| Age | Commit message (Collapse) | Author |
|---|
|
Currently, mlx5 driver does not enforce vector index to be lower than
the maximum number of supported completion vectors when requesting a
new completion EQ. Thus, mlx5_comp_eqn_get() fails when trying to
acquire an IRQ with an improper vector index.
To prevent the case above, enforce that vector index value is
valid and lower than maximum in mlx5_comp_eqn_get() before handling the
request.
Fixes: f14c1a14e632 ("net/mlx5: Allocate completion EQs dynamically")
Signed-off-by: Maher Sanalla <msanalla@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The HWS BWC API uses one lock per queue and usually acquires one of
them, except when doing changes which require locking all queues in
order. Naturally, lockdep isn't too happy about acquiring the same lock
class multiple times, so inform it that each queue lock is a different
class to avoid false positives.
Fixes: 2ca62599aa0b ("net/mlx5: HWS, added send engine and context handling")
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
hws_send_queues_bwc_locks_destroy destroyed more queue locks than
allocated, leading to memory corruption (occasionally) and warnings such
as DEBUG_LOCKS_WARN_ON(mutex_is_locked(lock)) in __mutex_destroy because
sometimes, the 'mutex' being destroyed was random memory.
The severity of this problem is proportional to the number of queues
configured because the code overreaches beyond the end of the
bwc_send_queue_locks array by 2x its length.
Fix that by using the correct number of bwc queues.
Fixes: 2ca62599aa0b ("net/mlx5: HWS, added send engine and context handling")
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Fix error flow bug that could lead to double free of a buffer
during a failure to calculate a suitable definer layout.
Fixes: 74a778b4a63f ("net/mlx5: HWS, added definers handling")
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Reviewed-by: Itamar Gozlan <igozlan@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Removed wrong access to the num_of_rules field of the matcher.
This is a usual u32 variable, but the access was as if it was atomic.
This fixes the following CI warnings:
mlx5hws_bwc.c:708:17: warning: large atomic operation may incur significant performance penalty;
the access size (4 bytes) exceeds the max lock-free size (0 bytes) [-Watomic-alignment]
Fixes: 510f9f61a112 ("net/mlx5: HWS, added API and enabled HWS support")
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202409291101.6NdtMFVC-lkp@intel.com/
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Reviewed-by: Itamar Gozlan <igozlan@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Use netif_napi_add_config to assign persistent per-NAPI config when
initializing RX CQ NAPIs.
Presently, struct napi_config only has support for two fields used for
RX, so there is no need to support them with TX CQs, yet.
Signed-off-by: Joe Damato <jdamato@fastly.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://patch.msgid.link/20241011184527.16393-10-jdamato@fastly.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Use netif_napi_add_config to assign persistent per-NAPI config when
initializing NAPIs.
Signed-off-by: Joe Damato <jdamato@fastly.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://patch.msgid.link/20241011184527.16393-9-jdamato@fastly.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Introduce a new function, mlx5_qos_tsar_type_supported(), to handle the
validation of TSAR types within QoS scheduling contexts.
Refactor the existing code to use this new function, replacing direct
checks for TSAR type support in the NIC scheduling hierarchy.
Signed-off-by: Carolina Jubran <cjubran@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Refactor the QoS element type support check by introducing a new
function, mlx5_qos_element_type_supported(), which handles element type
validation for both NIC and E-Switch schedulers.
This change removes the redundant esw_qos_element_type_supported()
function and unifies the element type checks into a single
implementation.
Signed-off-by: Carolina Jubran <cjubran@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
E-Switch qos changes used the esw state_lock to serialize qos changes.
With the introduction of cross-esw scheduling, multiple E-Switches might
be involved in a qos operation, so prepare for that by switching locking
to use a qos domain mutex.
Add three helper functions:
- esw_qos_lock
- esw_qos_unlock
- esw_assert_qos_lock_held
Convert existing direct lock/unlock/lockdep calls to them. Also call
esw_assert_qos_lock_held in a couple more places.
mlx5_esw_qos_set_vport_rate expected to be called with the esw
state_lock already held.
Change it to instead acquire the qos lock directly.
mlx5_eswitch_get_vport_config also accessed qos properties with the esw
state lock. Introduce a new function mlx5_esw_qos_get_vport_rate to
access those with the correct lock and change get_vport_config to use
it.
Finally, mlx5_vport_disable is called from the cleanup path with the esw
state_lock held, so have it additionally acquire the qos lock to make
sure there are no races.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Groups are currently maintained as a list in their corresponding
eswitch, protected by the esw state_lock.
The upcoming cross-eswitch scheduling feature cannot work with this
approach, as it would require acquiring multiple eswitch locks (in the
correct order) in order to maintain group membership.
This commit moves the rate groups into a new 'qos domain' struct and
adds explicit qos init/cleanup steps to the eswitch init/cleanup.
Upcoming patches will expand the qos domain struct and allow it to be
shared between eswitches. For now, qos domains are private to each esw
so there's only an extra indirection.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
'list' is not very descriptive, I prefer list membership to clearly
specify which list the entry belongs to. This commit renames the list
entry into the esw groups list as 'parent_entry' to make the code more
readable. This is a no-op change.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
vport qos trace calls used vport->dev implicitly as the device to which
the command was sent (and thus the device logged in traces).
But that will no longer be the case for cross-esw scheduling, where the
commands have to be sent to the group esw device instead.
This commit corrects that.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The rate groups are about to be moved out of eswitches, so store a
reference to the eswitch they belong to so things can still work
later.
This allows dropping the esw parameter from a couple of functions and
simplifying some of the code. Use this opportunity to make sure that
vport scheduling element commands are always sent to the group eswitch,
because that will be relevant for cross-esw scheduling. For now though,
the eswitches are not different.
There is no functionality change here.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The vport has a pointer to its own eswitch in vport->dev->priv.eswitch,
so passing the same eswitch as a parameter to the various functions
manipulating vport qos is superfluous at best and prone to errors at
worst.
More importantly, with the upcoming cross-esw scheduling changes, the
eswitch that should receive the various scheduling element commands is
NOT the same as the vport's eswitch, so the current code's assumptions
will break.
To avoid confusion and bugs, this commit drops the 'esw' parameter from
all vport qos functions and uses the vport's own eswitch pointer
instead.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Reviewed-by: Carolina Jubran <cjubran@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
All vports not explicitly members of a group with QoS enabled are part
of the internal esw group0, except when the hw reports that groups
aren't supported (log_esw_max_sched_depth == 0). This creates corner
cases in the code, which has to make sure that this case is supported.
Additionally, the groups are about to be moved out of eswitches, and
group0 being NULL creates additional complications there.
This patch makes sure to always create group0, even if max sched depth
is 0. In that case, a software-only group0 is created referencing the
root TSAR. Vports can point to this group when their QoS is enabled and
they'll be attached to the root TSAR directly. This eliminates corner
cases in the code by offering the guarantee that if qos is enabled,
vport->qos.group is non-NULL.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Previously, finding group members was done by iterating over all vports
of an eswitch and comparing their group with the required one, but that
approach will break down when a group can contain vports from multiple
eswitches.
Solve that by maintaining a list of vport members.
Instead of iterating over esw vports, loop over the members list.
Use this opportunity to provide two new functions to allocate and free a
group, so that the number of state transitions is smaller. This will
also be used in a future patch.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The previous function (esw_qos_calculate_group_min_rate_divider) had two
completely different modes of execution, depending on the 'group_level'
parameter. Split it into two separate functions:
- esw_qos_calculate_min_rate_divider - computes min across groups.
- esw_qos_calculate_group_min_rate_divider - computes min in a group.
Fold the divider calculation into the corresponding normalize functions
to avoid having the caller compute the corresponding divider.
Also rename the normalize functions to better indicate what level
they're operating on.
Finally, document everything so that this topic can more easily be
understood by future maintainers.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
The current mixture of 'vport' and 'evport' can be improved.
There is no functional change.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Vports do not use TSARs (Transmit Scheduling ARbiters), which are used
for grouping multiple entities together. Use the correct name in
variables and functions for clarity.
Also move the scheduling context to a local variable in the
esw_qos_sched_elem_config function instead of an empty parameter that
needs to be provided by all callers.
There is no functional change here.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
This is used for multiple purposes, depending on the scheduling element
created. There are a few helper struct defined a long time ago, but they
are not easy to find in the file and they are about to get new members.
This commit cleans up this area a bit by:
- moving the helper structs closer to where they are relevant.
- defining a helper union to include all of them to help
discoverability.
- making use of it everywhere element_attributes is used.
- using a consistent 'attr' name.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
'struct mlxsw_afk_element_inst' are not modified in these drivers.
Constifying these structures moves some data to a read-only section, so
increases overall security.
Update a few functions and struct mlxsw_afk_block accordingly.
On a x86_64, with allmodconfig, as an example:
Before:
======
text data bss dec hex filename
4278 4032 0 8310 2076 drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_flex_keys.o
After:
=====
text data bss dec hex filename
7934 352 0 8286 205e drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_flex_keys.o
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Link: https://patch.msgid.link/8ccfc7bfb2365dcee5b03c81ebe061a927d6da2e.1727541677.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
After commit 0edb555a65d1 ("platform: Make platform_driver::remove()
return void") .remove() is (again) the right callback to implement for
platform drivers.
Convert all platform drivers below drivers/net/ethernet to use
.remove(), with the eventual goal to drop struct
platform_driver::remove_new(). As .remove() and .remove_new() have the
same prototypes, conversion is done by just changing the structure
member name in the driver initializer.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@baylibre.com>
Link: https://patch.msgid.link/18f7c585a1a8a8ac8b03a2fca7de19bd5c52ac2b.1727949050.git.u.kleine-koenig@baylibre.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
It no longer serves any purpose and is identical to mlx5_fc_create upon
which it was originally based of.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://patch.msgid.link/20241001103709.58127-7-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
num_counters is only used for deciding whether to grow the bulk query
buffer, which is done once more counters than a small initial threshold
are present. After that, maintaining num_counters serves no purpose.
This commit replaces that with an actual xarray traversal to count the
counters. This appears expensive at first sight, but is only done when
the number of counters is less than the initial threshold (8) and only
once every sampling interval. Once the number of counters goes above the
threshold, the bulk query buffer is grown to max size and the xarray
traversal is never done again.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://patch.msgid.link/20241001103709.58127-6-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
The mlx5_fc struct has a cache for values queried from hw, which is
cacheline aligned. On x86_64, this results in:
struct mlx5_fc {
u32 id; /* 0 4 */
bool aging; /* 4 1 */
/* XXX 3 bytes hole, try to pack */
struct mlx5_fc_bulk * bulk; /* 8 8 */
/* XXX 48 bytes hole, try to pack */
/* --- cacheline 1 boundary (64 bytes) --- */
struct mlx5_fc_cache cache __attribute__((__aligned__(64)));
/* 64 24 */
u64 lastpackets; /* 88 8 */
u64 lastbytes; /* 96 8 */
/* size: 128, cachelines: 2, members: 6 */
/* sum members: 53, holes: 2, sum holes: 51 */
/* padding: 24 */
/* forced aligns: 1, forced holes: 1, sum forced holes: 48 */
} __attribute__((__aligned__(64)));
(output from pahole).
...So a 48+24=72 byte waste. As far as I can determine, this serves no
purpose other than maybe making sure that the values in the cache do not
span two cachelines in the worst case scenario, but that's not a valid
enough reason to waste 72 bytes per counter, especially since this code
is not performance-critical. There could potentially be hundreds of
thousands of counters (e.g. for connection-tracking), so this quickly
adds up to multiple MB wasted.
This commit removes the alignment, resulting in:
struct mlx5_fc {
[...]
/* size: 56, cachelines: 1, members: 6 */
/* sum members: 53, holes: 1, sum holes: 3 */
/* last cacheline: 56 bytes */
};
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://patch.msgid.link/20241001103709.58127-5-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Previously, managing counters was a complicated affair involving an IDR,
a sorted double linked list, two single linked lists and a complex dance
between a non-periodic wq task and users adding/deleting counters.
Adding was done by inserting new counters into the IDR and into a single
linked list, leaving the wq to process the list and actually add the
counters into the double linked list, maintained sorted with the IDR.
Deleting involved adding the counter into another single linked list,
leaving the wq to actually unlink the counter from the other structures
and release it.
Dumping the counters is done with the bulk query API, which relies on
the counter list being sorted and unmutable during querying to
efficiently retrieve cached counter values.
Finally, the IDR data struct is deprecated.
This commit replaces all of that with an xarray.
Adding is now done directly, by using xa_lock.
Deleting is also done directly, under the xa_lock.
Querying is done from a periodic task running every sampling_interval
(default 1s) and uses the bulk query API for efficiency.
It works by iterating over the xarray:
- when a new bulk needs to be started, the bulk information is computed
under the xa_lock.
- the xa iteration state is saved and the xa_lock dropped.
- the HW is queried for bulk counter values.
- the xa_lock is reacquired.
- counter caches with ids covered by the bulk response are updated.
Querying always requests the max bulk length, for simplicity.
Counters could be added/deleted while the HW is queried. This is safe,
as the HW API simply returns unknown values for counters not in HW, but
those values won't be accessed. Only counters present in xarray before
bulk query will actually read queried cache values.
This cuts down the size of mlx5_fc by 4 pointers (88->56 bytes), which
amounts to ~3MB / 100K counters.
But more importantly, this solves the wq spinlock congestion issue seen
happening on high-rate counter insertion+deletion.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://patch.msgid.link/20241001103709.58127-4-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
The bulk query buffer starts out small (see [1]) and as soon as the
number of counters goes past the initial threshold grows to max
size (32K entries, 512KB) with a retry scheme.
This commit switches to using kvmalloc for the buffer, which has a near
zero likelihood of failing, and thus the explicit retry scheme becomes
superfluous and is taken out. On the low chance the allocation fails, it
will still be retried every sampling_interval, when the wq task runs.
[1] commit b247f32aecad ("net/mlx5: Dynamically resize flow counters
query buffer")
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://patch.msgid.link/20241001103709.58127-3-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
The mlx5_fc_stats and mlx5_fc_pool structs are only used from
fs_counters.c. As such, make them private there.
mlx5_fc_pool is not used or referenced at all outside fs_counters.
mlx5_fc_stats is referenced from mlx5_core_dev, so instead of having it
as a direct member (which requires exporting it from fs_counters), store
a pointer to it, allocate it on init and clear it on destroy.
One caveat is that a simple container_of to get from a 'work' struct to
the outermost mlx5_core_dev struct directly no longer works, so an extra
pointer had to be added to mlx5_fc_stats back to the parent dev.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://patch.msgid.link/20241001103709.58127-2-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5 fixes 2024-09-25
* tag 'mlx5-fixes-2024-09-25' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
net/mlx5e: Fix crash caused by calling __xfrm_state_delete() twice
net/mlx5e: SHAMPO, Fix overflow of hd_per_wq
net/mlx5: HWS, changed E2BIG error to a negative return code
net/mlx5: HWS, fixed double-free in error flow of creating SQ
net/mlx5: Fix wrong reserved field in hca_cap_2 in mlx5_ifc
net/mlx5e: Fix NULL deref in mlx5e_tir_builder_alloc()
net/mlx5: Added cond_resched() to crdump collection
net/mlx5: Fix error path in multi-packet WQE transmit
====================
Link: https://patch.msgid.link/20240925202013.45374-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Pull virtio updates from Michael Tsirkin:
"Several new features here:
- virtio-balloon supports new stats
- vdpa supports setting mac address
- vdpa/mlx5 suspend/resume as well as MKEY ops are now faster
- virtio_fs supports new sysfs entries for queue info
- virtio/vsock performance has been improved
And fixes, cleanups all over the place"
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost: (34 commits)
vsock/virtio: avoid queuing packets when intermediate queue is empty
vsock/virtio: refactor virtio_transport_send_pkt_work
fw_cfg: Constify struct kobj_type
vdpa/mlx5: Postpone MR deletion
vdpa/mlx5: Introduce init/destroy for MR resources
vdpa/mlx5: Rename mr_mtx -> lock
vdpa/mlx5: Extract mr members in own resource struct
vdpa/mlx5: Rename function
vdpa/mlx5: Delete direct MKEYs in parallel
vdpa/mlx5: Create direct MKEYs in parallel
MAINTAINERS: add virtio-vsock driver in the VIRTIO CORE section
virtio_fs: add sysfs entries for queue information
virtio_fs: introduce virtio_fs_put_locked helper
vdpa: Remove unused declarations
vdpa/mlx5: Parallelize VQ suspend/resume for CVQ MQ command
vdpa/mlx5: Small improvement for change_num_qps()
vdpa/mlx5: Keep notifiers during suspend but ignore
vdpa/mlx5: Parallelize device resume
vdpa/mlx5: Parallelize device suspend
vdpa/mlx5: Use async API for vq modify commands
...
|
|
The km.state is not checked in driver's delayed work. When
xfrm_state_check_expire() is called, the state can be reset to
XFRM_STATE_EXPIRED, even if it is XFRM_STATE_DEAD already. This
happens when xfrm state is deleted, but not freed yet. As
__xfrm_state_delete() is called again in xfrm timer, the following
crash occurs.
To fix this issue, skip xfrm_state_check_expire() if km.state is not
XFRM_STATE_VALID.
Oops: general protection fault, probably for non-canonical address 0xdead000000000108: 0000 [#1] SMP
CPU: 5 UID: 0 PID: 7448 Comm: kworker/u102:2 Not tainted 6.11.0-rc2+ #1
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014
Workqueue: mlx5e_ipsec: eth%d mlx5e_ipsec_handle_sw_limits [mlx5_core]
RIP: 0010:__xfrm_state_delete+0x3d/0x1b0
Code: 0f 84 8b 01 00 00 48 89 fd c6 87 c8 00 00 00 05 48 8d bb 40 10 00 00 e8 11 04 1a 00 48 8b 95 b8 00 00 00 48 8b 85 c0 00 00 00 <48> 89 42 08 48 89 10 48 8b 55 10 48 b8 00 01 00 00 00 00 ad de 48
RSP: 0018:ffff88885f945ec8 EFLAGS: 00010246
RAX: dead000000000122 RBX: ffffffff82afa940 RCX: 0000000000000036
RDX: dead000000000100 RSI: 0000000000000000 RDI: ffffffff82afb980
RBP: ffff888109a20340 R08: ffff88885f945ea0 R09: 0000000000000000
R10: 0000000000000000 R11: ffff88885f945ff8 R12: 0000000000000246
R13: ffff888109a20340 R14: ffff88885f95f420 R15: ffff88885f95f400
FS: 0000000000000000(0000) GS:ffff88885f940000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f2163102430 CR3: 00000001128d6001 CR4: 0000000000370eb0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<IRQ>
? die_addr+0x33/0x90
? exc_general_protection+0x1a2/0x390
? asm_exc_general_protection+0x22/0x30
? __xfrm_state_delete+0x3d/0x1b0
? __xfrm_state_delete+0x2f/0x1b0
xfrm_timer_handler+0x174/0x350
? __xfrm_state_delete+0x1b0/0x1b0
__hrtimer_run_queues+0x121/0x270
hrtimer_run_softirq+0x88/0xd0
handle_softirqs+0xcc/0x270
do_softirq+0x3c/0x50
</IRQ>
<TASK>
__local_bh_enable_ip+0x47/0x50
mlx5e_ipsec_handle_sw_limits+0x7d/0x90 [mlx5_core]
process_one_work+0x137/0x2d0
worker_thread+0x28d/0x3a0
? rescuer_thread+0x480/0x480
kthread+0xb8/0xe0
? kthread_park+0x80/0x80
ret_from_fork+0x2d/0x50
? kthread_park+0x80/0x80
ret_from_fork_asm+0x11/0x20
</TASK>
Fixes: b2f7b01d36a9 ("net/mlx5e: Simulate missing IPsec TX limits hardware functionality")
Signed-off-by: Jianbo Liu <jianbol@nvidia.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
|
|
When having larger RQ sizes and small MTUs sizes, the hd_per_wq variable
can overflow. Like in the following case:
$> ethtool --set-ring eth1 rx 8192
$> ip link set dev eth1 mtu 144
$> ethtool --features eth1 rx-gro-hw on
... yields in dmesg:
mlx5_core 0000:08:00.1: mlx5_cmd_out_err:808:(pid 194797): CREATE_MKEY(0x200) op_mod(0x0) failed, status bad parameter(0x3), syndrome (0x3bf6f), err(-22)
because hd_per_wq is 64K which overflows to 0 and makes the command
fail.
This patch increases the variable size to 32 bit.
Fixes: 99be56171fa9 ("net/mlx5e: SHAMPO, Re-enable HW-GRO")
Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
|
|
Fixed all the 'E2BIG' returns in error flow of functions to
the negative '-E2BIG' as we are using negative error codes
everywhere in HWS code.
This also fixes the following smatch warnings:
"warn: was negative '-E2BIG' intended?"
Fixes: 74a778b4a63f ("net/mlx5: HWS, added definers handling")
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Closes: https://lore.kernel.org/all/f8c77688-7d83-4937-baba-ac844dfe2e0b@stanley.mountain/
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
|
|
When SQ creation fails, call the appropriate mlx5_core destroy function.
This fixes the following smatch warnings:
divers/net/ethernet/mellanox/mlx5/core/steering/hws/mlx5hws_send.c:739
hws_send_ring_open_sq() warn: 'sq->dep_wqe' double freed
hws_send_ring_open_sq() warn: 'sq->wq_ctrl.buf.frags' double freed
hws_send_ring_open_sq() warn: 'sq->wr_priv' double freed
Fixes: 2ca62599aa0b ("net/mlx5: HWS, added send engine and context handling")
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Closes: https://lore.kernel.org/all/e4ebc227-4b25-49bf-9e4c-14b7ea5c6a07@stanley.mountain/
Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
|
|
In mlx5e_tir_builder_alloc() kvzalloc() may return NULL
which is dereferenced on the next line in a reference
to the modify field.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: a6696735d694 ("net/mlx5e: Convert TIR to a dedicated object")
Signed-off-by: Elena Salomatkina <esalomatkina@ispras.ru>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
|
|
Collecting crdump involves reading vsc registers from pci config space
of mlx device, which can take long time to complete. This might result
in starving other threads waiting to run on the cpu.
Numbers I got from testing ConnectX-5 Ex MCX516A-CDAT in the lab:
- mlx5_vsc_gw_read_block_fast() was called with length = 1310716.
- mlx5_vsc_gw_read_fast() reads 4 bytes at a time. It was not used to
read the entire 1310716 bytes. It was called 53813 times because
there are jumps in read_addr.
- On average mlx5_vsc_gw_read_fast() took 35284.4ns.
- In total mlx5_vsc_wait_on_flag() called vsc_read() 54707 times.
The average time for each call was 17548.3ns. In some instances
vsc_read() was called more than one time when the flag was not set.
As expected the thread released the cpu after 16 iterations in
mlx5_vsc_wait_on_flag().
- Total time to read crdump was 35284.4ns * 53813 ~= 1.898s.
It was seen in the field that crdump can take more than 5 seconds to
complete. During that time mlx5_vsc_wait_on_flag() did not release the
cpu because it did not complete 16 iterations. It is believed that pci
config reads were slow. Adding cond_resched() every 128 register read
improves the situation. In the common case the, crdump takes ~1.8989s,
the thread yields the cpu every ~4.51ms. If crdump takes ~5s, the thread
yields the cpu every ~18.0ms.
Fixes: 8b9d8baae1de ("net/mlx5: Add Crdump support")
Reviewed-by: Yuanyuan Zhong <yzhong@purestorage.com>
Signed-off-by: Mohamed Khalfella <mkhalfella@purestorage.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
|
|
Remove the erroneous unmap in case no DMA mapping was established
The multi-packet WQE transmit code attempts to obtain a DMA mapping for
the skb. This could fail, e.g. under memory pressure, when the IOMMU
driver just can't allocate more memory for page tables. While the code
tries to handle this in the path below the err_unmap label it erroneously
unmaps one entry from the sq's FIFO list of active mappings. Since the
current map attempt failed this unmap is removing some random DMA mapping
that might still be required. If the PCI function now presents that IOVA,
the IOMMU may assumes a rogue DMA access and e.g. on s390 puts the PCI
function in error state.
The erroneous behavior was seen in a stress-test environment that created
memory pressure.
Fixes: 5af75c747e2a ("net/mlx5e: Enhanced TX MPWQE for SKBs")
Signed-off-by: Gerd Bayer <gbayer@linux.ibm.com>
Reviewed-by: Zhu Yanjun <yanjun.zhu@linux.dev>
Acked-by: Maxim Mikityanskiy <maxtram95@gmail.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
|
|
Currently, commands that qualify as throttled can't be used via the
async API. That's due to the fact that the throttle semaphore can sleep
but the async API can't.
This patch allows throttling in the async API by using the tentative
variant of the semaphore and upon failure (semaphore at 0) returns EBUSY
to signal to the caller that they need to wait for the completion of
previously issued commands.
Furthermore, make sure that the semaphore is released in the callback.
Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
Cc: Leon Romanovsky <leonro@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Message-Id: <20240816090159.1967650-2-dtatulea@nvidia.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Lei Yang <leiyang@redhat.com>
|
|
Pull rdma updates from Jason Gunthorpe:
"Usual collection of small improvements and fixes, nothing especially
stands out to me here.
The new multipath PCI feature is a sign of things to come, I think we
will see more of this in the next 10 years. Broadcom and HNS continue
to update their drivers for their new HW generations.
Summary:
- Bug fixes and minor improvments in cxgb4, siw, mlx5, rxe, efa, rts,
hfi, erdma, hns, irdma
- Code cleanups/typos/etc. Tidy alloc_ordered_workqueue() calls
- Multipath PCI for mlx5
- Variable size work queue, SRQ changes, and relaxed ordering for new
bnxt HW
- New ODP fault resolution FW protocol in mlx5
- New 'rdma monitor' netlink mechanism"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (99 commits)
RDMA/bnxt_re: Remove the unused variable en_dev
RDMA/nldev: Add missing break in rdma_nl_notify_err_msg()
RDMA/irdma: fix error message in irdma_modify_qp_roce()
RDMA/cxgb4: Added NULL check for lookup_atid
RDMA/hns: Fix ah error counter in sw stat not increasing
RDMA/bnxt_re: Recover the device when FW error is detected
RDMA/bnxt_re: Group all operations under add_device and remove_device
RDMA/bnxt_re: Use the aux device for L2 ULP callbacks
RDMA/bnxt_re: Change aux driver data to en_info to hold more information
RDMA/nldev: Expose whether RDMA monitoring is supported
RDMA/nldev: Add support for RDMA monitoring
RDMA/mlx5: Use IB set_netdev and get_netdev functions
RDMA/device: Remove optimization in ib_device_get_netdev()
RDMA/mlx5: Initialize phys_port_cnt earlier in RDMA device creation
RDMA/mlx5: Obtain upper net device only when needed
RDMA/mlx5: Check RoCE LAG status before getting netdev
RDMA/mlx5: Consider the query_vuid cap for data_direct
net/mlx5: Handle memory scheme ODP capabilities
RDMA/mlx5: Add implicit MR handling to ODP memory scheme
RDMA/mlx5: Add handling for memory scheme page fault events
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull thermal control updates from Rafael Wysocki:
"These mostly continue to rework the thermal core and the thermal zone
driver interface to make the code more straightforward and reduce
bloat
The most significant piece of this work is a change of the code
related to binding cooling devices to thermal zones which, among other
things, replaces two previously existing thermal zone operations with
one allowing driver implementations to be much simpler
There is also a new thermal core testing module allowing mock thermal
zones to be created and controlled via debugfs in order to exercise
the thermal core functionality. It is expected to be used for
implementing thermal core self tests in the future
Apart from the above, there are assorted thermal driver updates
Specifics:
- Update some thermal drivers to eliminate thermal_zone_get_trip()
calls from them and get rid of that function (Rafael Wysocki)
- Update the thermal sysfs code to store trip point attributes in
trip descriptors and get to trip points via attribute pointers
(Rafael Wysocki)
- Move the computation of the low and high boundaries for
thermal_zone_set_trips() to __thermal_zone_device_update() (Daniel
Lezcano)
- Introduce a debugfs-based facility for thermal core testing (Rafael
Wysocki)
- Replace the thermal zone .bind() and .unbind() callbacks for
binding cooling devices to thermal zones with one .should_bind()
callback used for deciding whether or not a given cooling devices
should be bound to a given trip point in a given thermal zone
(Rafael Wysocki)
- Eliminate code that has no more users after the other changes, drop
some redundant checks from the thermal core and clean it up (Rafael
Wysocki)
- Fix rounding of delay jiffies in the thermal core (Rafael Wysocki)
- Refuse to accept trip point temperature or hysteresis that would
lead to an invalid threshold value when setting them via sysfs
(Rafael Wysocki)
- Adjust states of all uninitialized instances in the .manage()
callback of the Bang-bang thermal governor (Rafael Wysocki)
- Drop a couple of redundant checks along with the code depending on
them from the thermal core (Rafael Wysocki)
- Rearrange the thermal core to avoid redundant checks and simplify
control flow in a couple of code paths (Rafael Wysocki)
- Add power domain DT bindings for new Amlogic SoCs (Georges Stark)
- Switch from CONFIG_PM_SLEEP guards to pm_sleep_ptr() in the ST
driver and add a Kconfig dependency on THERMAL_OF subsystem for the
STi driver (Raphael Gallais-Pou)
- Simplify the error code path in the probe functions in the brcmstb
driver with the helo of dev_err_probe() (Yan Zhen)
- Make imx_sc_thermal use dev_err_probe() (Alexander Stein)
- Remove trailing space after \n newline in the Renesas driver (Colin
Ian King)
- Add DT binding compatible string for the SA8255p to the tsens
thermal driver (Nikunj Kela)
- Use the devm_clk_get_enabled() helpers to simplify the init routine
in the sprd thermal driver (Huan Yang)
- Remove __maybe_unused notations for the functions by using the new
RUNTIME_PM_OPS() and SYSTEM_SLEEP_PM_OPS() macros on the IMx and
Qoriq drivers (Fabio Estevam)
- Remove unused declarations from the ti-soc-thermal driver's header
file as the functions in question were removed previously (Zhang
Zekun)"
* tag 'thermal-6.12-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (48 commits)
thermal: core: Drop thermal_zone_device_is_enabled()
thermal: core: Check passive delay in monitor_thermal_zone()
thermal: core: Drop dead code from monitor_thermal_zone()
thermal: core: Drop redundant lockdep_assert_held()
thermal: gov_bang_bang: Adjust states of all uninitialized instances
thermal: sysfs: Add sanity checks for trip temperature and hysteresis
thermal/drivers/imx_sc_thermal: Use dev_err_probe
thermal/drivers/ti-soc-thermal: Remove unused declarations
thermal/drivers/imx: Remove __maybe_unused notations
thermal/drivers/qoriq: Remove __maybe_unused notations
thermal/drivers/sprd: Use devm_clk_get_enabled() helpers
dt-bindings: thermal: tsens: document support on SA8255p
thermal/drivers/renesas: Remove trailing space after \n newline
thermal/drivers/brcmstb_thermal: Simplify with dev_err_probe()
thermal/drivers/sti: Depend on THERMAL_OF subsystem
thermal/drivers/st: Switch from CONFIG_PM_SLEEP guards to pm_sleep_ptr()
dt-bindings: thermal: amlogic,thermal: add optional power-domains
thermal: core: Drop tz field from struct thermal_instance
thermal: core: Drop redundant checks from thermal_bind_cdev_to_trip()
thermal: core: Rename cdev-to-thermal-zone bind/unbind functions
...
|
|
There is a copy and paste bug so this code checks "sq->dep_wqe" where
"sq->wr_priv" was intended. It could result in a NULL pointer
dereference.
Fixes: 2ca62599aa0b ("net/mlx5: HWS, added send engine and context handling")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/da822315-02b7-4f5b-9c86-0d5176c5069d@stanley.mountain
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
The IB layer provides a common interface to store and get net
devices associated to an IB device port (ib_device_set_netdev()
and ib_device_get_netdev()).
Previously, mlx5_ib stored and managed the associated net devices
internally.
Replace internal net device management in mlx5_ib with
ib_device_set_netdev() when attaching/detaching a net device and
ib_device_get_netdev() when retrieving the net device.
Export ib_device_get_netdev().
For mlx5 representors/PFs/VFs and lag creation we replace the netdev
assignments with the IB set/get netdev functions.
In active-backup mode lag the active slave net device is stored in the
lag itself. To assure the net device stored in a lag bond IB device is
the active slave we implement the following:
- mlx5_core: when modifying the slave of a bond we send the internal driver event
MLX5_DRIVER_EVENT_ACTIVE_BACKUP_LAG_CHANGE_LOWERSTATE.
- mlx5_ib: when catching the event call ib_device_set_netdev()
This patch also ensures the correct IB events are sent in switchdev lag.
While at it, when in multiport eswitch mode, only a single IB device is
created for all ports. The said IB device will receive all netdev events
of its VFs once loaded, thus to avoid overwriting the mapping of PF IB
device to PF netdev, ignore NETDEV_REGISTER events if the ib device has
already been mapped to a netdev.
Signed-off-by: Chiara Meiohas <cmeiohas@nvidia.com>
Signed-off-by: Michael Guralnik <michaelgur@nvidia.com>
Link: https://patch.msgid.link/20240909173025.30422-6-michaelgur@nvidia.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
|
|
When running over new FW that supports the new memory scheme ODP, set
the cap in the FW to signal the FW we are working in the new scheme.
In the memory scheme ODP the per_transport_service capabilities are RO
for the driver so we skip their setting.
Signed-off-by: Michael Guralnik <michaelgur@nvidia.com>
Link: https://patch.msgid.link/20240909100504.29797-9-michaelgur@nvidia.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>
|
|
mlx5e_free_rq previously cleaned resources in an order that was not the
reverse of the resource allocation order in mlx5e_alloc_rq.
Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20240911201757.1505453-16-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
When SHAMPO can't identify the protocol/header of a packet, it will
yield a packet that is not split - all the packet is in the data part.
Count this value in packets and bytes.
Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20240911201757.1505453-15-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Add a new command status MLX5_CMD_STAT_NOT_READY to handle cases
where the firmware is not ready.
Signed-off-by: Shay Drory <shayd@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Link: https://patch.msgid.link/20240911201757.1505453-14-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
SFs didn't allow to configure IRQ affinity for its vectors. Allow users
to configure the affinity of the SFs irqs.
Signed-off-by: Shay Drory <shayd@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Link: https://patch.msgid.link/20240911201757.1505453-13-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Sync reset request is nacked by the driver when PCIe bridge connected to
mlx5 device has HotPlug interrupt enabled. However, when using reset
method of hot reset this check can be skipped as Hotplug is supported on
this reset method.
Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20240911201757.1505453-12-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
On device that supports sync reset for firmware activate using hot
reset, the driver queries the required reset method while handling the
sync reset request. If the required reset method is hot reset, the
driver will use pci_reset_bus() to reset the PCI link instead of the
link toggle.
Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20240911201757.1505453-11-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
