linux-block.git
10 months agonet/mlx5e: Add local loopback counter to vport rep stats
Or Har-Toov [Mon, 16 Oct 2023 15:00:00 +0000 (18:00 +0300)]
net/mlx5e: Add local loopback counter to vport rep stats

Add counter for number of unicast, multicast and broadcast packets/
octets that were loopback.

Signed-off-by: Or Har-Toov <ohartoov@nvidia.com>
Reviewed-by: Patrisious Haddad <phaddad@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
10 months agonet/mlx5: Query maximum frequency adjustment of the PTP hardware clock
Rahul Rameshbabu [Thu, 30 Mar 2023 22:05:38 +0000 (15:05 -0700)]
net/mlx5: Query maximum frequency adjustment of the PTP hardware clock

Some mlx5 devices do not support the default advertised maximum frequency
adjustment value for the PTP hardware clock that is set by the driver.
These devices need to be queried when initializing the clock functionality
in order to get the maximum supported frequency adjustment value. This
value can be greater than the minimum supported frequency adjustment across
mlx5 devices (50 million ppb).

Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
10 months agonet/mlx5: Convert scaled ppm values outside the s32 range for PHC frequency adjustments
Rahul Rameshbabu [Fri, 25 Aug 2023 03:29:50 +0000 (20:29 -0700)]
net/mlx5: Convert scaled ppm values outside the s32 range for PHC frequency adjustments

Represent scaled ppm as ppb to the device when the value in scaled ppm is
not representable as a 32-bit signed integer. mlx5 devices only support a
32-bit field for the frequency adjustment value in units of either scaled
ppm or ppb.

Since mlx5 devices only support a 32-bit field for the frequency adjustment
value independent of unit used, limit the maximum frequency adjustment to
S32_MAX ppb.

Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
10 months agonet/mlx5: Initialize clock->ptp_info inside mlx5_init_timer_clock
Rahul Rameshbabu [Mon, 17 Jul 2023 19:10:42 +0000 (12:10 -0700)]
net/mlx5: Initialize clock->ptp_info inside mlx5_init_timer_clock

Configure the PHC inside mlx5_init_timer_clock for calling mlx5_ptp_settime
later in the function. Would previously use mlx5_ptp_clock_info instance to
invoke mlx5_ptp_settime to set the NIC real-time clock to be synchronized
with the host system clock.

Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
10 months agonet/mlx5: Refactor real time clock operation checks for PHC
Rahul Rameshbabu [Mon, 17 Jul 2023 18:40:21 +0000 (11:40 -0700)]
net/mlx5: Refactor real time clock operation checks for PHC

Check if the MTUTC register of the NIC can be modified before attempting to
execute a real-time clock operation. Previous implementation aborted the
real-time clock operation pre-emptively when the MTUTC register used to
control the real-time clock was not modifiable, indicating real-time clock
mode was not enabled on the NIC. The original control flow was confusing
since the noop-if-RTC-disabled branch looked similar to an error handling
guard clause. The purpose of this patch is purely for improving readability
and should lead to no functional change.

Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
10 months agonet/mlx5e: Access array with enum values instead of magic numbers
Gal Pressman [Tue, 24 Oct 2023 11:42:59 +0000 (14:42 +0300)]
net/mlx5e: Access array with enum values instead of magic numbers

Access the headers array using pedit_cmd enum values, and don't assume
anything about their values.

Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Vlad Buslov <vladbu@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
10 months agonet/mlx5: simplify mlx5_set_driver_version string assignments
Justin Stitt [Wed, 11 Oct 2023 21:29:57 +0000 (21:29 +0000)]
net/mlx5: simplify mlx5_set_driver_version string assignments

In total, just assigning this version string takes:
(1) strncpy()'s
(5) strlen()'s
(3) strncat()'s
(1) snprintf()'s
(4) max_t()'s

Moreover, `strncpy` is deprecated [1] and `strncat` really shouldn't be
used either [2]. With this in mind, let's simply use a single
`snprintf`.

Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings
Link: https://elixir.bootlin.com/linux/v6.6-rc5/source/include/linux/fortify-string.h#L448
Link: https://github.com/KSPP/linux/issues/90
Cc: linux-hardening@vger.kernel.org
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Justin Stitt <justinstitt@google.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
10 months agonet/mlx5: Annotate struct mlx5_flow_handle with __counted_by
Kees Cook [Tue, 3 Oct 2023 23:17:30 +0000 (16:17 -0700)]
net/mlx5: Annotate struct mlx5_flow_handle with __counted_by

Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time via CONFIG_UBSAN_BOUNDS (for
array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).

As found with Coccinelle[1], add __counted_by for struct mlx5_flow_handle.

Cc: Saeed Mahameed <saeedm@nvidia.com>
Cc: Leon Romanovsky <leon@kernel.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: netdev@vger.kernel.org
Cc: linux-rdma@vger.kernel.org
Link: https://github.com/kees/kernel-tools/blob/trunk/coccinelle/examples/counted_by.cocci
Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Justin Stitt <justinstitt@google.com>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
10 months agonet/mlx5: Annotate struct mlx5_fc_bulk with __counted_by
Kees Cook [Tue, 3 Oct 2023 23:17:22 +0000 (16:17 -0700)]
net/mlx5: Annotate struct mlx5_fc_bulk with __counted_by

Prepare for the coming implementation by GCC and Clang of the __counted_by
attribute. Flexible array members annotated with __counted_by can have
their accesses bounds-checked at run-time via CONFIG_UBSAN_BOUNDS (for
array indexing) and CONFIG_FORTIFY_SOURCE (for strcpy/memcpy-family
functions).

As found with Coccinelle[1], add __counted_by for struct mlx5_fc_bulk.

Cc: Saeed Mahameed <saeedm@nvidia.com>
Cc: Leon Romanovsky <leon@kernel.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: netdev@vger.kernel.org
Cc: linux-rdma@vger.kernel.org
Link: https://github.com/kees/kernel-tools/blob/trunk/coccinelle/examples/counted_by.cocci
Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Justin Stitt <justinstitt@google.com>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
10 months agonet/mlx5e: Some cleanup in mlx5e_tc_stats_matchall()
Amir Tzin [Wed, 6 Sep 2023 10:40:38 +0000 (13:40 +0300)]
net/mlx5e: Some cleanup in mlx5e_tc_stats_matchall()

Function mlx5e_tc_stats_matchall() is only called from one file:
drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c
Move it there and make it static as exposing it is unnecessary. Also
variable cur_stats is redundant. Remove it and avoid superfluous copy.

This patch does not change functionality.

Reviewed-by: Patrisious Haddad <phaddad@nvidia.com>
Signed-off-by: Amir Tzin <amirtz@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
10 months agonet/mlx5: Allow sync reset flow when BF MGT interface device is present
Moshe Shemesh [Mon, 12 Jun 2023 16:09:25 +0000 (19:09 +0300)]
net/mlx5: Allow sync reset flow when BF MGT interface device is present

In sync reset flow, PF is checking that only devices of same device ID
as itself present on the PCIe bridge, otherwise it will NACK the reset.
Since the PCIe bridge connection to NIC card has to be 1 to 1, this is
valid.

However, the BlueField device may also expose another sub-device to the
PCI called management interface, which only provides an ethernet channel
between the host and the smart NIC.

Allow sync reset flow also when management interface sub-device present
when checking devices on the PCIe bridge.

Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
10 months agonet/mlx5: print change on SW reset semaphore returns busy
Moshe Shemesh [Wed, 6 Sep 2023 09:18:05 +0000 (12:18 +0300)]
net/mlx5: print change on SW reset semaphore returns busy

While collecting crdump as part of fw_fatal health reporter dump the PF
may fail to lock the SW reset semaphore. Change the print to indicate if
it was due to another PF locked the semaphore already and so trying to
lock the semaphore returned -EBUSY.

Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Shay Drory <shayd@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
10 months agonet: don't dump stack on queue timeout
Jakub Kicinski [Tue, 14 Nov 2023 05:11:42 +0000 (00:11 -0500)]
net: don't dump stack on queue timeout

The top syzbot report for networking (#14 for the entire kernel)
is the queue timeout splat. We kept it around for a long time,
because in real life it provides pretty strong signal that
something is wrong with the driver or the device.

Removing it is also likely to break monitoring for those who
track it as a kernel warning.

Nevertheless, WARN()ings are best suited for catching kernel
programming bugs. If a Tx queue gets starved due to a pause
storm, priority configuration, or other weirdness - that's
obviously a problem, but not a problem we can fix at
the kernel level.

Bite the bullet and convert the WARN() to a print.

Before:

  NETDEV WATCHDOG: eni1np1 (netdevsim): transmit queue 0 timed out 1975 ms
  WARNING: CPU: 0 PID: 0 at net/sched/sch_generic.c:525 dev_watchdog+0x39e/0x3b0
  [... completely pointless stack trace of a timer follows ...]

Now:

  netdevsim netdevsim1 eni1np1: NETDEV WATCHDOG: CPU: 0: transmit queue 0 timed out 1769 ms

Alternatively we could mark the drivers which syzbot has
learned to abuse as "print-instead-of-WARN" selectively.

Reported-by: syzbot+d55372214aff0faa1f1f@syzkaller.appspotmail.com
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agoMerge branch 'bnxt_en-tx-improvements'
David S. Miller [Wed, 15 Nov 2023 10:07:40 +0000 (10:07 +0000)]
Merge branch 'bnxt_en-tx-improvements'

Michael Chan says:

====================
bnxt_en: TX path improvements

All patches in this patchset are related to improving the TX path.
There are 2 areas of improvements:

1. The TX interrupt logic currently counts the number of TX completions
to determine the number of TX SKBs to free.  We now change it so that
the TX completion will now contain the hardware consumer index
information.  The driver will keep track of the latest hardware
consumer index from the last TX completion and clean up all TX SKBs
up to that index.  This scheme aligns better with future chips and
allows xmit_more code path to be more optimized.

2. The current driver logic requires an additional MSIX for each
additional MQPRIO TX ring.  This scheme uses too many MSIX vectors if
the user enables a large number of MQPRIO TCs.  We now use a new scheme
that will use the same MSIX for all the MQPRIO TX rings for each
ethtool channel.  Each ethtool TX channel can have up to 8 MQPRIO
TX rings and now they all will share the same MSIX.

v2: Rebased

v1 posted on Oct 27 2023 right before the close of net-next:

https://lore.kernel.org/netdev/20231027232252.36111-1-michael.chan@broadcom.com/
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: Optimize xmit_more TX path
Michael Chan [Tue, 14 Nov 2023 00:16:21 +0000 (16:16 -0800)]
bnxt_en: Optimize xmit_more TX path

Now that we use the cumulative consumer index scheme for TX completion,
we don't need to have one TX completion per TX packet in the xmit_more
code path.  Set the TX_BD_FLAGS_NO_CMPL flag if xmit_more is true.
Fallback to one interrupt per packet if the ring is filled beyond
bp->tx_wake_thresh.

Also, move the wmb() to bnxt_txr_db_kick().  When xmit_more is true,
we'll skip the bnxt_txr_db_kick() call and there is no need to call
wmb() to sync. the TX BD data.

Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: Use existing MSIX vectors for all mqprio TX rings
Michael Chan [Tue, 14 Nov 2023 00:16:20 +0000 (16:16 -0800)]
bnxt_en: Use existing MSIX vectors for all mqprio TX rings

We can now fully support sharing the same MSIX for all mqprio TX rings
belonging to the same ethtool channel with the new infrastructure:

1. Allocate the proper entries for cp_ring_arr in struct bnxt_cp_ring_info
to support the additional TX rings.

2. Populate the tx_ring array in struct bnxt_napi for all TX rings
sharing the same NAPI.

3. bnxt_num_tx_to_cp() returns the proper NQ/completion rings to support
the TX rings in the input.

4. Adjust bnxt_get_num_ring_stats() for the reduced number of ring
counters with the new scheme.

Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: Add macros related to TC and TX rings
Michael Chan [Tue, 14 Nov 2023 00:16:19 +0000 (16:16 -0800)]
bnxt_en: Add macros related to TC and TX rings

Add 3 macros that handle to conversions between TC numbers and TX
ring numbers.  These will help to clarify the existing logic and the
new logic in the next patch.

Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: Add helper to get the number of CP rings required for TX rings
Michael Chan [Tue, 14 Nov 2023 00:16:18 +0000 (16:16 -0800)]
bnxt_en: Add helper to get the number of CP rings required for TX rings

Up until now, each TX ring always requires a completion ring/NQ/MSIX.
bnxt_trim_rings() and the assignment of bp->cp_nr_rings always make
this assumption.  This will no longer be true in the next patches, so
we refactor and add helper functions to determine the proper relationship
between TX rings and the required completion ring/NQ/MSIX.  This patch
does not change the 1:1 relationship yet.

Note that on P5 chips, each RX and TX ring still requires a completion
ring.  Only the number of NQs has been reduced.  We should no longer call
bnxt_trim_rings() to adjust the RX and TX rings on P5 chips.  Replace with
simple logic to check that RX + TX < CP and adjust accordingly.

bnxt_check_rings() should call _bnxt_get_max_rings() to get the raw
number of rings instead of bnxt_get_max_rings().  If we are about to
create TCs, bnxt_get_max_rings() would not be able to calculate the max
rings correctly.

Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: Support up to 8 TX rings per MSIX
Michael Chan [Tue, 14 Nov 2023 00:16:17 +0000 (16:16 -0800)]
bnxt_en: Support up to 8 TX rings per MSIX

For each mqprio TC, we allocate a set of TX rings to map to the new
hardware CoS queue.  Expand the tx_ring pointer in struct bnxt_napi
to an array of 8 to support up to 8 TX rings, one for each TC.
Only array entry 0 is used at this time.  The rest of the array
entries will be used in later patches.

Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: Refactor bnxt_hwrm_set_coal()
Michael Chan [Tue, 14 Nov 2023 00:16:16 +0000 (16:16 -0800)]
bnxt_en: Refactor bnxt_hwrm_set_coal()

Add 2 helper functions to set coalescing for each RX and TX rings.  This
will make it easier to expand the number of TX rings per MSIX in the
next patches.

Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: New encoding for the TX opaque field
Michael Chan [Tue, 14 Nov 2023 00:16:15 +0000 (16:16 -0800)]
bnxt_en: New encoding for the TX opaque field

In order to support multiple TX rings on the same MSIX, we'll use the
upper byte of the TX opaque field to store the ring index in the new
tx_napi_idx field.  This tx_napi_idx field is currently always 0 until
more infrastructure is added in later patches.

Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: Refactor bnxt_tx_int()
Michael Chan [Tue, 14 Nov 2023 00:16:14 +0000 (16:16 -0800)]
bnxt_en: Refactor bnxt_tx_int()

bnxt_tx_int() processes the only one TX ring from the bnxt_napi pointer.
To prepare for more TX rings associated with the bnxt_napi structure,
add a new __bnxt_tx_int() function that takes the bnxt_tx_ring_info
pointer to process that one TX ring.  No functional change.

Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: Remove BNXT_RX_HDL and BNXT_TX_HDL
Michael Chan [Tue, 14 Nov 2023 00:16:13 +0000 (16:16 -0800)]
bnxt_en: Remove BNXT_RX_HDL and BNXT_TX_HDL

These 2 constants were used for the one RX and one TX completion
ring pointer in the cpr->cp_ring_arr fixed array.  Now that we've
changed to allocating the array for the exact number of entries to
support more TX rings, we no longer use these constants.

The array index as well as the type of completion ring (RX/TX) are
now encoded in the handle for the completion ring.  This will allow
us to locate the completion ring during NAPI for any number of
completion rings sharing the same MSIX.  In the following patches,
we'll be adding support for more TX rings associated with the same
MSIX vector.

Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: Add completion ring pointer in TX and RX ring structures
Michael Chan [Tue, 14 Nov 2023 00:16:12 +0000 (16:16 -0800)]
bnxt_en: Add completion ring pointer in TX and RX ring structures

From the TX or RX ring structure, we need to find the corresponding
completion ring during initialization.  On P5 chips, we use the MSIX/napi
entry to locate the completion ring because there is only one RX/TX
ring per MSIX.  To allow multiple TX rings for each MSIX, we need
to add a direct pointer from the TX ring and RX ring structures.
This also simplifies the existing logic.

Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: Restructure cp_ring_arr in struct bnxt_cp_ring_info
Michael Chan [Tue, 14 Nov 2023 00:16:11 +0000 (16:16 -0800)]
bnxt_en: Restructure cp_ring_arr in struct bnxt_cp_ring_info

The cp_ring_arr is currently a fixed array of 2 pointers for the
TX and RX completion rings.  These pointers are allocated during
ring initialization.  Currntly, we support up to 2 completion rings
for each MSIX.  In order to support more completion rings, we change
this fixed array to a pointer and allocate the required entries
during ring initialization.  This patch keeps the current scheme of
allocating only 2 entries when needed.  Later patches will expand
and allocate more entries when required.

Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: Add completion ring pointer in TX and RX ring structures
Michael Chan [Tue, 14 Nov 2023 00:16:10 +0000 (16:16 -0800)]
bnxt_en: Add completion ring pointer in TX and RX ring structures

From the TX or RX ring structure, we need to find the corresponding
completion ring during initialization.  On P5 chips, we use the MSIX/napi
entry to locate the completion ring because there is only one RX/TX
ring per MSIX.  To allow multiple TX rings for each MSIX, we need
to add a direct pointer from the TX ring and RX ring structures.
This also simplifies the existing logic.

Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agobnxt_en: Put the TX producer information in the TX BD opaque field
Michael Chan [Tue, 14 Nov 2023 00:16:09 +0000 (16:16 -0800)]
bnxt_en: Put the TX producer information in the TX BD opaque field

Currently, the opaque field in the TX BD is only used for debugging.
The TX completion logic relies on getting one TX completion for each
packet and they always complete in order.

Improve this scheme by putting the producer information (ring index plus
number of BDs for the packet) in the opaque field.  This way, we can
handle TX completion processing by looking at the last TX completion
instead of counting the number of completions.

Since we no longer need to count the exact number of completions, we can
optimize xmit_more by disabling TX completion when the xmit_more
condition is true.  This will be done in later patches.

This patch is only initializing the opaque field in the TX BD and is
not changing the driver's TX completion logic yet.

Reviewed-by: Andy Gospodarek <gospo@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agoMerge branch 'intel-wired-lan-driver-updates-2023-11-13-i40e'
Jakub Kicinski [Wed, 15 Nov 2023 04:06:07 +0000 (20:06 -0800)]
Merge branch 'intel-wired-lan-driver-updates-2023-11-13-i40e'

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2023-11-13 (i40e)

This series contains updates to i40e driver only.

Justin Bronder increases number of allowable descriptors for XL710
devices.

Su Hui adds error check, and unroll, for RSS configuration.

Andrii changes module read error message to debug and makes it more
verbose.

Ivan Vecera performs numerous clean-ups and refactors to driver such as
removing unused defines and fields, converting use of flags to bitmaps,
adding helpers, re-organizing code, etc.
====================

Link: https://lore.kernel.org/r/20231113231047.548659-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Delete unused i40e_mac_info fields
Ivan Vecera [Mon, 13 Nov 2023 23:10:34 +0000 (15:10 -0800)]
i40e: Delete unused i40e_mac_info fields

From commit 9eed69a9147c ("i40e: Drop FCoE code from core driver files")
the field i40e_mac_info.san_addr is unused (never filled).
The field i40e_mac_info.max_fcoeq is unused from the beginning.
Remove both.

Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Co-developed-by: Michal Schmidt <mschmidt@redhat.com>
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Signed-off-by: Ivan Vecera <ivecera@redhat.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>
Link: https://lore.kernel.org/r/20231113231047.548659-16-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Move inline helpers to i40e_prototype.h
Ivan Vecera [Mon, 13 Nov 2023 23:10:33 +0000 (15:10 -0800)]
i40e: Move inline helpers to i40e_prototype.h

Move version check helper functions from i40e_type.h to
i40e_prototype.h as per discussion [1].

[1] https://lore.kernel.org/all/cdcd6b97-1138-4cd7-854f-b3faa1f475f8@intel.com/#t

Cc: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@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>
Link: https://lore.kernel.org/r/20231113231047.548659-15-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Remove VF MAC types
Ivan Vecera [Mon, 13 Nov 2023 23:10:32 +0000 (15:10 -0800)]
i40e: Remove VF MAC types

The i40e_hw.mac.type cannot to be equal to I40E_MAC_VF or
I40E_MAC_X722_VF so remove helper i40e_is_vf(), simplify
i40e_adminq_init_regs() and remove enums for these VF MAC types.

Signed-off-by: Ivan Vecera <ivecera@redhat.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>
Link: https://lore.kernel.org/r/20231113231047.548659-14-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Use helpers to check running FW and AQ API versions
Ivan Vecera [Mon, 13 Nov 2023 23:10:31 +0000 (15:10 -0800)]
i40e: Use helpers to check running FW and AQ API versions

Use new helpers to check versions of running FW and provided
AQ API to make the code more readable.

Signed-off-by: Ivan Vecera <ivecera@redhat.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>
Link: https://lore.kernel.org/r/20231113231047.548659-13-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Add other helpers to check version of running firmware and AQ API
Ivan Vecera [Mon, 13 Nov 2023 23:10:30 +0000 (15:10 -0800)]
i40e: Add other helpers to check version of running firmware and AQ API

Add another helper functions that will be used by subsequent
patch to refactor existing open-coded checks whether the version
of running firmware and AdminQ API is recent enough to provide
certain capabilities.

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Link: https://lore.kernel.org/r/20231113231047.548659-12-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Move i40e_is_aq_api_ver_ge helper
Ivan Vecera [Mon, 13 Nov 2023 23:10:29 +0000 (15:10 -0800)]
i40e: Move i40e_is_aq_api_ver_ge helper

Move i40e_is_aq_api_ver_ge helper function (used to check if AdminQ
API version is recent enough) to header so it can be used from
other .c files.

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@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>
Link: https://lore.kernel.org/r/20231113231047.548659-11-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Initialize hardware capabilities at single place
Ivan Vecera [Mon, 13 Nov 2023 23:10:28 +0000 (15:10 -0800)]
i40e: Initialize hardware capabilities at single place

Some i40e_hw.caps bits are set in i40e_set_hw_caps(), some of them
in i40e_init_adminq() and the rest of them in i40e_sw_init().
Consolidate the initialization to single proper place i40e_set_hw_caps().

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jacob Keller <jacob.e.keller@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>
Link: https://lore.kernel.org/r/20231113231047.548659-10-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Consolidate hardware capabilities
Ivan Vecera [Mon, 13 Nov 2023 23:10:27 +0000 (15:10 -0800)]
i40e: Consolidate hardware capabilities

Fields .caps in i40e_hw and .hw_features in i40e_pf both indicate
capabilities provided by hardware. Move and merge i40e_pf.hw_features
into i40e_hw.caps as this is more appropriate place for them and
adjust their names to I40E_HW_CAP_... convention.

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jacob Keller <jacob.e.keller@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>
Link: https://lore.kernel.org/r/20231113231047.548659-9-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Use DECLARE_BITMAP for flags field in i40e_hw
Ivan Vecera [Mon, 13 Nov 2023 23:10:26 +0000 (15:10 -0800)]
i40e: Use DECLARE_BITMAP for flags field in i40e_hw

Convert flags field in i40e_hw from u64 to bitmap and its usage
to use bit access functions and rename the field to 'caps' as
this field describes capabilities that are set once on init and
read many times later.

Changes:
- Convert "hw_ptr->flags & FLAG" to "test_bit(FLAG, ...)"
- Convert "hw_ptr->flags |= FLAG" to "set_bit(FLAG, ...)"
- Convert "hw_ptr->flags &= ~FLAG" to "clear_bit(FLAG, ...)"
- Rename i40e_hw.flags to i40e_hw.caps
- Rename i40e_set_hw_flags() to i40e_set_hw_caps()
- Adjust caps names so they are prefixed by I40E_HW_CAP_ and existing
  _CAPABLE suffixes are stripped

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jacob Keller <jacob.e.keller@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>
Link: https://lore.kernel.org/r/20231113231047.548659-8-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Use DECLARE_BITMAP for flags and hw_features fields in i40e_pf
Ivan Vecera [Mon, 13 Nov 2023 23:10:25 +0000 (15:10 -0800)]
i40e: Use DECLARE_BITMAP for flags and hw_features fields in i40e_pf

Convert flags and hw_features fields from i40e_pf from u32 to
bitmaps and their usage to use bit access functions.

Changes:
- Convert "pf_ptr->(flags|hw_features) & FL" to "test_bit(FL, ...)"
- Convert "pf_ptr->(flags|hw_features) |= FL" to "set_bit(FL, ...)"
- Convert "pf_ptr->(flags|hw_features) &= ~FL" to "clear_bit(FL, ...)"
- Rename flag field to bitno in i40e_priv_flags and adjust ethtool
  callbacks to work with flags bitmap
- Rename flag names where '_ENABLED'->'_ENA' and '_DISABLED'->'_DIS'
  like in ice driver

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jacob Keller <jacob.e.keller@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>
Link: https://lore.kernel.org/r/20231113231047.548659-7-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Remove _t suffix from enum type names
Ivan Vecera [Mon, 13 Nov 2023 23:10:24 +0000 (15:10 -0800)]
i40e: Remove _t suffix from enum type names

Enum type names should not be suffixed by '_t'. Either to use
'typedef enum name name_t' to so plain 'name_t var' instead of
'enum name_t var'.

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jacob Keller <jacob.e.keller@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>
Link: https://lore.kernel.org/r/20231113231047.548659-6-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Remove unused flags
Ivan Vecera [Mon, 13 Nov 2023 23:10:23 +0000 (15:10 -0800)]
i40e: Remove unused flags

The flag I40E_FLAG_RX_CSUM_ENABLED and I40E_HW_FLAG_DROP_MODE are
set and never read. Remove them.

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jacob Keller <jacob.e.keller@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>
Link: https://lore.kernel.org/r/20231113231047.548659-5-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: Change user notification of non-SFP module in i40e_get_module_info()
Andrii Staikov [Mon, 13 Nov 2023 23:10:22 +0000 (15:10 -0800)]
i40e: Change user notification of non-SFP module in i40e_get_module_info()

Make the driver not produce an error message on
"ethtool -m ethX" command when a non-SFP module is encountered
hence there is no possibility to read the EEPROM.
Make the message to appear in the debug log rather
than as an error string.

Change the return code to -EOPNOTSUPP and the string to make
it more verbose.

Signed-off-by: Andrii Staikov <andrii.staikov@intel.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
CC: Stefan Assmann <sassmann@redhat.com>
CC: Michal Schmidt <mschmidt@redhat.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>
Link: https://lore.kernel.org/r/20231113231047.548659-4-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: add an error code check in i40e_vsi_setup
Su Hui [Mon, 13 Nov 2023 23:10:21 +0000 (15:10 -0800)]
i40e: add an error code check in i40e_vsi_setup

check the value of 'ret' after calling 'i40e_vsi_config_rss'.

Signed-off-by: Su Hui <suhui@nfschina.com>
Reviewed-by: Dan Carpenter <dan.carpenter@linaro.org>
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>
Link: https://lore.kernel.org/r/20231113231047.548659-3-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoi40e: increase max descriptors for XL710
Justin Bronder [Mon, 13 Nov 2023 23:10:20 +0000 (15:10 -0800)]
i40e: increase max descriptors for XL710

In Tables 8-12 and 8-22 in the X710/XXV710/XL710 datasheet, the QLEN
description states that the maximum size of the descriptor queue is 8k
minus 32, or 8160.

Signed-off-by: Justin Bronder <jsbronder@cold-front.org>
Reviewed-by: Jacob Keller <jacob.e.keller@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>
Link: https://lore.kernel.org/r/20231113231047.548659-2-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agonet: dsa: tag_rtl4_a: Use existing ETH_P_REALTEK constant
Florian Fainelli [Mon, 13 Nov 2023 16:50:30 +0000 (08:50 -0800)]
net: dsa: tag_rtl4_a: Use existing ETH_P_REALTEK constant

No functional change, uses the existing ETH_P_REALTEK constant already
defined in if_ether.h.

Signed-off-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20231113165030.2440083-1-florian.fainelli@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoMerge tag 'net-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Fri, 10 Nov 2023 01:09:35 +0000 (17:09 -0800)]
Merge tag 'net-6.7-rc1' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Including fixes from netfilter and bpf.

  Current release - regressions:

   - sched: fix SKB_NOT_DROPPED_YET splat under debug config

  Current release - new code bugs:

   - tcp:
       - fix usec timestamps with TCP fastopen
       - fix possible out-of-bounds reads in tcp_hash_fail()
       - fix SYN option room calculation for TCP-AO

   - tcp_sigpool: fix some off by one bugs

   - bpf: fix compilation error without CGROUPS

   - ptp:
       - ptp_read() should not release queue
       - fix tsevqs corruption

  Previous releases - regressions:

   - llc: verify mac len before reading mac header

  Previous releases - always broken:

   - bpf:
       - fix check_stack_write_fixed_off() to correctly spill imm
       - fix precision tracking for BPF_ALU | BPF_TO_BE | BPF_END
       - check map->usercnt after timer->timer is assigned

   - dsa: lan9303: consequently nested-lock physical MDIO

   - dccp/tcp: call security_inet_conn_request() after setting IP addr

   - tg3: fix the TX ring stall due to incorrect full ring handling

   - phylink: initialize carrier state at creation

   - ice: fix direction of VF rules in switchdev mode

  Misc:

   - fill in a bunch of missing MODULE_DESCRIPTION()s, more to come"

* tag 'net-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (84 commits)
  net: ti: icss-iep: fix setting counter value
  ptp: fix corrupted list in ptp_open
  ptp: ptp_read should not release queue
  net_sched: sch_fq: better validate TCA_FQ_WEIGHTS and TCA_FQ_PRIOMAP
  net: kcm: fill in MODULE_DESCRIPTION()
  net/sched: act_ct: Always fill offloading tuple iifidx
  netfilter: nat: fix ipv6 nat redirect with mapped and scoped addresses
  netfilter: xt_recent: fix (increase) ipv6 literal buffer length
  ipvs: add missing module descriptions
  netfilter: nf_tables: remove catchall element in GC sync path
  netfilter: add missing module descriptions
  drivers/net/ppp: use standard array-copy-function
  net: enetc: shorten enetc_setup_xdp_prog() error message to fit NETLINK_MAX_FMTMSG_LEN
  virtio/vsock: Fix uninit-value in virtio_transport_recv_pkt()
  r8169: respect userspace disabling IFF_MULTICAST
  selftests/bpf: get trusted cgrp from bpf_iter__cgroup directly
  bpf: Let verifier consider {task,cgroup} is trusted in bpf_iter_reg
  net: phylink: initialize carrier state at creation
  test/vsock: add dobule bind connect test
  test/vsock: refactor vsock_accept
  ...

10 months agoMerge tag 'v6.7-p2' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Fri, 10 Nov 2023 01:04:58 +0000 (17:04 -0800)]
Merge tag 'v6.7-p2' of git://git./linux/kernel/git/herbert/crypto-2.6

Pull crypto fixes from Herbert Xu:
 "This fixes a regression in ahash and hides the Kconfig sub-options for
  the jitter RNG"

* tag 'v6.7-p2' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: ahash - Set using_shash for cloned ahash wrapper over shash
  crypto: jitterentropy - Hide esoteric Kconfig options under FIPS and EXPERT

10 months agoMerge tag 'input-for-v6.7-rc0' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor...
Linus Torvalds [Thu, 9 Nov 2023 22:18:42 +0000 (14:18 -0800)]
Merge tag 'input-for-v6.7-rc0' of git://git./linux/kernel/git/dtor/input

Pull input updates from Dmitry Torokhov:

 - a number of input drivers has been converted to use facilities
   provided by the device core to instantiate driver-specific attributes
   instead of using devm_device_add_group() and similar APIs

 - platform input devices have been converted to use remove() callback
   returning void

 - a fix for use-after-free when tearing down a Synaptics RMI device

 - a few flexible arrays in input structures have been annotated with
   __counted_by to help hardening efforts

 - handling of vddio supply in cyttsp5 driver

 - other miscellaneous fixups

* tag 'input-for-v6.7-rc0' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input: (86 commits)
  Input: walkera0701 - use module_parport_driver macro to simplify the code
  Input: synaptics-rmi4 - fix use after free in rmi_unregister_function()
  dt-bindings: input: fsl,scu-key: Document wakeup-source
  Input: cyttsp5 - add handling for vddio regulator
  dt-bindings: input: cyttsp5: document vddio-supply
  Input: tegra-kbc - use device_get_match_data()
  Input: Annotate struct ff_device with __counted_by
  Input: axp20x-pek - avoid needless newline removal
  Input: mt - annotate struct input_mt with __counted_by
  Input: leds - annotate struct input_leds with __counted_by
  Input: evdev - annotate struct evdev_client with __counted_by
  Input: synaptics-rmi4 - replace deprecated strncpy
  Input: wm97xx-core - convert to platform remove callback returning void
  Input: wm831x-ts - convert to platform remove callback returning void
  Input: ti_am335x_tsc - convert to platform remove callback returning void
  Input: sun4i-ts - convert to platform remove callback returning void
  Input: stmpe-ts - convert to platform remove callback returning void
  Input: pcap_ts - convert to platform remove callback returning void
  Input: mc13783_ts - convert to platform remove callback returning void
  Input: mainstone-wm97xx - convert to platform remove callback returning void
  ...

10 months agoMerge tag 'for-6.7-rc1-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Thu, 9 Nov 2023 22:10:38 +0000 (14:10 -0800)]
Merge tag 'for-6.7-rc1-part2' of git://git./linux/kernel/git/wsa/linux

Pull more i2c updates from Wolfram Sang:
 "This contains one patch which slipped through the cracks (iproc), a
  core sanitizing improvement as the new memdup_array_user() helper went
  upstream (i2c-dev), and two driver bugfixes (designware, cp2615)"

* tag 'for-6.7-rc1-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: cp2615: Fix 'assignment to __be16' warning
  i2c: dev: copy userspace array safely
  i2c: designware: Disable TX_EMPTY irq while waiting for block length byte
  i2c: iproc: handle invalid slave state

10 months agoMerge tag 'linux-watchdog-6.7-rc1' of git://www.linux-watchdog.org/linux-watchdog
Linus Torvalds [Thu, 9 Nov 2023 21:54:25 +0000 (13:54 -0800)]
Merge tag 'linux-watchdog-6.7-rc1' of git://linux-watchdog.org/linux-watchdog

Pull watchdog updates from Wim Van Sebroeck:

 - add support for Amlogic C3 and S4 SoCs

 - add IT8613 ID

 - add MSM8226 and MSM8974 compatibles

 - other small fixes and improvements

* tag 'linux-watchdog-6.7-rc1' of git://www.linux-watchdog.org/linux-watchdog: (24 commits)
  dt-bindings: watchdog: Add support for Amlogic C3 and S4 SoCs
  watchdog: mlx-wdt: Parameter desctiption warning fix
  watchdog: aspeed: Add support for aspeed,reset-mask DT property
  dt-bindings: watchdog: aspeed-wdt: Add aspeed,reset-mask property
  watchdog: apple: Deactivate on suspend
  dt-bindings: watchdog: qcom-wdt: Add MSM8226 and MSM8974 compatibles
  dt-bindings: watchdog: fsl-imx7ulp-wdt: Add 'fsl,ext-reset-output'
  wdog: imx7ulp: Enable wdog int_en bit for watchdog any reset
  drivers: watchdog: marvell_gti: Program the max_hw_heartbeat_ms
  drivers: watchdog: marvell_gti: fix zero pretimeout handling
  watchdog: marvell_gti: Replace of_platform.h with explicit includes
  watchdog: imx_sc_wdt: continue if the wdog already enabled
  watchdog: st_lpc: Use device_get_match_data()
  watchdog: wdat_wdt: Add timeout value as a param in ping method
  watchdog: gpio_wdt: Make use of device properties
  sbsa_gwdt: Calculate timeout with 64-bit math
  watchdog: ixp4xx: Make sure restart always works
  watchdog: it87_wdt: add IT8613 ID
  watchdog: marvell_gti_wdt: Fix error code in probe()
  Watchdog: marvell_gti_wdt: Remove redundant dev_err_probe() for platform_get_irq()
  ...

10 months agoMerge tag 'pwm/for-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/thierry...
Linus Torvalds [Thu, 9 Nov 2023 21:47:52 +0000 (13:47 -0800)]
Merge tag 'pwm/for-6.7-rc1' of git://git./linux/kernel/git/thierry.reding/linux-pwm

Pull pwm updates from Thierry Reding:
 "This contains a few fixes and a bunch of cleanups, a lot of which is
  in preparation for Uwe's character device support that may be ready in
  time for the next merge window"

* tag 'pwm/for-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/thierry.reding/linux-pwm: (37 commits)
  pwm: samsung: Document new member .channel in struct samsung_pwm_chip
  pwm: bcm2835: Add support for suspend/resume
  pwm: brcmstb: Checked clk_prepare_enable() return value
  pwm: brcmstb: Utilize appropriate clock APIs in suspend/resume
  pwm: pxa: Explicitly include correct DT includes
  pwm: cros-ec: Simplify using devm_pwmchip_add() and dev_err_probe()
  pwm: samsung: Consistently use the same name for driver data
  pwm: vt8500: Simplify using devm functions
  pwm: sprd: Simplify using devm_pwmchip_add() and dev_err_probe()
  pwm: sprd: Provide a helper to cast a chip to driver data
  pwm: spear: Simplify using devm functions
  pwm: mtk-disp: Simplify using devm_pwmchip_add()
  pwm: imx-tpm: Simplify using devm functions
  pwm: brcmstb: Simplify using devm functions
  pwm: bcm2835: Simplify using devm functions
  pwm: bcm-iproc: Simplify using devm functions
  pwm: Adapt sysfs API documentation to reality
  pwm: dwc: add PWM bit unset in get_state call
  pwm: dwc: make timer clock configurable
  pwm: dwc: split pci out of core driver
  ...

10 months agoMerge tag 'iommu-updates-v6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/joro...
Linus Torvalds [Thu, 9 Nov 2023 21:37:28 +0000 (13:37 -0800)]
Merge tag 'iommu-updates-v6.7' of git://git./linux/kernel/git/joro/iommu

Pull iommu updates from Joerg Roedel:
 "Core changes:
   - Make default-domains mandatory for all IOMMU drivers
   - Remove group refcounting
   - Add generic_single_device_group() helper and consolidate drivers
   - Cleanup map/unmap ops
   - Scaling improvements for the IOVA rcache depot
   - Convert dart & iommufd to the new domain_alloc_paging()

  ARM-SMMU:
   - Device-tree binding update:
       - Add qcom,sm7150-smmu-v2 for Adreno on SM7150 SoC
   - SMMUv2:
       - Support for Qualcomm SDM670 (MDSS) and SM7150 SoCs
   - SMMUv3:
       - Large refactoring of the context descriptor code to move the CD
         table into the master, paving the way for '->set_dev_pasid()'
         support on non-SVA domains
   - Minor cleanups to the SVA code

  Intel VT-d:
   - Enable debugfs to dump domain attached to a pasid
   - Remove an unnecessary inline function

  AMD IOMMU:
   - Initial patches for SVA support (not complete yet)

  S390 IOMMU:
   - DMA-API conversion and optimized IOTLB flushing

  And some smaller fixes and improvements"

* tag 'iommu-updates-v6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu: (102 commits)
  iommu/dart: Remove the force_bypass variable
  iommu/dart: Call apple_dart_finalize_domain() as part of alloc_paging()
  iommu/dart: Convert to domain_alloc_paging()
  iommu/dart: Move the blocked domain support to a global static
  iommu/dart: Use static global identity domains
  iommufd: Convert to alloc_domain_paging()
  iommu/vt-d: Use ops->blocked_domain
  iommu/vt-d: Update the definition of the blocking domain
  iommu: Move IOMMU_DOMAIN_BLOCKED global statics to ops->blocked_domain
  Revert "iommu/vt-d: Remove unused function"
  iommu/amd: Remove DMA_FQ type from domain allocation path
  iommu: change iommu_map_sgtable to return signed values
  iommu/virtio: Add __counted_by for struct viommu_request and use struct_size()
  iommu/vt-d: debugfs: Support dumping a specified page table
  iommu/vt-d: debugfs: Create/remove debugfs file per {device, pasid}
  iommu/vt-d: debugfs: Dump entry pointing to huge page
  iommu/vt-d: Remove unused function
  iommu/arm-smmu-v3-sva: Remove bond refcount
  iommu/arm-smmu-v3-sva: Remove unused iommu_sva handle
  iommu/arm-smmu-v3: Rename cdcfg to cd_table
  ...

10 months agonet: ti: icss-iep: fix setting counter value
Diogo Ivo [Tue, 7 Nov 2023 12:00:36 +0000 (12:00 +0000)]
net: ti: icss-iep: fix setting counter value

Currently icss_iep_set_counter() writes the upper 32-bits of the
counter value to both the lower and upper counter registers, so
fix this by writing the appropriate value to the lower register.

Fixes: c1e0230eeaab ("net: ti: icss-iep: Add IEP driver")
Signed-off-by: Diogo Ivo <diogo.ivo@siemens.com>
Link: https://lore.kernel.org/r/20231107120037.1513546-1-diogo.ivo@siemens.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
10 months agoptp: fix corrupted list in ptp_open
Edward Adam Davis [Tue, 7 Nov 2023 08:00:41 +0000 (16:00 +0800)]
ptp: fix corrupted list in ptp_open

There is no lock protection when writing ptp->tsevqs in ptp_open() and
ptp_release(), which can cause data corruption, use spin lock to avoid this
issue.

Moreover, ptp_release() should not be used to release the queue in ptp_read(),
and it should be deleted altogether.

Acked-by: Richard Cochran <richardcochran@gmail.com>
Reported-and-tested-by: syzbot+df3f3ef31f60781fa911@syzkaller.appspotmail.com
Fixes: 8f5de6fb2453 ("ptp: support multiple timestamp event readers")
Signed-off-by: Edward Adam Davis <eadavis@qq.com>
Link: https://lore.kernel.org/r/tencent_CD19564FFE8DA8A5918DFE92325D92DD8107@qq.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoptp: ptp_read should not release queue
Edward Adam Davis [Tue, 7 Nov 2023 08:00:40 +0000 (16:00 +0800)]
ptp: ptp_read should not release queue

Firstly, queue is not the memory allocated in ptp_read;
Secondly, other processes may block at ptp_read and wait for conditions to be
met to perform read operations.

Acked-by: Richard Cochran <richardcochran@gmail.com>
Reported-and-tested-by: syzbot+df3f3ef31f60781fa911@syzkaller.appspotmail.com
Fixes: 8f5de6fb2453 ("ptp: support multiple timestamp event readers")
Signed-off-by: Edward Adam Davis <eadavis@qq.com>
Link: https://lore.kernel.org/r/tencent_18747D76F1675A3C633772960237544AAA09@qq.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoMerge branch '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net...
Jakub Kicinski [Thu, 9 Nov 2023 02:43:52 +0000 (18:43 -0800)]
Merge branch '100GbE' of git://git./linux/kernel/git/tnguy/net-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2023-11-06 (ice)

This series contains updates to ice driver only.

Dave removes SR-IOV LAG attribute for only the interface being disabled
to allow for proper unwinding of all interfaces.

Michal Schmidt changes some LAG allocations from GFP_KERNEL to GFP_ATOMIC
due to non-allowed sleeping.

Aniruddha and Marcin fix redirection and drop rules for switchdev by
properly setting and marking egress/ingress type.

* '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
  ice: Fix VF-VF direction matching in drop rule in switchdev
  ice: Fix VF-VF filter rules in switchdev mode
  ice: lag: in RCU, use atomic allocation
  ice: Fix SRIOV LAG disable on non-compliant aggregate
====================

Link: https://lore.kernel.org/r/20231107004844.655549-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agonet_sched: sch_fq: better validate TCA_FQ_WEIGHTS and TCA_FQ_PRIOMAP
Eric Dumazet [Tue, 7 Nov 2023 16:04:40 +0000 (16:04 +0000)]
net_sched: sch_fq: better validate TCA_FQ_WEIGHTS and TCA_FQ_PRIOMAP

syzbot was able to trigger the following report while providing
too small TCA_FQ_WEIGHTS attribute [1]

Fix is to use NLA_POLICY_EXACT_LEN() to ensure user space
provided correct sizes.

Apply the same fix to TCA_FQ_PRIOMAP.

[1]
BUG: KMSAN: uninit-value in fq_load_weights net/sched/sch_fq.c:960 [inline]
BUG: KMSAN: uninit-value in fq_change+0x1348/0x2fe0 net/sched/sch_fq.c:1071
fq_load_weights net/sched/sch_fq.c:960 [inline]
fq_change+0x1348/0x2fe0 net/sched/sch_fq.c:1071
fq_init+0x68e/0x780 net/sched/sch_fq.c:1159
qdisc_create+0x12f3/0x1be0 net/sched/sch_api.c:1326
tc_modify_qdisc+0x11ef/0x2c20
rtnetlink_rcv_msg+0x16a6/0x1840 net/core/rtnetlink.c:6558
netlink_rcv_skb+0x371/0x650 net/netlink/af_netlink.c:2545
rtnetlink_rcv+0x34/0x40 net/core/rtnetlink.c:6576
netlink_unicast_kernel net/netlink/af_netlink.c:1342 [inline]
netlink_unicast+0xf47/0x1250 net/netlink/af_netlink.c:1368
netlink_sendmsg+0x1238/0x13d0 net/netlink/af_netlink.c:1910
sock_sendmsg_nosec net/socket.c:730 [inline]
__sock_sendmsg net/socket.c:745 [inline]
____sys_sendmsg+0x9c2/0xd60 net/socket.c:2588
___sys_sendmsg+0x28d/0x3c0 net/socket.c:2642
__sys_sendmsg net/socket.c:2671 [inline]
__do_sys_sendmsg net/socket.c:2680 [inline]
__se_sys_sendmsg net/socket.c:2678 [inline]
__x64_sys_sendmsg+0x307/0x490 net/socket.c:2678
do_syscall_x64 arch/x86/entry/common.c:51 [inline]
do_syscall_64+0x44/0x110 arch/x86/entry/common.c:82
entry_SYSCALL_64_after_hwframe+0x63/0x6b

Uninit was created at:
slab_post_alloc_hook+0x129/0xa70 mm/slab.h:768
slab_alloc_node mm/slub.c:3478 [inline]
kmem_cache_alloc_node+0x5e9/0xb10 mm/slub.c:3523
kmalloc_reserve+0x13d/0x4a0 net/core/skbuff.c:560
__alloc_skb+0x318/0x740 net/core/skbuff.c:651
alloc_skb include/linux/skbuff.h:1286 [inline]
netlink_alloc_large_skb net/netlink/af_netlink.c:1214 [inline]
netlink_sendmsg+0xb34/0x13d0 net/netlink/af_netlink.c:1885
sock_sendmsg_nosec net/socket.c:730 [inline]
__sock_sendmsg net/socket.c:745 [inline]
____sys_sendmsg+0x9c2/0xd60 net/socket.c:2588
___sys_sendmsg+0x28d/0x3c0 net/socket.c:2642
__sys_sendmsg net/socket.c:2671 [inline]
__do_sys_sendmsg net/socket.c:2680 [inline]
__se_sys_sendmsg net/socket.c:2678 [inline]
__x64_sys_sendmsg+0x307/0x490 net/socket.c:2678
do_syscall_x64 arch/x86/entry/common.c:51 [inline]
do_syscall_64+0x44/0x110 arch/x86/entry/common.c:82
entry_SYSCALL_64_after_hwframe+0x63/0x6b

CPU: 1 PID: 5001 Comm: syz-executor300 Not tainted 6.6.0-syzkaller-12401-g8f6f76a6a29f #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/09/2023

Fixes: 29f834aa326e ("net_sched: sch_fq: add 3 bands and WRR scheduling")
Fixes: 49e7265fd098 ("net_sched: sch_fq: add TCA_FQ_WEIGHTS attribute")
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Jamal Hadi Salim<jhs@mojatatu.com>
Link: https://lore.kernel.org/r/20231107160440.1992526-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoMerge branch '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue
Jakub Kicinski [Thu, 9 Nov 2023 02:20:13 +0000 (18:20 -0800)]
Merge branch '40GbE' of git://git./linux/kernel/git/tnguy/net-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2023-11-06 (i40e)

This series contains updates to i40e driver only.

Ivan Vecera resolves a couple issues with devlink; removing a call to
devlink_port_type_clear() and ensuring devlink port is unregistered
after the net device.

* '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
  i40e: Fix devlink port unregistering
  i40e: Do not call devlink_port_type_clear()
====================

Link: https://lore.kernel.org/r/20231107003600.653796-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agonet: kcm: fill in MODULE_DESCRIPTION()
Jakub Kicinski [Wed, 8 Nov 2023 02:03:05 +0000 (18:03 -0800)]
net: kcm: fill in MODULE_DESCRIPTION()

W=1 builds now warn if module is built without a MODULE_DESCRIPTION().

Link: https://lore.kernel.org/r/20231108020305.537293-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoMerge tag 'nf-23-11-08' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf
Jakub Kicinski [Thu, 9 Nov 2023 02:14:59 +0000 (18:14 -0800)]
Merge tag 'nf-23-11-08' of git://git./linux/kernel/git/netfilter/nf

Pablo Neira Ayuso says:

====================
Netfilter fixes for net

The following patchset contains Netfilter fixes for net:

1) Add missing netfilter modules description to fix W=1, from Florian Westphal.

2) Fix catch-all element GC with timeout when use with the pipapo set
   backend, this remained broken since I tried to fix it this summer,
   then another attempt to fix it recently.

3) Add missing IPVS modules descriptions to fix W=1, also from Florian.

4) xt_recent allocated a too small buffer to store an IPv4-mapped IPv6
   address which can be parsed by in6_pton(), from Maciej Zenczykowski.
   Broken for many releases.

5) Skip IPv4-mapped IPv6, IPv4-compat IPv6, site/link local scoped IPv6
   addressses to set up IPv6 NAT redirect, also from Florian. This is
   broken since 2012.

* tag 'nf-23-11-08' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
  netfilter: nat: fix ipv6 nat redirect with mapped and scoped addresses
  netfilter: xt_recent: fix (increase) ipv6 literal buffer length
  ipvs: add missing module descriptions
  netfilter: nf_tables: remove catchall element in GC sync path
  netfilter: add missing module descriptions
====================

Link: https://lore.kernel.org/r/20231108155802.84617-1-pablo@netfilter.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoMerge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
Jakub Kicinski [Thu, 9 Nov 2023 01:56:13 +0000 (17:56 -0800)]
Merge tag 'for-netdev' of https://git./linux/kernel/git/bpf/bpf

Daniel Borkmann says:

====================
pull-request: bpf 2023-11-08

We've added 16 non-merge commits during the last 6 day(s) which contain
a total of 30 files changed, 341 insertions(+), 130 deletions(-).

The main changes are:

1) Fix a BPF verifier issue in precision tracking for BPF_ALU | BPF_TO_BE |
   BPF_END where the source register was incorrectly marked as precise,
   from Shung-Hsi Yu.

2) Fix a concurrency issue in bpf_timer where the former could still have
   been alive after an application releases or unpins the map, from Hou Tao.

3) Fix a BPF verifier issue where immediates are incorrectly cast to u32
   before being spilled and therefore losing sign information, from Hao Sun.

4) Fix a misplaced BPF_TRACE_ITER in check_css_task_iter_allowlist which
   incorrectly compared bpf_prog_type with bpf_attach_type, from Chuyi Zhou.

5) Add __bpf_hook_{start,end} as well as __bpf_kfunc_{start,end}_defs macros,
   migrate all BPF-related __diag callsites over to it, and add a new
   __diag_ignore_all for -Wmissing-declarations to the macros to address
   recent build warnings, from Dave Marchevsky.

6) Fix broken BPF selftest build of xdp_hw_metadata test on architectures
   where char is not signed, from Björn Töpel.

7) Fix test_maps selftest to properly use LIBBPF_OPTS() macro to initialize
   the bpf_map_create_opts, from Andrii Nakryiko.

8) Fix bpffs selftest to avoid unmounting /sys/kernel/debug as it may have
   been mounted and used by other applications already, from Manu Bretelle.

9) Fix a build issue without CONFIG_CGROUPS wrt css_task open-coded
   iterators, from Matthieu Baerts.

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
  selftests/bpf: get trusted cgrp from bpf_iter__cgroup directly
  bpf: Let verifier consider {task,cgroup} is trusted in bpf_iter_reg
  selftests/bpf: Fix broken build where char is unsigned
  selftests/bpf: precision tracking test for BPF_NEG and BPF_END
  bpf: Fix precision tracking for BPF_ALU | BPF_TO_BE | BPF_END
  selftests/bpf: Add test for using css_task iter in sleepable progs
  selftests/bpf: Add tests for css_task iter combining with cgroup iter
  bpf: Relax allowlist for css_task iter
  selftests/bpf: fix test_maps' use of bpf_map_create_opts
  bpf: Check map->usercnt after timer->timer is assigned
  bpf: Add __bpf_hook_{start,end} macros
  bpf: Add __bpf_kfunc_{start,end}_defs macros
  selftests/bpf: fix test_bpffs
  selftests/bpf: Add test for immediate spilled to stack
  bpf: Fix check_stack_write_fixed_off() to correctly spill imm
  bpf: fix compilation error without CGROUPS
====================

Link: https://lore.kernel.org/r/20231108132448.1970-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agonet/sched: act_ct: Always fill offloading tuple iifidx
Vlad Buslov [Fri, 3 Nov 2023 15:14:10 +0000 (16:14 +0100)]
net/sched: act_ct: Always fill offloading tuple iifidx

Referenced commit doesn't always set iifidx when offloading the flow to
hardware. Fix the following cases:

- nf_conn_act_ct_ext_fill() is called before extension is created with
nf_conn_act_ct_ext_add() in tcf_ct_act(). This can cause rule offload with
unspecified iifidx when connection is offloaded after only single
original-direction packet has been processed by tc data path. Always fill
the new nf_conn_act_ct_ext instance after creating it in
nf_conn_act_ct_ext_add().

- Offloading of unidirectional UDP NEW connections is now supported, but ct
flow iifidx field is not updated when connection is promoted to
bidirectional which can result reply-direction iifidx to be zero when
refreshing the connection. Fill in the extension and update flow iifidx
before calling flow_offload_refresh().

Fixes: 9795ded7f924 ("net/sched: act_ct: Fill offloading tuple iifidx")
Reviewed-by: Paul Blakey <paulb@nvidia.com>
Signed-off-by: Vlad Buslov <vladbu@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Fixes: 6a9bad0069cf ("net/sched: act_ct: offload UDP NEW connections")
Link: https://lore.kernel.org/r/20231103151410.764271-1-vladbu@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoMerge tag 'nfs-for-6.7-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Linus Torvalds [Wed, 8 Nov 2023 21:39:16 +0000 (13:39 -0800)]
Merge tag 'nfs-for-6.7-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client updates from Trond Myklebust:
 "Bugfixes:

   - SUNRPC:
       - re-probe the target RPC port after an ECONNRESET error
       - handle allocation errors from rpcb_call_async()
       - fix a use-after-free condition in rpc_pipefs
       - fix up various checks for timeouts

   - NFSv4.1:
       - Handle NFS4ERR_DELAY errors during session trunking
       - fix SP4_MACH_CRED protection for pnfs IO

   - NFSv4:
       - Ensure that we test all delegations when the server notifies
         us that it may have revoked some of them

  Features:

   - Allow knfsd processes to break out of NFS4ERR_DELAY loops when
     re-exporting NFSv4.x by setting appropriate values for the
     'delay_retrans' module parameter

   - nfs: Convert nfs_symlink() to use a folio"

* tag 'nfs-for-6.7-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  nfs: Convert nfs_symlink() to use a folio
  SUNRPC: Fix RPC client cleaned up the freed pipefs dentries
  NFSv4.1: fix SP4_MACH_CRED protection for pnfs IO
  SUNRPC: Add an IS_ERR() check back to where it was
  NFSv4.1: fix handling NFS4ERR_DELAY when testing for session trunking
  nfs41: drop dependency between flexfiles layout driver and NFSv3 modules
  NFSv4: fairly test all delegations on a SEQ4_ revocation
  SUNRPC: SOFTCONN tasks should time out when on the sending list
  SUNRPC: Force close the socket when a hard error is reported
  SUNRPC: Don't skip timeout checks in call_connect_status()
  SUNRPC: ECONNRESET might require a rebind
  NFSv4/pnfs: Allow layoutget to return EAGAIN for softerr mounts
  NFSv4: Add a parameter to limit the number of retries after NFS4ERR_DELAY

10 months agoMerge tag 'exfat-for-6.7-rc1-part2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 8 Nov 2023 21:33:47 +0000 (13:33 -0800)]
Merge tag 'exfat-for-6.7-rc1-part2' of git://git./linux/kernel/git/linkinjeon/exfat

Pull exfat updates from Namjae Jeon:

 - Fix an issue that exfat timestamps are not updated caused by new
   timestamp accessor function patch

* tag 'exfat-for-6.7-rc1-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/linkinjeon/exfat:
  exfat: fix ctime is not updated
  exfat: fix setting uninitialized time to ctime/atime

10 months agoMerge tag 'xfs-6.7-merge-2' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Linus Torvalds [Wed, 8 Nov 2023 21:22:16 +0000 (13:22 -0800)]
Merge tag 'xfs-6.7-merge-2' of git://git./fs/xfs/xfs-linux

Pull xfs updates from Chandan Babu:

 - Realtime device subsystem:
    - Cleanup usage of xfs_rtblock_t and xfs_fsblock_t data types
    - Replace open coded conversions between rt blocks and rt extents
      with calls to static inline helpers
    - Replace open coded realtime geometry compuation and macros with
      helper functions
    - CPU usage optimizations for realtime allocator
    - Misc bug fixes associated with Realtime device

 - Allow read operations to execute while an FICLONE ioctl is being
   serviced

 - Misc bug fixes:
    - Alert user when xfs_droplink() encounters an inode with a link
      count of zero
    - Handle the case where the allocator could return zero extents when
      servicing an fallocate request

* tag 'xfs-6.7-merge-2' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (40 commits)
  xfs: allow read IO and FICLONE to run concurrently
  xfs: handle nimaps=0 from xfs_bmapi_write in xfs_alloc_file_space
  xfs: introduce protection for drop nlink
  xfs: don't look for end of extent further than necessary in xfs_rtallocate_extent_near()
  xfs: don't try redundant allocations in xfs_rtallocate_extent_near()
  xfs: limit maxlen based on available space in xfs_rtallocate_extent_near()
  xfs: return maximum free size from xfs_rtany_summary()
  xfs: invert the realtime summary cache
  xfs: simplify rt bitmap/summary block accessor functions
  xfs: simplify xfs_rtbuf_get calling conventions
  xfs: cache last bitmap block in realtime allocator
  xfs: use accessor functions for summary info words
  xfs: consolidate realtime allocation arguments
  xfs: create helpers for rtsummary block/wordcount computations
  xfs: use accessor functions for bitmap words
  xfs: create helpers for rtbitmap block/wordcount computations
  xfs: create a helper to handle logging parts of rt bitmap/summary blocks
  xfs: convert rt summary macros to helpers
  xfs: convert open-coded xfs_rtword_t pointer accesses to helper
  xfs: remove XFS_BLOCKWSIZE and XFS_BLOCKWMASK macros
  ...

10 months agoMAINTAINERS: update lists.linuxfoundation.org migrated lists
Konstantin Ryabitsev [Tue, 7 Nov 2023 20:00:32 +0000 (15:00 -0500)]
MAINTAINERS: update lists.linuxfoundation.org migrated lists

The mailman-2 system behind lists.linux[-]foundation.org is being
retired, so the lists are being migrated to lists.linux.dev.

Since both domains belong to LF and setting up proper forwards is
possible, the old addresses will continue to work for a while, but all
new patches should be sent to the new canonical addresses for each list.

Signed-off-by: Konstantin Ryabitsev <konstantin@linuxfoundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agoMerge tag 's390-6.7-2' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
Linus Torvalds [Wed, 8 Nov 2023 20:39:54 +0000 (12:39 -0800)]
Merge tag 's390-6.7-2' of git://git./linux/kernel/git/s390/linux

Pull more s390 updates from Vasily Gorbik:

 - Get rid of s390 specific use of two PTEs per 4KB page with complex
   half-used pages tracking. Using full 4KB pages for 2KB PTEs increases
   the memory footprint of page tables but drastically simplify mm code,
   removing a common blocker for common code changes and adaptations

 - Simplify and rework "cmma no-dat" handling. This is a follow up for
   recent fixes which prevent potential incorrect guest TLB flushes

 - Add perf user stack unwinding as well as USER_STACKTRACE support for
   user space built with -mbackchain compile option

 - Add few missing conversion from tlb_remove_table to tlb_remove_ptdesc

 - Fix crypto cards vanishing in a secure execution environment due to
   asynchronous errors

 - Avoid reporting crypto cards or queues in check-stop state as online

 - Fix null-ptr deference in AP bus code triggered by early config
   change via SCLP

 - Couple of stability improvements in AP queue interrupt handling

* tag 's390-6.7-2' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
  s390/mm: make pte_free_tlb() similar to pXd_free_tlb()
  s390/mm: use compound page order to distinguish page tables
  s390/mm: use full 4KB page for 2KB PTE
  s390/cmma: rework no-dat handling
  s390/cmma: move arch_set_page_dat() to header file
  s390/cmma: move set_page_stable() and friends to header file
  s390/cmma: move parsing of cmma kernel parameter to early boot code
  s390/cmma: cleanup inline assemblies
  s390/ap: fix vanishing crypto cards in SE environment
  s390/zcrypt: don't report online if card or queue is in check-stop state
  s390: add USER_STACKTRACE support
  s390/perf: implement perf_callchain_user()
  s390/ap: fix AP bus crash on early config change callback invocation
  s390/ap: re-enable interrupt for AP queues
  s390/ap: rework to use irq info from ap queue status
  s390/mm: add missing conversion to use ptdescs

10 months agoMerge tag 'rcu-fixes-v6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/frederic...
Linus Torvalds [Wed, 8 Nov 2023 17:47:52 +0000 (09:47 -0800)]
Merge tag 'rcu-fixes-v6.7' of git://git./linux/kernel/git/frederic/linux-dynticks

Pull RCU fixes from Frederic Weisbecker:

 - Fix a lock inversion between scheduler and RCU introduced in
   v6.2-rc4. The scenario could trigger on any user of RCU_NOCB
   (mostly Android but also nohz_full)

 - Fix PF_IDLE semantic changes introduced in v6.6-rc3 breaking
   some RCU-Tasks and RCU-Tasks-Trace expectations as to what
   exactly is an idle task. This resulted in potential spurious
   stalls and warnings.

* tag 'rcu-fixes-v6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/frederic/linux-dynticks:
  rcu/tasks-trace: Handle new PF_IDLE semantics
  rcu/tasks: Handle new PF_IDLE semantics
  rcu: Introduce rcu_cpu_online()
  rcu: Break rcu_node_0 --> &rq->__lock order

10 months agoMerge tag 'memblock-v6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rppt...
Linus Torvalds [Wed, 8 Nov 2023 17:40:13 +0000 (09:40 -0800)]
Merge tag 'memblock-v6.7-rc1' of git://git./linux/kernel/git/rppt/memblock

Pull memblock update from Mike Rapoport:
 "Report failures when memblock_can_resize is not set.

  Numerous memblock reservations at early boot may exhaust static
  memblock.reserved array and it is unnoticed because most of the
  callers don't check memblock_reserve() return value.

  In this case the system will crash later, but the reason is hard to
  identify.

  Replace return of an error with panic() when memblock.reserved is
  exhausted before it can be resized"

* tag 'memblock-v6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rppt/memblock:
  memblock: report failures when memblock_can_resize is not set

10 months agoMerge tag 'kgdb-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/danielt...
Linus Torvalds [Wed, 8 Nov 2023 17:28:38 +0000 (09:28 -0800)]
Merge tag 'kgdb-6.7-rc1' of git://git./linux/kernel/git/danielt/linux

Pull kgdb updates from Daniel Thompson:
 "Just two patches for you this time!

   - During a panic, flush the console before entering kgdb.

     This makes things a little easier to comprehend, especially if an
     NMI backtrace was triggered on all CPUs just before we enter the
     panic routines

   - Correcting a couple of misleading (a.k.a. plain wrong) comments"

* tag 'kgdb-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/danielt/linux:
  kdb: Corrects comment for kdballocenv
  kgdb: Flush console before entering kgdb on panic

10 months agoMerge tag 'riscv-for-linus-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 8 Nov 2023 17:21:18 +0000 (09:21 -0800)]
Merge tag 'riscv-for-linus-6.7-rc1' of git://git./linux/kernel/git/riscv/linux

Pull RISC-V updates from Palmer Dabbelt:

 - Support for cbo.zero in userspace

 - Support for CBOs on ACPI-based systems

 - A handful of improvements for the T-Head cache flushing ops

 - Support for software shadow call stacks

 - Various cleanups and fixes

* tag 'riscv-for-linus-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux: (31 commits)
  RISC-V: hwprobe: Fix vDSO SIGSEGV
  riscv: configs: defconfig: Enable configs required for RZ/Five SoC
  riscv: errata: prefix T-Head mnemonics with th.
  riscv: put interrupt entries into .irqentry.text
  riscv: mm: Update the comment of CONFIG_PAGE_OFFSET
  riscv: Using TOOLCHAIN_HAS_ZIHINTPAUSE marco replace zihintpause
  riscv/mm: Fix the comment for swap pte format
  RISC-V: clarify the QEMU workaround in ISA parser
  riscv: correct pt_level name via pgtable_l5/4_enabled
  RISC-V: Provide pgtable_l5_enabled on rv32
  clocksource: timer-riscv: Increase rating of clock_event_device for Sstc
  clocksource: timer-riscv: Don't enable/disable timer interrupt
  lkdtm: Fix CFI_BACKWARD on RISC-V
  riscv: Use separate IRQ shadow call stacks
  riscv: Implement Shadow Call Stack
  riscv: Move global pointer loading to a macro
  riscv: Deduplicate IRQ stack switching
  riscv: VMAP_STACK overflow detection thread-safe
  RISC-V: cacheflush: Initialize CBO variables on ACPI systems
  RISC-V: ACPI: RHCT: Add function to get CBO block sizes
  ...

10 months agonetfilter: nat: fix ipv6 nat redirect with mapped and scoped addresses
Florian Westphal [Wed, 8 Nov 2023 12:18:53 +0000 (13:18 +0100)]
netfilter: nat: fix ipv6 nat redirect with mapped and scoped addresses

The ipv6 redirect target was derived from the ipv4 one, i.e. its
identical to a 'dnat' with the first (primary) address assigned to the
network interface.  The code has been moved around to make it usable
from nf_tables too, but its still the same as it was back when this
was added in 2012.

IPv6, however, has different types of addresses, if the 'wrong' address
comes first the redirection does not work.

In Daniels case, the addresses are:
  inet6 ::ffff:192 ...
  inet6 2a01: ...

... so the function attempts to redirect to the mapped address.

Add more checks before the address is deemed correct:
1. If the packets' daddr is scoped, search for a scoped address too
2. skip tentative addresses
3. skip mapped addresses

Use the first address that appears to match our needs.

Reported-by: Daniel Huhardeaux <tech@tootai.net>
Closes: https://lore.kernel.org/netfilter/71be06b8-6aa0-4cf9-9e0b-e2839b01b22f@tootai.net/
Fixes: 115e23ac78f8 ("netfilter: ip6tables: add REDIRECT target")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
10 months agonetfilter: xt_recent: fix (increase) ipv6 literal buffer length
Maciej Żenczykowski [Sun, 5 Nov 2023 19:56:00 +0000 (11:56 -0800)]
netfilter: xt_recent: fix (increase) ipv6 literal buffer length

in6_pton() supports 'low-32-bit dot-decimal representation'
(this is useful with DNS64/NAT64 networks for example):

  # echo +aaaa:bbbb:cccc:dddd:eeee:ffff:1.2.3.4 > /proc/self/net/xt_recent/DEFAULT
  # cat /proc/self/net/xt_recent/DEFAULT
  src=aaaa:bbbb:cccc:dddd:eeee:ffff:0102:0304 ttl: 0 last_seen: 9733848829 oldest_pkt: 1 9733848829

but the provided buffer is too short:

  # echo +aaaa:bbbb:cccc:dddd:eeee:ffff:255.255.255.255 > /proc/self/net/xt_recent/DEFAULT
  -bash: echo: write error: Invalid argument

Fixes: 079aa88fe717 ("netfilter: xt_recent: IPv6 support")
Signed-off-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
10 months agoipvs: add missing module descriptions
Florian Westphal [Tue, 7 Nov 2023 09:48:04 +0000 (10:48 +0100)]
ipvs: add missing module descriptions

W=1 builds warn on missing MODULE_DESCRIPTION, add them.

Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
10 months agonetfilter: nf_tables: remove catchall element in GC sync path
Pablo Neira Ayuso [Mon, 6 Nov 2023 09:53:09 +0000 (10:53 +0100)]
netfilter: nf_tables: remove catchall element in GC sync path

The expired catchall element is not deactivated and removed from GC sync
path. This path holds mutex so just call nft_setelem_data_deactivate()
and nft_setelem_catchall_remove() before queueing the GC work.

Fixes: 4a9e12ea7e70 ("netfilter: nft_set_pipapo: call nft_trans_gc_queue_sync() in catchall GC")
Reported-by: lonial con <kongln9170@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
10 months agonetfilter: add missing module descriptions
Florian Westphal [Sat, 4 Nov 2023 10:14:05 +0000 (11:14 +0100)]
netfilter: add missing module descriptions

W=1 builds warn on missing MODULE_DESCRIPTION, add them.

Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
10 months agodrivers/net/ppp: use standard array-copy-function
Philipp Stanner [Mon, 6 Nov 2023 09:16:00 +0000 (10:16 +0100)]
drivers/net/ppp: use standard array-copy-function

In ppp_generic.c, memdup_user() is utilized to copy a userspace array.
This is done without an overflow-check, which is, however, not critical
because the multiplicands are an unsigned short and struct sock_filter,
which is currently of size 8.

Regardless, string.h now provides memdup_array_user(), a wrapper for
copying userspace arrays in a standardized manner, which has the
advantage of making it more obvious to the reader that an array is being
copied.
The wrapper additionally performs an obligatory overflow check, saving
the reader the effort of analyzing the potential for overflow, and
making the code a bit more robust in case of future changes to the
multiplicands len * size.

Replace memdup_user() with memdup_array_user().

Suggested-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Philipp Stanner <pstanner@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agoi2c: cp2615: Fix 'assignment to __be16' warning
Bence Csókás [Mon, 30 Oct 2023 17:19:10 +0000 (17:19 +0000)]
i2c: cp2615: Fix 'assignment to __be16' warning

While the preamble field _is_ technically big-endian, its value is always 0x2A2A,
which is the same in either endianness. However, to avoid generating a warning,
we should still call `htons()` explicitly.

Signed-off-by: Bence Csókás <bence98@sch.bme.hu>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
10 months agoi2c: dev: copy userspace array safely
Philipp Stanner [Thu, 2 Nov 2023 19:26:13 +0000 (20:26 +0100)]
i2c: dev: copy userspace array safely

i2c-dev.c utilizes memdup_user() to copy a userspace array. This is done
without an overflow check.

Use the new wrapper memdup_array_user() to copy the array more safely.

Suggested-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Philipp Stanner <pstanner@redhat.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
10 months agoi2c: designware: Disable TX_EMPTY irq while waiting for block length byte
Tam Nguyen [Thu, 2 Nov 2023 03:30:08 +0000 (10:30 +0700)]
i2c: designware: Disable TX_EMPTY irq while waiting for block length byte

During SMBus block data read process, we have seen high interrupt rate
because of TX_EMPTY irq status while waiting for block length byte (the
first data byte after the address phase). The interrupt handler does not
do anything because the internal state is kept as STATUS_WRITE_IN_PROGRESS.
Hence, we should disable TX_EMPTY IRQ until I2C DesignWare receives
first data byte from I2C device, then re-enable it to resume SMBus
transaction.

It takes 0.789 ms for host to receive data length from slave.
Without the patch, i2c_dw_isr() is called 99 times by TX_EMPTY interrupt.
And it is none after applying the patch.

Cc: stable@vger.kernel.org
Co-developed-by: Chuong Tran <chuong@os.amperecomputing.com>
Signed-off-by: Chuong Tran <chuong@os.amperecomputing.com>
Signed-off-by: Tam Nguyen <tamnguyenchi@os.amperecomputing.com>
Acked-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Reviewed-by: Serge Semin <fancer.lancer@gmail.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
10 months agoi2c: iproc: handle invalid slave state
Roman Bacik [Thu, 24 Aug 2023 21:23:51 +0000 (14:23 -0700)]
i2c: iproc: handle invalid slave state

Add the code to handle an invalid state when both bits S_RX_EVENT
(indicating a transaction) and S_START_BUSY (indicating the end
of transaction - transition of START_BUSY from 1 to 0) are set in
the interrupt status register during a slave read.

Signed-off-by: Roman Bacik <roman.bacik@broadcom.com>
Fixes: 1ca1b4516088 ("i2c: iproc: handle Master aborted error")
Acked-by: Ray Jui <ray.jui@broadcom.com>
Signed-off-by: Wolfram Sang <wsa@kernel.org>
10 months agonet: enetc: shorten enetc_setup_xdp_prog() error message to fit NETLINK_MAX_FMTMSG_LEN
Vladimir Oltean [Mon, 6 Nov 2023 16:03:11 +0000 (18:03 +0200)]
net: enetc: shorten enetc_setup_xdp_prog() error message to fit NETLINK_MAX_FMTMSG_LEN

NETLINK_MAX_FMTMSG_LEN is currently hardcoded to 80, and we provide an
error printf-formatted string having 96 characters including the
terminating \0. Assuming each %d (representing a queue) gets replaced by
a number having at most 2 digits (a reasonable assumption), the final
string is also 96 characters wide, which is too much.

Reduce the verbiage a bit by removing some (partially) redundant words,
which makes the new printf-formatted string be 73 characters wide with
the trailing newline.

Fixes: 800db2d125c2 ("net: enetc: ensure we always have a minimum number of TXQs for stack")
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/lkml/202311061336.4dsWMT1h-lkp@intel.com/
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20231106160311.616118-1-vladimir.oltean@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agovirtio/vsock: Fix uninit-value in virtio_transport_recv_pkt()
Shigeru Yoshida [Sat, 4 Nov 2023 15:05:31 +0000 (00:05 +0900)]
virtio/vsock: Fix uninit-value in virtio_transport_recv_pkt()

KMSAN reported the following uninit-value access issue:

=====================================================
BUG: KMSAN: uninit-value in virtio_transport_recv_pkt+0x1dfb/0x26a0 net/vmw_vsock/virtio_transport_common.c:1421
 virtio_transport_recv_pkt+0x1dfb/0x26a0 net/vmw_vsock/virtio_transport_common.c:1421
 vsock_loopback_work+0x3bb/0x5a0 net/vmw_vsock/vsock_loopback.c:120
 process_one_work kernel/workqueue.c:2630 [inline]
 process_scheduled_works+0xff6/0x1e60 kernel/workqueue.c:2703
 worker_thread+0xeca/0x14d0 kernel/workqueue.c:2784
 kthread+0x3cc/0x520 kernel/kthread.c:388
 ret_from_fork+0x66/0x80 arch/x86/kernel/process.c:147
 ret_from_fork_asm+0x11/0x20 arch/x86/entry/entry_64.S:304

Uninit was stored to memory at:
 virtio_transport_space_update net/vmw_vsock/virtio_transport_common.c:1274 [inline]
 virtio_transport_recv_pkt+0x1ee8/0x26a0 net/vmw_vsock/virtio_transport_common.c:1415
 vsock_loopback_work+0x3bb/0x5a0 net/vmw_vsock/vsock_loopback.c:120
 process_one_work kernel/workqueue.c:2630 [inline]
 process_scheduled_works+0xff6/0x1e60 kernel/workqueue.c:2703
 worker_thread+0xeca/0x14d0 kernel/workqueue.c:2784
 kthread+0x3cc/0x520 kernel/kthread.c:388
 ret_from_fork+0x66/0x80 arch/x86/kernel/process.c:147
 ret_from_fork_asm+0x11/0x20 arch/x86/entry/entry_64.S:304

Uninit was created at:
 slab_post_alloc_hook+0x105/0xad0 mm/slab.h:767
 slab_alloc_node mm/slub.c:3478 [inline]
 kmem_cache_alloc_node+0x5a2/0xaf0 mm/slub.c:3523
 kmalloc_reserve+0x13c/0x4a0 net/core/skbuff.c:559
 __alloc_skb+0x2fd/0x770 net/core/skbuff.c:650
 alloc_skb include/linux/skbuff.h:1286 [inline]
 virtio_vsock_alloc_skb include/linux/virtio_vsock.h:66 [inline]
 virtio_transport_alloc_skb+0x90/0x11e0 net/vmw_vsock/virtio_transport_common.c:58
 virtio_transport_reset_no_sock net/vmw_vsock/virtio_transport_common.c:957 [inline]
 virtio_transport_recv_pkt+0x1279/0x26a0 net/vmw_vsock/virtio_transport_common.c:1387
 vsock_loopback_work+0x3bb/0x5a0 net/vmw_vsock/vsock_loopback.c:120
 process_one_work kernel/workqueue.c:2630 [inline]
 process_scheduled_works+0xff6/0x1e60 kernel/workqueue.c:2703
 worker_thread+0xeca/0x14d0 kernel/workqueue.c:2784
 kthread+0x3cc/0x520 kernel/kthread.c:388
 ret_from_fork+0x66/0x80 arch/x86/kernel/process.c:147
 ret_from_fork_asm+0x11/0x20 arch/x86/entry/entry_64.S:304

CPU: 1 PID: 10664 Comm: kworker/1:5 Not tainted 6.6.0-rc3-00146-g9f3ebbef746f #3
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-1.fc38 04/01/2014
Workqueue: vsock-loopback vsock_loopback_work
=====================================================

The following simple reproducer can cause the issue described above:

int main(void)
{
  int sock;
  struct sockaddr_vm addr = {
    .svm_family = AF_VSOCK,
    .svm_cid = VMADDR_CID_ANY,
    .svm_port = 1234,
  };

  sock = socket(AF_VSOCK, SOCK_STREAM, 0);
  connect(sock, (struct sockaddr *)&addr, sizeof(addr));
  return 0;
}

This issue occurs because the `buf_alloc` and `fwd_cnt` fields of the
`struct virtio_vsock_hdr` are not initialized when a new skb is allocated
in `virtio_transport_init_hdr()`. This patch resolves the issue by
initializing these fields during allocation.

Fixes: 71dc9ec9ac7d ("virtio/vsock: replace virtio_vsock_pkt with sk_buff")
Reported-and-tested-by: syzbot+0c8ce1da0ac31abbadcd@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=0c8ce1da0ac31abbadcd
Signed-off-by: Shigeru Yoshida <syoshida@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20231104150531.257952-1-syoshida@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agor8169: respect userspace disabling IFF_MULTICAST
Heiner Kallweit [Sun, 5 Nov 2023 22:43:36 +0000 (23:43 +0100)]
r8169: respect userspace disabling IFF_MULTICAST

So far we ignore the setting of IFF_MULTICAST. Fix this and clear bit
AcceptMulticast if IFF_MULTICAST isn't set.

Note: Based on the implementations I've seen it doesn't seem to be 100% clear
what a driver is supposed to do if IFF_ALLMULTI is set but IFF_MULTICAST
is not. This patch is based on the understanding that IFF_MULTICAST has
precedence.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://lore.kernel.org/r/4a57ba02-d52d-4369-9f14-3565e6c1f7dc@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
10 months agoMerge tag 'pm-6.7-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Wed, 8 Nov 2023 01:16:23 +0000 (17:16 -0800)]
Merge tag 'pm-6.7-rc1-2' of git://git./linux/kernel/git/rafael/linux-pm

Pull more power management updates from Rafael Wysocki:
 "These add new hardware support to a cpufreq driver and fix cpupower
  utility documentation:

   - Add support for several Qualcomm SoC versions to the Qualcomm
     cpufreq driver (Robert Marko, Varadarajan Narayanan)

   - Fix a reference to a removed document in the cpupower utility
     documentation (Vegard Nossum)"

* tag 'pm-6.7-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  cpufreq: qcom-nvmem: Introduce cpufreq for ipq95xx
  cpufreq: qcom-nvmem: Enable cpufreq for ipq53xx
  cpufreq: qcom-nvmem: add support for IPQ8074
  cpupower: fix reference to nonexistent document

10 months agoMerge tag 'drm-next-2023-11-07' of git://anongit.freedesktop.org/drm/drm
Linus Torvalds [Wed, 8 Nov 2023 01:10:02 +0000 (17:10 -0800)]
Merge tag 'drm-next-2023-11-07' of git://anongit.freedesktop.org/drm/drm

Pull more drm updates from Dave Airlie:
 "Geert pointed out I missed the renesas reworks in my main pull, so
  this pull contains the renesas next work for atomic conversion and DT
  support.

  It also contains a bunch of amdgpu and some small ssd13xx fixes.

  renesas:
   - atomic conversion
   - DT support

  ssd13xx:
   - dt binding fix for ssd132x
   - Initialize ssd130x crtc_state to NULL.

  amdgpu:
   - Fix RAS support check
   - RAS fixes
   - MES fixes
   - SMU13 fixes
   - Contiguous memory allocation fix
   - BACO fixes
   - GPU reset fixes
   - Min power limit fixes
   - GFX11 fixes
   - USB4/TB hotplug fixes
   - ARM regression fix
   - GFX9.4.3 fixes
   - KASAN/KCSAN stack size check fixes
   - SR-IOV fixes
   - SMU14 fixes
   - PSP13 fixes
   - Display blend fixes
   - Flexible array size fixes

  amdkfd:
   - GPUVM fix

  radeon:
   - Flexible array size fixes"

* tag 'drm-next-2023-11-07' of git://anongit.freedesktop.org/drm/drm: (83 commits)
  drm/amd/display: Enable fast update on blendTF change
  drm/amd/display: Fix blend LUT programming
  drm/amd/display: Program plane color setting correctly
  drm/amdgpu: Query and report boot status
  drm/amdgpu: Add psp v13 function to query boot status
  drm/amd/swsmu: remove fw version check in sw_init.
  drm/amd/swsmu: update smu v14_0_0 driver if and metrics table
  drm/amdgpu: Add C2PMSG_109/126 reg field shift/masks
  drm/amdgpu: Optimize the asic type fix code
  drm/amdgpu: fix GRBM read timeout when do mes_self_test
  drm/amdgpu: check recovery status of xgmi hive in ras_reset_error_count
  drm/amd/pm: only check sriov vf flag once when creating hwmon sysfs
  drm/amdgpu: Attach eviction fence on alloc
  drm/amdkfd: Improve amdgpu_vm_handle_moved
  drm/amd/display: Increase frame warning limit with KASAN or KCSAN in dml2
  drm/amd/display: Avoid NULL dereference of timing generator
  drm/amdkfd: Update cache info for GFX 9.4.3
  drm/amdkfd: Populate cache info for GFX 9.4.3
  drm/amdgpu: don't put MQDs in VRAM on ARM | ARM64
  drm/amdgpu/smu13: drop compute workload workaround
  ...

10 months agoMerge tag 'regmap-fix-v6.7-merge-window' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 8 Nov 2023 00:56:10 +0000 (16:56 -0800)]
Merge tag 'regmap-fix-v6.7-merge-window' of git://git./linux/kernel/git/broonie/regmap

Pull regmap fix from Mark Brown:
 "One fix here, for an interaction between noinc registers and caches.

  If a device uses noinc registers (which is rare) then we could corrupt
  registers after the noinc register in the cache"

* tag 'regmap-fix-v6.7-merge-window' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap:
  regmap: prevent noinc writes from clobbering cache

10 months agoMerge tag 'rproc-v6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/remoteproc...
Linus Torvalds [Wed, 8 Nov 2023 00:53:28 +0000 (16:53 -0800)]
Merge tag 'rproc-v6.7' of git://git./linux/kernel/git/remoteproc/linux

Pull remoteproc updates from Bjorn Andersson:
 "Support for controlling the second core in Mediatek's SCP dual-core
  setup is introduced.

  Support for audio, compute and modem DSPs on Qualcomm SM6375, and the
  audio DSP in SC7180 are introduced. The peripheral NoC clock is
  dropped from MSM8996 modem DSP, as this is handled through the
  interconnect provider.

  In the zynqmp driver the setup for TCM memory, and device address
  translation thereof, when operating in lockstep mode is corrected.

  A few bug fixes and cleanups are introduces across the ST and STM32
  remoteproc drivers"

* tag 'rproc-v6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/remoteproc/linux: (28 commits)
  remoteproc: st: Fix sometimes uninitialized ret in st_rproc_probe()
  remoteproc: st: Use device_get_match_data()
  remoteproc: zynqmp: Change tcm address translation method
  remoteproc: mediatek: Refactor single core check and fix retrocompatibility
  remoteproc: qcom: q6v5-mss: Remove PNoC clock from 8996 MSS
  dt-bindings: remoteproc: qcom,msm8996-mss-pil: Remove PNoC clock
  dt-bindings: remoteproc: qcom,adsp: Remove AGGRE2 clock
  remoteproc: qcom: pas: Add SM6375 MPSS
  remoteproc: qcom: pas: Add SM6375 ADSP & CDSP
  dt-bindings: remoteproc: qcom,sm6375-pas: Document remoteprocs
  dt-bindings: remoteproc: pru: Add Interrupt property
  remoteproc: qcom: pas: Add sc7180 adsp
  dt-bindings: remoteproc: qcom: sc7180-pas: Add ADSP compatible
  arm64: dts: mediatek: Update the node name of SCP rpmsg subnode
  remoteproc: zynqmp: fix TCM carveouts in lockstep mode
  remoteproc: mediatek: Refine ipi handler error message
  remoteproc: mediatek: Report watchdog crash to all cores
  remoteproc: mediatek: Handle MT8195 SCP core 1 watchdog timeout
  remoteproc: mediatek: Setup MT8195 SCP core 1 SRAM offset
  remoteproc: mediatek: Remove dependency of MT8195 SCP L2TCM power control on dual-core SCP
  ...

10 months agoMerge tag 'rpmsg-v6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/remoteproc...
Linus Torvalds [Wed, 8 Nov 2023 00:50:51 +0000 (16:50 -0800)]
Merge tag 'rpmsg-v6.7' of git://git./linux/kernel/git/remoteproc/linux

Pull rpmsg updates from Bjorn Andersson:
 "This replaces a number of strncpy() instances with strscpy() or
  strscpy_pad() through the rpmsg core and virtio implementation"

* tag 'rpmsg-v6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/remoteproc/linux:
  rpmsg: virtio: Replace deprecated strncpy with strscpy/_pad
  rpmsg: Replace deprecated strncpy with strscpy_pad
  rpmsg: core: Replace deprecated strncpy with strscpy

10 months agoMerge tag 'pcmcia-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brodo...
Linus Torvalds [Wed, 8 Nov 2023 00:40:42 +0000 (16:40 -0800)]
Merge tag 'pcmcia-6.7-rc1' of git://git./linux/kernel/git/brodo/linux

Pull PCMCIA updates from Dominik Brodowski:
 "Cleanups and fixes.

  Yang Yingliang has fixed a number of resource leaks and Dongliang Mu
  contributed a spelling fix for the PCMCIA core. Also included is a
  tiny clenaup to the tcic PCMCIA socket driver provided by lizhe"

* tag 'pcmcia-6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brodo/linux:
  pcmcia: tcic: remove unneeded "&" in call to setup_timer()
  pcmcia: typo fix
  pcmcia: ds: fix possible name leak in error path in pcmcia_device_add()
  pcmcia: ds: fix refcount leak in pcmcia_device_add()
  pcmcia: cs: fix possible hung task and memory leak pccardd()

10 months agoMerge tag 'gpio-pinctrl-updates-for-v6.7-rc1' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Wed, 8 Nov 2023 00:36:25 +0000 (16:36 -0800)]
Merge tag 'gpio-pinctrl-updates-for-v6.7-rc1' of git://git./linux/kernel/git/brgl/linux

Pull gpio/pinctrl updates from Bartosz Golaszewski:
 "The bulk of it is a rework of the glue layer between pinctrl and GPIO.
  We changed the signature of GPIO helpers for pinctrl to taking the
  gpio_chip/offset pair as arguments instead of using the deprecated
  global GPIO numberspace.

  The last little bit is removing the gpiochip_find() function as it now
  has no more users in-tree.

  Summary:

   - rework the GPIO-to-pinctrl glue code to stop using the deprecated
     global GPIO numberspace

   - remove now unused wrappers around pinctrl GPIO helpers from drivers

   - remove gpiochip_find() as it has no more users"

* tag 'gpio-pinctrl-updates-for-v6.7-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux: (70 commits)
  pinctrl: tegra: drop the wrapper around pinctrl_gpio_request()
  pinctrl: em: drop the wrapper around pinctrl_gpio_request()
  pinctrl: nuvoton: npcm8xx: drop wrappers around pinctrl_gpio_request/free()
  pinctrl: nuvoton: npcm7xx: drop wrappers around pinctrl_gpio_request/free()
  pinctrl: stm32: drop wrappers around pinctrl_gpio_free/input()
  pinctrl: starfive: jh7110: drop wrappers around pinctrl_gpio_request/free()
  pinctrl: starfive: jh7100: drop wrappers around pinctrl_gpio_request/free()
  pinctrl: ocelot: drop the wrapper around pinctrl_gpio_direction_input()
  pinctrl: cirrus: drop the wrapper around pinctrl_gpio_direction_input()
  pinctrl: mediatek: common: drop the wrappers around pinctrl_gpio_direction_input()
  pinctrl: mediatek: moore: drop the wrappers around pinctrl_gpio_direction_input()
  pinctrl: rk805: drop the wrapper around pinctrl_gpio_direction_input()
  pinctrl: axp209: drop the wrapper around pinctrl_gpio_direction_input()
  pinctrl: vt8500: drop the wrapper around pinctrl_gpio_direction_input()
  pinctrl: as3722: drop the wrapper around pinctrl_gpio_direction_input()
  pinctrl: ingenic: drop the wrapper around pinctrl_gpio_direction_input()
  pinctrl: st: drop the wrapper around pinctrl_gpio_direction_input()
  pinctrl: change the signature of pinctrl_ready_for_gpio_range()
  pinctrl: change the signature of gpio_to_pin()
  pinctrl: change the signature of pinctrl_match_gpio_range()
  ...

10 months agoMerge branch 'Let BPF verifier consider {task,cgroup} is trusted in bpf_iter_reg'
Martin KaFai Lau [Tue, 7 Nov 2023 21:26:03 +0000 (13:26 -0800)]
Merge branch 'Let BPF verifier consider {task,cgroup} is trusted in bpf_iter_reg'

Chuyi Zhou says:

====================
The patchset aims to let the BPF verivier consider
bpf_iter__cgroup->cgroup and bpf_iter__task->task is trusted suggested by
Alexei[1].

Please see individual patches for more details. And comments are always
welcome.

Link[1]:https://lore.kernel.org/bpf/20231022154527.229117-1-zhouchuyi@bytedance.com/T/#mb57725edc8ccdd50a1b165765c7619b4d65ed1b0

v2->v1:
 * Patch #1: Add Yonghong's ack and add description of similar case in
   log.
 * Patch #2: Add Yonghong's ack
====================

Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
10 months agoselftests/bpf: get trusted cgrp from bpf_iter__cgroup directly
Chuyi Zhou [Tue, 7 Nov 2023 13:22:04 +0000 (21:22 +0800)]
selftests/bpf: get trusted cgrp from bpf_iter__cgroup directly

Commit f49843afde (selftests/bpf: Add tests for css_task iter combining
with cgroup iter) added a test which demonstrates how css_task iter can be
combined with cgroup iter. That test used bpf_cgroup_from_id() to convert
bpf_iter__cgroup->cgroup to a trusted ptr which is pointless now, since
with the previous fix, we can get a trusted cgroup directly from
bpf_iter__cgroup.

Signed-off-by: Chuyi Zhou <zhouchuyi@bytedance.com>
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20231107132204.912120-3-zhouchuyi@bytedance.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
10 months agobpf: Let verifier consider {task,cgroup} is trusted in bpf_iter_reg
Chuyi Zhou [Tue, 7 Nov 2023 13:22:03 +0000 (21:22 +0800)]
bpf: Let verifier consider {task,cgroup} is trusted in bpf_iter_reg

BTF_TYPE_SAFE_TRUSTED(struct bpf_iter__task) in verifier.c wanted to
teach BPF verifier that bpf_iter__task -> task is a trusted ptr. But it
doesn't work well.

The reason is, bpf_iter__task -> task would go through btf_ctx_access()
which enforces the reg_type of 'task' is ctx_arg_info->reg_type, and in
task_iter.c, we actually explicitly declare that the
ctx_arg_info->reg_type is PTR_TO_BTF_ID_OR_NULL.

Actually we have a previous case like this[1] where PTR_TRUSTED is added to
the arg flag for map_iter.

This patch sets ctx_arg_info->reg_type is PTR_TO_BTF_ID_OR_NULL |
PTR_TRUSTED in task_reg_info.

Similarly, bpf_cgroup_reg_info -> cgroup is also PTR_TRUSTED since we are
under the protection of cgroup_mutex and we would check cgroup_is_dead()
in __cgroup_iter_seq_show().

This patch is to improve the user experience of the newly introduced
bpf_iter_css_task kfunc before hitting the mainline. The Fixes tag is
pointing to the commit introduced the bpf_iter_css_task kfunc.

Link[1]:https://lore.kernel.org/all/20230706133932.45883-3-aspsk@isovalent.com/

Fixes: 9c66dc94b62a ("bpf: Introduce css_task open-coded iterator kfuncs")
Signed-off-by: Chuyi Zhou <zhouchuyi@bytedance.com>
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20231107132204.912120-2-zhouchuyi@bytedance.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
10 months agonet: phylink: initialize carrier state at creation
Klaus Kudielka [Tue, 7 Nov 2023 17:44:02 +0000 (18:44 +0100)]
net: phylink: initialize carrier state at creation

Background: Turris Omnia (Armada 385); eth2 (mvneta) connected to SFP bus;
SFP module is present, but no fiber connected, so definitely no carrier.

After booting, eth2 is down, but netdev LED trigger surprisingly reports
link active. Then, after "ip link set eth2 up", the link indicator goes
away - as I would have expected it from the beginning.

It turns out, that the default carrier state after netdev creation is
"carrier ok". Some ethernet drivers explicitly call netif_carrier_off
during probing, others (like mvneta) don't - which explains the current
behaviour: only when the device is brought up, phylink_start calls
netif_carrier_off.

Fix this for all drivers using phylink, by calling netif_carrier_off in
phylink_create.

Fixes: 089381b27abe ("leds: initial support for Turris Omnia LEDs")
Cc: stable@vger.kernel.org
Suggested-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Klaus Kudielka <klaus.kudielka@gmail.com>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agoMerge branch 'vsock-fixes'
David S. Miller [Tue, 7 Nov 2023 22:27:07 +0000 (22:27 +0000)]
Merge branch 'vsock-fixes'

Filippo Storniolo says:

====================
vsock: fix server prevents clients from reconnecting

This patch series introduce fix and tests for the following vsock bug:
If the same remote peer, using the same port, tries to connect
to a server on a listening port more than once, the server will
reject the connection, causing a "connection reset by peer"
error on the remote peer. This is due to the presence of a
dangling socket from a previous connection in both the connected
and bound socket lists.
The inconsistency of the above lists only occurs when the remote
peer disconnects and the server remains active.
This bug does not occur when the server socket is closed.

More details on the first patch changelog.
The remaining patches are refactoring and test.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agotest/vsock: add dobule bind connect test
Filippo Storniolo [Fri, 3 Nov 2023 17:55:51 +0000 (18:55 +0100)]
test/vsock: add dobule bind connect test

This add bind connect test which creates a listening server socket
and tries to connect a client with a bound local port to it twice.

Co-developed-by: Luigi Leonardi <luigi.leonardi@outlook.com>
Signed-off-by: Luigi Leonardi <luigi.leonardi@outlook.com>
Signed-off-by: Filippo Storniolo <f.storniolo95@gmail.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agotest/vsock: refactor vsock_accept
Filippo Storniolo [Fri, 3 Nov 2023 17:55:50 +0000 (18:55 +0100)]
test/vsock: refactor vsock_accept

This is a preliminary patch to introduce SOCK_STREAM bind connect test.
vsock_accept() is split into vsock_listen() and vsock_accept().

Co-developed-by: Luigi Leonardi <luigi.leonardi@outlook.com>
Signed-off-by: Luigi Leonardi <luigi.leonardi@outlook.com>
Signed-off-by: Filippo Storniolo <f.storniolo95@gmail.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agotest/vsock fix: add missing check on socket creation
Filippo Storniolo [Fri, 3 Nov 2023 17:55:49 +0000 (18:55 +0100)]
test/vsock fix: add missing check on socket creation

Add check on socket() return value in vsock_listen()
and vsock_connect()

Co-developed-by: Luigi Leonardi <luigi.leonardi@outlook.com>
Signed-off-by: Luigi Leonardi <luigi.leonardi@outlook.com>
Signed-off-by: Filippo Storniolo <f.storniolo95@gmail.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agovsock/virtio: remove socket from connected/bound list on shutdown
Filippo Storniolo [Fri, 3 Nov 2023 17:55:48 +0000 (18:55 +0100)]
vsock/virtio: remove socket from connected/bound list on shutdown

If the same remote peer, using the same port, tries to connect
to a server on a listening port more than once, the server will
reject the connection, causing a "connection reset by peer"
error on the remote peer. This is due to the presence of a
dangling socket from a previous connection in both the connected
and bound socket lists.
The inconsistency of the above lists only occurs when the remote
peer disconnects and the server remains active.

This bug does not occur when the server socket is closed:
virtio_transport_release() will eventually schedule a call to
virtio_transport_do_close() and the latter will remove the socket
from the bound and connected socket lists and clear the sk_buff.

However, virtio_transport_do_close() will only perform the above
actions if it has been scheduled, and this will not happen
if the server is processing the shutdown message from a remote peer.

To fix this, introduce a call to vsock_remove_sock()
when the server is handling a client disconnect.
This is to remove the socket from the bound and connected socket
lists without clearing the sk_buff.

Fixes: 06a8fc78367d ("VSOCK: Introduce virtio_vsock_common.ko")
Reported-by: Daan De Meyer <daan.j.demeyer@gmail.com>
Tested-by: Daan De Meyer <daan.j.demeyer@gmail.com>
Co-developed-by: Luigi Leonardi <luigi.leonardi@outlook.com>
Signed-off-by: Luigi Leonardi <luigi.leonardi@outlook.com>
Signed-off-by: Filippo Storniolo <f.storniolo95@gmail.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 months agotcp: Fix -Wc23-extensions in tcp_options_write()
Nathan Chancellor [Mon, 6 Nov 2023 21:14:16 +0000 (14:14 -0700)]
tcp: Fix -Wc23-extensions in tcp_options_write()

Clang warns (or errors with CONFIG_WERROR=y) when CONFIG_TCP_AO is set:

  net/ipv4/tcp_output.c:663:2: error: label at end of compound statement is a C23 extension [-Werror,-Wc23-extensions]
    663 |         }
        |         ^
  1 error generated.

On earlier releases (such as clang-11, the current minimum supported
version for building the kernel) that do not support C23, this was a
hard error unconditionally:

  net/ipv4/tcp_output.c:663:2: error: expected statement
          }
          ^
  1 error generated.

While adding a semicolon after the label would resolve this, it is more
in line with the kernel as a whole to refactor this block into a
standalone function, which means the goto a label construct can just be
replaced with a return statement. Do so to resolve the warning.

Closes: https://github.com/ClangBuiltLinux/linux/issues/1953
Fixes: 1e03d32bea8e ("net/tcp: Add TCP-AO sign to outgoing packets")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Dmitry Safonov <dima@arista.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>