* FAILED: patch "[PATCH] idpf: disable DIM work before freeing q_vectors" failed to apply to 6.12-stable tree
@ 2026-09-16 15:43 gregkh
2026-09-17 1:13 ` [PATCH 6.12.y 1/3] idpf: Fix kernel-doc descriptions to avoid warnings Sasha Levin
0 siblings, 1 reply; 4+ messages in thread
From: gregkh @ 2026-09-16 15:43 UTC (permalink / raw)
To: mhun512, Samuel.salin, ae878000, anthony.l.nguyen, stable; +Cc: stable
The patch below does not apply to the 6.12-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable@vger.kernel.org>.
To reproduce the conflict and resubmit, you may use the following commands:
git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.12.y
git checkout FETCH_HEAD
git cherry-pick -x 7dd4c829bac2916be98a3e34b41daaba7f42b4c4
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<stable@vger.kernel.org>' --in-reply-to '2026091649-mourner-moonlit-9772@gregkh' --subject-prefix 'PATCH 6.12.y' 'HEAD^..'
Possible dependencies:
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
From 7dd4c829bac2916be98a3e34b41daaba7f42b4c4 Mon Sep 17 00:00:00 2001
From: Myeonghun Pak <mhun512@gmail.com>
Date: Mon, 20 Jul 2026 22:58:46 +0900
Subject: [PATCH] idpf: disable DIM work before freeing q_vectors
idpf never drains the Tx/Rx DIM works before freeing the memory they
live in. tx_dim and rx_dim are embedded in struct idpf_q_vector, they
are queued from the NAPI poll via net_dim(), and idpf_vport_intr_rel()
ends with kfree(rsrc->q_vectors). Nothing in the driver cancels them.
idpf_tx_dim_work() and idpf_rx_dim_work() then run on freed memory:
idpf_vport_intr_write_itr() writes the ITR register through
q_vector->intr_reg.tx_itr / rx_itr, void __iomem pointers loaded out of
the freed q_vector. No configuration is needed to get there --
IDPF_ITR_IS_DYNAMIC() is defined as (itr_mode) and idpf_vport_alloc()
initialises both modes to IDPF_ITR_DYNAMIC.
Draining after idpf_vport_intr_napi_dis_all() is not enough on its own.
idpf_net_dim() is called from inside the
"if (napi_complete_done(napi, work_done))" branch of the poll, and
napi_complete_done() has already cleared NAPIF_STATE_SCHED by then.
napi_disable_locked() waits only while (val & (NAPIF_STATE_SCHED |
NAPIF_STATE_NPSVC)), so napi_disable() can return while the poll tail is
still queueing the work, and a plain cancel_work_sync() would be
re-armed behind the drain.
Use disable_work_sync(): schedule_work() on a work with a non-zero
disable count is dropped by clear_pending_if_disabled() before
__queue_work() is reached.
Move idpf_init_dim() to idpf_vport_intr_alloc() so the works are
initialised on every path that can reach the drain -- the three
"goto intr_deinit" sites between idpf_vport_intr_init() and
idpf_vport_intr_ena() get there without the enable side having run.
Nothing re-enables them: rsrc->q_vectors is freed on every exit from
idpf_vport_open() and on every idpf_vport_stop(), so the count dies with
the object.
It is a race, not a deterministic failure -- net_dim() only schedules
once DIM_NEVENTS events have accumulated and the profile index changes.
A KASAN ifup/ifdown loop under load is the way to see it.
Fixes: c2d548cad150 ("idpf: add TX splitq napi poll support")
Fixes: 3a8845af66ed ("idpf: add RX splitq napi poll support")
Cc: <stable@vger.kernel.org> # see patch description, needs adjustments for <= 6.9
Co-developed-by: Ijae Kim <ae878000@gmail.com>
Signed-off-by: Ijae Kim <ae878000@gmail.com>
Signed-off-by: Myeonghun Pak <mhun512@gmail.com>
Tested-by: Samuel Salin <Samuel.salin@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
diff --git a/drivers/net/ethernet/intel/idpf/idpf_txrx.c b/drivers/net/ethernet/intel/idpf/idpf_txrx.c
index 24b91be25676..9ba9c2952d78 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_txrx.c
+++ b/drivers/net/ethernet/intel/idpf/idpf_txrx.c
@@ -4145,6 +4145,26 @@ static void idpf_vport_intr_ena_irq_all(struct idpf_vport *vport,
writel(rsrc->noirq_dyn_ctl_ena, rsrc->noirq_dyn_ctl);
}
+/**
+ * idpf_vport_intr_dis_dim_all - Disable DIM work for all q_vectors
+ * @rsrc: pointer to queue and vector resources
+ *
+ * The DIM works are embedded in the q_vector array that
+ * idpf_vport_intr_rel() frees, and the poll arms them after
+ * napi_complete_done() has already cleared NAPI_STATE_SCHED. Disable
+ * rather than just cancel, so that a poll tail still running past
+ * napi_disable() cannot queue them again behind the drain.
+ */
+static void idpf_vport_intr_dis_dim_all(struct idpf_q_vec_rsrc *rsrc)
+{
+ for (u16 v_idx = 0; v_idx < rsrc->num_q_vectors; v_idx++) {
+ struct idpf_q_vector *q_vector = &rsrc->q_vectors[v_idx];
+
+ disable_work_sync(&q_vector->tx_dim.work);
+ disable_work_sync(&q_vector->rx_dim.work);
+ }
+}
+
/**
* idpf_vport_intr_deinit - Release all vector associations for the vport
* @vport: main vport structure
@@ -4155,6 +4175,7 @@ void idpf_vport_intr_deinit(struct idpf_vport *vport,
{
idpf_vport_intr_dis_irq_all(rsrc);
idpf_vport_intr_napi_dis_all(rsrc);
+ idpf_vport_intr_dis_dim_all(rsrc);
idpf_vport_intr_napi_del_all(rsrc);
idpf_vport_intr_rel_irq(vport, rsrc);
}
@@ -4235,7 +4256,6 @@ static void idpf_vport_intr_napi_ena_all(struct idpf_q_vec_rsrc *rsrc)
for (u16 q_idx = 0; q_idx < rsrc->num_q_vectors; q_idx++) {
struct idpf_q_vector *q_vector = &rsrc->q_vectors[q_idx];
- idpf_init_dim(q_vector);
napi_enable(&q_vector->napi);
}
}
@@ -4578,6 +4598,8 @@ int idpf_vport_intr_alloc(struct idpf_vport *vport,
q_coal = &user_config->q_coalesce[v_idx];
q_vector->vport = vport;
+ idpf_init_dim(q_vector);
+
q_vector->tx_itr_value = q_coal->tx_coalesce_usecs;
q_vector->tx_intr_mode = q_coal->tx_intr_mode;
q_vector->tx_itr_idx = VIRTCHNL2_ITR_IDX_1;
^ permalink raw reply related [flat|nested] 4+ messages in thread
* [PATCH 6.12.y 1/3] idpf: Fix kernel-doc descriptions to avoid warnings
2026-09-16 15:43 FAILED: patch "[PATCH] idpf: disable DIM work before freeing q_vectors" failed to apply to 6.12-stable tree gregkh
@ 2026-09-17 1:13 ` Sasha Levin
2026-09-17 1:13 ` [PATCH 6.12.y 2/3] idpf: introduce idpf_q_vec_rsrc struct and move vector resources to it Sasha Levin
2026-09-17 1:13 ` [PATCH 6.12.y 3/3] idpf: disable DIM work before freeing q_vectors Sasha Levin
0 siblings, 2 replies; 4+ messages in thread
From: Sasha Levin @ 2026-09-17 1:13 UTC (permalink / raw)
To: stable
Cc: Andy Shevchenko, Przemek Kitszel, Aleksandr Loktionov,
Paul Menzel, Krishneil Singh, Tony Nguyen, Sasha Levin
From: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
[ Upstream commit 7fe9c81aa24a2374b1a9a2160d44f4afbf8ab80d ]
In many functions the Return section is missing. Fix kernel-doc
descriptions to address that and other warnings.
Before the change:
$ scripts/kernel-doc -none -Wreturn drivers/net/ethernet/intel/idpf/idpf_txrx.c 2>&1 | wc -l
85
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Stable-dep-of: 7dd4c829bac2 ("idpf: disable DIM work before freeing q_vectors")
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
drivers/net/ethernet/intel/idpf/idpf_txrx.c | 94 +++++++++++++--------
1 file changed, 58 insertions(+), 36 deletions(-)
diff --git a/drivers/net/ethernet/intel/idpf/idpf_txrx.c b/drivers/net/ethernet/intel/idpf/idpf_txrx.c
index afb922e8704d0..370d574f1f9b7 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_txrx.c
+++ b/drivers/net/ethernet/intel/idpf/idpf_txrx.c
@@ -19,6 +19,8 @@ LIBETH_SQE_CHECK_PRIV(u32);
* Make sure we don't exceed maximum scatter gather buffers for a single
* packet.
* TSO case has been handled earlier from idpf_features_check().
+ *
+ * Return: %true if skb exceeds max descriptors per packet, %false otherwise.
*/
static bool idpf_chk_linearize(const struct sk_buff *skb,
unsigned int max_bufs,
@@ -148,7 +150,7 @@ static void idpf_tx_desc_rel_all(struct idpf_vport *vport)
* idpf_tx_buf_alloc_all - Allocate memory for all buffer resources
* @tx_q: queue for which the buffers are allocated
*
- * Returns 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
static int idpf_tx_buf_alloc_all(struct idpf_tx_queue *tx_q)
{
@@ -172,7 +174,7 @@ static int idpf_tx_buf_alloc_all(struct idpf_tx_queue *tx_q)
* @vport: vport to allocate resources for
* @tx_q: the tx ring to set up
*
- * Returns 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
static int idpf_tx_desc_alloc(const struct idpf_vport *vport,
struct idpf_tx_queue *tx_q)
@@ -264,7 +266,7 @@ static int idpf_compl_desc_alloc(const struct idpf_vport *vport,
* idpf_tx_desc_alloc_all - allocate all queues Tx resources
* @vport: virtual port private structure
*
- * Returns 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
static int idpf_tx_desc_alloc_all(struct idpf_vport *vport)
{
@@ -509,7 +511,7 @@ static void idpf_rx_buf_hw_update(struct idpf_buf_queue *bufq, u32 val)
* idpf_rx_hdr_buf_alloc_all - Allocate memory for header buffers
* @bufq: ring to use
*
- * Returns 0 on success, negative on failure.
+ * Return: 0 on success, negative on failure.
*/
static int idpf_rx_hdr_buf_alloc_all(struct idpf_buf_queue *bufq)
{
@@ -560,7 +562,7 @@ static void idpf_post_buf_refill(struct idpf_sw_queue *refillq, u16 buf_id)
* @bufq: buffer queue to post to
* @buf_id: buffer id to post
*
- * Returns false if buffer could not be allocated, true otherwise.
+ * Return: %false if buffer could not be allocated, %true otherwise.
*/
static bool idpf_rx_post_buf_desc(struct idpf_buf_queue *bufq, u16 buf_id)
{
@@ -609,7 +611,7 @@ static bool idpf_rx_post_buf_desc(struct idpf_buf_queue *bufq, u16 buf_id)
* @bufq: buffer queue to post working set to
* @working_set: number of buffers to put in working set
*
- * Returns true if @working_set bufs were posted successfully, false otherwise.
+ * Return: %true if @working_set bufs were posted successfully, %false otherwise.
*/
static bool idpf_rx_post_init_bufs(struct idpf_buf_queue *bufq,
u16 working_set)
@@ -678,7 +680,7 @@ static int idpf_rx_bufs_init_singleq(struct idpf_rx_queue *rxq)
* idpf_rx_buf_alloc_all - Allocate memory for all buffer resources
* @rxbufq: queue for which the buffers are allocated
*
- * Returns 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
static int idpf_rx_buf_alloc_all(struct idpf_buf_queue *rxbufq)
{
@@ -706,7 +708,7 @@ static int idpf_rx_buf_alloc_all(struct idpf_buf_queue *rxbufq)
* @bufq: buffer queue to create page pool for
* @type: type of Rx buffers to allocate
*
- * Returns 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
static int idpf_rx_bufs_init(struct idpf_buf_queue *bufq,
enum libeth_fqe_type type)
@@ -737,7 +739,7 @@ static int idpf_rx_bufs_init(struct idpf_buf_queue *bufq,
* idpf_rx_bufs_init_all - Initialize all RX bufs
* @vport: virtual port struct
*
- * Returns 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
int idpf_rx_bufs_init_all(struct idpf_vport *vport)
{
@@ -790,7 +792,7 @@ int idpf_rx_bufs_init_all(struct idpf_vport *vport)
* @vport: vport to allocate resources for
* @rxq: Rx queue for which the resources are setup
*
- * Returns 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
static int idpf_rx_desc_alloc(const struct idpf_vport *vport,
struct idpf_rx_queue *rxq)
@@ -849,7 +851,7 @@ static int idpf_bufq_desc_alloc(const struct idpf_vport *vport,
* idpf_rx_desc_alloc_all - allocate all RX queues resources
* @vport: virtual port structure
*
- * Returns 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
static int idpf_rx_desc_alloc_all(struct idpf_vport *vport)
{
@@ -1038,7 +1040,7 @@ void idpf_vport_queues_rel(struct idpf_vport *vport)
* dereference the queue from queue groups. This allows us to quickly pull a
* txq based on a queue index.
*
- * Returns 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
static int idpf_vport_init_fast_path_txqs(struct idpf_vport *vport)
{
@@ -1151,7 +1153,7 @@ void idpf_vport_calc_num_q_desc(struct idpf_vport *vport)
* @vport_msg: message to fill with data
* @max_q: vport max queue info
*
- * Return 0 on success, error value on failure.
+ * Return: 0 on success, error value on failure.
*/
int idpf_vport_calc_total_qs(struct idpf_adapter *adapter, u16 vport_idx,
struct virtchnl2_create_vport *vport_msg,
@@ -1274,7 +1276,7 @@ static void idpf_rxq_set_descids(const struct idpf_vport *vport,
* @vport: vport to allocate txq groups for
* @num_txq: number of txqs to allocate for each group
*
- * Returns 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
static int idpf_txq_group_alloc(struct idpf_vport *vport, u16 num_txq)
{
@@ -1365,7 +1367,7 @@ static int idpf_txq_group_alloc(struct idpf_vport *vport, u16 num_txq)
* @vport: vport to allocate rxq groups for
* @num_rxq: number of rxqs to allocate for each group
*
- * Returns 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
static int idpf_rxq_group_alloc(struct idpf_vport *vport, u16 num_rxq)
{
@@ -1495,7 +1497,7 @@ static int idpf_rxq_group_alloc(struct idpf_vport *vport, u16 num_rxq)
* idpf_vport_queue_grp_alloc_all - Allocate all queue groups/resources
* @vport: vport with qgrps to allocate
*
- * Returns 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
static int idpf_vport_queue_grp_alloc_all(struct idpf_vport *vport)
{
@@ -1524,8 +1526,9 @@ static int idpf_vport_queue_grp_alloc_all(struct idpf_vport *vport)
* idpf_vport_queues_alloc - Allocate memory for all queues
* @vport: virtual port
*
- * Allocate memory for queues associated with a vport. Returns 0 on success,
- * negative on failure.
+ * Allocate memory for queues associated with a vport.
+ *
+ * Return: 0 on success, negative on failure.
*/
int idpf_vport_queues_alloc(struct idpf_vport *vport)
{
@@ -1737,7 +1740,7 @@ static void idpf_tx_handle_rs_completion(struct idpf_tx_queue *txq,
* @budget: Used to determine if we are in netpoll
* @cleaned: returns number of packets cleaned
*
- * Returns true if there's any budget left (e.g. the clean is finished)
+ * Return: %true if there's any budget left (e.g. the clean is finished)
*/
static bool idpf_tx_clean_complq(struct idpf_compl_queue *complq, int budget,
int *cleaned)
@@ -1903,7 +1906,7 @@ void idpf_tx_splitq_build_flow_desc(union idpf_tx_flex_desc *desc,
}
/**
- * idpf_tx_splitq_has_room - check if enough Tx splitq resources are available
+ * idpf_txq_has_room - check if enough Tx splitq resources are available
* @tx_q: the queue to be checked
* @descs_needed: number of descriptors required for this packet
* @bufs_needed: number of Tx buffers required for this packet
@@ -2034,6 +2037,8 @@ unsigned int idpf_tx_res_count_required(struct idpf_tx_queue *txq,
* idpf_tx_splitq_bump_ntu - adjust NTU and generation
* @txq: the tx ring to wrap
* @ntu: ring index to bump
+ *
+ * Return: the next ring index hopping to 0 when wraps around
*/
static unsigned int idpf_tx_splitq_bump_ntu(struct idpf_tx_queue *txq, u16 ntu)
{
@@ -2302,7 +2307,7 @@ static void idpf_tx_splitq_map(struct idpf_tx_queue *tx_q,
* @skb: pointer to skb
* @off: pointer to struct that holds offload parameters
*
- * Returns error (negative) if TSO was requested but cannot be applied to the
+ * Return: error (negative) if TSO was requested but cannot be applied to the
* given skb, 0 if TSO does not apply to the given skb, or 1 otherwise.
*/
int idpf_tso(struct sk_buff *skb, struct idpf_tx_offload_params *off)
@@ -2380,6 +2385,8 @@ int idpf_tso(struct sk_buff *skb, struct idpf_tx_offload_params *off)
*
* Since the TX buffer rings mimics the descriptor ring, update the tx buffer
* ring entry to reflect that this index is a context descriptor
+ *
+ * Return: pointer to the next descriptor
*/
static struct idpf_flex_tx_ctx_desc *
idpf_tx_splitq_get_ctx_desc(struct idpf_tx_queue *txq)
@@ -2398,6 +2405,8 @@ idpf_tx_splitq_get_ctx_desc(struct idpf_tx_queue *txq)
* idpf_tx_drop_skb - free the SKB and bump tail if necessary
* @tx_q: queue to send buffer on
* @skb: pointer to skb
+ *
+ * Return: always NETDEV_TX_OK
*/
netdev_tx_t idpf_tx_drop_skb(struct idpf_tx_queue *tx_q, struct sk_buff *skb)
{
@@ -2432,7 +2441,7 @@ static bool idpf_tx_splitq_need_re(struct idpf_tx_queue *tx_q)
* @skb: send buffer
* @tx_q: queue to send buffer on
*
- * Returns NETDEV_TX_OK if sent, else an error code
+ * Return: NETDEV_TX_OK if sent, else an error code
*/
static netdev_tx_t idpf_tx_splitq_frame(struct sk_buff *skb,
struct idpf_tx_queue *tx_q)
@@ -2549,7 +2558,7 @@ static netdev_tx_t idpf_tx_splitq_frame(struct sk_buff *skb,
* @skb: send buffer
* @netdev: network interface device structure
*
- * Returns NETDEV_TX_OK if sent, else an error code
+ * Return: NETDEV_TX_OK if sent, else an error code
*/
netdev_tx_t idpf_tx_start(struct sk_buff *skb, struct net_device *netdev)
{
@@ -2698,10 +2707,10 @@ idpf_rx_splitq_extract_csum_bits(const struct virtchnl2_rx_flex_desc_adv_nic_3 *
* @rx_desc: Receive descriptor
* @decoded: Decoded Rx packet type related fields
*
- * Return 0 on success and error code on failure
- *
* Populate the skb fields with the total number of RSC segments, RSC payload
* length and packet type.
+ *
+ * Return: 0 on success and error code on failure
*/
static int idpf_rx_rsc(struct idpf_rx_queue *rxq, struct sk_buff *skb,
const struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc,
@@ -2772,6 +2781,8 @@ static int idpf_rx_rsc(struct idpf_rx_queue *rxq, struct sk_buff *skb,
* This function checks the ring, descriptor, and packet information in
* order to populate the hash, checksum, protocol, and
* other fields within the skb.
+ *
+ * Return: 0 on success and error code on failure
*/
static int
idpf_rx_process_skb_fields(struct idpf_rx_queue *rxq, struct sk_buff *skb,
@@ -2890,6 +2901,7 @@ struct sk_buff *idpf_rx_build_skb(const struct libeth_fqe *buf, u32 size)
* @stat_err_field: field from descriptor to test bits in
* @stat_err_bits: value to mask
*
+ * Return: %true if any of given @stat_err_bits are set, %false otherwise.
*/
static bool idpf_rx_splitq_test_staterr(const u8 stat_err_field,
const u8 stat_err_bits)
@@ -2901,8 +2913,8 @@ static bool idpf_rx_splitq_test_staterr(const u8 stat_err_field,
* idpf_rx_splitq_is_eop - process handling of EOP buffers
* @rx_desc: Rx descriptor for current buffer
*
- * If the buffer is an EOP buffer, this function exits returning true,
- * otherwise return false indicating that this is in fact a non-EOP buffer.
+ * Return: %true if the buffer is an EOP buffer, %false otherwise, indicating
+ * that this is in fact a non-EOP buffer.
*/
static bool idpf_rx_splitq_is_eop(struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc)
{
@@ -2921,7 +2933,7 @@ static bool idpf_rx_splitq_is_eop(struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_de
* expensive overhead for IOMMU access this provides a means of avoiding
* it by maintaining the mapping of the page to the system.
*
- * Returns amount of work completed
+ * Return: amount of work completed
*/
static int idpf_rx_splitq_clean(struct idpf_rx_queue *rxq, int budget)
{
@@ -3083,7 +3095,7 @@ static int idpf_rx_splitq_clean(struct idpf_rx_queue *rxq, int budget)
* @buf_id: buffer ID
* @buf_desc: Buffer queue descriptor
*
- * Return 0 on success and negative on failure.
+ * Return: 0 on success and negative on failure.
*/
static int idpf_rx_update_bufq_desc(struct idpf_buf_queue *bufq, u32 buf_id,
struct virtchnl2_splitq_rx_buf_desc *buf_desc)
@@ -3210,6 +3222,7 @@ static void idpf_rx_clean_refillq_all(struct idpf_buf_queue *bufq, int nid)
* @irq: interrupt number
* @data: pointer to a q_vector
*
+ * Return: always IRQ_HANDLED
*/
static irqreturn_t idpf_vport_intr_clean_queues(int __always_unused irq,
void *data)
@@ -3316,6 +3329,8 @@ static void idpf_vport_intr_dis_irq_all(struct idpf_vport *vport)
/**
* idpf_vport_intr_buildreg_itr - Enable default interrupt generation settings
* @q_vector: pointer to q_vector
+ *
+ * Return: value to be written back to HW to enable interrupt generation
*/
static u32 idpf_vport_intr_buildreg_itr(struct idpf_q_vector *q_vector)
{
@@ -3447,6 +3462,8 @@ void idpf_vport_intr_update_itr_ena_irq(struct idpf_q_vector *q_vector)
/**
* idpf_vport_intr_req_irq - get MSI-X vectors from the OS for the vport
* @vport: main vport structure
+ *
+ * Return: 0 on success, negative on failure
*/
static int idpf_vport_intr_req_irq(struct idpf_vport *vport)
{
@@ -3655,7 +3672,7 @@ static void idpf_vport_intr_napi_ena_all(struct idpf_vport *vport)
* @budget: Used to determine if we are in netpoll
* @cleaned: returns number of packets cleaned
*
- * Returns false if clean is not complete else returns true
+ * Return: %false if clean is not complete else returns %true
*/
static bool idpf_tx_splitq_clean_all(struct idpf_q_vector *q_vec,
int budget, int *cleaned)
@@ -3682,7 +3699,7 @@ static bool idpf_tx_splitq_clean_all(struct idpf_q_vector *q_vec,
* @budget: Used to determine if we are in netpoll
* @cleaned: returns number of packets cleaned
*
- * Returns false if clean is not complete else returns true
+ * Return: %false if clean is not complete else returns %true
*/
static bool idpf_rx_splitq_clean_all(struct idpf_q_vector *q_vec, int budget,
int *cleaned)
@@ -3721,6 +3738,8 @@ static bool idpf_rx_splitq_clean_all(struct idpf_q_vector *q_vec, int budget,
* idpf_vport_splitq_napi_poll - NAPI handler
* @napi: struct from which you get q_vector
* @budget: budget provided by stack
+ *
+ * Return: how many packets were cleaned
*/
static int idpf_vport_splitq_napi_poll(struct napi_struct *napi, int budget)
{
@@ -3857,7 +3876,9 @@ static void idpf_vport_intr_map_vector_to_qs(struct idpf_vport *vport)
* idpf_vport_intr_init_vec_idx - Initialize the vector indexes
* @vport: virtual port
*
- * Initialize vector indexes with values returened over mailbox
+ * Initialize vector indexes with values returned over mailbox.
+ *
+ * Return: 0 on success, negative on failure
*/
static int idpf_vport_intr_init_vec_idx(struct idpf_vport *vport)
{
@@ -3918,8 +3939,9 @@ static void idpf_vport_intr_napi_add_all(struct idpf_vport *vport)
* idpf_vport_intr_alloc - Allocate memory for interrupt vectors
* @vport: virtual port
*
- * We allocate one q_vector per queue interrupt. If allocation fails we
- * return -ENOMEM.
+ * Allocate one q_vector per queue interrupt.
+ *
+ * Return: 0 on success, if allocation fails we return -ENOMEM.
*/
int idpf_vport_intr_alloc(struct idpf_vport *vport)
{
@@ -4000,7 +4022,7 @@ int idpf_vport_intr_alloc(struct idpf_vport *vport)
* idpf_vport_intr_init - Setup all vectors for the given vport
* @vport: virtual port
*
- * Returns 0 on success or negative on failure
+ * Return: 0 on success or negative on failure
*/
int idpf_vport_intr_init(struct idpf_vport *vport)
{
@@ -4039,7 +4061,7 @@ void idpf_vport_intr_ena(struct idpf_vport *vport)
* idpf_config_rss - Send virtchnl messages to configure RSS
* @vport: virtual port
*
- * Return 0 on success, negative on failure
+ * Return: 0 on success, negative on failure
*/
int idpf_config_rss(struct idpf_vport *vport)
{
--
2.53.0
^ permalink raw reply related [flat|nested] 4+ messages in thread
* [PATCH 6.12.y 2/3] idpf: introduce idpf_q_vec_rsrc struct and move vector resources to it
2026-09-17 1:13 ` [PATCH 6.12.y 1/3] idpf: Fix kernel-doc descriptions to avoid warnings Sasha Levin
@ 2026-09-17 1:13 ` Sasha Levin
2026-09-17 1:13 ` [PATCH 6.12.y 3/3] idpf: disable DIM work before freeing q_vectors Sasha Levin
1 sibling, 0 replies; 4+ messages in thread
From: Sasha Levin @ 2026-09-17 1:13 UTC (permalink / raw)
To: stable
Cc: Pavan Kumar Linga, Anton Nadezhdin, Joshua Hay, Samuel Salin,
Tony Nguyen, Sasha Levin
From: Pavan Kumar Linga <pavan.kumar.linga@intel.com>
[ Upstream commit d1061502f353d5e1a8937181b84b722e75fabf11 ]
To group all the vector and queue resources, introduce idpf_q_vec_rsrc
structure. This helps to reuse the same config path functions by other
features. For example, PTP implementation can use the existing config
infrastructure to configure secondary mailbox by passing its queue and
vector info. It also helps to avoid any duplication of code.
Existing queue and vector resources are grouped as default resources.
This patch moves vector info to the newly introduced structure.
Following patch moves the queue resources.
While at it, declare the loop iterator for 'num_q_vectors' in loop and
use the correct type.
Include idpf_q_vec_rsrc backpointer in idpf_alloc_queue_set along with
vport.
Reviewed-by: Anton Nadezhdin <anton.nadezhdin@intel.com>
Signed-off-by: Pavan Kumar Linga <pavan.kumar.linga@intel.com>
Signed-off-by: Joshua Hay <joshua.a.hay@intel.com>
Tested-by: Samuel Salin <Samuel.salin@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
[Backport to 6.12: retain the vector-resource grouping and convert the
existing PF/VF register, interrupt, vport lifecycle and vector-allocation
paths. Also convert the stable-only NAPI scheduling in
idpf_send_disable_queues_msg().
Drop changes to the unavailable queue-set, AF_XDP and NOIRQ support;
retain the stable queue/vector counts, mailbox interface and vport state
handling. Keep the resource member at the original vector fields' location.
No functions are added. This prepares the existing interrupt lifecycle
for 7dd4c829bac2 ("idpf: disable DIM work before freeing q_vectors").]
Stable-dep-of: 7dd4c829bac2 ("idpf: disable DIM work before freeing q_vectors")
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
drivers/net/ethernet/intel/idpf/idpf.h | 24 ++-
drivers/net/ethernet/intel/idpf/idpf_dev.c | 10 +-
drivers/net/ethernet/intel/idpf/idpf_lib.c | 35 ++--
drivers/net/ethernet/intel/idpf/idpf_txrx.c | 183 +++++++++---------
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 14 +-
drivers/net/ethernet/intel/idpf/idpf_vf_dev.c | 10 +-
.../net/ethernet/intel/idpf/idpf_virtchnl.c | 20 +-
.../net/ethernet/intel/idpf/idpf_virtchnl.h | 3 +-
8 files changed, 167 insertions(+), 132 deletions(-)
diff --git a/drivers/net/ethernet/intel/idpf/idpf.h b/drivers/net/ethernet/intel/idpf/idpf.h
index c420f6b712839..55b43a27f106d 100644
--- a/drivers/net/ethernet/intel/idpf/idpf.h
+++ b/drivers/net/ethernet/intel/idpf/idpf.h
@@ -8,6 +8,7 @@
struct idpf_adapter;
struct idpf_vport;
struct idpf_vport_max_q;
+struct idpf_q_vec_rsrc;
#include <net/pkt_sched.h>
#include <linux/aer.h>
@@ -196,7 +197,8 @@ struct idpf_vport_max_q {
*/
struct idpf_reg_ops {
void (*ctlq_reg_init)(struct idpf_ctlq_create_info *cq);
- int (*intr_reg_init)(struct idpf_vport *vport);
+ int (*intr_reg_init)(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc);
void (*mb_intr_reg_init)(struct idpf_adapter *adapter);
void (*reset_reg_init)(struct idpf_adapter *adapter);
void (*trigger_reset)(struct idpf_adapter *adapter,
@@ -251,6 +253,18 @@ struct idpf_port_stats {
struct virtchnl2_vport_stats vport_stats;
};
+/**
+ * struct idpf_q_vec_rsrc - handle for queue and vector resources
+ * @q_vectors: array of queue vectors
+ * @q_vector_idxs: starting index of queue vectors
+ * @num_q_vectors: number of IRQ vectors allocated
+ */
+struct idpf_q_vec_rsrc {
+ struct idpf_q_vector *q_vectors;
+ u16 *q_vector_idxs;
+ u16 num_q_vectors;
+};
+
/**
* struct idpf_vport - Handle for netdevices and queue resources
* @num_txq: Number of allocated TX queues
@@ -284,9 +298,7 @@ struct idpf_port_stats {
* @idx: Software index in adapter vports struct
* @default_vport: Use this vport if one isn't specified
* @base_rxd: True if the driver should use base descriptors instead of flex
- * @num_q_vectors: Number of IRQ vectors allocated
- * @q_vectors: Array of queue vectors
- * @q_vector_idxs: Starting index of queue vectors
+ * @dflt_qv_rsrc: contains default queue and vector resources
* @max_mtu: device given max possible MTU
* @default_mac_addr: device will give a default MAC to use
* @rx_itr_profile: RX profiles for Dynamic Interrupt Moderation
@@ -326,9 +338,7 @@ struct idpf_vport {
bool default_vport;
bool base_rxd;
- u16 num_q_vectors;
- struct idpf_q_vector *q_vectors;
- u16 *q_vector_idxs;
+ struct idpf_q_vec_rsrc dflt_qv_rsrc;
u16 max_mtu;
u8 default_mac_addr[ETH_ALEN];
u16 rx_itr_profile[IDPF_DIM_PROFILE_SLOTS];
diff --git a/drivers/net/ethernet/intel/idpf/idpf_dev.c b/drivers/net/ethernet/intel/idpf/idpf_dev.c
index 41e4bd49402a8..45272bd9e472b 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_dev.c
+++ b/drivers/net/ethernet/intel/idpf/idpf_dev.c
@@ -66,11 +66,13 @@ static void idpf_mb_intr_reg_init(struct idpf_adapter *adapter)
/**
* idpf_intr_reg_init - Initialize interrupt registers
* @vport: virtual port structure
+ * @rsrc: pointer to queue and vector resources
*/
-static int idpf_intr_reg_init(struct idpf_vport *vport)
+static int idpf_intr_reg_init(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc)
{
struct idpf_adapter *adapter = vport->adapter;
- int num_vecs = vport->num_q_vectors;
+ u16 num_vecs = rsrc->num_q_vectors;
struct idpf_vec_regs *reg_vals;
int num_regs, i, err = 0;
u32 rx_itr, tx_itr;
@@ -89,8 +91,8 @@ static int idpf_intr_reg_init(struct idpf_vport *vport)
}
for (i = 0; i < num_vecs; i++) {
- struct idpf_q_vector *q_vector = &vport->q_vectors[i];
- u16 vec_id = vport->q_vector_idxs[i] - IDPF_MBX_Q_VEC;
+ struct idpf_q_vector *q_vector = &rsrc->q_vectors[i];
+ u16 vec_id = rsrc->q_vector_idxs[i] - IDPF_MBX_Q_VEC;
struct idpf_intr_reg *intr = &q_vector->intr_reg;
u32 spacing;
diff --git a/drivers/net/ethernet/intel/idpf/idpf_lib.c b/drivers/net/ethernet/intel/idpf/idpf_lib.c
index c2101f7f6e1bf..36777ca10f32a 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_lib.c
+++ b/drivers/net/ethernet/intel/idpf/idpf_lib.c
@@ -908,6 +908,7 @@ static void idpf_remove_features(struct idpf_vport *vport)
static void idpf_vport_stop(struct idpf_vport *vport)
{
struct idpf_netdev_priv *np = netdev_priv(vport->netdev);
+ struct idpf_q_vec_rsrc *rsrc = &vport->dflt_qv_rsrc;
if (np->state <= __IDPF_VPORT_DOWN)
return;
@@ -929,9 +930,9 @@ static void idpf_vport_stop(struct idpf_vport *vport)
idpf_remove_features(vport);
vport->link_up = false;
- idpf_vport_intr_deinit(vport);
+ idpf_vport_intr_deinit(vport, rsrc);
idpf_vport_queues_rel(vport);
- idpf_vport_intr_rel(vport);
+ idpf_vport_intr_rel(rsrc);
np->state = __IDPF_VPORT_DOWN;
}
@@ -991,6 +992,7 @@ static void idpf_decfg_netdev(struct idpf_vport *vport)
*/
static void idpf_vport_rel(struct idpf_vport *vport)
{
+ struct idpf_q_vec_rsrc *rsrc = &vport->dflt_qv_rsrc;
struct idpf_adapter *adapter = vport->adapter;
struct idpf_vport_config *vport_config;
struct idpf_vector_info vec_info;
@@ -1015,13 +1017,13 @@ static void idpf_vport_rel(struct idpf_vport *vport)
/* Release all the allocated vectors on the stack */
vec_info.num_req_vecs = 0;
- vec_info.num_curr_vecs = vport->num_q_vectors;
+ vec_info.num_curr_vecs = rsrc->num_q_vectors;
vec_info.default_vport = vport->default_vport;
- idpf_req_rel_vector_indexes(adapter, vport->q_vector_idxs, &vec_info);
+ idpf_req_rel_vector_indexes(adapter, rsrc->q_vector_idxs, &vec_info);
- kfree(vport->q_vector_idxs);
- vport->q_vector_idxs = NULL;
+ kfree(rsrc->q_vector_idxs);
+ rsrc->q_vector_idxs = NULL;
kfree(adapter->vport_params_recvd[idx]);
adapter->vport_params_recvd[idx] = NULL;
@@ -1146,6 +1148,7 @@ static struct idpf_vport *idpf_vport_alloc(struct idpf_adapter *adapter,
{
struct idpf_rss_data *rss_data;
u16 idx = adapter->next_vport;
+ struct idpf_q_vec_rsrc *rsrc;
struct idpf_vport *vport;
u16 num_max_q;
int err;
@@ -1193,8 +1196,9 @@ static struct idpf_vport *idpf_vport_alloc(struct idpf_adapter *adapter,
vport->default_vport = adapter->num_alloc_vports <
idpf_get_default_vports(adapter);
- vport->q_vector_idxs = kcalloc(num_max_q, sizeof(u16), GFP_KERNEL);
- if (!vport->q_vector_idxs)
+ rsrc = &vport->dflt_qv_rsrc;
+ rsrc->q_vector_idxs = kcalloc(num_max_q, sizeof(u16), GFP_KERNEL);
+ if (!rsrc->q_vector_idxs)
goto free_vport;
idpf_vport_init(vport, max_q);
@@ -1232,7 +1236,7 @@ static struct idpf_vport *idpf_vport_alloc(struct idpf_adapter *adapter,
kfree(rss_data->rss_key);
rss_data->rss_key = NULL;
free_vector_idxs:
- kfree(vport->q_vector_idxs);
+ kfree(rsrc->q_vector_idxs);
free_vport:
kfree(vport);
@@ -1405,6 +1409,7 @@ static void idpf_rx_init_buf_tail(struct idpf_vport *vport)
static int idpf_vport_open(struct idpf_vport *vport)
{
struct idpf_netdev_priv *np = netdev_priv(vport->netdev);
+ struct idpf_q_vec_rsrc *rsrc = &vport->dflt_qv_rsrc;
struct idpf_adapter *adapter = vport->adapter;
int err;
@@ -1414,7 +1419,7 @@ static int idpf_vport_open(struct idpf_vport *vport)
/* we do not allow interface up just yet */
netif_carrier_off(vport->netdev);
- err = idpf_vport_intr_alloc(vport);
+ err = idpf_vport_intr_alloc(vport, rsrc);
if (err) {
dev_err(&adapter->pdev->dev, "Failed to allocate interrupts for vport %u: %d\n",
vport->vport_id, err);
@@ -1432,7 +1437,7 @@ static int idpf_vport_open(struct idpf_vport *vport)
goto queues_rel;
}
- err = idpf_vport_intr_init(vport);
+ err = idpf_vport_intr_init(vport, rsrc);
if (err) {
dev_err(&adapter->pdev->dev, "Failed to initialize interrupts for vport %u: %d\n",
vport->vport_id, err);
@@ -1454,7 +1459,7 @@ static int idpf_vport_open(struct idpf_vport *vport)
}
idpf_rx_init_buf_tail(vport);
- idpf_vport_intr_ena(vport);
+ idpf_vport_intr_ena(vport, rsrc);
err = idpf_send_config_queues_msg(vport);
if (err) {
@@ -1510,11 +1515,11 @@ static int idpf_vport_open(struct idpf_vport *vport)
unmap_queue_vectors:
idpf_send_map_unmap_queue_vector_msg(vport, false);
intr_deinit:
- idpf_vport_intr_deinit(vport);
+ idpf_vport_intr_deinit(vport, rsrc);
queues_rel:
idpf_vport_queues_rel(vport);
intr_rel:
- idpf_vport_intr_rel(vport);
+ idpf_vport_intr_rel(rsrc);
return err;
}
@@ -1958,7 +1963,7 @@ int idpf_initiate_soft_reset(struct idpf_vport *vport,
memcpy(vport, new_vport, offsetof(struct idpf_vport, link_up));
if (reset_cause == IDPF_SR_Q_CHANGE)
- idpf_vport_alloc_vec_indexes(vport);
+ idpf_vport_alloc_vec_indexes(vport, &vport->dflt_qv_rsrc);
err = idpf_set_real_num_queues(vport);
if (err)
diff --git a/drivers/net/ethernet/intel/idpf/idpf_txrx.c b/drivers/net/ethernet/intel/idpf/idpf_txrx.c
index 370d574f1f9b7..dc7b482ac0484 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_txrx.c
+++ b/drivers/net/ethernet/intel/idpf/idpf_txrx.c
@@ -3237,39 +3237,34 @@ static irqreturn_t idpf_vport_intr_clean_queues(int __always_unused irq,
/**
* idpf_vport_intr_napi_del_all - Unregister napi for all q_vectors in vport
- * @vport: virtual port structure
- *
+ * @rsrc: pointer to queue and vector resources
*/
-static void idpf_vport_intr_napi_del_all(struct idpf_vport *vport)
+static void idpf_vport_intr_napi_del_all(struct idpf_q_vec_rsrc *rsrc)
{
- u16 v_idx;
-
- for (v_idx = 0; v_idx < vport->num_q_vectors; v_idx++)
- netif_napi_del(&vport->q_vectors[v_idx].napi);
+ for (u16 v_idx = 0; v_idx < rsrc->num_q_vectors; v_idx++)
+ netif_napi_del(&rsrc->q_vectors[v_idx].napi);
}
/**
* idpf_vport_intr_napi_dis_all - Disable NAPI for all q_vectors in the vport
- * @vport: main vport structure
+ * @rsrc: pointer to queue and vector resources
*/
-static void idpf_vport_intr_napi_dis_all(struct idpf_vport *vport)
+static void idpf_vport_intr_napi_dis_all(struct idpf_q_vec_rsrc *rsrc)
{
- int v_idx;
-
- for (v_idx = 0; v_idx < vport->num_q_vectors; v_idx++)
- napi_disable(&vport->q_vectors[v_idx].napi);
+ for (u16 v_idx = 0; v_idx < rsrc->num_q_vectors; v_idx++)
+ napi_disable(&rsrc->q_vectors[v_idx].napi);
}
/**
* idpf_vport_intr_rel - Free memory allocated for interrupt vectors
- * @vport: virtual port
+ * @rsrc: pointer to queue and vector resources
*
* Free the memory allocated for interrupt vectors associated to a vport
*/
-void idpf_vport_intr_rel(struct idpf_vport *vport)
+void idpf_vport_intr_rel(struct idpf_q_vec_rsrc *rsrc)
{
- for (u32 v_idx = 0; v_idx < vport->num_q_vectors; v_idx++) {
- struct idpf_q_vector *q_vector = &vport->q_vectors[v_idx];
+ for (u16 v_idx = 0; v_idx < rsrc->num_q_vectors; v_idx++) {
+ struct idpf_q_vector *q_vector = &rsrc->q_vectors[v_idx];
kfree(q_vector->complq);
q_vector->complq = NULL;
@@ -3283,28 +3278,29 @@ void idpf_vport_intr_rel(struct idpf_vport *vport)
free_cpumask_var(q_vector->affinity_mask);
}
- kfree(vport->q_vectors);
- vport->q_vectors = NULL;
+ kfree(rsrc->q_vectors);
+ rsrc->q_vectors = NULL;
}
/**
* idpf_vport_intr_rel_irq - Free the IRQ association with the OS
* @vport: main vport structure
+ * @rsrc: pointer to queue and vector resources
*/
-static void idpf_vport_intr_rel_irq(struct idpf_vport *vport)
+static void idpf_vport_intr_rel_irq(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc)
{
struct idpf_adapter *adapter = vport->adapter;
- int vector;
- for (vector = 0; vector < vport->num_q_vectors; vector++) {
- struct idpf_q_vector *q_vector = &vport->q_vectors[vector];
+ for (u16 vector = 0; vector < rsrc->num_q_vectors; vector++) {
+ struct idpf_q_vector *q_vector = &rsrc->q_vectors[vector];
int irq_num, vidx;
/* free only the irqs that were actually requested */
if (!q_vector)
continue;
- vidx = vport->q_vector_idxs[vector];
+ vidx = rsrc->q_vector_idxs[vector];
irq_num = adapter->msix_entries[vidx].vector;
/* clear the affinity_mask in the IRQ descriptor */
@@ -3315,14 +3311,13 @@ static void idpf_vport_intr_rel_irq(struct idpf_vport *vport)
/**
* idpf_vport_intr_dis_irq_all - Disable all interrupt
- * @vport: main vport structure
+ * @rsrc: pointer to queue and vector resources
*/
-static void idpf_vport_intr_dis_irq_all(struct idpf_vport *vport)
+static void idpf_vport_intr_dis_irq_all(struct idpf_q_vec_rsrc *rsrc)
{
- struct idpf_q_vector *q_vector = vport->q_vectors;
- int q_idx;
+ struct idpf_q_vector *q_vector = rsrc->q_vectors;
- for (q_idx = 0; q_idx < vport->num_q_vectors; q_idx++)
+ for (u16 q_idx = 0; q_idx < rsrc->num_q_vectors; q_idx++)
writel(0, q_vector[q_idx].intr_reg.dyn_ctl);
}
@@ -3462,10 +3457,12 @@ void idpf_vport_intr_update_itr_ena_irq(struct idpf_q_vector *q_vector)
/**
* idpf_vport_intr_req_irq - get MSI-X vectors from the OS for the vport
* @vport: main vport structure
+ * @rsrc: pointer to queue and vector resources
*
* Return: 0 on success, negative on failure
*/
-static int idpf_vport_intr_req_irq(struct idpf_vport *vport)
+static int idpf_vport_intr_req_irq(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc)
{
struct idpf_adapter *adapter = vport->adapter;
const char *drv_name, *if_name, *vec_name;
@@ -3474,11 +3471,11 @@ static int idpf_vport_intr_req_irq(struct idpf_vport *vport)
drv_name = dev_driver_string(&adapter->pdev->dev);
if_name = netdev_name(vport->netdev);
- for (vector = 0; vector < vport->num_q_vectors; vector++) {
- struct idpf_q_vector *q_vector = &vport->q_vectors[vector];
+ for (vector = 0; vector < rsrc->num_q_vectors; vector++) {
+ struct idpf_q_vector *q_vector = &rsrc->q_vectors[vector];
char *name;
- vidx = vport->q_vector_idxs[vector];
+ vidx = rsrc->q_vector_idxs[vector];
irq_num = adapter->msix_entries[vidx].vector;
if (q_vector->num_rxq && q_vector->num_txq)
@@ -3508,9 +3505,9 @@ static int idpf_vport_intr_req_irq(struct idpf_vport *vport)
free_q_irqs:
while (--vector >= 0) {
- vidx = vport->q_vector_idxs[vector];
+ vidx = rsrc->q_vector_idxs[vector];
irq_num = adapter->msix_entries[vidx].vector;
- kfree(free_irq(irq_num, &vport->q_vectors[vector]));
+ kfree(free_irq(irq_num, &rsrc->q_vectors[vector]));
}
return err;
@@ -3539,15 +3536,16 @@ void idpf_vport_intr_write_itr(struct idpf_q_vector *q_vector, u16 itr, bool tx)
/**
* idpf_vport_intr_ena_irq_all - Enable IRQ for the given vport
* @vport: main vport structure
+ * @rsrc: pointer to queue and vector resources
*/
-static void idpf_vport_intr_ena_irq_all(struct idpf_vport *vport)
+static void idpf_vport_intr_ena_irq_all(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc)
{
bool dynamic;
- int q_idx;
u16 itr;
- for (q_idx = 0; q_idx < vport->num_q_vectors; q_idx++) {
- struct idpf_q_vector *qv = &vport->q_vectors[q_idx];
+ for (u16 q_idx = 0; q_idx < rsrc->num_q_vectors; q_idx++) {
+ struct idpf_q_vector *qv = &rsrc->q_vectors[q_idx];
/* Set the initial ITR values */
if (qv->num_txq) {
@@ -3574,13 +3572,15 @@ static void idpf_vport_intr_ena_irq_all(struct idpf_vport *vport)
/**
* idpf_vport_intr_deinit - Release all vector associations for the vport
* @vport: main vport structure
+ * @rsrc: pointer to queue and vector resources
*/
-void idpf_vport_intr_deinit(struct idpf_vport *vport)
+void idpf_vport_intr_deinit(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc)
{
- idpf_vport_intr_dis_irq_all(vport);
- idpf_vport_intr_napi_dis_all(vport);
- idpf_vport_intr_napi_del_all(vport);
- idpf_vport_intr_rel_irq(vport);
+ idpf_vport_intr_dis_irq_all(rsrc);
+ idpf_vport_intr_napi_dis_all(rsrc);
+ idpf_vport_intr_napi_del_all(rsrc);
+ idpf_vport_intr_rel_irq(vport, rsrc);
}
/**
@@ -3652,14 +3652,12 @@ static void idpf_init_dim(struct idpf_q_vector *qv)
/**
* idpf_vport_intr_napi_ena_all - Enable NAPI for all q_vectors in the vport
- * @vport: main vport structure
+ * @rsrc: pointer to queue and vector resources
*/
-static void idpf_vport_intr_napi_ena_all(struct idpf_vport *vport)
+static void idpf_vport_intr_napi_ena_all(struct idpf_q_vec_rsrc *rsrc)
{
- int q_idx;
-
- for (q_idx = 0; q_idx < vport->num_q_vectors; q_idx++) {
- struct idpf_q_vector *q_vector = &vport->q_vectors[q_idx];
+ for (u16 q_idx = 0; q_idx < rsrc->num_q_vectors; q_idx++) {
+ struct idpf_q_vector *q_vector = &rsrc->q_vectors[q_idx];
idpf_init_dim(q_vector);
napi_enable(&q_vector->napi);
@@ -3788,10 +3786,12 @@ static int idpf_vport_splitq_napi_poll(struct napi_struct *napi, int budget)
/**
* idpf_vport_intr_map_vector_to_qs - Map vectors to queues
* @vport: virtual port
+ * @rsrc: pointer to queue and vector resources
*
* Mapping for vectors to queues
*/
-static void idpf_vport_intr_map_vector_to_qs(struct idpf_vport *vport)
+static void idpf_vport_intr_map_vector_to_qs(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc)
{
bool split = idpf_is_queue_model_split(vport->rxq_model);
u16 num_txq_grp = vport->num_txq_grp;
@@ -3802,7 +3802,7 @@ static void idpf_vport_intr_map_vector_to_qs(struct idpf_vport *vport)
for (i = 0, qv_idx = 0; i < vport->num_rxq_grp; i++) {
u16 num_rxq;
- if (qv_idx >= vport->num_q_vectors)
+ if (qv_idx >= rsrc->num_q_vectors)
qv_idx = 0;
rx_qgrp = &vport->rxq_grps[i];
@@ -3818,7 +3818,7 @@ static void idpf_vport_intr_map_vector_to_qs(struct idpf_vport *vport)
q = &rx_qgrp->splitq.rxq_sets[j]->rxq;
else
q = rx_qgrp->singleq.rxqs[j];
- q->q_vector = &vport->q_vectors[qv_idx];
+ q->q_vector = &rsrc->q_vectors[qv_idx];
q_index = q->q_vector->num_rxq;
q->q_vector->rx[q_index] = q;
q->q_vector->num_rxq++;
@@ -3832,7 +3832,7 @@ static void idpf_vport_intr_map_vector_to_qs(struct idpf_vport *vport)
struct idpf_buf_queue *bufq;
bufq = &rx_qgrp->splitq.bufq_sets[j].bufq;
- bufq->q_vector = &vport->q_vectors[qv_idx];
+ bufq->q_vector = &rsrc->q_vectors[qv_idx];
q_index = bufq->q_vector->num_bufq;
bufq->q_vector->bufq[q_index] = bufq;
bufq->q_vector->num_bufq++;
@@ -3847,7 +3847,7 @@ static void idpf_vport_intr_map_vector_to_qs(struct idpf_vport *vport)
for (i = 0, qv_idx = 0; i < num_txq_grp; i++) {
u16 num_txq;
- if (qv_idx >= vport->num_q_vectors)
+ if (qv_idx >= rsrc->num_q_vectors)
qv_idx = 0;
tx_qgrp = &vport->txq_grps[i];
@@ -3857,14 +3857,14 @@ static void idpf_vport_intr_map_vector_to_qs(struct idpf_vport *vport)
struct idpf_tx_queue *q;
q = tx_qgrp->txqs[j];
- q->q_vector = &vport->q_vectors[qv_idx];
+ q->q_vector = &rsrc->q_vectors[qv_idx];
q->q_vector->tx[q->q_vector->num_txq++] = q;
}
if (split) {
struct idpf_compl_queue *q = tx_qgrp->complq;
- q->q_vector = &vport->q_vectors[qv_idx];
+ q->q_vector = &rsrc->q_vectors[qv_idx];
q->q_vector->complq[q->q_vector->num_complq++] = q;
}
@@ -3875,12 +3875,14 @@ static void idpf_vport_intr_map_vector_to_qs(struct idpf_vport *vport)
/**
* idpf_vport_intr_init_vec_idx - Initialize the vector indexes
* @vport: virtual port
+ * @rsrc: pointer to queue and vector resources
*
* Initialize vector indexes with values returned over mailbox.
*
* Return: 0 on success, negative on failure
*/
-static int idpf_vport_intr_init_vec_idx(struct idpf_vport *vport)
+static int idpf_vport_intr_init_vec_idx(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc)
{
struct idpf_adapter *adapter = vport->adapter;
struct virtchnl2_alloc_vectors *ac;
@@ -3889,8 +3891,8 @@ static int idpf_vport_intr_init_vec_idx(struct idpf_vport *vport)
ac = adapter->req_vec_chunks;
if (!ac) {
- for (i = 0; i < vport->num_q_vectors; i++)
- vport->q_vectors[i].v_idx = vport->q_vector_idxs[i];
+ for (i = 0; i < rsrc->num_q_vectors; i++)
+ rsrc->q_vectors[i].v_idx = rsrc->q_vector_idxs[i];
return 0;
}
@@ -3902,8 +3904,8 @@ static int idpf_vport_intr_init_vec_idx(struct idpf_vport *vport)
idpf_get_vec_ids(adapter, vecids, total_vecs, &ac->vchunks);
- for (i = 0; i < vport->num_q_vectors; i++)
- vport->q_vectors[i].v_idx = vecids[vport->q_vector_idxs[i]];
+ for (i = 0; i < rsrc->num_q_vectors; i++)
+ rsrc->q_vectors[i].v_idx = vecids[rsrc->q_vector_idxs[i]];
kfree(vecids);
@@ -3913,19 +3915,20 @@ static int idpf_vport_intr_init_vec_idx(struct idpf_vport *vport)
/**
* idpf_vport_intr_napi_add_all- Register napi handler for all qvectors
* @vport: virtual port structure
+ * @rsrc: pointer to queue and vector resources
*/
-static void idpf_vport_intr_napi_add_all(struct idpf_vport *vport)
+static void idpf_vport_intr_napi_add_all(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc)
{
int (*napi_poll)(struct napi_struct *napi, int budget);
- u16 v_idx;
if (idpf_is_queue_model_split(vport->txq_model))
napi_poll = idpf_vport_splitq_napi_poll;
else
napi_poll = idpf_vport_singleq_napi_poll;
- for (v_idx = 0; v_idx < vport->num_q_vectors; v_idx++) {
- struct idpf_q_vector *q_vector = &vport->q_vectors[v_idx];
+ for (u16 v_idx = 0; v_idx < rsrc->num_q_vectors; v_idx++) {
+ struct idpf_q_vector *q_vector = &rsrc->q_vectors[v_idx];
netif_napi_add(vport->netdev, &q_vector->napi, napi_poll);
@@ -3938,38 +3941,41 @@ static void idpf_vport_intr_napi_add_all(struct idpf_vport *vport)
/**
* idpf_vport_intr_alloc - Allocate memory for interrupt vectors
* @vport: virtual port
+ * @rsrc: pointer to queue and vector resources
*
* Allocate one q_vector per queue interrupt.
*
* Return: 0 on success, if allocation fails we return -ENOMEM.
*/
-int idpf_vport_intr_alloc(struct idpf_vport *vport)
+int idpf_vport_intr_alloc(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc)
{
u16 txqs_per_vector, rxqs_per_vector, bufqs_per_vector;
struct idpf_vport_user_config_data *user_config;
struct idpf_q_vector *q_vector;
struct idpf_q_coalesce *q_coal;
- u32 complqs_per_vector, v_idx;
+ u32 complqs_per_vector;
u16 idx = vport->idx;
user_config = &vport->adapter->vport_config[idx]->user_config;
- vport->q_vectors = kcalloc(vport->num_q_vectors,
- sizeof(struct idpf_q_vector), GFP_KERNEL);
- if (!vport->q_vectors)
+
+ rsrc->q_vectors = kcalloc(rsrc->num_q_vectors,
+ sizeof(struct idpf_q_vector), GFP_KERNEL);
+ if (!rsrc->q_vectors)
return -ENOMEM;
txqs_per_vector = DIV_ROUND_UP(vport->num_txq_grp,
- vport->num_q_vectors);
+ rsrc->num_q_vectors);
rxqs_per_vector = DIV_ROUND_UP(vport->num_rxq_grp,
- vport->num_q_vectors);
+ rsrc->num_q_vectors);
bufqs_per_vector = vport->num_bufqs_per_qgrp *
DIV_ROUND_UP(vport->num_rxq_grp,
- vport->num_q_vectors);
+ rsrc->num_q_vectors);
complqs_per_vector = DIV_ROUND_UP(vport->num_txq_grp,
- vport->num_q_vectors);
+ rsrc->num_q_vectors);
- for (v_idx = 0; v_idx < vport->num_q_vectors; v_idx++) {
- q_vector = &vport->q_vectors[v_idx];
+ for (u16 v_idx = 0; v_idx < rsrc->num_q_vectors; v_idx++) {
+ q_vector = &rsrc->q_vectors[v_idx];
q_coal = &user_config->q_coalesce[v_idx];
q_vector->vport = vport;
@@ -4013,7 +4019,7 @@ int idpf_vport_intr_alloc(struct idpf_vport *vport)
return 0;
error:
- idpf_vport_intr_rel(vport);
+ idpf_vport_intr_rel(rsrc);
return -ENOMEM;
}
@@ -4021,40 +4027,41 @@ int idpf_vport_intr_alloc(struct idpf_vport *vport)
/**
* idpf_vport_intr_init - Setup all vectors for the given vport
* @vport: virtual port
+ * @rsrc: pointer to queue and vector resources
*
* Return: 0 on success or negative on failure
*/
-int idpf_vport_intr_init(struct idpf_vport *vport)
+int idpf_vport_intr_init(struct idpf_vport *vport, struct idpf_q_vec_rsrc *rsrc)
{
int err;
- err = idpf_vport_intr_init_vec_idx(vport);
+ err = idpf_vport_intr_init_vec_idx(vport, rsrc);
if (err)
return err;
- idpf_vport_intr_map_vector_to_qs(vport);
- idpf_vport_intr_napi_add_all(vport);
+ idpf_vport_intr_map_vector_to_qs(vport, rsrc);
+ idpf_vport_intr_napi_add_all(vport, rsrc);
- err = vport->adapter->dev_ops.reg_ops.intr_reg_init(vport);
+ err = vport->adapter->dev_ops.reg_ops.intr_reg_init(vport, rsrc);
if (err)
goto unroll_vectors_alloc;
- err = idpf_vport_intr_req_irq(vport);
+ err = idpf_vport_intr_req_irq(vport, rsrc);
if (err)
goto unroll_vectors_alloc;
return 0;
unroll_vectors_alloc:
- idpf_vport_intr_napi_del_all(vport);
+ idpf_vport_intr_napi_del_all(rsrc);
return err;
}
-void idpf_vport_intr_ena(struct idpf_vport *vport)
+void idpf_vport_intr_ena(struct idpf_vport *vport, struct idpf_q_vec_rsrc *rsrc)
{
- idpf_vport_intr_napi_ena_all(vport);
- idpf_vport_intr_ena_irq_all(vport);
+ idpf_vport_intr_napi_ena_all(rsrc);
+ idpf_vport_intr_ena_irq_all(vport, rsrc);
}
/**
diff --git a/drivers/net/ethernet/intel/idpf/idpf_txrx.h b/drivers/net/ethernet/intel/idpf/idpf_txrx.h
index 2d9960dd51ad5..9faa19e2e1744 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_txrx.h
+++ b/drivers/net/ethernet/intel/idpf/idpf_txrx.h
@@ -1011,12 +1011,16 @@ int idpf_vport_calc_total_qs(struct idpf_adapter *adapter, u16 vport_index,
void idpf_vport_calc_num_q_groups(struct idpf_vport *vport);
int idpf_vport_queues_alloc(struct idpf_vport *vport);
void idpf_vport_queues_rel(struct idpf_vport *vport);
-void idpf_vport_intr_rel(struct idpf_vport *vport);
-int idpf_vport_intr_alloc(struct idpf_vport *vport);
+void idpf_vport_intr_rel(struct idpf_q_vec_rsrc *rsrc);
+int idpf_vport_intr_alloc(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc);
void idpf_vport_intr_update_itr_ena_irq(struct idpf_q_vector *q_vector);
-void idpf_vport_intr_deinit(struct idpf_vport *vport);
-int idpf_vport_intr_init(struct idpf_vport *vport);
-void idpf_vport_intr_ena(struct idpf_vport *vport);
+void idpf_vport_intr_deinit(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc);
+int idpf_vport_intr_init(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc);
+void idpf_vport_intr_ena(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc);
void idpf_fill_dflt_rss_lut(struct idpf_vport *vport);
int idpf_config_rss(struct idpf_vport *vport);
int idpf_init_rss_lut(struct idpf_vport *vport);
diff --git a/drivers/net/ethernet/intel/idpf/idpf_vf_dev.c b/drivers/net/ethernet/intel/idpf/idpf_vf_dev.c
index aba828abcb171..61d6f774e2f66 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_vf_dev.c
+++ b/drivers/net/ethernet/intel/idpf/idpf_vf_dev.c
@@ -66,11 +66,13 @@ static void idpf_vf_mb_intr_reg_init(struct idpf_adapter *adapter)
/**
* idpf_vf_intr_reg_init - Initialize interrupt registers
* @vport: virtual port structure
+ * @rsrc: pointer to queue and vector resources
*/
-static int idpf_vf_intr_reg_init(struct idpf_vport *vport)
+static int idpf_vf_intr_reg_init(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc)
{
struct idpf_adapter *adapter = vport->adapter;
- int num_vecs = vport->num_q_vectors;
+ u16 num_vecs = rsrc->num_q_vectors;
struct idpf_vec_regs *reg_vals;
int num_regs, i, err = 0;
u32 rx_itr, tx_itr;
@@ -89,8 +91,8 @@ static int idpf_vf_intr_reg_init(struct idpf_vport *vport)
}
for (i = 0; i < num_vecs; i++) {
- struct idpf_q_vector *q_vector = &vport->q_vectors[i];
- u16 vec_id = vport->q_vector_idxs[i] - IDPF_MBX_Q_VEC;
+ struct idpf_q_vector *q_vector = &rsrc->q_vectors[i];
+ u16 vec_id = rsrc->q_vector_idxs[i] - IDPF_MBX_Q_VEC;
struct idpf_intr_reg *intr = &q_vector->intr_reg;
u32 spacing;
diff --git a/drivers/net/ethernet/intel/idpf/idpf_virtchnl.c b/drivers/net/ethernet/intel/idpf/idpf_virtchnl.c
index c4ae7d62aaf8d..4e49e4673dd63 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_virtchnl.c
+++ b/drivers/net/ethernet/intel/idpf/idpf_virtchnl.c
@@ -1993,6 +1993,7 @@ int idpf_send_enable_queues_msg(struct idpf_vport *vport)
*/
int idpf_send_disable_queues_msg(struct idpf_vport *vport)
{
+ struct idpf_q_vec_rsrc *rsrc = &vport->dflt_qv_rsrc;
int err, i;
err = idpf_send_ena_dis_queues_msg(vport, false);
@@ -2007,8 +2008,8 @@ int idpf_send_disable_queues_msg(struct idpf_vport *vport)
/* schedule the napi to receive all the marker packets */
local_bh_disable();
- for (i = 0; i < vport->num_q_vectors; i++)
- napi_schedule(&vport->q_vectors[i].napi);
+ for (u16 i = 0; i < rsrc->num_q_vectors; i++)
+ napi_schedule(&rsrc->q_vectors[i].napi);
local_bh_enable();
return idpf_wait_for_marker_event(vport);
@@ -3118,25 +3119,27 @@ void idpf_vc_core_deinit(struct idpf_adapter *adapter)
/**
* idpf_vport_alloc_vec_indexes - Get relative vector indexes
* @vport: virtual port data struct
+ * @rsrc: pointer to queue and vector resources
*
* This function requests the vector information required for the vport and
* stores the vector indexes received from the 'global vector distribution'
* in the vport's queue vectors array.
*
- * Return 0 on success, error on failure
+ * Return: 0 on success, error on failure
*/
-int idpf_vport_alloc_vec_indexes(struct idpf_vport *vport)
+int idpf_vport_alloc_vec_indexes(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc)
{
struct idpf_vector_info vec_info;
int num_alloc_vecs;
- vec_info.num_curr_vecs = vport->num_q_vectors;
+ vec_info.num_curr_vecs = rsrc->num_q_vectors;
vec_info.num_req_vecs = max(vport->num_txq, vport->num_rxq);
vec_info.default_vport = vport->default_vport;
vec_info.index = vport->idx;
num_alloc_vecs = idpf_req_rel_vector_indexes(vport->adapter,
- vport->q_vector_idxs,
+ rsrc->q_vector_idxs,
&vec_info);
if (num_alloc_vecs <= 0) {
dev_err(&vport->adapter->pdev->dev, "Vector distribution failed: %d\n",
@@ -3144,7 +3147,7 @@ int idpf_vport_alloc_vec_indexes(struct idpf_vport *vport)
return -EINVAL;
}
- vport->num_q_vectors = num_alloc_vecs;
+ rsrc->num_q_vectors = num_alloc_vecs;
return 0;
}
@@ -3158,6 +3161,7 @@ int idpf_vport_alloc_vec_indexes(struct idpf_vport *vport)
*/
void idpf_vport_init(struct idpf_vport *vport, struct idpf_vport_max_q *max_q)
{
+ struct idpf_q_vec_rsrc *rsrc = &vport->dflt_qv_rsrc;
struct idpf_adapter *adapter = vport->adapter;
struct virtchnl2_create_vport *vport_msg;
struct idpf_vport_config *vport_config;
@@ -3196,7 +3200,7 @@ void idpf_vport_init(struct idpf_vport *vport, struct idpf_vport_max_q *max_q)
idpf_vport_init_num_qs(vport, vport_msg);
idpf_vport_calc_num_q_desc(vport);
idpf_vport_calc_num_q_groups(vport);
- idpf_vport_alloc_vec_indexes(vport);
+ idpf_vport_alloc_vec_indexes(vport, rsrc);
vport->crc_enable = adapter->crc_enable;
}
diff --git a/drivers/net/ethernet/intel/idpf/idpf_virtchnl.h b/drivers/net/ethernet/intel/idpf/idpf_virtchnl.h
index 23271cf0a2160..0bfe44fa0385c 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_virtchnl.h
+++ b/drivers/net/ethernet/intel/idpf/idpf_virtchnl.h
@@ -45,7 +45,8 @@ int idpf_send_enable_queues_msg(struct idpf_vport *vport);
int idpf_send_disable_queues_msg(struct idpf_vport *vport);
int idpf_send_config_queues_msg(struct idpf_vport *vport);
-int idpf_vport_alloc_vec_indexes(struct idpf_vport *vport);
+int idpf_vport_alloc_vec_indexes(struct idpf_vport *vport,
+ struct idpf_q_vec_rsrc *rsrc);
int idpf_get_vec_ids(struct idpf_adapter *adapter,
u16 *vecids, int num_vecids,
struct virtchnl2_vector_chunks *chunks);
--
2.53.0
^ permalink raw reply related [flat|nested] 4+ messages in thread
* [PATCH 6.12.y 3/3] idpf: disable DIM work before freeing q_vectors
2026-09-17 1:13 ` [PATCH 6.12.y 1/3] idpf: Fix kernel-doc descriptions to avoid warnings Sasha Levin
2026-09-17 1:13 ` [PATCH 6.12.y 2/3] idpf: introduce idpf_q_vec_rsrc struct and move vector resources to it Sasha Levin
@ 2026-09-17 1:13 ` Sasha Levin
1 sibling, 0 replies; 4+ messages in thread
From: Sasha Levin @ 2026-09-17 1:13 UTC (permalink / raw)
To: stable; +Cc: Myeonghun Pak, Ijae Kim, Samuel Salin, Tony Nguyen, Sasha Levin
From: Myeonghun Pak <mhun512@gmail.com>
[ Upstream commit 7dd4c829bac2916be98a3e34b41daaba7f42b4c4 ]
idpf never drains the Tx/Rx DIM works before freeing the memory they
live in. tx_dim and rx_dim are embedded in struct idpf_q_vector, they
are queued from the NAPI poll via net_dim(), and idpf_vport_intr_rel()
ends with kfree(rsrc->q_vectors). Nothing in the driver cancels them.
idpf_tx_dim_work() and idpf_rx_dim_work() then run on freed memory:
idpf_vport_intr_write_itr() writes the ITR register through
q_vector->intr_reg.tx_itr / rx_itr, void __iomem pointers loaded out of
the freed q_vector. No configuration is needed to get there --
IDPF_ITR_IS_DYNAMIC() is defined as (itr_mode) and idpf_vport_alloc()
initialises both modes to IDPF_ITR_DYNAMIC.
Draining after idpf_vport_intr_napi_dis_all() is not enough on its own.
idpf_net_dim() is called from inside the
"if (napi_complete_done(napi, work_done))" branch of the poll, and
napi_complete_done() has already cleared NAPIF_STATE_SCHED by then.
napi_disable_locked() waits only while (val & (NAPIF_STATE_SCHED |
NAPIF_STATE_NPSVC)), so napi_disable() can return while the poll tail is
still queueing the work, and a plain cancel_work_sync() would be
re-armed behind the drain.
Use disable_work_sync(): schedule_work() on a work with a non-zero
disable count is dropped by clear_pending_if_disabled() before
__queue_work() is reached.
Move idpf_init_dim() to idpf_vport_intr_alloc() so the works are
initialised on every path that can reach the drain -- the three
"goto intr_deinit" sites between idpf_vport_intr_init() and
idpf_vport_intr_ena() get there without the enable side having run.
Nothing re-enables them: rsrc->q_vectors is freed on every exit from
idpf_vport_open() and on every idpf_vport_stop(), so the count dies with
the object.
It is a race, not a deterministic failure -- net_dim() only schedules
once DIM_NEVENTS events have accumulated and the profile index changes.
A KASAN ifup/ifdown loop under load is the way to see it.
Fixes: c2d548cad150 ("idpf: add TX splitq napi poll support")
Fixes: 3a8845af66ed ("idpf: add RX splitq napi poll support")
Cc: <stable@vger.kernel.org> # see patch description, needs adjustments for <= 6.9
Co-developed-by: Ijae Kim <ae878000@gmail.com>
Signed-off-by: Ijae Kim <ae878000@gmail.com>
Signed-off-by: Myeonghun Pak <mhun512@gmail.com>
Tested-by: Samuel Salin <Samuel.salin@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
drivers/net/ethernet/intel/idpf/idpf_txrx.c | 24 ++++++++++++++++++++-
1 file changed, 23 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/intel/idpf/idpf_txrx.c b/drivers/net/ethernet/intel/idpf/idpf_txrx.c
index dc7b482ac0484..2597e86b9adab 100644
--- a/drivers/net/ethernet/intel/idpf/idpf_txrx.c
+++ b/drivers/net/ethernet/intel/idpf/idpf_txrx.c
@@ -3569,6 +3569,26 @@ static void idpf_vport_intr_ena_irq_all(struct idpf_vport *vport,
}
}
+/**
+ * idpf_vport_intr_dis_dim_all - Disable DIM work for all q_vectors
+ * @rsrc: pointer to queue and vector resources
+ *
+ * The DIM works are embedded in the q_vector array that
+ * idpf_vport_intr_rel() frees, and the poll arms them after
+ * napi_complete_done() has already cleared NAPI_STATE_SCHED. Disable
+ * rather than just cancel, so that a poll tail still running past
+ * napi_disable() cannot queue them again behind the drain.
+ */
+static void idpf_vport_intr_dis_dim_all(struct idpf_q_vec_rsrc *rsrc)
+{
+ for (u16 v_idx = 0; v_idx < rsrc->num_q_vectors; v_idx++) {
+ struct idpf_q_vector *q_vector = &rsrc->q_vectors[v_idx];
+
+ disable_work_sync(&q_vector->tx_dim.work);
+ disable_work_sync(&q_vector->rx_dim.work);
+ }
+}
+
/**
* idpf_vport_intr_deinit - Release all vector associations for the vport
* @vport: main vport structure
@@ -3579,6 +3599,7 @@ void idpf_vport_intr_deinit(struct idpf_vport *vport,
{
idpf_vport_intr_dis_irq_all(rsrc);
idpf_vport_intr_napi_dis_all(rsrc);
+ idpf_vport_intr_dis_dim_all(rsrc);
idpf_vport_intr_napi_del_all(rsrc);
idpf_vport_intr_rel_irq(vport, rsrc);
}
@@ -3659,7 +3680,6 @@ static void idpf_vport_intr_napi_ena_all(struct idpf_q_vec_rsrc *rsrc)
for (u16 q_idx = 0; q_idx < rsrc->num_q_vectors; q_idx++) {
struct idpf_q_vector *q_vector = &rsrc->q_vectors[q_idx];
- idpf_init_dim(q_vector);
napi_enable(&q_vector->napi);
}
}
@@ -3979,6 +3999,8 @@ int idpf_vport_intr_alloc(struct idpf_vport *vport,
q_coal = &user_config->q_coalesce[v_idx];
q_vector->vport = vport;
+ idpf_init_dim(q_vector);
+
q_vector->tx_itr_value = q_coal->tx_coalesce_usecs;
q_vector->tx_intr_mode = q_coal->tx_intr_mode;
q_vector->tx_itr_idx = VIRTCHNL2_ITR_IDX_1;
--
2.53.0
^ permalink raw reply related [flat|nested] 4+ messages in thread
end of thread, other threads:[~2026-09-17 1:13 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-09-16 15:43 FAILED: patch "[PATCH] idpf: disable DIM work before freeing q_vectors" failed to apply to 6.12-stable tree gregkh
2026-09-17 1:13 ` [PATCH 6.12.y 1/3] idpf: Fix kernel-doc descriptions to avoid warnings Sasha Levin
2026-09-17 1:13 ` [PATCH 6.12.y 2/3] idpf: introduce idpf_q_vec_rsrc struct and move vector resources to it Sasha Levin
2026-09-17 1:13 ` [PATCH 6.12.y 3/3] idpf: disable DIM work before freeing q_vectors Sasha Levin
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).