[PATCH ath-next v3 0/2] wifi: ath12k: accept what newer QCN9274 firmware requests at boot

[PATCH ath-next v3 0/2] wifi: ath12k: accept what newer QCN9274 firmware requests at boot

[Goran Sustek]

QCN9274 firmware WLAN.WBE.1.6-01243 requests two things at boot that
ath12k does not implement: a target memory segment of type 0xa (AFC)
and direct buffer rings for module IDs 2 and 3. Neither is needed for
operation. Today the first gets a wrong answer and the second aborts
the parse of the service ready ext event.

Patch 1 answers every unsupported memory type with size 0. Patch 2
skips unknown ring module IDs instead of failing the parse. Both keep
a warning, so the gap between firmware and driver stays visible.

v3:
- 1/2: clear the size in the default case, no AFC case; keep the
  warning; describe only the response change (Rameshkumar Sundaram).
- 2/2: warn for a skipped entry instead of a debug message.
- Rebased on ath-next.

v2:
- 1/2: do not allocate the region, answer with size 0 (Rameshkumar
  Sundaram); former 3/3 dropped.
- 2/2: commit message and comment reworded.

v2: https://lore.kernel.org/r/20261102200235.48973-1-gsustek@gmail.com
v1: https://lore.kernel.org/r/20261030220836.14164-1-gsustek@gmail.com

Goran Sustek (2):
  wifi: ath12k: report size 0 for unsupported target memory types
  wifi: ath12k: skip unknown direct buffer ring module IDs

 drivers/net/wireless/ath/ath12k/qmi.c |  1 +
 drivers/net/wireless/ath/ath12k/wmi.c | 43 +++++++++++++++++----------
 2 files changed, 29 insertions(+), 15 deletions(-)


base-commit: d0fbb39d65dc9dcb4175fa6b23e473dec3670f4d
-- 
2.56.0

[PATCH ath-next v3 1/2] wifi: ath12k: report size 0 for unsupported target memory types

[Goran Sustek]

For a target memory segment of an unsupported type,
ath12k_qmi_alloc_target_mem_chunk() warns, clears the DMA address in
chunk->paddr and keeps the requested size. The response then carries
address 0 with a non-zero size for that segment.

QCN9274 firmware WLAN.WBE.1.6-01243 requests such a segment at every
boot: type 0xa (QMI_WLFW_AFC_MEM_V01), 8 KiB for an Automated Frequency
Coordination response, which ath12k does not support:

  ath12k_wifi7_pci 0000:06:00.0: memory type 10 not supported

The firmware treats a non-zero size as a provided region and maps the
given address. Nothing visible follows with this firmware: memory ready,
firmware ready and all radios come up with or without this change. The
response is still wrong, since it claims a region the host did not
provide.

Clear the size as well, so that an unsupported type is answered with
address 0 and size 0, which the firmware treats as not provided. The
warning is unchanged.

Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.6-01243-QCAHKSWPL_SILICONZ-1
Signed-off-by: Goran Sustek <gsustek@gmail.com>
---
v3: size cleared in the default case, no AFC case; warning kept;
    message reworded (Rameshkumar Sundaram)
v2: no allocation, size 0 (Rameshkumar Sundaram)

 drivers/net/wireless/ath/ath12k/qmi.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/wireless/ath/ath12k/qmi.c b/drivers/net/wireless/ath/ath12k/qmi.c
index 52a925fc44ee..4aaee5870ae9 100644
--- a/drivers/net/wireless/ath/ath12k/qmi.c
+++ b/drivers/net/wireless/ath/ath12k/qmi.c
@@ -2792,6 +2792,7 @@ static int ath12k_qmi_alloc_target_mem_chunk(struct ath12k_base *ab)
 				    chunk->type);
 			chunk->paddr = 0;
 			chunk->v.addr = NULL;
+			chunk->size = 0;
 			break;
 		}
 	}
-- 
2.56.0

[PATCH ath-next v3 2/2] wifi: ath12k: skip unknown direct buffer ring module IDs

[Goran Sustek]

QCN9274 firmware WLAN.WBE.1.6-01243 advertises direct buffer rings for
module IDs 2 (CV upload) and 3 (Wi-Fi radar). ath12k_wmi_dma_ring_caps()
treats any ID at or above WMI_DIRECT_BUF_MAX as fatal:

  ath12k_wifi7_pci 0000:06:00.0: Invalid module id 2
  ath12k_wifi7_pci 0000:06:00.0: failed to parse tlv -22

It frees the capabilities stored so far and aborts the parse of the
service ready ext event; initialisation continues only because the
service ready ext2 event completes it.

An unknown ID is not an error: the firmware uses a ring only after the
host configures it, and the host does that only for a feature it
implements. Skip such entries with a warning, as for an unsupported
memory type in ath12k_qmi_alloc_target_mem_chunk(). Store the accepted
entries contiguously and set num_db_cap to the stored count, so that a
skipped entry cannot read as a zero-sized spectral ring.

Based on Nazar Mokrynskyi's version, with the separate count that Jeff
Johnson suggested in its review.

Link: https://lore.kernel.org/linux-wireless/20260505172415.566328-3-nazar@mokrynskyi.com/
Suggested-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.6-01243-QCAHKSWPL_SILICONZ-1
Signed-off-by: Goran Sustek <gsustek@gmail.com>
---
v3: warning instead of a debug message for a skipped entry
v2: commit message and comment reworded

 drivers/net/wireless/ath/ath12k/wmi.c | 43 +++++++++++++++++----------
 1 file changed, 28 insertions(+), 15 deletions(-)

diff --git a/drivers/net/wireless/ath/ath12k/wmi.c b/drivers/net/wireless/ath/ath12k/wmi.c
index f41138a7c68e..9720bed21631 100644
--- a/drivers/net/wireless/ath/ath12k/wmi.c
+++ b/drivers/net/wireless/ath/ath12k/wmi.c
@@ -4988,8 +4988,8 @@ static int ath12k_wmi_dma_ring_caps(struct ath12k_base *ab,
 	struct ath12k_wmi_dma_ring_caps_parse *dma_caps_parse = data;
 	struct ath12k_wmi_dma_ring_caps_params *dma_caps;
 	struct ath12k_dbring_cap *dir_buff_caps;
+	u32 i, n = 0, module_id;
 	int ret;
-	u32 i;
 
 	dma_caps_parse->n_dma_ring_caps = 0;
 	dma_caps = (struct ath12k_wmi_dma_ring_caps_params *)ptr;
@@ -5015,26 +5015,39 @@ static int ath12k_wmi_dma_ring_caps(struct ath12k_base *ab,
 
 	dir_buff_caps = ab->db_caps;
 	for (i = 0; i < dma_caps_parse->n_dma_ring_caps; i++) {
-		if (le32_to_cpu(dma_caps[i].module_id) >= WMI_DIRECT_BUF_MAX) {
-			ath12k_warn(ab, "Invalid module id %d\n",
-				    le32_to_cpu(dma_caps[i].module_id));
-			ret = -EINVAL;
-			goto free_dir_buff;
+		module_id = le32_to_cpu(dma_caps[i].module_id);
+
+		/* Firmware may advertise rings for modules this driver
+		 * does not implement. It uses a ring only after the host
+		 * configures it, so skipping the entry is safe. Warn, as
+		 * for an unsupported memory type, so the gap is visible.
+		 */
+		if (module_id >= WMI_DIRECT_BUF_MAX) {
+			ath12k_warn(ab, "dma ring caps for module id %u not supported, skipping\n",
+				    module_id);
+			continue;
 		}
 
-		dir_buff_caps[i].id = le32_to_cpu(dma_caps[i].module_id);
-		dir_buff_caps[i].pdev_id =
+		dir_buff_caps[n].id = module_id;
+		dir_buff_caps[n].pdev_id =
 			DP_HW2SW_MACID(le32_to_cpu(dma_caps[i].pdev_id));
-		dir_buff_caps[i].min_elem = le32_to_cpu(dma_caps[i].min_elem);
-		dir_buff_caps[i].min_buf_sz = le32_to_cpu(dma_caps[i].min_buf_sz);
-		dir_buff_caps[i].min_buf_align = le32_to_cpu(dma_caps[i].min_buf_align);
+		dir_buff_caps[n].min_elem = le32_to_cpu(dma_caps[i].min_elem);
+		dir_buff_caps[n].min_buf_sz = le32_to_cpu(dma_caps[i].min_buf_sz);
+		dir_buff_caps[n].min_buf_align = le32_to_cpu(dma_caps[i].min_buf_align);
+		n++;
 	}
 
-	return 0;
+	if (!n) {
+		ath12k_wmi_free_dbring_caps(ab);
+		return 0;
+	}
 
-free_dir_buff:
-	ath12k_wmi_free_dbring_caps(ab);
-	return ret;
+	/* count only the entries stored, so no zeroed entry looks like
+	 * a spectral ring
+	 */
+	ab->num_db_cap = n;
+
+	return 0;
 }
 
 static void
-- 
2.56.0