondly,stmmac as DSA master,the dsa_tag_8021q_port_setup() function
will callback stmmac_vlan_rx_add_vid when DSA dirver starts. However,
The system is hanged for the stmmac_vlan_rx_add_vid() accesses its
registers after stmmac's clock is closed.

I would suggest adding the pm_runtime_resume_and_get() to the
stmmac_vlan_rx_add_vid().This guarantees that resuming clock output
while in use.

Fixes: b3dcb3127786 ("net: stmmac: correct clocks enabled in stmmac_vlan_rx_kill_vid()")
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Yan Wang <rk.code@outlook.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/mlx5e: Nullify table pointer when failing to create

[ Upstream commit 1b540decd03acd736693aabb6cb46c71fca38ae6 ]

On failing to create promisc flow steering table, the pointer is
returned with an error. Nullify it so unloading the driver won't try to
destroy a non existing table.

Failing to create promisc table may happen over BF devices when the ARM
side is going through a firmware tear down. The host side start a
reload flow. While the driver unloads, it tries to remove the promisc
table. Remove WARN in this state as it is a valid error flow.

Fixes: 1c46d7409f30 ("net/mlx5e: Optimize promiscuous mode")
Signed-off-by: Aya Levin <ayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/mlx5: Use recovery timeout on sync reset flow

[ Upstream commit dfad99750c0f83b0242572a573afa2c055f85b36 ]

Use the same timeout for sync reset flow and health recovery flow, since
the former involves driver's recovery from firmware reset, which is
similar to health recovery. Otherwise, in some cases, such as a firmware
upgrade on the DPU, the firmware pre-init bit may not be ready within
current timeout and the driver will abort loading back after reset.

Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Fixes: 37ca95e62ee2 ("net/mlx5: Increase FW pre-init timeout for health recovery")
Reviewed-by: Maher Sanalla <msanalla@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

Revert "net/mlx5: Remove "recovery" arg from mlx5_load_one() function"

[ Upstream commit 21608a2cf38e9f743636b5f86507ffbca18ecee7 ]

This reverts commit 5977ac3910f1cbaf44dca48179118b25c206ac29.

Revert this patch as we need the "recovery" arg back in mlx5_load_one()
function. This arg will be used in the next patch for using recovery
timeout during sync reset flow.

Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Maher Sanalla <msanalla@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Stable-dep-of: dfad99750c0f ("net/mlx5: Use recovery timeout on sync reset flow")
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/mlx5: Suspend auxiliary devices only in case of PCI device suspend

[ Upstream commit 72ed5d5624af384eaf74d84915810d54486a75e2 ]

The original behavior introduced by commit c6acd629eec7 ("net/mlx5e: Add
support for devlink-port in non-representors mode") correctly
re-instantiated uplink devlink port and related netdevice during devlink
reload. However with migration to auxiliary devices, this behaviour
changed.

Restore the original behaviour and tear down auxiliary devices
completely during devlink reload.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Stable-dep-of: dfad99750c0f ("net/mlx5: Use recovery timeout on sync reset flow")
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/mlx5: Remove "recovery" arg from mlx5_load_one() function

[ Upstream commit 5977ac3910f1cbaf44dca48179118b25c206ac29 ]

mlx5_load_one() is always called with recovery==false, so remove the
unneeded function arg.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Stable-dep-of: dfad99750c0f ("net/mlx5: Use recovery timeout on sync reset flow")
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/mlx5e: Fix error flow in representor failing to add vport rx rule

[ Upstream commit 0a6b069cc60d68d33b4f6e7dd7f1adc3ec749766 ]

On representor init rx error flow the flow steering pointer is being
released so mlx5e_attach_netdev() doesn't have a valid fs pointer
in its error flow. Make sure the pointer is nullified when released
and add a check in mlx5e_fs_cleanup() to verify fs is not null
as representor cleanup callback would be called anyway.

Fixes: af8bbf730068 ("net/mlx5e: Convert mlx5e_flow_steering member of mlx5e_priv to pointer")
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Maor Dickman <maord@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/mlx5: E-switch, Don't destroy indirect table in split rule

[ Upstream commit 4c8189302567f75099a336b0efcff8291ec86ff4 ]

Source port rewrite (forward to ovs internal port or statck device) isn't
supported in the rule of split action. So there is no indirect table in
split rule. The cited commit destroyes indirect table in split rule. The
indirect table for other rules will be destroyed wrongly. It will cause
traffic loss.

Fix it by removing the destroy function in split rule. And also remove
the destroy function in error flow.

Fixes: 10742efc20a4 ("net/mlx5e: VF tunnel TX traffic offloading")
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Maor Dickman <maord@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/mlx5: E-switch, Create per vport table based on devlink encap mode

[ Upstream commit fd745f4c0abe41ebb09d11bf622b054a0f3e7b49 ]

Currently when creating per vport table, create flags are hardcoded.
Devlink encap mode is set based on user input and HW capability.
Create per vport table based on devlink encap mode.

Fixes: c796bb7cd230 ("net/mlx5: E-switch, Generalize per vport table API")
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Maor Dickman <maord@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net/mlx5e: Don't clone flow post action attributes second time

[ Upstream commit e9fce818fe003b6c527f25517b9ac08eb4661b5d ]

The code already clones post action attributes in
mlx5e_clone_flow_attr_for_post_act(). Creating another copy in
mlx5e_tc_post_act_add() is a erroneous leftover from original
implementation. Instead, assign handle->attribute to post_attr provided by
the caller. Note that cloning the attribute second time is not just
wasteful but also causes issues like second copy not being properly updated
in neigh update code which leads to following use-after-free:

Feb 21 09:02:00 c-237-177-40-045 kernel: BUG: KASAN: use-after-free in mlx5_cmd_set_fte+0x200d/0x24c0 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  kasan_report+0xbb/0x1a0
Feb 21 09:02:00 c-237-177-40-045 kernel:  kasan_save_stack+0x1e/0x40
Feb 21 09:02:00 c-237-177-40-045 kernel:  kasan_set_track+0x21/0x30
Feb 21 09:02:00 c-237-177-40-045 kernel:  __kasan_kmalloc+0x7a/0x90
Feb 21 09:02:00 c-237-177-40-045 kernel:  kasan_save_stack+0x1e/0x40
Feb 21 09:02:00 c-237-177-40-045 kernel:  kasan_set_track+0x21/0x30
Feb 21 09:02:00 c-237-177-40-045 kernel:  kasan_save_free_info+0x2a/0x40
Feb 21 09:02:00 c-237-177-40-045 kernel:  ____kasan_slab_free+0x11a/0x1b0
Feb 21 09:02:00 c-237-177-40-045 kernel: page dumped because: kasan: bad access detected
Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5_core 0000:08:00.0: mlx5_cmd_out_err:803:(pid 8833): SET_FLOW_TABLE_ENTRY(0x936) op_mod(0x0) failed, status bad resource state(0x9), syndrome (0xf2ff71), err(-22)
Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5_core 0000:08:00.0 enp8s0f0: Failed to add post action rule
Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5_core 0000:08:00.0: mlx5e_tc_encap_flows_add:190:(pid 8833): Failed to update flow post acts, -22
Feb 21 09:02:00 c-237-177-40-045 kernel: Call Trace:
Feb 21 09:02:00 c-237-177-40-045 kernel:  <TASK>
Feb 21 09:02:00 c-237-177-40-045 kernel:  dump_stack_lvl+0x57/0x7d
Feb 21 09:02:00 c-237-177-40-045 kernel:  print_report+0x170/0x471
Feb 21 09:02:00 c-237-177-40-045 kernel:  ? mlx5_cmd_set_fte+0x200d/0x24c0 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  kasan_report+0xbb/0x1a0
Feb 21 09:02:00 c-237-177-40-045 kernel:  ? mlx5_cmd_set_fte+0x200d/0x24c0 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5_cmd_set_fte+0x200d/0x24c0 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  ? __module_address.part.0+0x62/0x200
Feb 21 09:02:00 c-237-177-40-045 kernel:  ? mlx5_cmd_stub_create_flow_table+0xd0/0xd0 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  ? __raw_spin_lock_init+0x3b/0x110
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5_cmd_create_fte+0x80/0xb0 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  add_rule_fg+0xe80/0x19c0 [mlx5_core]
--
Feb 21 09:02:00 c-237-177-40-045 kernel: Allocated by task 13476:
Feb 21 09:02:00 c-237-177-40-045 kernel:  kasan_save_stack+0x1e/0x40
Feb 21 09:02:00 c-237-177-40-045 kernel:  kasan_set_track+0x21/0x30
Feb 21 09:02:00 c-237-177-40-045 kernel:  __kasan_kmalloc+0x7a/0x90
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5_packet_reformat_alloc+0x7b/0x230 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5e_tc_tun_create_header_ipv4+0x977/0xf10 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5e_attach_encap+0x15b4/0x1e10 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  post_process_attr+0x305/0xa30 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5e_tc_add_fdb_flow+0x4c0/0xcf0 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  __mlx5e_add_fdb_flow+0x7cf/0xe90 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5e_configure_flower+0xcaa/0x4b90 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5e_rep_setup_tc_cls_flower+0x99/0x1b0 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5e_rep_setup_tc_cb+0x133/0x1e0 [mlx5_core]
--
Feb 21 09:02:00 c-237-177-40-045 kernel: Freed by task 8833:
Feb 21 09:02:00 c-237-177-40-045 kernel:  kasan_save_stack+0x1e/0x40
Feb 21 09:02:00 c-237-177-40-045 kernel:  kasan_set_track+0x21/0x30
Feb 21 09:02:00 c-237-177-40-045 kernel:  kasan_save_free_info+0x2a/0x40
Feb 21 09:02:00 c-237-177-40-045 kernel:  ____kasan_slab_free+0x11a/0x1b0
Feb 21 09:02:00 c-237-177-40-045 kernel:  __kmem_cache_free+0x1de/0x400
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5_packet_reformat_dealloc+0xad/0x100 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5e_tc_encap_flows_del+0x3c0/0x500 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5e_rep_update_flows+0x40c/0xa80 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  mlx5e_rep_neigh_update+0x473/0x7a0 [mlx5_core]
Feb 21 09:02:00 c-237-177-40-045 kernel:  process_one_work+0x7c2/0x1310
Feb 21 09:02:00 c-237-177-40-045 kernel:  worker_thread+0x59d/0xec0
Feb 21 09:02:00 c-237-177-40-045 kernel:  kthread+0x28f/0x330

Fixes: 8300f225268b ("net/mlx5e: Create new flow attr for multi table actions")
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ixgbe: Enable setting RSS table to default values

[ Upstream commit e85d3d55875f7a1079edfbc4e4e98d6f8aea9ac7 ]

ethtool uses `ETHTOOL_GRXRINGS` to compute how many queues are supported
by RSS. The driver should return the smaller of either:
  - The maximum number of RSS queues the device supports, OR
  - The number of RX queues configured

Prior to this change, running `ethtool -X $iface default` fails if the
number of queues configured is larger than the number supported by RSS,
even though changing the queue count correctly resets the flowhash to
use all supported queues.

Other drivers (for example, i40e) will succeed but the flow hash will
reset to support the maximum number of queues supported by RSS, even if
that amount is smaller than the configured amount.

Prior to this change:

$ sudo ethtool -L eth1 combined 20
$ sudo ethtool -x eth1
RX flow hash indirection table for eth1 with 20 RX ring(s):
    0:      0     1     2     3     4     5     6     7
    8:      8     9    10    11    12    13    14    15
   16:      0     1     2     3     4     5     6     7
   24:      8     9    10    11    12    13    14    15
   32:      0     1     2     3     4     5     6     7
...

You can see that the flowhash was correctly set to use the maximum
number of queues supported by the driver (16).

However, asking the NIC to reset to "default" fails:

$ sudo ethtool -X eth1 default
Cannot set RX flow hash configuration: Invalid argument

After this change, the flowhash can be reset to default which will use
all of the available RSS queues (16) or the configured queue count,
whichever is smaller.

Starting with eth1 which has 10 queues and a flowhash distributing to
all 10 queues:

$ sudo ethtool -x eth1
RX flow hash indirection table for eth1 with 10 RX ring(s):
    0:      0     1     2     3     4     5     6     7
    8:      8     9     0     1     2     3     4     5
   16:      6     7     8     9     0     1     2     3
...

Increasing the queue count to 48 resets the flowhash to distribute to 16
queues, as it did before this patch:

$ sudo ethtool -L eth1 combined 48
$ sudo ethtool -x eth1
RX flow hash indirection table for eth1 with 16 RX ring(s):
    0:      0     1     2     3     4     5     6     7
    8:      8     9    10    11    12    13    14    15
   16:      0     1     2     3     4     5     6     7
...

Due to the other bugfix in this series, the flowhash can be set to use
queues 0-5:

$ sudo ethtool -X eth1 equal 5
$ sudo ethtool -x eth1
RX flow hash indirection table for eth1 with 16 RX ring(s):
    0:      0     1     2     3     4     0     1     2
    8:      3     4     0     1     2     3     4     0
   16:      1     2     3     4     0     1     2     3
...

Due to this bugfix, the flowhash can be reset to default and use 16
queues:

$ sudo ethtool -X eth1 default
$ sudo ethtool -x eth1
RX flow hash indirection table for eth1 with 16 RX ring(s):
    0:      0     1     2     3     4     5     6     7
    8:      8     9    10    11    12    13    14    15
   16:      0     1     2     3     4     5     6     7
...

Fixes: 91cd94bfe4f0 ("ixgbe: add basic support for setting and getting nfc controls")
Signed-off-by: Joe Damato <jdamato@fastly.com>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ixgbe: Allow flow hash to be set via ethtool

[ Upstream commit 4f3ed1293feb9502dc254b05802faf1ad3317ac6 ]

ixgbe currently returns `EINVAL` whenever the flowhash it set by ethtool
because the ethtool code in the kernel passes a non-zero value for hfunc
that ixgbe should allow.

When ethtool is called with `ETHTOOL_SRXFHINDIR`,
`ethtool_set_rxfh_indir` will call ixgbe's set_rxfh function
with `ETH_RSS_HASH_NO_CHANGE`. This value should be accepted.

When ethtool is called with `ETHTOOL_SRSSH`, `ethtool_set_rxfh` will
call ixgbe's set_rxfh function with `rxfh.hfunc`, which appears to be
hardcoded in ixgbe to always be `ETH_RSS_HASH_TOP`. This value should
also be accepted.

Before this patch:

$ sudo ethtool -L eth1 combined 10
$ sudo ethtool -X eth1 default
Cannot set RX flow hash configuration: Invalid argument

After this patch:

$ sudo ethtool -L eth1 combined 10
$ sudo ethtool -X eth1 default
$ sudo ethtool -x eth1
RX flow hash indirection table for eth1 with 10 RX ring(s):
    0:      0     1     2     3     4     5     6     7
    8:      8     9     0     1     2     3     4     5
   16:      6     7     8     9     0     1     2     3
   24:      4     5     6     7     8     9     0     1
   ...

Fixes: 1c7cf0784e4d ("ixgbe: support for ethtool set_rxfh")
Signed-off-by: Joe Damato <jdamato@fastly.com>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: fw: fix memory leak in debugfs

[ Upstream commit 3d90d2f4a018fe8cfd65068bc6350b6222be4852 ]

Fix a memory leak that occurs when reading the fw_info
file all the way, since we return NULL indicating no
more data, but don't free the status tracking object.

Fixes: 36dfe9ac6e8b ("iwlwifi: dump api version in yaml format")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230418122405.239e501b3b8d.I4268f87809ef91209cbcd748eee0863195e70fa2@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

netfilter: conntrack: fix wrong ct->timeout value

[ Upstream commit 73db1b8f2bb6725b7391e85aab41fdf592b3c0c1 ]

(struct nf_conn)->timeout is an interval before the conntrack
confirmed.  After confirmed, it becomes a timestamp.

It is observed that timeout of an unconfirmed conntrack:
- Set by calling ctnetlink_change_timeout(). As a result,
  `nfct_time_stamp` was wrongly added to `ct->timeout` twice.
- Get by calling ctnetlink_dump_timeout(). As a result,
  `nfct_time_stamp` was wrongly subtracted.

Call Trace:
 <TASK>
 dump_stack_lvl
 ctnetlink_dump_timeout
 __ctnetlink_glue_build
 ctnetlink_glue_build
 __nfqnl_enqueue_packet
 nf_queue
 nf_hook_slow
 ip_mc_output
 ? __pfx_ip_finish_output
 ip_send_skb
 ? __pfx_dst_output
 udp_send_skb
 udp_sendmsg
 ? __pfx_ip_generic_getfrag
 sock_sendmsg

Separate the 2 cases in:
- Setting `ct->timeout` in __nf_ct_set_timeout().
- Getting `ct->timeout` in ctnetlink_dump_timeout().

Pablo appends:

Update ctnetlink to set up the timeout _after_ the IPS_CONFIRMED flag is
set on, otherwise conntrack creation via ctnetlink breaks.

Note that the problem described in this patch occurs since the
introduction of the nfnetlink_queue conntrack support, select a
sufficiently old Fixes: tag for -stable kernel to pick up this fix.

Fixes: a4b4766c3ceb ("netfilter: nfnetlink_queue: rename related to nfqueue attaching conntrack info")
Signed-off-by: Tzung-Bi Shih <tzungbi@kernel.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

netfilter: conntrack: restore IPS_CONFIRMED out of nf_conntrack_hash_check_insert()

[ Upstream commit 2cdaa3eefed83082923cf219c8b6a314e622da74 ]

e6d57e9ff0ae ("netfilter: conntrack: fix rmmod double-free race")
consolidates IPS_CONFIRMED bit set in nf_conntrack_hash_check_insert().
However, this breaks ctnetlink:

 # conntrack -I -p tcp --timeout 123 --src 1.2.3.4 --dst 5.6.7.8 --state ESTABLISHED --sport 1 --dport 4 -u SEEN_REPLY
 conntrack v1.4.6 (conntrack-tools): Operation failed: Device or resource busy

This is a partial revert of the aforementioned commit to restore
IPS_CONFIRMED.

Fixes: e6d57e9ff0ae ("netfilter: conntrack: fix rmmod double-free race")
Reported-by: Stéphane Graber <stgraber@stgraber.org>
Tested-by: Stéphane Graber <stgraber@stgraber.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: mvm: check firmware response size

[ Upstream commit 13513cec93ac9902d0b896976d8bab3758a9881c ]

Check the firmware response size for responses to the
memory read/write command in debugfs before using it.

Fixes: 2b55f43f8e47 ("iwlwifi: mvm: Add mem debugfs entry")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230417113648.0d56fcaf68ee.I70e9571f3ed7263929b04f8fabad23c9b999e4ea@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: connac: fix txd multicast rate setting

[ Upstream commit 3d2892e05086d09aecf14ea64b2debbf495e313c ]

The vif->bss_conf.mcast_rate should be applied to multicast data frame
only.

Fixes: 182071cdd594 ("mt76: connac: move connac2_mac_write_txwi in mt76_connac module")
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: mt7921e: stop chip reset worker in unregister hook

[ Upstream commit 3d78c46423c6567ed25ca033e086865b1b4d5ae1 ]

If the chip reset worker is triggered during the remove process, the chip
DMA may not be properly pushed back to the idle state. This can lead to
corruption of the DMA flow due to the chip reset. Therefore, it is
necessary to stop the chip reset before the DMA is finalized.

To avoid resetting the chip after the reset worker is cancelled, use
__mt7921_mcu_drv_pmctrl() instead of mt7921_mcu_drv_pmctrl(). It is safe to
ignore the pm mutex because the pm worker and wake worker have already been
cancelled.

Fixes: 033ae79b3830 ("mt76: mt7921: refactor init.c to be bus independent")
Co-developed-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Co-developed-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Co-developed-by: Wang Zhao <wang.zhao@mediatek.com>
Signed-off-by: Wang Zhao <wang.zhao@mediatek.com>
Signed-off-by: Quan Zhou <quan.zhou@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: mt7921e: improve reliability of dma reset

[ Upstream commit 87714bf6ed1589813e473db5471e6e9857755764 ]

The hardware team has advised the driver that it is necessary to first put
WFDMA into an idle state before resetting the WFDMA. Otherwise, the WFDMA
may enter an unknown state where it cannot be polled with the right state
successfully. To ensure that the DMA can work properly while a stressful
cold reboot test was being made, we have reordered the programming sequence
in the driver based on the hardware team's guidance.

The patch would modify the WFDMA disabling flow from

"DMA reset -> disabling DMASHDL -> disabling WFDMA -> polling and waiting
until DMA idle" to "disabling WFDMA -> polling and waiting for DMA idle ->
disabling DMASHDL -> DMA reset.

Where he polling and waiting until WFDMA is idle is coordinated with the
operation of disabling WFDMA. Even while WFDMA is being disabled, it can
still handle Tx/Rx requests. The additional polling allows sufficient time
for WFDMA to process the last T/Rx request. When the idle state of WFDMA is
reached, it is a reliable indication that DMASHDL is also idle to ensure it
is safe to disable it and perform the DMA reset.

Fixes: 0a1059d0f060 ("mt76: mt7921: move mt7921_dma_reset in dma.c")
Co-developed-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Co-developed-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Co-developed-by: Wang Zhao <wang.zhao@mediatek.com>
Signed-off-by: Wang Zhao <wang.zhao@mediatek.com>
Signed-off-by: Quan Zhou <quan.zhou@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: mt7921: fix missing unwind goto in `mt7921u_probe`

[ Upstream commit 5c47cdebbaeb7724df6f9f46917c93e53f791547 ]

`mt7921u_dma_init` can only return zero or negative number according to its
definition. When it returns non-zero number, there exists an error and this
function should handle this error rather than return directly.

Fixes: 0d2afe09fad5 ("mt76: mt7921: add mt7921u driver")
Signed-off-by: Jiefeng Li <jiefeng_li@hust.edu.cn>
Reviewed-by: Dongliang Mu <dzm91@hust.edu.cn>
Reviewed-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

mt76: mt7921: fix kernel panic by accessing unallocated eeprom.data

[ Upstream commit 12db28c3ef31f719bd18fa186a40bb152e6a527c ]

The MT7921 driver no longer uses eeprom.data, but the relevant code has not
been removed completely since
commit 16d98b548365 ("mt76: mt7921: rely on mcu_get_nic_capability").
This could result in potential invalid memory access.

To fix the kernel panic issue in mt7921, it is necessary to avoid accessing
unallocated eeprom.data which can lead to invalid memory access.

Furthermore, it is possible to entirely eliminate the
mt7921_mcu_parse_eeprom function and solely depend on
mt7921_mcu_parse_response to divide the RxD header.

[2.702735] BUG: kernel NULL pointer dereference, address: 0000000000000550
[2.702740] #PF: supervisor write access in kernel mode
[2.702741] #PF: error_code(0x0002) - not-present page
[2.702743] PGD 0 P4D 0
[2.702747] Oops: 0002 [#1] PREEMPT SMP NOPTI
[2.702755] RIP: 0010:mt7921_mcu_parse_response+0x147/0x170 [mt7921_common]
[2.702758] RSP: 0018:ffffae7c00fef828 EFLAGS: 00010286
[2.702760] RAX: ffffa367f57be024 RBX: ffffa367cc7bf500 RCX: 0000000000000000
[2.702762] RDX: 0000000000000550 RSI: 0000000000000000 RDI: ffffa367cc7bf500
[2.702763] RBP: ffffae7c00fef840 R08: ffffa367cb167000 R09: 0000000000000005
[2.702764] R10: 0000000000000000 R11: ffffffffc04702e4 R12: ffffa367e8329f40
[2.702766] R13: 0000000000000000 R14: 0000000000000001 R15: ffffa367e8329f40
[2.702768] FS:  000079ee6cf20c40(0000) GS:ffffa36b2f940000(0000) knlGS:0000000000000000
[2.702769] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[2.702775] CR2: 0000000000000550 CR3: 00000001233c6004 CR4: 0000000000770ee0
[2.702776] PKRU: 55555554
[2.702777] Call Trace:
[2.702782]  mt76_mcu_skb_send_and_get_msg+0xc3/0x11e [mt76 <HASH:1bc4 5>]
[2.702785]  mt7921_run_firmware+0x241/0x853 [mt7921_common <HASH:6a2f 6>]
[2.702789]  mt7921e_mcu_init+0x2b/0x56 [mt7921e <HASH:d290 7>]
[2.702792]  mt7921_register_device+0x2eb/0x5a5 [mt7921_common <HASH:6a2f 6>]
[2.702795]  ? mt7921_irq_tasklet+0x1d4/0x1d4 [mt7921e <HASH:d290 7>]
[2.702797]  mt7921_pci_probe+0x2d6/0x319 [mt7921e <HASH:d290 7>]
[2.702799]  pci_device_probe+0x9f/0x12a

Fixes: 16d98b548365 ("mt76: mt7921: rely on mcu_get_nic_capability")
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: fix 6GHz high channel not be scanned

[ Upstream commit 23792cedaff02351b57bddd8957fc917fa88f2e0 ]

mt76 scan command only support 64 channels currently. If the
channel count is larger than 64(for 2+5+6GHz), some channels will
not be scanned. Hence change the scan type to full channel scan
in case of the command cannot include proper list for chip.

Fixes: 399090ef9605 ("mt76: mt76_connac: move hw_scan and sched_scan routine in mt76_connac_mcu module")
Reported-by: Ben Greear <greearb@candelatech.com>
Tested-by: Isaac Konikoff <konikofi@candelatech.com>
Signed-off-by: Ming Yen Hsieh <mingyen.hsieh@mediatek.com>
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: mt7921e: fix probe timeout after reboot

[ Upstream commit c397fc1e6365a2a9e5540a85b2c1d4ea412aa0e2 ]

In system warm reboot scene, due to the polling timeout(now 1000us)
is too short to wait dma idle in time, it may make driver probe fail
with error code -ETIMEDOUT. Meanwhile, we also found the dma may take
around 70ms to enter idle state. Change the polling idle timeout to
100ms to avoid the probabilistic probe fail.

Tested pass with 5000 times warm reboot on x86 platform.

[4.477496] pci 0000:01:00.0: attach allowed to drvr mt7921e [internal device]
[4.478306] mt7921e 0000:01:00.0: ASIC revision: 79610010
[4.480063] mt7921e: probe of 0000:01:00.0 failed with error -110

Fixes: 0a1059d0f060 ("mt76: mt7921: move mt7921_dma_reset in dma.c")
Signed-off-by: Quan Zhou <quan.zhou@mediatek.com>
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: add flexible polling wait-interval support

[ Upstream commit 35effe6c0c24adcf0f732bb1c3d75573d4c88e63 ]

The default waiting unit is 10ms and the value is too much for
data path related control. Provide a new API mt76_poll_msec_tick()
to support different cases, such as 1ms polling waiting kick.

Reviewed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Stable-dep-of: c397fc1e6365 ("wifi: mt76: mt7921e: fix probe timeout after reboot")
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: handle failure of vzalloc in mt7615_coredump_work

[ Upstream commit 9e47dd9f64a47ae00ca0123017584c37209ee900 ]

vzalloc may fails, dump might be null and will cause
illegal address access later.

Link: https://lore.kernel.org/all/Y%2Fy5Asxw3T3m4jCw@lore-desk
Fixes: d2bf7959d9c0 ("mt76: mt7663: introduce coredump support")
Signed-off-by: Kang Chen <void0red@gmail.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: mt76: mt7915: expose device tree match table

[ Upstream commit 90fb69212c60e26ef70ed0e8532b116c7649ac88 ]

On MT7986 the WiFi driver currently does not get automatically loaded,
requiring manual modprobing because the device tree compatibles are not
exported into metadata.

Add the missing MODULE_DEVICE_TABLE macro to fix this.

Fixes: 99ad32a4ca3a2 ("mt76: mt7915: add support for MT7986")
Signed-off-by: Lorenz Brun <lorenz@brun.one>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: make the loop for card preparation effective

[ Upstream commit 28965ec0b5d9112585f725660e2ff13218505ace ]

Since we didn't reset t to 0, only the first iteration of the loop
did checked the ready bit several times.
From the second iteration and on, we just tested the bit once and
continued to the next iteration.

Reported-and-tested-by: Lorenzo Zolfanelli <lorenzo@zolfa.nl>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=216452
Fixes: 289e5501c314 ("iwlwifi: fix the preparation of the card")
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230416154301.615b683ab9c8.Ic52c3229d3345b0064fa34263293db095d88daf8@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

io_uring/rsrc: use nospec'ed indexes

[ Upstream commit 953c37e066f05a3dca2d74643574b8dfe8a83983 ]

We use array_index_nospec() for registered buffer indexes, but don't use
it while poking into rsrc tags, fix that.

Fixes: 634d00df5e1cf ("io_uring: add full-fledged dynamic buffers support")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/f02fafc5a9c0dd69be2b0618c38831c078232ff0.1681395792.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

jdb2: Don't refuse invalidation of already invalidated buffers

[ Upstream commit bd159398a2d2234de07d310132865706964aaaa7 ]

When invalidating buffers under the partial tail page,
jbd2_journal_invalidate_folio() returns -EBUSY if the buffer is part of
the committing transaction as we cannot safely modify buffer state.
However if the buffer is already invalidated (due to previous
invalidation attempts from ext4_wait_for_tail_page_commit()), there's
nothing to do and there's no point in returning -EBUSY. This fixes
occasional warnings from ext4_journalled_invalidate_folio() triggered by
generic/051 fstest when blocksize < pagesize.

Fixes: 53e872681fed ("ext4: fix deadlock in journal_unmap_buffer()")
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20230329154950.19720-1-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: fw: move memset before early return

[ Upstream commit 8ce437dd5b2e4adef13aa4ecce07392f9966b1ab ]

Clang static analysis reports this representative issue
dbg.c:1455:6: warning: Branch condition evaluates to
a garbage value
  if (!rxf_data.size)
       ^~~~~~~~~~~~~~

This check depends on iwl_ini_get_rxf_data() to clear
rxf_data but the function can return early without
doing the clear.  So move the memset before the early
return.

Fixes: cc9b6012d34b ("iwlwifi: yoyo: use hweight_long instead of bit manipulating")
Signed-off-by: Tom Rix <trix@redhat.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230414130637.872a7175f1ff.I33802a77a91998276992b088fbe25f61c87c33ac@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: mvm: initialize seq variable

[ Upstream commit 11e94d2bcd88dea5d9ce99555b6b172f5232d3e2 ]

Clang static analysis reports this issue
d3.c:567:22: warning: The left operand of '>' is
  a garbage value
  if (seq.tkip.iv32 > cur_rx_iv32)
      ~~~~~~~~~~~~~ ^

seq is never initialized. Call ieee80211_get_key_rx_seq() to
initialize seq.

Fixes: 0419e5e672d6 ("iwlwifi: mvm: d3: separate TKIP data from key iteration")
Signed-off-by: Tom Rix <trix@redhat.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230414130637.6dd372f84f93.If1f708c90e6424a935b4eba3917dfb7582e0dd0a@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: yoyo: Fix possible division by zero

[ Upstream commit ba30415118eee374a08b39a0460a1d1e52c24a25 ]

Don't allow buffer allocation TLV with zero req_size since it
leads later to division by zero in iwl_dbg_tlv_alloc_fragments().
Also, NPK/SRAM locations are allowed to have zero buffer req_size,
don't discard them.

Fixes: a9248de42464 ("iwlwifi: dbg_ini: add TLV allocation new API support")
Signed-off-by: Daniel Gabay <daniel.gabay@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230413213309.5d6688ed74d8.I5c2f3a882b50698b708d54f4524dc5bdf11e3d32@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: yoyo: skip dump correctly on hw error

[ Upstream commit 11195ab0d6f3202cf7af1a4c69570f59c377d8ad ]

When NIC is in a bad state, reading data will return 28 bits as
0xa5a5a5a and the lowest 4 bits are not fixed value.

Mask these bits in a few places to skip the dump correctly.

Fixes: 89639e06d0f3 ("iwlwifi: yoyo: support for new DBGI_SRAM region")
Signed-off-by: Daniel Gabay <daniel.gabay@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230413213309.df6c0663179d.I36d8487b2419c6fefa65e5514855d94327c3b1eb@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: mvm: don't drop unencrypted MCAST frames

[ Upstream commit 8e5a26360cbe29b896b6758518280d41c3704d43 ]

MCAST frames are filtered out by the driver if we are not
authorized yet.
Fix it to filter out only protected frames.

Fixes: 147eb05f24e6 ("iwlwifi: mvm: always tell the firmware to accept MCAST frames in BSS")
Signed-off-by: Ayala Beker <ayala.beker@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230413213309.9cedcc27db60.I8fb7057981392660da482dd215e85c15946d3f4b@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

md/raid10: don't call bio_start_io_acct twice for bio which experienced read error

[ Upstream commit 7cddb055bfda5f7b0be931e8ea750fc28bc18a27 ]

handle_read_error() will resumit r10_bio by raid10_read_request(), which
will call bio_start_io_acct() again, while bio_end_io_acct() will only
be called once.

Fix the problem by don't account io again from handle_read_error().

Fixes: 528bc2cf2fcc ("md/raid10: enable io accounting")
Suggested-by: Song Liu <song@kernel.org>
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Signed-off-by: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/r/20230314012258.2395894-1-yukuai1@huaweicloud.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

md/raid10: fix memleak of md thread

[ Upstream commit f0ddb83da3cbbf8a1f9087a642c448ff52ee9abd ]

In raid10_run(), if setup_conf() succeed and raid10_run() failed before
setting 'mddev->thread', then in the error path 'conf->thread' is not
freed.

Fix the problem by setting 'mddev->thread' right after setup_conf().

Fixes: 43a521238aca ("md-cluster: choose correct label when clustered layout is not supported")
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Signed-off-by: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/r/20230310073855.1337560-7-yukuai1@huaweicloud.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

md/raid10: fix memleak for 'conf->bio_split'

[ Upstream commit c9ac2acde53f5385de185bccf6aaa91cf9ac1541 ]

In the error path of raid10_run(), 'conf' need be freed, however,
'conf->bio_split' is missed and memory will be leaked.

Since there are 3 places to free 'conf', factor out a helper to fix the
problem.

Fixes: fc9977dd069e ("md/raid10: simplify the splitting of requests.")
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Signed-off-by: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/r/20230310073855.1337560-6-yukuai1@huaweicloud.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

md/raid10: fix leak of 'r10bio->remaining' for recovery

[ Upstream commit 26208a7cffd0c7cbf14237ccd20c7270b3ffeb7e ]

raid10_sync_request() will add 'r10bio->remaining' for both rdev and
replacement rdev. However, if the read io fails, recovery_request_write()
returns without issuing the write io, in this case, end_sync_request()
is only called once and 'remaining' is leaked, cause an io hang.

Fix the problem by decreasing 'remaining' according to if 'bio' and
'repl_bio' is valid.

Fixes: 24afd80d99f8 ("md/raid10: handle recovery of replacement devices.")
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Signed-off-by: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/r/20230310073855.1337560-5-yukuai1@huaweicloud.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

md/raid10: fix task hung in raid10d

[ Upstream commit 72c215ed8731c88b2d7e09afc51fffc207ae47b8 ]

commit fe630de009d0 ("md/raid10: avoid deadlock on recovery.") allowed
normal io and sync io to exist at the same time. Task hung will occur as
below:

T1                      T2 T3 T4
raid10d
 handle_read_error
  allow_barrier
   conf->nr_pending--
    -> 0
                        //submit sync io
                        raid10_sync_request
                         raise_barrier
  ->will not be blocked
  ...
//submit to drivers
  raid10_read_request
   wait_barrier
    conf->nr_pending++
     -> 1
//retry read fail
raid10_end_read_request
 reschedule_retry
  add to retry_list
  conf->nr_queued++
   -> 1
//sync io fail
end_sync_read
 __end_sync_read
  reschedule_retry
   add to retry_list
                    conf->nr_queued++
     -> 2
 ...
 handle_read_error
 get form retry_list
 conf->nr_queued--
  freeze_array
   wait nr_pending == nr_queued+1
        ->1       ->2
   //task hung

retry read and sync io will be added to retry_list(nr_queued->2) if they
fails. raid10d() called handle_read_error() and hung in freeze_array().
nr_queued will not decrease because raid10d is blocked, nr_pending will
not increase because conf->barrier is not released.

Fix it by moving allow_barrier() after raid10_read_request().
raise_barrier() will wait for nr_waiting to become 0. Therefore, sync io
and regular io will not be issued at the same time.

Also remove the check of nr_queued in stop_waiting_barrier. It can be 0
but don't need to be blocking. Remove the check for MD_RECOVERY_RUNNING as
the check is redundent.

Fixes: fe630de009d0 ("md/raid10: avoid deadlock on recovery.")
Signed-off-by: Li Nan <linan122@huawei.com>
Signed-off-by: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/r/20230222041000.3341651-2-linan666@huaweicloud.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

f2fs: fix to check return value of inc_valid_block_count()

[ Upstream commit 935fc6fa6466cf18dd72dd1518ebc7bfc4cd58a4 ]

In __replace_atomic_write_block(), we missed to check return value
of inc_valid_block_count(), for extreme testcase that f2fs image is
run out of space, it may cause inconsistent status in between SIT
table and total valid block count.

Cc: Daeho Jeong <daehojeong@google.com>
Fixes: 3db1de0e582c ("f2fs: change the current atomic write way")
Signed-off-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

f2fs: fix to check return value of f2fs_do_truncate_blocks()

[ Upstream commit b851ee6ba3cc212641e622ebcf92b950c7bafa07 ]

Otherwise, if truncation on cow_inode failed, remained data may
pollute current transaction of atomic write.

Cc: Daeho Jeong <daehojeong@google.com>
Fixes: a46bebd502fe ("f2fs: synchronize atomic write aborts")
Signed-off-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpf, sockmap: Revert buggy deadlock fix in the sockhash and sockmap

[ Upstream commit 8c5c2a4898e3d6bad86e29d471e023c8a19ba799 ]

syzbot reported a splat and bisected it to recent commit ed17aa92dc56 ("bpf,
sockmap: fix deadlocks in the sockhash and sockmap"):

  [...]
  WARNING: CPU: 1 PID: 9280 at kernel/softirq.c:376 __local_bh_enable_ip+0xbe/0x130 kernel/softirq.c:376
  Modules linked in:
  CPU: 1 PID: 9280 Comm: syz-executor.1 Not tainted 6.2.0-syzkaller-13249-gd319f344561d #0
  Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 03/30/2023
  RIP: 0010:__local_bh_enable_ip+0xbe/0x130 kernel/softirq.c:376
  [...]
  Call Trace:
  <TASK>
  spin_unlock_bh include/linux/spinlock.h:395 [inline]
  sock_map_del_link+0x2ea/0x510 net/core/sock_map.c:165
  sock_map_unref+0xb0/0x1d0 net/core/sock_map.c:184
  sock_hash_delete_elem+0x1ec/0x2a0 net/core/sock_map.c:945
  map_delete_elem kernel/bpf/syscall.c:1536 [inline]
  __sys_bpf+0x2edc/0x53e0 kernel/bpf/syscall.c:5053
  __do_sys_bpf kernel/bpf/syscall.c:5166 [inline]
  __se_sys_bpf kernel/bpf/syscall.c:5164 [inline]
  __x64_sys_bpf+0x79/0xc0 kernel/bpf/syscall.c:5164
  do_syscall_x64 arch/x86/entry/common.c:50 [inline]
  do_syscall_64+0x39/0xb0 arch/x86/entry/common.c:80
  entry_SYSCALL_64_after_hwframe+0x63/0xcd
  RIP: 0033:0x7fe8f7c8c169
  </TASK>
  [...]

Revert for now until we have a proper solution.

Fixes: ed17aa92dc56 ("bpf, sockmap: fix deadlocks in the sockhash and sockmap")
Reported-by: syzbot+49f6cef45247ff249498@syzkaller.appspotmail.com
Cc: Hsin-Wei Hung <hsinweih@uci.edu>
Cc: Xin Liu <liuxin350@huawei.com>
Cc: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/000000000000f1db9605f939720e@google.com/
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: mvm: don't set CHECKSUM_COMPLETE for unsupported protocols

[ Upstream commit 217f3c52f00d3419ecdd38a99a7eceecb11679b2 ]

On Bz devices, CHECKSUM_COMPLETE was set for unsupported protocols
which results in a warning. Fix it.

Fixes: b6f5b647f694 ("iwlwifi: mvm: handle RX checksum on Bz devices")
Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230413102635.a2a35286f0ca.I50daa9445a6465514c44f5096c32adef64beba5f@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: trans: don't trigger d3 interrupt twice

[ Upstream commit 277f56a141fc54ef7f9e09dba65fb2e12021411d ]

When the IPC registers are used for sleep control, setting
the IPC sleep bit already triggers an interrupt to the fw, so
there is no need to also set the doorbell. Setting also the
doorbell triggers the sleep interrupt twice which lead to
an assert.

Fixes: af08571d3925 ("iwlwifi: pcie: support Bz suspend/resume trigger")
Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230413102635.b5f2f6e44d38.I4cb5b6ad4914db47a714e731c4c8b4db679cabce@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: debug: fix crash in __iwl_err()

[ Upstream commit 634c7b1bd08ca322537ab389f8cbd7bb543b5e45 ]

In __iwl_err(), if we rate-limit the message away, then
vaf.va is still NULL-initialized by the time we get to
the tracing code, which then crashes. When it doesn't
get rate-limited out, it's still wrong to reuse the old
args2 that was already printed, which is why we bother
making a copy in the first place.

Assign vaf.va properly to fix this.

Fixes: e5f1cc98cc1b ("iwlwifi: allow rate-limited error messages")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230413102635.e27134c6bcd4.Ib3894cd2ba7a5ad5e75912a7634f146ceaa569e2@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

blk-mq: don't plug for head insertions in blk_execute_rq_nowait

[ Upstream commit 50947d7fe9fa6abe3ddc40769dfb02a51c58edb6 ]

Plugs never insert at head, so don't plug for head insertions.

Fixes: 1c2d2fff6dc0 ("block: wire-up support for passthrough plugging")
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Link: https://lore.kernel.org/r/20230413064057.707578-2-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/bpf: Fix leaked bpf_link in get_stackid_cannot_attach

[ Upstream commit c1e07a80cf23d3a6e96172bc9a73bfa912a9fcbc ]

skel->links.oncpu is leaked in one case. This causes test perf_branches
fails when it runs after get_stackid_cannot_attach:

./test_progs -t get_stackid_cannot_attach,perf_branches
84      get_stackid_cannot_attach:OK
test_perf_branches_common:PASS:test_perf_branches_load 0 nsec
test_perf_branches_common:PASS:attach_perf_event 0 nsec
test_perf_branches_common:PASS:set_affinity 0 nsec
check_good_sample:FAIL:output not valid no valid sample from prog
146/1   perf_branches/perf_branches_hw:FAIL
146/2   perf_branches/perf_branches_no_hw:OK
146     perf_branches:FAIL

All error logs:
test_perf_branches_common:PASS:test_perf_branches_load 0 nsec
test_perf_branches_common:PASS:attach_perf_event 0 nsec
test_perf_branches_common:PASS:set_affinity 0 nsec
check_good_sample:FAIL:output not valid no valid sample from prog
146/1   perf_branches/perf_branches_hw:FAIL
146     perf_branches:FAIL
Summary: 1/1 PASSED, 0 SKIPPED, 1 FAILED

Fix this by adding the missing bpf_link__destroy().

Fixes: 346938e9380c ("selftests/bpf: Add get_stackid_cannot_attach")
Signed-off-by: Song Liu <song@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230412210423.900851-3-song@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/bpf: Use read_perf_max_sample_freq() in perf_event_stackmap

[ Upstream commit de6d014a09bf12a9a8959d60c0a1d4a41d394a89 ]

Currently, perf_event sample period in perf_event_stackmap is set too low
that the test fails randomly. Fix this by using the max sample frequency,
from read_perf_max_sample_freq().

Move read_perf_max_sample_freq() to testing_helpers.c. Replace the CHECK()
with if-printf, as CHECK is not available in testing_helpers.c.

Fixes: 1da4864c2b20 ("selftests/bpf: Add callchain_stackid")
Signed-off-by: Song Liu <song@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230412210423.900851-2-song@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

nvme-fcloop: fix "inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage"

[ Upstream commit 4f86a6ff6fbd891232dda3ca97fd1b9630b59809 ]

fcloop_fcp_op() could be called from flush request's ->end_io(flush_end_io) in
which the spinlock of fq->mq_flush_lock is grabbed with irq saved/disabled.

So fcloop_fcp_op() can't call spin_unlock_irq(&tfcp_req->reqlock) simply
which enables irq unconditionally.

Fixes the warning by switching to spin_lock_irqsave()/spin_unlock_irqrestore()

Fixes: c38dbbfab1bc ("nvme-fcloop: fix inconsistent lock state warnings")
Reported-by: Yi Zhang <yi.zhang@redhat.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Ewan D. Milne <emilne@redhat.com>
Tested-by: Yi Zhang <yi.zhang@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nvme: fix async event trace event

[ Upstream commit 6622b76fe922b94189499a90ccdb714a4a8d0773 ]

Mixing AER Event Type and Event Info has masking clashes. Just print the
event type, but also include the event info of the AER result in the
trace.

Fixes: 09bd1ff4b15143b ("nvme-core: add async event trace helper")
Reported-by: Nate Thornton <nate.thornton@samsung.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Minwoo Im <minwoo.im@samsung.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nvmet: fix I/O Command Set specific Identify Controller

[ Upstream commit a5a6ab0950b46ab1ef4a5c83c80234018b81b38a ]

For an identify command with cns set to NVME_ID_CNS_CS_CTRL, the NVMe
2.0 specification states that:

If the I/O Command Set specified by the CSI field does not have an
Identify Controller data structure, then the controller shall return
a zero filled data structure. If the host requests a data structure for
an I/O Command Set that the controller does not support, the controller
shall abort the command with a status code of Invalid Field in Command.

However, the current implementation of this identify command in
nvmet_execute_identify() only handles the ZNS command set, returning an
error for the NVM command set, which is not compliant with the
specifications as we do support this command set.

Fix this by:
1) Renaming nvmet_execute_identify_cns_cs_ctrl() to
   nvmet_execute_identify_ctrl_zns() to continue handling the
   ZNS command set as is.
2) Introduce a nvmet_execute_identify_ctrl_ns() helper to handle the
   NVM command set, returning a zero filled nvme_id_ctrl_nvm data
   structure.
3) Modify nvmet_execute_identify() to call these helpers based on
   the csi specified, returning an error for unsupported command sets.

Fixes: aaf2e048af27 ("nvmet: add ZBD over ZNS backend support")
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Tested-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nvmet: fix Identify Active Namespace ID list handling

[ Upstream commit 97416f67d55fb8b866ff1815ca7ef26b6dfa6a5e ]

The identify command with cns set to NVME_ID_CNS_NS_ACTIVE_LIST does
not depend on the command set. The execution of this command should
thus not look at the csi field specified in the command. Simplify
nvmet_execute_identify() to directly call
nvmet_execute_identify_nslist() without the csi switch-case.

Fixes: ab5d0b38c047 ("nvmet: add Command Set Identifier support")
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Tested-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nvmet: fix Identify Controller handling

[ Upstream commit 62904b3b333e7f3c0f879dc3513295eee5765c9f ]

The identify command with cns set to NVME_ID_CNS_CTRL does not depend on
the command set. The execution of this command should thus not look at
the csi specified in the command. Simplify nvmet_execute_identify() to
directly call nvmet_execute_identify_ctrl() without the csi switch-case.

Fixes: ab5d0b38c047 ("nvmet: add Command Set Identifier support")
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Tested-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nvmet: fix Identify Namespace handling

[ Upstream commit 8c098aa00118c35108f0c19bd3cdc45e11574948 ]

The identify command with cns set to NVME_ID_CNS_NS does not directly
depend on the command set. The NVMe specifications is rather confusing
here as it appears that this command only applies to the NVM command
set. However, footnote 8 of Figure 273 in the NVMe 2.0 base
specifications clearly state that this command applies to NVM command
sets that support logical blocks, that is, NVM and ZNS. Both the NVM and
ZNS command set specifications also list this identify as mandatory.

The command handling should thus not look at the csi field since it is
defined as unused for this command. Given that we do not support the
KV command set, simply remove the csi switch-case for that command
handling and call directly nvmet_execute_identify_ns() in
nvmet_execute_identify().

Fixes: ab5d0b38c047 ("nvmet: add Command Set Identifier support")
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Tested-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nvmet: fix error handling in nvmet_execute_identify_cns_cs_ns()

[ Upstream commit ab76e7206b672b2e8818cb121a04506956d6b223 ]

Nvme specifications state that:

If the I/O Command Set associated with the namespace identified by the
NSID field does not support the Identify Namespace data structure
specified by the CSI field, the controller shall abort the command with
a status code of Invalid Field in Command.

In other words, if nvmet_execute_identify_cns_cs_ns() is called for a
target with a block device that is not zoned, we should not return any
data and set the status to NVME_SC_INVALID_FIELD.

While at it, it is also better to revalidate the ns block devie *before*
checking if the block device is zoned, to ensure that
nvmet_execute_identify_cns_cs_ns() operates against updated device
characteristics.

Fixes: aaf2e048af27 ("nvmet: add ZBD over ZNS backend support")
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpf, sockmap: fix deadlocks in the sockhash and sockmap

[ Upstream commit ed17aa92dc56b6d8883e4b7a8f1c6fbf5ed6cd29 ]

When huang uses sched_switch tracepoint, the tracepoint
does only one thing in the mounted ebpf program, which
deletes the fixed elements in sockhash ([0])

It seems that elements in sockhash are rarely actively
deleted by users or ebpf program. Therefore, we do not
pay much attention to their deletion. Compared with hash
maps, sockhash only provides spin_lock_bh protection.
This causes it to appear to have self-locking behavior
in the interrupt context.

  [0]:https://lore.kernel.org/all/CABcoxUayum5oOqFMMqAeWuS8+EzojquSOSyDA3J_2omY=2EeAg@mail.gmail.com/

Reported-by: Hsin-Wei Hung <hsinweih@uci.edu>
Fixes: 604326b41a6f ("bpf, sockmap: convert to generic sk_msg interface")
Signed-off-by: Xin Liu <liuxin350@huawei.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/r/20230406122622.109978-1-liuxin350@huawei.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: ath11k: fix writing to unintended memory region

[ Upstream commit 756a7f90878f0866fd2fe167ef37e90b47326b96 ]

While initializing spectral, the magic value is getting written to the
invalid memory address leading to random boot-up crash. This occurs
due to the incorrect index increment in ath11k_dbring_fill_magic_value
function. Fix it by replacing the existing logic with memset32 to ensure
there is no invalid memory access.

Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.4.0.1-01838-QCAHKSWPL_SILICONZ-1

Fixes: d3d358efc553 ("ath11k: add spectral/CFR buffer validation support")
Signed-off-by: P Praneesh <quic_ppranees@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://lore.kernel.org/r/20230321052900.16895-1-quic_ppranees@quicinc.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: ethernet: stmmac: dwmac-rk: fix optional phy regulator handling

[ Upstream commit db21973263f8c56750cb610f1d5e8bee00a513b9 ]

The usual devm_regulator_get() call already handles "optional"
regulators by returning a valid dummy and printing a warning
that the dummy regulator should be described properly. This
code open coded the same behaviour, but masked any errors that
are not -EPROBE_DEFER and is quite noisy.

This change effectively unmasks and propagates regulators errors
not involving -ENODEV, downgrades the error print to warning level
if no regulator is specified and captures the probe defer message
for /sys/kernel/debug/devices_deferred.

Fixes: 2e12f536635f ("net: stmmac: dwmac-rk: Use standard devicetree property for phy regulator")
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: ethernet: stmmac: dwmac-rk: rework optional clock handling

[ Upstream commit ea449f7fa0bf3fcd02e04a770b9ff707bf5e8f96 ]

The clock requesting code is quite repetitive. Fix this by requesting
the clocks via devm_clk_bulk_get_optional. The optional variant has been
used, since this is effectively what the old code did. The exact clocks
required depend on the platform and configuration. As a side effect
this change adds correct -EPROBE_DEFER handling.

Suggested-by: Jakub Kicinski <kuba@kernel.org>
Suggested-by: Andrew Lunn <andrew@lunn.ch>
Fixes: 7ad269ea1a2b ("GMAC: add driver for Rockchip RK3288 SoCs integrated GMAC")
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: lpfc: Fix ioremap issues in lpfc_sli4_pci_mem_setup()

[ Upstream commit 91a0c0c1413239d0548b5aac4c82f38f6d53a91e ]

When if_type equals zero and pci_resource_start(pdev, PCI_64BIT_BAR4)
returns false, drbl_regs_memmap_p is not remapped. This passes a NULL
pointer to iounmap(), which can trigger a WARN() on certain arches.

When if_type equals six and pci_resource_start(pdev, PCI_64BIT_BAR4)
returns true, drbl_regs_memmap_p may has been remapped and
ctrl_regs_memmap_p is not remapped. This is a resource leak and passes a
NULL pointer to iounmap().

To fix these issues, we need to add null checks before iounmap(), and
change some goto labels.

Fixes: 1351e69fc6db ("scsi: lpfc: Add push-to-adapter support to sli4")
Signed-off-by: Shuchang Li <lishuchang@hust.edu.cn>
Link: https://lore.kernel.org/r/20230404072133.1022-1-lishuchang@hust.edu.cn
Reviewed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpf/btf: Fix is_int_ptr()

[ Upstream commit 91f2dc6838c19342f7f2993627c622835cc24890 ]

When tracing a kernel function with arg type is u32*, btf_ctx_access()
would report error: arg2 type INT is not a struct.

The commit bb6728d75611 ("bpf: Allow access to int pointer arguments
in tracing programs") added support for int pointer, but did not skip
modifiers before checking it's type. This patch fixes it.

Fixes: bb6728d75611 ("bpf: Allow access to int pointer arguments in tracing programs")
Co-developed-by: Chengming Zhou <zhouchengming@bytedance.com>
Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
Signed-off-by: Feng Zhou <zhoufeng.zf@bytedance.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/bpf/20230410085908.98493-2-zhoufeng.zf@bytedance.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: iwlwifi: fix duplicate entry in iwl_dev_info_table

[ Upstream commit fc3c2f0ed86b65dff4b6844c59c597b977cae533 ]

There're two identical entries for ax1650 device in
iwl_dev_info_table. Remove one of the duplicate entries.

Fixes: 953e66a7238b ("iwlwifi: add new ax1650 killer device")
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Link: https://lore.kernel.org/r/20230410140721.897683-2-gregory.greenman@intel.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

f2fs: fix to avoid use-after-free for cached IPU bio

[ Upstream commit 5cdb422c839134273866208dad5360835ddb9794 ]

xfstest generic/019 reports a bug:

kernel BUG at mm/filemap.c:1619!
RIP: 0010:folio_end_writeback+0x8a/0x90
Call Trace:
 end_page_writeback+0x1c/0x60
 f2fs_write_end_io+0x199/0x420
 bio_endio+0x104/0x180
 submit_bio_noacct+0xa5/0x510
 submit_bio+0x48/0x80
 f2fs_submit_write_bio+0x35/0x300
 f2fs_submit_merged_ipu_write+0x2a0/0x2b0
 f2fs_write_single_data_page+0x838/0x8b0
 f2fs_write_cache_pages+0x379/0xa30
 f2fs_write_data_pages+0x30c/0x340
 do_writepages+0xd8/0x1b0
 __writeback_single_inode+0x44/0x370
 writeback_sb_inodes+0x233/0x4d0
 __writeback_inodes_wb+0x56/0xf0
 wb_writeback+0x1dd/0x2d0
 wb_workfn+0x367/0x4a0
 process_one_work+0x21d/0x430
 worker_thread+0x4e/0x3c0
 kthread+0x103/0x130
 ret_from_fork+0x2c/0x50

The root cause is: after cp_error is set, f2fs_submit_merged_ipu_write()
in f2fs_write_single_data_page() tries to flush IPU bio in cache, however
f2fs_submit_merged_ipu_write() missed to check validity of @bio parameter,
result in submitting random cached bio which belong to other IO context,
then it will cause use-after-free issue, fix it by adding additional
validity check.

Fixes: 0b20fcec8651 ("f2fs: cache global IPU bio")
Signed-off-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

xsk: Fix unaligned descriptor validation

[ Upstream commit d769ccaf957fe7391f357c0a923de71f594b8a2b ]

Make sure unaligned descriptors that straddle the end of the UMEM are
considered invalid. Currently, descriptor validation is broken for
zero-copy mode which only checks descriptors at page granularity.
For example, descriptors in zero-copy mode that overrun the end of the
UMEM but not a page boundary are (incorrectly) considered valid. The
UMEM boundary check needs to happen before the page boundary and
contiguity checks in xp_desc_crosses_non_contig_pg(). Do this check in
xp_unaligned_validate_desc() instead like xp_check_unaligned() already
does.

Fixes: 2b43470add8c ("xsk: Introduce AF_XDP buffer allocation API")
Signed-off-by: Kal Conley <kal.conley@dectris.com>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://lore.kernel.org/r/20230405235920.7305-2-kal.conley@dectris.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: drbg - Only fail when jent is unavailable in FIPS mode

[ Upstream commit 686cd976b6ddedeeb1a1fb09ba53a891d3cc9a03 ]

When jent initialisation fails for any reason other than ENOENT,
the entire drbg fails to initialise, even when we're not in FIPS
mode.  This is wrong because we can still use the kernel RNG when
we're not in FIPS mode.

Change it so that it only fails when we are in FIPS mode.

Fixes: 57225e679788 ("crypto: drbg - Use callback API for random readiness")
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Reviewed-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpftool: Fix bug for long instructions in program CFG dumps

[ Upstream commit 67cf52cdb6c8fa6365d29106555dacf95c9fd374 ]

When dumping the control flow graphs for programs using the 16-byte long
load instruction, we need to skip the second part of this instruction
when looking for the next instruction to process. Otherwise, we end up
printing "BUG_ld_00" from the kernel disassembler in the CFG.

Fixes: efcef17a6d65 ("tools: bpftool: generate .dot graph from CFG information")
Signed-off-by: Quentin Monnet <quentin@isovalent.com>
Link: https://lore.kernel.org/r/20230405132120.59886-3-quentin@isovalent.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/bpf: Wait for receive in cg_storage_multi test

[ Upstream commit 5af607a861d43ffff830fc1890033e579ec44799 ]

In some cases the loopback latency might be large enough, causing
the assertion on invocations to be run before ingress prog getting
executed. The assertion would fail and the test would flake.

This can be reliably reproduced by arbitrarily increasing the
loopback latency (thanks to [1]):
  tc qdisc add dev lo root handle 1: htb default 12
  tc class add dev lo parent 1:1 classid 1:12 htb rate 20kbps ceil 20kbps
  tc qdisc add dev lo parent 1:12 netem delay 100ms

Fix this by waiting on the receive end, instead of instantly
returning to the assert. The call to read() will wait for the
default SO_RCVTIMEO timeout of 3 seconds provided by
start_server().

[1] https://gist.github.com/kstevens715/4598301

Reported-by: Martin KaFai Lau <martin.lau@linux.dev>
Link: https://lore.kernel.org/bpf/9c5c8b7e-1d89-a3af-5400-14fde81f4429@linux.dev/
Fixes: 3573f384014f ("selftests/bpf: Test CGROUP_STORAGE behavior on shared egress + ingress")
Acked-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: YiFei Zhu <zhuyifei@google.com>
Link: https://lore.kernel.org/r/20230405193354.1956209-1-zhuyifei@google.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests: xsk: Deflakify STATS_RX_DROPPED test

[ Upstream commit 68e7322142f5e731af222892d384d311835db0f1 ]

Fix flaky STATS_RX_DROPPED test. The receiver calls getsockopt after
receiving the last (valid) packet which is not the final packet sent in
the test (valid and invalid packets are sent in alternating fashion with
the final packet being invalid). Since the last packet may or may not
have been dropped already, both outcomes must be allowed.

This issue could also be fixed by making sure the last packet sent is
valid. This alternative is left as an exercise to the reader (or the
benevolent maintainers of this file).

This problem was quite visible on certain setups. On one machine this
failure was observed 50% of the time.

Also, remove a redundant assignment of pkt_stream->nb_pkts. This field
is already initialized by __pkt_stream_alloc.

Fixes: 27e934bec35b ("selftests: xsk: make stat tests not spin on getsockopt")
Signed-off-by: Kal Conley <kal.conley@dectris.com>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://lore.kernel.org/r/20230403120400.31018-1-kal.conley@dectris.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests: xsk: Disable IPv6 on VETH1

[ Upstream commit f2b50f17268390567bc0e95642170d88f336c8f4 ]

This change fixes flakiness in the BIDIRECTIONAL test:

    # [is_pkt_valid] expected length [60], got length [90]
    not ok 1 FAIL: SKB BUSY-POLL BIDIRECTIONAL

When IPv6 is enabled, the interface will periodically send MLDv1 and
MLDv2 packets. These packets can cause the BIDIRECTIONAL test to fail
since it uses VETH0 for RX.

For other tests, this was not a problem since they only receive on VETH1
and IPv6 was already disabled on VETH0.

Fixes: a89052572ebb ("selftests/bpf: Xsk selftests framework")
Signed-off-by: Kal Conley <kal.conley@dectris.com>
Link: https://lore.kernel.org/r/20230405082905.6303-1-kal.conley@dectris.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests: xsk: Use correct UMEM size in testapp_invalid_desc

[ Upstream commit 7a2050df244e2c9a4e90882052b7907450ad10ed ]

Avoid UMEM_SIZE macro in testapp_invalid_desc which is incorrect when
the frame size is not XSK_UMEM__DEFAULT_FRAME_SIZE. Also remove the
macro since it's no longer being used.

Fixes: 909f0e28207c ("selftests: xsk: Add tests for 2K frame size")
Signed-off-by: Kal Conley <kal.conley@dectris.com>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://lore.kernel.org/r/20230403145047.33065-2-kal.conley@dectris.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: qrtr: correct types of trace event parameters

[ Upstream commit 054fbf7ff8143d35ca7d3bb5414bb44ee1574194 ]

The arguments passed to the trace events are of type unsigned int,
however the signature of the events used __le32 parameters.

I may be missing the point here, but sparse flagged this and it
does seem incorrect to me.

  net/qrtr/ns.c: note: in included file (through include/trace/trace_events.h, include/trace/define_trace.h, include/trace/events/qrtr.h):
  ./include/trace/events/qrtr.h:11:1: warning: cast to restricted __le32
  ./include/trace/events/qrtr.h:11:1: warning: restricted __le32 degrades to integer
  ./include/trace/events/qrtr.h:11:1: warning: restricted __le32 degrades to integer
  ... (a lot more similar warnings)
  net/qrtr/ns.c:115:47:    expected restricted __le32 [usertype] service
  net/qrtr/ns.c:115:47:    got unsigned int service
  net/qrtr/ns.c:115:61: warning: incorrect type in argument 2 (different base types)
  ... (a lot more similar warnings)

Fixes: dfddb54043f0 ("net: qrtr: Add tracepoint support")
Reviewed-by: Manivannan Sadhasivam <mani@kernel.org>
Signed-off-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20230402-qrtr-trace-types-v1-1-92ad55008dd3@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

f2fs: fix iostat lock protection

[ Upstream commit 144f1cd40bf91fb3ac1d41806470756ce774f389 ]

Made iostat lock irq safe to avoid potentinal deadlock.

Deadlock scenario:
f2fs_attr_store
  -> f2fs_sbi_store
  -> _sbi_store
  -> spin_lock(sbi->iostat_lock)
    <interrupt request>
    -> scsi_end_request
    -> bio_endio
    -> f2fs_dio_read_end_io
    -> f2fs_update_iostat
    -> spin_lock_irqsave(sbi->iostat_lock)  ===> Dead lock here

Fixes: 61803e984307 ("f2fs: fix iostat related lock protection")
Fixes: a1e09b03e6f5 ("f2fs: use iomap for direct I/O")
Signed-off-by: Qilin Tan <qilin.tan@mediatek.com>
Reviewed-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: rt2x00: Fix memory leak when handling surveys

[ Upstream commit cbef9a83c51dfcb07f77cfa6ac26f53a1ea86f49 ]

When removing a rt2x00 device, its associated channel surveys
are not freed, causing a memory leak observable with kmemleak:

unreferenced object 0xffff9620f0881a00 (size 512):
  comm "systemd-udevd", pid 2290, jiffies 4294906974 (age 33.768s)
  hex dump (first 32 bytes):
    70 44 12 00 00 00 00 00 92 8a 00 00 00 00 00 00  pD..............
    00 00 00 00 00 00 00 00 ab 87 01 00 00 00 00 00  ................
  backtrace:
    [<ffffffffb0ed858b>] __kmalloc+0x4b/0x130
    [<ffffffffc1b0f29b>] rt2800_probe_hw+0xc2b/0x1380 [rt2800lib]
    [<ffffffffc1a9496e>] rt2800usb_probe_hw+0xe/0x60 [rt2800usb]
    [<ffffffffc1ae491a>] rt2x00lib_probe_dev+0x21a/0x7d0 [rt2x00lib]
    [<ffffffffc1b3b83e>] rt2x00usb_probe+0x1be/0x980 [rt2x00usb]
    [<ffffffffc05981e2>] usb_probe_interface+0xe2/0x310 [usbcore]
    [<ffffffffb13be2d5>] really_probe+0x1a5/0x410
    [<ffffffffb13be5c8>] __driver_probe_device+0x78/0x180
    [<ffffffffb13be6fe>] driver_probe_device+0x1e/0x90
    [<ffffffffb13be972>] __driver_attach+0xd2/0x1c0
    [<ffffffffb13bbc57>] bus_for_each_dev+0x77/0xd0
    [<ffffffffb13bd2a2>] bus_add_driver+0x112/0x210
    [<ffffffffb13bfc6c>] driver_register+0x5c/0x120
    [<ffffffffc0596ae8>] usb_register_driver+0x88/0x150 [usbcore]
    [<ffffffffb0c011c4>] do_one_initcall+0x44/0x220
    [<ffffffffb0d6134c>] do_init_module+0x4c/0x220

Fix this by freeing the channel surveys on device removal.

Tested with a RT3070 based USB wireless adapter.

Fixes: 5447626910f5 ("rt2x00: save survey for every channel visited")
Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Acked-by: Stanislaw Gruszka <stf_xl@wp.pl>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20230330215637.4332-1-W_Armin@gmx.de
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: hisi_sas: Handle NCQ error when IPTT is valid

[ Upstream commit bb544224da77b96b2c11a13872bf91ede1e015be ]

If an NCQ error occurs when the IPTT is valid and slot->abort flag is set
in completion path, sas_task_abort() will be called to abort only one NCQ
command now, and the host would be set to SHOST_RECOVERY state. But this
may not kick-off EH Immediately until other outstanding QCs timeouts. As a
result, the host may remain in the SHOST_RECOVERY state for up to 30
seconds, such as follows:

[7972317.645234] hisi_sas_v3_hw 0000:74:04.0: erroneous completion iptt=3264 task=00000000466116b8 dev id=2 sas_addr=0x5000000000000502 CQ hdr: 0x1883 0x20cc0 0x40000 0x20420000 Error info: 0x0 0x0 0x200000 0x0
[7972341.508264] sas: Enter sas_scsi_recover_host busy: 32 failed: 32
[7972341.984731] sas: --- Exit sas_scsi_recover_host: busy: 0 failed: 32 tries: 1

All NCQ commands that are in the queue should be aborted when an NCQ error
occurs in this scenario.

Fixes: 05d91b557af9 ("scsi: hisi_sas: Directly trigger SCSI error handling for completion errors")
Signed-off-by: Xingui Yang <yangxingui@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Link: https://lore.kernel.org/r/1679283265-115066-3-git-send-email-chenxiang66@hisilicon.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: libsas: Add sas_ata_device_link_abort()

[ Upstream commit 44112922674b94a7d699dfff6307fc830018df7c ]

Similar to how AHCI handles NCQ errors in ahci_error_intr() ->
ata_port_abort() -> ata_do_link_abort(), add an NCQ error handler for LLDDs
to call to initiate a link abort.

This will mark all outstanding QCs as failed and kick-off EH.

Note:

A "force reset" argument is added for drivers which require the ATA error
handling to always reset the device.

A driver may require this feature for when SATA device per-SCSI cmnd
resources are only released during reset for ATA EH. As such, we need an
option to force reset to be done, regardless of what any EH autopsy
decides.

The SATA device FIS fields are set to indicate a device error from
ata_eh_analyze_tf().

Suggested-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Suggested-by: Niklas Cassel <niklas.cassel@wdc.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Link: https://lore.kernel.org/r/1665998435-199946-2-git-send-email-john.garry@huawei.com
Tested-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Tested-by: Niklas Cassel <niklas.cassel@wdc.com> # pm80xx
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Stable-dep-of: bb544224da77 ("scsi: hisi_sas: Handle NCQ error when IPTT is valid")
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: rtlwifi: fix incorrect error codes in rtl_debugfs_set_write_reg()

[ Upstream commit 5dbe1f8eb8c5ac69394400a5b86fd81775e96c43 ]

If there is a failure during copy_from_user or user-provided data buffer is
invalid, rtl_debugfs_set_write_reg should return negative error code instead
of a positive value count.

Fix this bug by returning correct error code. Moreover, the check of buffer
against null is removed since it will be handled by copy_from_user.

Fixes: 610247f46feb ("rtlwifi: Improve debugging by using debugfs")
Signed-off-by: Wei Chen <harperchen1110@gmail.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20230326054217.93492-1-harperchen1110@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

wifi: rtlwifi: fix incorrect error codes in rtl_debugfs_set_write_rfreg()

[ Upstream commit 905a9241e4e8c15d2c084fee916280514848fe35 ]

If there is a failure during copy_from_user or user-provided data buffer
is invalid, rtl_debugfs_set_write_rfreg should return negative error code
instead of a positive value count.

Fix this bug by returning correct error code. Moreover, the check of buffer
against null is removed since it will be handled by copy_from_user.

Fixes: 610247f46feb ("rtlwifi: Improve debugging by using debugfs")
Signed-off-by: Wei Chen <harperchen1110@gmail.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20230326053138.91338-1-harperchen1110@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: sa2ul - Select CRYPTO_DES

[ Upstream commit 8832023efd20966e29944dac92118dfbf1fa1bc0 ]

The SA2UL Crypto driver provides support for couple of
DES3 algos "cbc(des3_ede)" and "ecb(des3_ede)", and enabling
the crypto selftest throws the following errors (as seen on
K3 J721E SoCs):
  saul-crypto 4e00000.crypto: Error allocating fallback algo cbc(des3_ede)
  alg: skcipher: failed to allocate transform for cbc-des3-sa2ul: -2
  saul-crypto 4e00000.crypto: Error allocating fallback algo ecb(des3_ede)
  alg: skcipher: failed to allocate transform for ecb-des3-sa2ul: -2

Fix this by selecting CRYPTO_DES which was missed while
adding base driver support.

Fixes: 7694b6ca649f ("crypto: sa2ul - Add crypto driver")
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Jayesh Choudhary <j-choudhary@ti.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

crypto: caam - Clear some memory in instantiate_rng

[ Upstream commit 9c19fb86a8cb2ee82a832c95e139f29ea05c4d08 ]

According to the comment at the end of the 'for' loop just a few lines
below, it looks needed to clear 'desc'.

So it should also be cleared for the first iteration.

Move the memset() to the beginning of the loop to be safe.

Fixes: 281922a1d4f5 ("crypto: caam - add support for SEC v5.x RNG4")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Gaurav Jain <gaurav.jain@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sashal@kernel.org>

f2fs: fix scheduling while atomic in decompression path

[ Upstream commit 1aa161e43106d46ca8e9a86f4aa28d420258134b ]

[   16.945668][    C0] Call trace:
[   16.945678][    C0]  dump_backtrace+0x110/0x204
[   16.945706][    C0]  dump_stack_lvl+0x84/0xbc
[   16.945735][    C0]  __schedule_bug+0xb8/0x1ac
[   16.945756][    C0]  __schedule+0x724/0xbdc
[   16.945778][    C0]  schedule+0x154/0x258
[   16.945793][    C0]  bit_wait_io+0x48/0xa4
[   16.945808][    C0]  out_of_line_wait_on_bit+0x114/0x198
[   16.945824][    C0]  __sync_dirty_buffer+0x1f8/0x2e8
[   16.945853][    C0]  __f2fs_commit_super+0x140/0x1f4
[   16.945881][    C0]  f2fs_commit_super+0x110/0x28c
[   16.945898][    C0]  f2fs_handle_error+0x1f4/0x2f4
[   16.945917][    C0]  f2fs_decompress_cluster+0xc4/0x450
[   16.945942][    C0]  f2fs_end_read_compressed_page+0xc0/0xfc
[   16.945959][    C0]  f2fs_handle_step_decompress+0x118/0x1cc
[   16.945978][    C0]  f2fs_read_end_io+0x168/0x2b0
[   16.945993][    C0]  bio_endio+0x25c/0x2c8
[   16.946015][    C0]  dm_io_dec_pending+0x3e8/0x57c
[   16.946052][    C0]  clone_endio+0x134/0x254
[   16.946069][    C0]  bio_endio+0x25c/0x2c8
[   16.946084][    C0]  blk_update_request+0x1d4/0x478
[   16.946103][    C0]  scsi_end_request+0x38/0x4cc
[   16.946129][    C0]  scsi_io_completion+0x94/0x184
[   16.946147][    C0]  scsi_finish_command+0xe8/0x154
[   16.946164][    C0]  scsi_complete+0x90/0x1d8
[   16.946181][    C0]  blk_done_softirq+0xa4/0x11c
[   16.946198][    C0]  _stext+0x184/0x614
[   16.946214][    C0]  __irq_exit_rcu+0x78/0x144
[   16.946234][    C0]  handle_domain_irq+0xd4/0x154
[   16.946260][    C0]  gic_handle_irq.33881+0x5c/0x27c
[   16.946281][    C0]  call_on_irq_stack+0x40/0x70
[   16.946298][    C0]  do_interrupt_handler+0x48/0xa4
[   16.946313][    C0]  el1_interrupt+0x38/0x68
[   16.946346][    C0]  el1h_64_irq_handler+0x20/0x30
[   16.946362][    C0]  el1h_64_irq+0x78/0x7c
[   16.946377][    C0]  finish_task_switch+0xc8/0x3d8
[   16.946394][    C0]  __schedule+0x600/0xbdc
[   16.946408][    C0]  preempt_schedule_common+0x34/0x5c
[   16.946423][    C0]  preempt_schedule+0x44/0x48
[   16.946438][    C0]  process_one_work+0x30c/0x550
[   16.946456][    C0]  worker_thread+0x414/0x8bc
[   16.946472][    C0]  kthread+0x16c/0x1e0
[   16.946486][    C0]  ret_from_fork+0x10/0x20

Fixes: bff139b49d9f ("f2fs: handle decompress only post processing in softirq")
Fixes: 95fa90c9e5a7 ("f2fs: support recording errors into superblock")
Reviewed-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

f2fs: compress: fix to call f2fs_wait_on_page_writeback() in f2fs_write_raw_pages()

[ Upstream commit babedcbac164cec970872b8097401ca913a80e61 ]

BUG_ON() will be triggered when writing files concurrently,
because the same page is writtenback multiple times.

1597 void folio_end_writeback(struct folio *folio)
1598 {
......
1618     if (!__folio_end_writeback(folio))
1619         BUG();
......
1625 }

kernel BUG at mm/filemap.c:1619!
Call Trace:
 <TASK>
 f2fs_write_end_io+0x1a0/0x370
 blk_update_request+0x6c/0x410
 blk_mq_end_request+0x15/0x130
 blk_complete_reqs+0x3c/0x50
 __do_softirq+0xb8/0x29b
 ? sort_range+0x20/0x20
 run_ksoftirqd+0x19/0x20
 smpboot_thread_fn+0x10b/0x1d0
 kthread+0xde/0x110
 ? kthread_complete_and_exit+0x20/0x20
 ret_from_fork+0x22/0x30
 </TASK>

Below is the concurrency scenario:

[Process A] [Process B] [Process C]
f2fs_write_raw_pages()
  - redirty_page_for_writepage()
  - unlock page()
f2fs_do_write_data_page()
  - lock_page()
  - clear_page_dirty_for_io()
  - set_page_writeback() [1st writeback]
    .....
    - unlock page()

generic_perform_write()
  - f2fs_write_begin()
    - wait_for_stable_page()

  - f2fs_write_end()
    - set_page_dirty()

  - lock_page()
    - f2fs_do_write_data_page()
      - set_page_writeback() [2st writeback]

This problem was introduced by the previous commit 7377e853967b ("f2fs:
compress: fix potential deadlock of compress file"). All pagelocks were
released in f2fs_write_raw_pages(), but whether the page was
in the writeback state was ignored in the subsequent writing process.
Let's fix it by waiting for the page to writeback before writing.

Cc: Christoph Hellwig <hch@lst.de>
Fixes: 4c8ff7095bef ("f2fs: support data compression")
Fixes: 7377e853967b ("f2fs: compress: fix potential deadlock of compress file")
Signed-off-by: Qi Han <hanqi@vivo.com>
Signed-off-by: Yangtao Li <frank.li@vivo.com>
Reviewed-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

f2fs: apply zone capacity to all zone type

[ Upstream commit 0b37ed21e3367539b79284e0b0af2246ffcf0dca ]

If we manage the zone capacity per zone type, it'll break the GC assumption.
And, the current logic complains valid block count mismatch.
Let's apply zone capacity to all zone type, if specified.

Fixes: de881df97768 ("f2fs: support zone capacity less than zone size")
Reviewed-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

f2fs: fix uninitialized skipped_gc_rwsem

[ Upstream commit c17caf0ba3aa3411b96c71b4ce24be1040b8f3e8 ]

When f2fs skipped a gc round during victim migration, there was a bug which
would skip all upcoming gc rounds unconditionally because skipped_gc_rwsem
was not initialized. It fixes the bug by correctly initializing the
skipped_gc_rwsem inside the gc loop.

Fixes: 6f8d4455060d ("f2fs: avoid fi->i_gc_rwsem[WRITE] lock in f2fs_gc")
Signed-off-by: Yonggil Song <yonggil.song@samsung.com>
Reviewed-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

f2fs: handle dqget error in f2fs_transfer_project_quota()

[ Upstream commit 8051692f5f23260215bfe9a72e712d93606acc5f ]

We should set the error code when dqget() failed.

Fixes: 2c1d03056991 ("f2fs: support F2FS_IOC_FS{GET,SET}XATTR")
Signed-off-by: Yangtao Li <frank.li@vivo.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: sunhme: Fix uninitialized return code

[ Upstream commit d61157414d0a591d10d27d0ce5873916614e5e31 ]

Fix an uninitialized return code if we never found a qfe slot. It would be
a bug if we ever got into this situation, but it's good to return something
tracable.

Fixes: acb3f35f920b ("sunhme: forward the error code from pci_enable_device()")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Sean Anderson <seanga2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: megaraid: Fix mega_cmd_done() CMDID_INT_CMDS

[ Upstream commit 75cb113cd43f06aaf4f1bda0069cfd5b98e909eb ]

When cmdid == CMDID_INT_CMDS, the 'cmds' pointer is NULL but is
dereferenced below.

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Fixes: 0f2bb84d2a68 ("[SCSI] megaraid: simplify internal command handling")
Signed-off-by: Danila Chernetsov <listdansp@mail.ru>
Link: https://lore.kernel.org/r/20230317175109.18585-1-listdansp@mail.ru
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: target: iscsit: Fix TAS handling during conn cleanup

[ Upstream commit cc79da306ebb2edb700c3816b90219223182ac3c ]

Fix a bug added in commit f36199355c64 ("scsi: target: iscsi: Fix cmd abort
fabric stop race").

If CMD_T_TAS is set on the se_cmd we must call iscsit_free_cmd() to do the
last put on the cmd and free it, because the connection is down and we will
not up sending the response and doing the put from the normal I/O
path.

Add a check for CMD_T_TAS in iscsit_release_commands_from_conn() so we now
detect this case and run iscsit_free_cmd().

Fixes: f36199355c64 ("scsi: target: iscsi: Fix cmd abort fabric stop race")
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20230319015620.96006-9-michael.christie@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: target: Fix multiple LUN_RESET handling

[ Upstream commit 673db054d7a2b5a470d7a25baf65956d005ad729 ]

This fixes a bug where an initiator thinks a LUN_RESET has cleaned up
running commands when it hasn't. The bug was added in commit 51ec502a3266
("target: Delete tmr from list before processing").

The problem occurs when:

 1. We have N I/O cmds running in the target layer spread over 2 sessions.

 2. The initiator sends a LUN_RESET for each session.

 3. session1's LUN_RESET loops over all the running commands from both
    sessions and moves them to its local drain_task_list.

 4. session2's LUN_RESET does not see the LUN_RESET from session1 because
    the commit above has it remove itself. session2 also does not see any
    commands since the other reset moved them off the state lists.

 5. sessions2's LUN_RESET will then complete with a successful response.

 6. sessions2's inititor believes the running commands on its session are
    now cleaned up due to the successful response and cleans up the running
    commands from its side. It then restarts them.

 7. The commands do eventually complete on the backend and the target
    starts to return aborted task statuses for them. The initiator will
    either throw a invalid ITT error or might accidentally lookup a new
    task if the ITT has been reallocated already.

Fix the bug by reverting the patch, and serialize the execution of
LUN_RESETs and Preempt and Aborts.

Also prevent us from waiting on LUN_RESETs in core_tmr_drain_tmr_list,
because it turns out the original patch fixed a bug that was not
mentioned. For LUN_RESET1 core_tmr_drain_tmr_list can see a second
LUN_RESET and wait on it. Then the second reset will run
core_tmr_drain_tmr_list and see the first reset and wait on it resulting in
a deadlock.

Fixes: 51ec502a3266 ("target: Delete tmr from list before processing")
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20230319015620.96006-8-michael.christie@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: target: iscsit: Stop/wait on cmds during conn close

[ Upstream commit 395cee83d02de3073211b04fc85724f4abc663ad ]

This fixes a bug added in commit f36199355c64 ("scsi: target: iscsi: Fix
cmd abort fabric stop race").

If we have multiple sessions to the same se_device we can hit a race where
a LUN_RESET on one session cleans up the se_cmds from under another
session which is being closed. This results in the closing session freeing
its conn/session structs while they are still in use.

The bug is:

 1. Session1 has IO se_cmd1.

 2. Session2 can also have se_cmds for I/O and optionally TMRs for ABORTS
    but then gets a LUN_RESET.

 3. The LUN_RESET on session2 sees the se_cmds on session1 and during the
    drain stages marks them all with CMD_T_ABORTED.

 4. session1 is now closed so iscsit_release_commands_from_conn() only sees
    se_cmds with the CMD_T_ABORTED bit set and returns immediately even
    though we have outstanding commands.

 5. session1's connection and session are freed.

 6. The backend request for se_cmd1 completes and it accesses the freed
    connection/session.

This hooks the iscsit layer into the cmd counter code, so we can wait for
all outstanding se_cmds before freeing the connection.

Fixes: f36199355c64 ("scsi: target: iscsi: Fix cmd abort fabric stop race")
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20230319015620.96006-6-michael.christie@oracle.com
Reviewed-by: Maurizio Lombardi <mlombard@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: target: iscsit: isert: Alloc per conn cmd counter

[ Upstream commit 6d256bee602b131bd4fbc92863b6a1210bcf6325 ]

This has iscsit allocate a per conn cmd counter and converts iscsit/isert
to use it instead of the per session one.

Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20230319015620.96006-5-michael.christie@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Stable-dep-of: 395cee83d02d ("scsi: target: iscsit: Stop/wait on cmds during conn close")
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: target: Pass in cmd counter to use during cmd setup

[ Upstream commit 8e288be8606ad87c1726618eacfb8fbd3ab4b806 ]

Allow target_get_sess_cmd() users to pass in the cmd counter they want to
use. Right now we pass in the session's cmd counter but in a subsequent
commit iSCSI will switch from per session to per conn.

Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20230319015620.96006-4-michael.christie@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Stable-dep-of: 395cee83d02d ("scsi: target: iscsit: Stop/wait on cmds during conn close")
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: target: Move cmd counter allocation

[ Upstream commit 4edba7e4a8f39112398d3cda94128a8e13a7d527 ]

iSCSI needs to allocate its cmd counter per connection for MCS support
where we need to stop and wait on commands running on a connection instead
of per session. This moves the cmd counter allocation to
target_setup_session() which is used by drivers that need the stop+wait
behavior per session.

xcopy doesn't need stop+wait at all, so we will be OK moving the cmd
counter allocation outside of transport_init_session().

Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20230319015620.96006-3-michael.christie@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Stable-dep-of: 395cee83d02d ("scsi: target: iscsit: Stop/wait on cmds during conn close")
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: target: Move sess cmd counter to new struct

[ Upstream commit becd9be6069e7b183c084f460f0eb363e43cc487 ]

iSCSI needs to wait on outstanding commands like how SRP and the FC/FCoE
drivers do. It can't use target_stop_session() because for MCS support we
can't stop the entire session during recovery because if other connections
are OK then we want to be able to continue to execute I/O on them.

Move the per session cmd counters to a new struct so iSCSI can allocate
them per connection. The xcopy code can also just not allocate in the
future since it doesn't need to track commands.

Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20230319015620.96006-2-michael.christie@oracle.com
Reviewed-by: Maurizio Lombardi <mlombard@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Stable-dep-of: 395cee83d02d ("scsi: target: iscsit: Stop/wait on cmds during conn close")
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: target: core: Change the way target_xcopy_do_work() sets restiction on max I/O

[ Upstream commit 689d94ec208cfdf95101d99319cb4bdc5f55774d ]

To determine how many blocks sends in one command, the minimum value is
selected from the hw_max_sectors of both devices. In target_xcopy_do_work,
hw_max_sectors are used as blocks, not sectors; it also ignores the fact
that sectors can be of different sizes, for example 512 and 4096
bytes. Because of this, a number of blocks can be transmitted that the
device will not be able to accept.

Change the selection of max transmission size into bytes.

Reviewed-by: Konstantin Shelekhin <k.shelekhin@yadro.com>
Reviewed-by: Dmitriy Bogdanov <d.bogdanov@yadro.com>
Signed-off-by: Anastasia Kovaleva <a.kovaleva@yadro.com>
Link: https://lore.kernel.org/r/20221114102500.88892-4-a.kovaleva@yadro.com
Reviewed-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Stable-dep-of: 395cee83d02d ("scsi: target: iscsit: Stop/wait on cmds during conn close")
Signed-off-by: Sasha Levin <sashal@kernel.org>

bpf: Fix __reg_bound_offset 64->32 var_off subreg propagation

[ Upstream commit 7be14c1c9030f73cc18b4ff23b78a0a081f16188 ]

Xu reports that after commit 3f50f132d840 ("bpf: Verifier, do explicit ALU32
bounds tracking"), the following BPF program is rejected by the verifier:

   0: (61) r2 = *(u32 *)(r1 +0)          ; R2_w=pkt(off=0,r=0,imm=0)
   1: (61) r3 = *(u32 *)(r1 +4)          ; R3_w=pkt_end(off=0,imm=0)
   2: (bf) r1 = r2
   3: (07) r1 += 1
   4: (2d) if r1 > r3 goto pc+8
   5: (71) r1 = *(u8 *)(r2 +0)           ; R1_w=scalar(umax=255,var_off=(0x0; 0xff))
   6: (18) r0 = 0x7fffffffffffff10
   8: (0f) r1 += r0                      ; R1_w=scalar(umin=0x7fffffffffffff10,umax=0x800000000000000f)
   9: (18) r0 = 0x8000000000000000
  11: (07) r0 += 1
  12: (ad) if r0 < r1 goto pc-2
  13: (b7) r0 = 0
  14: (95) exit

And the verifier log says:

  func#0 @0
  0: R1=ctx(off=0,imm=0) R10=fp0
  0: (61) r2 = *(u32 *)(r1 +0)          ; R1=ctx(off=0,imm=0) R2_w=pkt(off=0,r=0,imm=0)
  1: (61) r3 = *(u32 *)(r1 +4)          ; R1=ctx(off=0,imm=0) R3_w=pkt_end(off=0,imm=0)
  2: (bf) r1 = r2                       ; R1_w=pkt(off=0,r=0,imm=0) R2_w=pkt(off=0,r=0,imm=0)
  3: (07) r1 += 1                       ; R1_w=pkt(off=1,r=0,imm=0)
  4: (2d) if r1 > r3 goto pc+8          ; R1_w=pkt(off=1,r=1,imm=0) R3_w=pkt_end(off=0,imm=0)
  5: (71) r1 = *(u8 *)(r2 +0)           ; R1_w=scalar(umax=255,var_off=(0x0; 0xff)) R2_w=pkt(off=0,r=1,imm=0)
  6: (18) r0 = 0x7fffffffffffff10       ; R0_w=9223372036854775568
  8: (0f) r1 += r0                      ; R0_w=9223372036854775568 R1_w=scalar(umin=9223372036854775568,umax=9223372036854775823,s32_min=-240,s32_max=15)
  9: (18) r0 = 0x8000000000000000       ; R0_w=-9223372036854775808
  11: (07) r0 += 1                      ; R0_w=-9223372036854775807
  12: (ad) if r0 < r1 goto pc-2         ; R0_w=-9223372036854775807 R1_w=scalar(umin=9223372036854775568,umax=9223372036854775809)
  13: (b7) r0 = 0                       ; R0_w=0
  14: (95) exit

  from 12 to 11: R0_w=-9223372036854775807 R1_w=scalar(umin=9223372036854775810,umax=9223372036854775823,var_off=(0x8000000000000000; 0xffffffff)) R2_w=pkt(off=0,r=1,imm=0) R3_w=pkt_end(off=0,imm=0) R10=fp0
  11: (07) r0 += 1                      ; R0_w=-9223372036854775806
  12: (ad) if r0 < r1 goto pc-2         ; R0_w=-9223372036854775806 R1_w=scalar(umin=9223372036854775810,umax=9223372036854775810,var_off=(0x8000000000000000; 0xffffffff))
  13: safe

  [...]

  from 12 to 11: R0_w=-9223372036854775795 R1=scalar(umin=9223372036854775822,umax=9223372036854775823,var_off=(0x8000000000000000; 0xffffffff)) R2=pkt(off=0,r=1,imm=0) R3=pkt_end(off=0,imm=0) R10=fp0
  11: (07) r0 += 1                      ; R0_w=-9223372036854775794
  12: (ad) if r0 < r1 goto pc-2         ; R0_w=-9223372036854775794 R1=scalar(umin=9223372036854775822,umax=9223372036854775822,var_off=(0x8000000000000000; 0xffffffff))
  13: safe

  from 12 to 11: R0_w=-9223372036854775794 R1=scalar(umin=9223372036854775823,umax=9223372036854775823,var_off=(0x8000000000000000; 0xffffffff)) R2=pkt(off=0,r=1,imm=0) R3=pkt_end(off=0,imm=0) R10=fp0
  11: (07) r0 += 1                      ; R0_w=-9223372036854775793
  12: (ad) if r0 < r1 goto pc-2         ; R0_w=-9223372036854775793 R1=scalar(umin=9223372036854775823,umax=9223372036854775823,var_off=(0x8000000000000000; 0xffffffff))
  13: safe

  from 12 to 11: R0_w=-9223372036854775793 R1=scalar(umin=9223372036854775824,umax=9223372036854775823,var_off=(0x8000000000000000; 0xffffffff)) R2=pkt(off=0,r=1,imm=0) R3=pkt_end(off=0,imm=0) R10=fp0
  11: (07) r0 += 1                      ; R0_w=-9223372036854775792
  12: (ad) if r0 < r1 goto pc-2         ; R0_w=-9223372036854775792 R1=scalar(umin=9223372036854775824,umax=9223372036854775823,var_off=(0x8000000000000000; 0xffffffff))
  13: safe

  [...]

The 64bit umin=9223372036854775810 bound continuously bumps by +1 while
umax=9223372036854775823 stays as-is until the verifier complexity limit
is reached and the program gets finally rejected. During this simulation,
the umin also eventually surpasses umax. Looking at the first 'from 12
to 11' output line from the loop, R1 has the following state:

  R1_w=scalar(umin=0x8000000000000002 (9223372036854775810),
              umax=0x800000000000000f (9223372036854775823),
          var_off=(0x8000000000000000;
                           0xffffffff))

The var_off has technically not an inconsistent state but it's very
imprecise and far off surpassing 64bit umax bounds whereas the expected
output with refined known bits in var_off should have been like:

  R1_w=scalar(umin=0x8000000000000002 (9223372036854775810),
              umax=0x800000000000000f (9223372036854775823),
          var_off=(0x8000000000000000;
                                  0xf))

In the above log, var_off stays as var_off=(0x8000000000000000; 0xffffffff)
and does not converge into a narrower mask where more bits become known,
eventually transforming R1 into a constant upon umin=9223372036854775823,
umax=9223372036854775823 case where the verifier would have terminated and
let the program pass.

The __reg_combine_64_into_32() marks the subregister unknown and propagates
64bit {s,u}min/{s,u}max bounds to their 32bit equivalents iff they are within
the 32bit universe. The question came up whether __reg_combine_64_into_32()
should special case the situation that when 64bit {s,u}min bounds have
the same value as 64bit {s,u}max bounds to then assign the latter as
well to the 32bit reg->{s,u}32_{min,max}_value. As can be seen from the
above example however, that is just /one/ special case and not a /generic/
solution given above example would still not be addressed this way and
remain at an imprecise var_off=(0x8000000000000000; 0xffffffff).

The improvement is needed in __reg_bound_offset() to refine var32_off with
the updated var64_off instead of the prior reg->var_off. The reg_bounds_sync()
code first refines information about the register's min/max bounds via
__update_reg_bounds() from the current var_off, then in __reg_deduce_bounds()
from sign bit and with the potentially learned bits from bounds it'll
update the var_off tnum in __reg_bound_offset(). For example, intersecting
with the old var_off might have improved bounds slightly, e.g. if umax
was 0x7f...f and var_off was (0; 0xf...fc), then new var_off will then
result in (0; 0x7f...fc). The intersected var64_off holds then the
universe which is a superset of var32_off. The point for the latter is
not to broaden, but to further refine known bits based on the intersection
of var_off with 32 bit bounds, so that we later construct the final var_off
from upper and lower 32 bits. The final __update_reg_bounds() can then
potentially still slightly refine bounds if more bits became known from the
new var_off.

After the improvement, we can see R1 converging successively:

  func#0 @0
  0: R1=ctx(off=0,imm=0) R10=fp0
  0: (61) r2 = *(u32 *)(r1 +0)          ; R1=ctx(off=0,imm=0) R2_w=pkt(off=0,r=0,imm=0)
  1: (61) r3 = *(u32 *)(r1 +4)          ; R1=ctx(off=0,imm=0) R3_w=pkt_end(off=0,imm=0)
  2: (bf) r1 = r2                       ; R1_w=pkt(off=0,r=0,imm=0) R2_w=pkt(off=0,r=0,imm=0)
  3: (07) r1 += 1                       ; R1_w=pkt(off=1,r=0,imm=0)
  4: (2d) if r1 > r3 goto pc+8          ; R1_w=pkt(off=1,r=1,imm=0) R3_w=pkt_end(off=0,imm=0)
  5: (71) r1 = *(u8 *)(r2 +0)           ; R1_w=scalar(umax=255,var_off=(0x0; 0xff)) R2_w=pkt(off=0,r=1,imm=0)
  6: (18) r0 = 0x7fffffffffffff10       ; R0_w=9223372036854775568
  8: (0f) r1 += r0                      ; R0_w=9223372036854775568 R1_w=scalar(umin=9223372036854775568,umax=9223372036854775823,s32_min=-240,s32_max=15)
  9: (18) r0 = 0x8000000000000000       ; R0_w=-9223372036854775808
  11: (07) r0 += 1                      ; R0_w=-9223372036854775807
  12: (ad) if r0 < r1 goto pc-2         ; R0_w=-9223372036854775807 R1_w=scalar(umin=9223372036854775568,umax=9223372036854775809)
  13: (b7) r0 = 0                       ; R0_w=0
  14: (95) exit

  from 12 to 11: R0_w=-9223372036854775807 R1_w=scalar(umin=9223372036854775810,umax=9223372036854775823,var_off=(0x8000000000000000; 0xf),s32_min=0,s32_max=15,u32_max=15) R2_w=pkt(off=0,r=1,imm=0) R3_w=pkt_end(off=0,imm=0) R10=fp0
  11: (07) r0 += 1                      ; R0_w=-9223372036854775806
  12: (ad) if r0 < r1 goto pc-2         ; R0_w=-9223372036854775806 R1_w=-9223372036854775806
  13: safe

  from 12 to 11: R0_w=-9223372036854775806 R1_w=scalar(umin=9223372036854775811,umax=9223372036854775823,var_off=(0x8000000000000000; 0xf),s32_min=0,s32_max=15,u32_max=15) R2_w=pkt(off=0,r=1,imm=0) R3_w=pkt_end(off=0,imm=0) R10=fp0
  11: (07) r0 += 1                      ; R0_w=-9223372036854775805
  12: (ad) if r0 < r1 goto pc-2         ; R0_w=-9223372036854775805 R1_w=-9223372036854775805
  13: safe

  [...]

  from 12 to 11: R0_w=-9223372036854775798 R1=scalar(umin=9223372036854775819,umax=9223372036854775823,var_off=(0x8000000000000008; 0x7),s32_min=8,s32_max=15,u32_min=8,u32_max=15) R2=pkt(off=0,r=1,imm=0) R3=pkt_end(off=0,imm=0) R10=fp0
  11: (07) r0 += 1                      ; R0_w=-9223372036854775797
  12: (ad) if r0 < r1 goto pc-2         ; R0_w=-9223372036854775797 R1=-9223372036854775797
  13: safe

  from 12 to 11: R0_w=-9223372036854775797 R1=scalar(umin=9223372036854775820,umax=9223372036854775823,var_off=(0x800000000000000c; 0x3),s32_min=12,s32_max=15,u32_min=12,u32_max=15) R2=pkt(off=0,r=1,imm=0) R3=pkt_end(off=0,imm=0) R10=fp0
  11: (07) r0 += 1                      ; R0_w=-9223372036854775796
  12: (ad) if r0 < r1 goto pc-2         ; R0_w=-9223372036854775796 R1=-9223372036854775796
  13: safe

  from 12 to 11: R0_w=-9223372036854775796 R1=scalar(umin=9223372036854775821,umax=9223372036854775823,var_off=(0x800000000000000c; 0x3),s32_min=12,s32_max=15,u32_min=12,u32_max=15) R2=pkt(off=0,r=1,imm=0) R3=pkt_end(off=0,imm=0) R10=fp0
  11: (07) r0 += 1                      ; R0_w=-9223372036854775795
  12: (ad) if r0 < r1 goto pc-2         ; R0_w=-9223372036854775795 R1=-9223372036854775795
  13: safe

  from 12 to 11: R0_w=-9223372036854775795 R1=scalar(umin=9223372036854775822,umax=9223372036854775823,var_off=(0x800000000000000e; 0x1),s32_min=14,s32_max=15,u32_min=14,u32_max=15) R2=pkt(off=0,r=1,imm=0) R3=pkt_end(off=0,imm=0) R10=fp0
  11: (07) r0 += 1                      ; R0_w=-9223372036854775794
  12: (ad) if r0 < r1 goto pc-2         ; R0_w=-9223372036854775794 R1=-9223372036854775794
  13: safe

  from 12 to 11: R0_w=-9223372036854775794 R1=-9223372036854775793 R2=pkt(off=0,r=1,imm=0) R3=pkt_end(off=0,imm=0) R10=fp0
  11: (07) r0 += 1                      ; R0_w=-9223372036854775793
  12: (ad) if r0 < r1 goto pc-2
  last_idx 12 first_idx 12
  parent didn't have regs=1 stack=0 marks: R0_rw=P-9223372036854775801 R1_r=scalar(umin=9223372036854775815,umax=9223372036854775823,var_off=(0x8000000000000000; 0xf),s32_min=0,s32_max=15,u32_max=15) R2=pkt(off=0,r=1,imm=0) R3=pkt_end(off=0,imm=0) R10=fp0
  last_idx 11 first_idx 11
  regs=1 stack=0 before 11: (07) r0 += 1
  parent didn't have regs=1 stack=0 marks: R0_rw=P-9223372036854775805 R1_rw=scalar(umin=9223372036854775812,umax=9223372036854775823,var_off=(0x8000000000000000; 0xf),s32_min=0,s32_max=15,u32_max=15) R2_w=pkt(off=0,r=1,imm=0) R3_w=pkt_end(off=0,imm=0) R10=fp0
  last_idx 12 first_idx 0
  regs=1 stack=0 before 12: (ad) if r0 < r1 goto pc-2
  regs=1 stack=0 before 11: (07) r0 += 1
  regs=1 stack=0 before 12: (ad) if r0 < r1 goto pc-2
  regs=1 stack=0 before 11: (07) r0 += 1
  regs=1 stack=0 before 12: (ad) if r0 < r1 goto pc-2
  regs=1 stack=0 before 11: (07) r0 += 1
  regs=1 stack=0 before 9: (18) r0 = 0x8000000000000000
  last_idx 12 first_idx 12
  parent didn't have regs=2 stack=0 marks: R0_rw=P-9223372036854775801 R1_r=Pscalar(umin=9223372036854775815,umax=9223372036854775823,var_off=(0x8000000000000000; 0xf),s32_min=0,s32_max=15,u32_max=15) R2=pkt(off=0,r=1,imm=0) R3=pkt_end(off=0,imm=0) R10=fp0
  last_idx 11 first_idx 11
  regs=2 stack=0 before 11: (07) r0 += 1
  parent didn't have regs=2 stack=0 marks: R0_rw=P-9223372036854775805 R1_rw=Pscalar(umin=9223372036854775812,umax=9223372036854775823,var_off=(0x8000000000000000; 0xf),s32_min=0,s32_max=15,u32_max=15) R2_w=pkt(off=0,r=1,imm=0) R3_w=pkt_end(off=0,imm=0) R10=fp0
  last_idx 12 first_idx 0
  regs=2 stack=0 before 12: (ad) if r0 < r1 goto pc-2
  regs=2 stack=0 before 11: (07) r0 += 1
  regs=2 stack=0 before 12: (ad) if r0 < r1 goto pc-2
  regs=2 stack=0 before 11: (07) r0 += 1
  regs=2 stack=0 before 12: (ad) if r0 < r1 goto pc-2
  regs=2 stack=0 before 11: (07) r0 += 1
  regs=2 stack=0 before 9: (18) r0 = 0x8000000000000000
  regs=2 stack=0 before 8: (0f) r1 += r0
  regs=3 stack=0 before 6: (18) r0 = 0x7fffffffffffff10
  regs=2 stack=0 before 5: (71) r1 = *(u8 *)(r2 +0)
  13: safe

  from 4 to 13: safe
  verification time 322 usec
  stack depth 0
  processed 56 insns (limit 1000000) max_states_per_insn 1 total_states 3 peak_states 3 mark_read 1

This also fixes up a test case along with this improvement where we match
on the verifier log. The updated log now has a refined var_off, too.

Fixes: 3f50f132d840 ("bpf: Verifier, do explicit ALU32 bounds tracking")
Reported-by: Xu Kuohai <xukuohai@huaweicloud.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20230314203424.4015351-2-xukuohai@huaweicloud.com
Link: https://lore.kernel.org/bpf/20230322213056.2470-1-daniel@iogearbox.net
Signed-off-by: Sasha Levin <sashal@kernel.org>

netfilter: keep conntrack reference until IPsecv6 policy checks are done

[ Upstream commit b0e214d212030fe497d4d150bb3474e50ad5d093 ]

Keep the conntrack reference until policy checks have been performed for
IPsec V6 NAT support, just like ipv4.

The reference needs to be dropped before a packet is
queued to avoid having the conntrack module unloadable.

Fixes: 58a317f1061c ("netfilter: ipv6: add IPv6 NAT support")
Signed-off-by: Madhu Koriginja <madhu.koriginja@nxp.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

net: dsa: qca8k: remove assignment of an_enabled in pcs_get_state()

[ Upstream commit 9ef70d0130f282638b28cfce24222f71ada00c9c ]

pcs_get_state() implementations are not supposed to alter an_enabled.
Remove this assignment.

Fixes: b3591c2a3661 ("net: dsa: qca8k: Switch to PHYLINK instead of PHYLIB")
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://lore.kernel.org/r/E1pdsE5-00Dl2l-8F@rmk-PC.armlinux.org.uk
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

libbpf: Fix ld_imm64 copy logic for ksym in light skeleton.

[ Upstream commit a506d6ce1dd184051037dc9d26c3eb187c9fe625 ]

Unlike normal libbpf the light skeleton 'loader' program is doing
btf_find_by_name_kind() call at run-time to find ksym in the kernel and
populate its {btf_id, btf_obj_fd} pair in ld_imm64 insn. To avoid doing the
search multiple times for the same ksym it remembers the first patched ld_imm64
insn and copies {btf_id, btf_obj_fd} from it into subsequent ld_imm64 insn.
Fix a bug in copying logic, since it may incorrectly clear BPF_PSEUDO_BTF_ID flag.

Also replace always true if (btf_obj_fd >= 0) check with unconditional JMP_JA
to clarify the code.

Fixes: d995816b77eb ("libbpf: Avoid reload of imm for weak, unresolved, repeating ksym")
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20230319203014.55866-1-alexei.starovoitov@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>