In order to use queue_stop/queue_start to support the new Steering
Tags, we need to free the TX ring and TX completion ring if it is a
combined channel with TX/RX sharing the same NAPI. Otherwise
TX completions will not have the updated Steering Tag. If TPH is
not enabled, we just stop the TX ring without freeing the TX/TX cmpl
rings. With that we can now add napi_disable() and napi_enable()
during queue_stop()/ queue_start(). This will guarantee that NAPI
will stop processing the completion entries in case there are
additional pending entries in the completion rings after queue_stop().
There could be some NQEs sitting unprocessed while NAPI is disabled
thereby leaving the NQ unarmed. Explicitly re-arm the NQ after
napi_enable() in queue start so that NAPI will resume properly.
Error handling in bnxt_queue_start() requires a reset. If a TX
ring cannot be allocated or initialized properly, it will cause
TX timeout. The reset will also free any partially allocated
rings. We don't expect to hit this error path because re-allocating
previously reserved and allocated rings with the same parameters
should never fail.
Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20250213011240.1640031-11-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In order to program the correct Steering Tag during an IRQ affinity
change, we need to free/re-allocate the RX completion ring during
queue_restart. If TPH is enabled, call FW to free the Rx completion
ring and clear the ring entries in queue_stop(). Re-allocate it in
queue_start() if TPH is enabled. Note that TPH mode is not enabled
in this patch and will be enabled later in the patch series.
While modifying bnxt_queue_start(), remove the unnecessary zeroing of
rxr->rx_next_cons. It gets overwritten by the clone in
bnxt_queue_start(). Remove the rx_reset counter increment since
restart is not reset. Add comment to clarify that the ring
allocations in queue_start should never fail.
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Link: https://patch.msgid.link/20250213011240.1640031-9-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
GCC can see that the value range for "order" is capped, but this leads
it to consider that it might be negative, leading to a false positive
warning (with GCC 15 with -Warray-bounds -fdiagnostics-details):
../drivers/net/ethernet/mellanox/mlx4/alloc.c:691:47: error: array subscript -1 is below array bounds of 'long unsigned int *[2]' [-Werror=array-bounds=]
691 | i = find_first_bit(pgdir->bits[o], MLX4_DB_PER_PAGE >> o);
| ~~~~~~~~~~~^~~
'mlx4_alloc_db_from_pgdir': events 1-2
691 | i = find_first_bit(pgdir->bits[o], MLX4_DB_PER_PAGE >> o); | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| | | | | (2) out of array bounds here
| (1) when the condition is evaluated to true In file included from ../drivers/net/ethernet/mellanox/mlx4/mlx4.h:53,
from ../drivers/net/ethernet/mellanox/mlx4/alloc.c:42:
../include/linux/mlx4/device.h:664:33: note: while referencing 'bits'
664 | unsigned long *bits[2];
| ^~~~
Switch the argument to unsigned int, which removes the compiler needing
to consider negative values.
Signed-off-by: Kees Cook <kees@kernel.org>
Link: https://patch.msgid.link/20250210174504.work.075-kees@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently disabled EEE modes are shown as supported in ethtool.
Change this by filtering them out when populating data->supported
in genphy_c45_ethtool_get_eee.
Disabled EEE modes are silently filtered out by genphy_c45_write_eee_adv.
This is planned to be removed, therefore ensure in
genphy_c45_ethtool_set_eee that disabled EEE modes are removed from the
user space provided EEE advertisement. For now keep the current behavior
to do this silently.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://patch.msgid.link/5187c86d-9a5a-482c-974f-cc103ce9738c@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
If pci_alloc_irq_vectors() can't allocate the minimum number of vectors
then it returns -ENOSPC so there is no need to check for that in the
caller. In fact, because pf->msix.min is an unsigned int, it means that
any negative error codes are type promoted to high positive values and
treated as success. So here, the "return -ENOMEM;" is unreachable code.
Check for negatives instead.
Now that we're only dealing with error codes, it's easier to propagate
the error code from pci_alloc_irq_vectors() instead of hardcoding
-ENOMEM.
Fixes: 79d97b8cf9 ("ice: remove splitting MSI-X between features")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Link: https://patch.msgid.link/b16e4f01-4c85-46e2-b602-fce529293559@stanley.mountain
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
As all PCS are using the neg_mode parameter rather than the legacy
an_mode, remove the ability to use the legacy an_mode. We remove the
tests in the phylink code, unconditionally passing the PCS neg_mode
parameter to PCS methods, and remove setting the flag from drivers.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://patch.msgid.link/E1tidPn-0040hd-2R@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Heiner Kallweit says:
====================
net: phy: realtek: improve MMD register access for internal PHY's
The integrated PHYs on chip versions from RTL8168g allow to address
MDIO_MMD_VEND2 registers. All c22 standard registers are mapped to
MDIO_MMD_VEND2 registers. So far the paging mechanism is used to
address PHY registers. Add support for c45 ops to address MDIO_MMD_VEND2
registers directly, w/o the paging.
====================
Link: https://patch.msgid.link/c6a969ef-fd7f-48d6-8c48-4bc548831a8d@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The integrated PHYs on chip versions from RTL8168g allow to address
MDIO_MMD_VEND2 registers. All c22 standard registers are mapped to
MDIO_MMD_VEND2 registers. So far the paging mechanism is used to
address PHY registers. Add support for c45 ops to address MDIO_MMD_VEND2
registers directly, w/o the paging.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/d6f97eaa-0f13-468f-89cb-75a41087bc4a@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Russell King says:
====================
net: phylink,xpcs,stmmac: support PCS EEE configuration
This series adds support for phylink managed EEE at the PCS level,
allowing xpcs_config_eee() to be removed. Sadly, we still end up with
a XPCS specific function to configure the clock multiplier.
====================
Link: https://patch.msgid.link/Z6naiPpxfxGr1Ic6@shell.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
There is now no need to pass the mult_fact into xpcs_config_eee(), so
let's remove that argument and use xpcs->eee_mult_fact directly. While
changing the function signature, as we pass true/false for enable, use
"bool" instead of "int" for this.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://patch.msgid.link/E1thRQY-003w7U-IG@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add a function to separate out the EEE clock multiplying factor. This
will be called by the stmmac driver to configure this value.
It would have been better had the driver used the CLK API to retrieve
this clock, get its rate and calculate the appropriate multiplier, but
that door has closed.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://patch.msgid.link/E1thRQ8-003w70-VT@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
There are hooks in the stmmac driver into XPCS to control the EEE
settings when LPI is configured at the MAC. This bypasses the layering.
To allow this to be removed from the stmmac driver, add two new
methods for PCS to inform them when the LPI/EEE enablement state
changes at the MAC.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://patch.msgid.link/E1thRQ3-003w6u-RH@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Eric Dumazet says:
====================
inet: better inet_sock_set_state() for passive flows
Small series to make inet_sock_set_state() more interesting for
LISTEN -> TCP_SYN_RECV changes : The 4-tuple parts are now correct.
First patch is a cleanup.
====================
Link: https://patch.msgid.link/20250212131328.1514243-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Current inet_sock_set_state trace from inet_csk_clone_lock() is missing
many details :
... sock:inet_sock_set_state: family=AF_INET6 protocol=IPPROTO_TCP \
sport=4901 dport=0 \
saddr=127.0.0.6 daddr=0.0.0.0 \
saddrv6=:: daddrv6=:: \
oldstate=TCP_LISTEN newstate=TCP_SYN_RECV
Only the sport gives the listener port, no other parts of the n-tuple are correct.
In this patch, I initialize relevant fields of the new socket before
calling inet_sk_set_state(newsk, TCP_SYN_RECV).
We now have a trace including all the source/destination bits.
... sock:inet_sock_set_state: family=AF_INET6 protocol=IPPROTO_TCP \
sport=4901 dport=47648 \
saddr=127.0.0.6 daddr=127.0.0.6 \
saddrv6=2002:a05:6830:1f85:: daddrv6=2001:4860:f803:65::3 \
oldstate=TCP_LISTEN newstate=TCP_SYN_RECV
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://patch.msgid.link/20250212131328.1514243-3-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
For GSO packets, skb_cow_head() will reallocate the skb for TSO header
cloned skbs in airoha_dev_xmit(). For this reason, sinfo pointer can be
no more valid. Fix the issue relying on skb_shinfo() macro directly in
airoha_dev_xmit().
The problem exists since
commit 23020f0493 ("net: airoha: Introduce ethernet support for EN7581 SoC")
but it is not a user visible, since we can't currently enable TSO
for DSA user ports since we are missing to initialize net_device
vlan_features field.
Reviewed-by: Mateusz Polchlopek <mateusz.polchlopek@intel.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Link: https://patch.msgid.link/20250213-airoha-en7581-flowtable-offload-v4-1-b69ca16d74db@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pull networking fixes from Jakub Kicinski:
"Including fixes from netfilter, wireless and bluetooth.
Kalle Valo steps down after serving as the WiFi driver maintainer for
over a decade.
Current release - fix to a fix:
- vsock: orphan socket after transport release, avoid null-deref
- Bluetooth: L2CAP: fix corrupted list in hci_chan_del
Current release - regressions:
- eth:
- stmmac: correct Rx buffer layout when SPH is enabled
- iavf: fix a locking bug in an error path
- rxrpc: fix alteration of headers whilst zerocopy pending
- s390/qeth: move netif_napi_add_tx() and napi_enable() from under BH
- Revert "netfilter: flowtable: teardown flow if cached mtu is stale"
Current release - new code bugs:
- rxrpc: fix ipv6 path MTU discovery, only ipv4 worked
- pse-pd: fix deadlock in current limit functions
Previous releases - regressions:
- rtnetlink: fix netns refleak with rtnl_setlink()
- wifi: brcmfmac: use random seed flag for BCM4355 and BCM4364
firmware
Previous releases - always broken:
- add missing RCU protection of struct net throughout the stack
- can: rockchip: bail out if skb cannot be allocated
- eth: ti: am65-cpsw: base XDP support fixes
Misc:
- ethtool: tsconfig: update the format of hwtstamp flags, changes the
uAPI but this uAPI was not in any release yet"
* tag 'net-6.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (72 commits)
net: pse-pd: Fix deadlock in current limit functions
rxrpc: Fix ipv6 path MTU discovery
Reapply "net: skb: introduce and use a single page frag cache"
s390/qeth: move netif_napi_add_tx() and napi_enable() from under BH
mlxsw: Add return value check for mlxsw_sp_port_get_stats_raw()
ipv6: mcast: add RCU protection to mld_newpack()
team: better TEAM_OPTION_TYPE_STRING validation
Bluetooth: L2CAP: Fix corrupted list in hci_chan_del
Bluetooth: btintel_pcie: Fix a potential race condition
Bluetooth: L2CAP: Fix slab-use-after-free Read in l2cap_send_cmd
net: ethernet: ti: am65_cpsw: fix tx_cleanup for XDP case
net: ethernet: ti: am65-cpsw: fix RX & TX statistics for XDP_TX case
net: ethernet: ti: am65-cpsw: fix memleak in certain XDP cases
vsock/test: Add test for SO_LINGER null ptr deref
vsock: Orphan socket after transport release
MAINTAINERS: Add sctp headers to the general netdev entry
Revert "netfilter: flowtable: teardown flow if cached mtu is stale"
iavf: Fix a locking bug in an error path
rxrpc: Fix alteration of headers whilst zerocopy pending
net: phylink: make configuring clock-stop dependent on MAC support
...
Pull btrfs fixes from David Sterba:
- fix stale page cache after race between readahead and direct IO write
- fix hole expansion when writing at an offset beyond EOF, the range
will not be zeroed
- use proper way to calculate offsets in folio ranges
* tag 'for-6.14-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
btrfs: fix hole expansion when writing at an offset beyond EOF
btrfs: fix stale page cache after race between readahead and direct IO write
btrfs: fix two misuses of folio_shift()
Pull bcachefs fixes from Kent Overstreet:
"Just small stuff.
As a general announcement, on disk format is now frozen in my master
branch - future on disk format changes will be optional, not required.
- More fixes for going read-only: the previous fix was insufficient,
but with more work on ordering journal reclaim flushing (and a
btree node accounting fix so we don't split until we have to) the
tiering_replication test now consistently goes read-only in less
than a second.
- fix for fsck when we have reflink pointers to missing indirect
extents
- some transaction restart handling fixes from Alan; the "Pass
_orig_restart_count to trans_was_restarted" likely fixes some rare
undefined behaviour heisenbugs"
* tag 'bcachefs-2025-02-12' of git://evilpiepirate.org/bcachefs:
bcachefs: Reuse transaction
bcachefs: Pass _orig_restart_count to trans_was_restarted
bcachefs: CONFIG_BCACHEFS_INJECT_TRANSACTION_RESTARTS
bcachefs: Fix want_new_bset() so we write until the end of the btree node
bcachefs: Split out journal pins by btree level
bcachefs: Fix use after free
bcachefs: Fix marking reflink pointers to missing indirect extents
