Skip to contentSkip to conversation

ethernet: intel: fix building with large NR_CPUS

PATCH New Sasha Levin 1 participant 3 messages Jul 23, 2025 Stable branch

lore
Share & tools
lore

Sasha Levin

author maintainer
#0 lore raw

PATCH 1/3 ethernet: intel: fix building with large NR_CPUS

From: Arnd Bergmann <[email protected]>

[ Upstream commit 24171a5a4a952c26568ff0d2a0bc8c4708a95e1d ]

With large values of CONFIG_NR_CPUS, three Intel ethernet drivers fail to compile like:

In function ‘i40e_free_q_vector’,

    inlined from ‘i40e_vsi_alloc_q_vectors’ at drivers/net/ethernet/intel/i40e/i40e_main.c:12112:3:

571 | _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
include/linux/rcupdate.h:1084:17: note: in expansion of macro ‘BUILD_BUG_ON’
1084 | BUILD_BUG_ON(offsetof(typeof(*(ptr)), rhf) >= 4096); \
drivers/net/ethernet/intel/i40e/i40e_main.c:5113:9: note: in expansion of macro ‘kfree_rcu’

 5113 |         kfree_rcu(q_vector, rcu);
      |         ^~~~~~~~~

The problem is that the 'rcu' member in 'q_vector' is too far from the start of the structure. Move this member before the CPU mask instead, in all three drivers.

---

LLM Generated explanations, may be completely bogus:

**YES**

This commit should be backported to stable kernel trees for the following reasons:

1. **Compilation Failure Fix**: This fixes a build-time compilation failure that occurs with large values of CONFIG_NR_CPUS. The error prevents the kernel from building successfully, which is a significant issue for systems configured with high CPU counts.

2. **Root Cause**: The issue stems from the kfree_rcu() macro in include/linux/rcupdate.h:1084 which includes a BUILD_BUG_ON(offsetof(typeof(*(ptr)), rhf) >= 4096) check. This compile-time assertion ensures that the RCU head field is within the first 4096 bytes of the structure.

3. **Simple and Safe Fix**: The fix is straightforward - it moves the struct rcu_head rcu member earlier in the structure, placing it before the cpumask_t affinity_mask member. This is a safe change because:
- It only reorders structure members
- The RCU head doesn't need to be at any specific location functionally
- Moving it earlier reduces the offset from the structure start
- No functional code changes are required

4. **Affects Multiple Drivers**: The issue affects three Intel ethernet drivers (i40e, ixgbe, fm10k), all of which are fixed identically by moving the rcu member earlier in their respective q_vector structures.

5. **Configuration-Dependent Bug**: This bug only manifests with large CONFIG_NR_CPUS values (likely >= 512 or higher), where the cpumask_t type becomes large enough to push the rcu member beyond the 4096-byte offset limit. Systems with high core counts are becoming more common in production environments.

6. **No Risk of Regression**: The change is minimal and doesn't alter any functionality. It's purely a structural reordering that maintains all existing behavior while fixing the compilation issue.

7. **Meets Stable Criteria**: This fix clearly meets the stable kernel criteria as it:
- Fixes a real bug (compilation failure)
- Is minimal and contained
- Has no risk of introducing new issues
- Affects users with legitimate configurations

The commit is an excellent candidate for stable backporting as it fixes a real compilation issue that prevents kernel builds on systems with large CPU counts, and the fix is trivial with no risk of regression.

 drivers/net/ethernet/intel/fm10k/fm10k.h | 3 ++-
 drivers/net/ethernet/intel/i40e/i40e.h   | 2 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe.h | 3 ++-
 3 files changed, 5 insertions(+), 3 deletions(-)
drivers/net/ethernet/intel/fm10k/fm10k.h +2 -1
index 6119a41088381..65a2816142d96 100644
@@ -189,13 +189,14 @@ struct fm10k_q_vector {
 	struct fm10k_ring_container rx, tx;
 
 	struct napi_struct napi;
+	struct rcu_head rcu;	/* to avoid race with update stats on free */
+
 	cpumask_t affinity_mask;
 	char name[IFNAMSIZ + 9];
 
 #ifdef CONFIG_DEBUG_FS
 	struct dentry *dbg_q_vector;
 #endif /* CONFIG_DEBUG_FS */
-	struct rcu_head rcu;	/* to avoid race with update stats on free */
 
 	/* for dynamic allocation of rings associated with this q_vector */
 	struct fm10k_ring ring[] ____cacheline_internodealigned_in_smp;
drivers/net/ethernet/intel/i40e/i40e.h +1 -1
index 22ac8c48ca340..61590e92f3abc 100644
@@ -980,6 +980,7 @@ struct i40e_q_vector {
 	u16 reg_idx;		/* register index of the interrupt */
 
 	struct napi_struct napi;
+	struct rcu_head rcu;	/* to avoid race with update stats on free */
 
 	struct i40e_ring_container rx;
 	struct i40e_ring_container tx;
@@ -990,7 +991,6 @@ struct i40e_q_vector {
 	cpumask_t affinity_mask;
 	struct irq_affinity_notify affinity_notify;
 
-	struct rcu_head rcu;	/* to avoid race with update stats on free */
 	char name[I40E_INT_NAME_STR_LEN];
 	bool arm_wb_state;
 	bool in_busy_poll;
drivers/net/ethernet/intel/ixgbe/ixgbe.h +2 -1
index 2bf387e52e202..f49b99b175ef4 100644
@@ -498,9 +498,10 @@ struct ixgbe_q_vector {
 	struct ixgbe_ring_container rx, tx;
 
 	struct napi_struct napi;
+	struct rcu_head rcu;	/* to avoid race with update stats on free */
+
 	cpumask_t affinity_mask;
 	int numa_node;
-	struct rcu_head rcu;	/* to avoid race with update stats on free */
 	char name[IFNAMSIZ + 9];
 
 	/* for dynamic allocation of rings associated with this q_vector */
Signature · 1 line
2.39.5
author maintainer
#1 lore raw

PATCH 2/3 ASoC: amd: yc: Add DMI entries to support HP 15-fb1xxx

From: Adam Queler <[email protected]>

[ Upstream commit 949ddec3728f3a793a13c1c9003028b9b159aefc ]

This model requires an additional detection quirk to enable the internal microphone.

---

LLM Generated explanations, may be completely bogus:

**YES**

This commit should be backported to stable kernel trees. Here's my extensive analysis:

## Nature of the Change

This commit adds a simple DMI (Desktop Management Interface) quirk entry for the HP Victus Gaming Laptop 15-fb1xxx model to enable internal microphone support. The change is:
- Adding a new entry to the yc_acp_quirk_table[] array
- The entry matches DMI_BOARD_VENDOR "HP" and DMI_PRODUCT_NAME "Victus by HP Gaming Laptop 15-fb1xxx"
- Associates the match with &acp6x_card driver data

## Comparison with Similar Commits

All 5 similar commits with "Backport Status: YES" follow the exact same pattern:
1. They add DMI entries for HP laptops (HP OMEN, HP Victus) or other brands
2. They enable internal microphone functionality for specific hardware models
3. They are minimal, self-contained changes adding only DMI match entries
4. They fix hardware enablement issues that affect users

## Backport Suitability Analysis

**Reasons this qualifies for stable backporting:**

1. **Bug Fix**: This fixes a hardware functionality issue where the internal microphone doesn't work on HP Victus 15-fb1xxx laptops without this quirk

2. **User Impact**: Without this patch, users of this specific laptop model cannot use their internal microphone, which is a significant functionality regression

3. **Minimal Risk**: The change is:
- Extremely localized (only adds one DMI entry)
- Cannot affect other hardware (DMI matching is specific to this model)
- Uses the same pattern as dozens of other entries in the same table
- No algorithmic changes or new features

4. **Self-Contained**: The patch is completely self-contained with no dependencies on other changes

5. **Hardware Enablement**: This falls under the stable tree criteria for "hardware that is broken by design and needs a quirk"

6. **Precedent**: The git history shows numerous similar DMI quirk additions for this driver have been backported to stable

## Code Safety Analysis

The code change is safe because:
- It only adds a static data structure entry
- The DMI matching system is well-established and reliable
- The acp6x_card driver data is already used by many other entries
- No memory allocations, no new logic paths, no behavioral changes for existing hardware

This is a textbook example of a stable-appropriate patch: it fixes a real user-facing bug with minimal risk and follows established patterns.

 sound/soc/amd/yc/acp6x-mach.c | 7 +++++++
 1 file changed, 7 insertions(+)
sound/soc/amd/yc/acp6x-mach.c +7 -0
index 1f4c43bf817e4..fce918a089e37 100644
@@ -458,6 +458,13 @@ static const struct dmi_system_id yc_acp_quirk_table[] = {
 			DMI_MATCH(DMI_PRODUCT_NAME, "OMEN by HP Gaming Laptop 16z-n000"),
 		}
 	},
+	{
+		.driver_data = &acp6x_card,
+		.matches = {
+			DMI_MATCH(DMI_BOARD_VENDOR, "HP"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "Victus by HP Gaming Laptop 15-fb1xxx"),
+		}
+	},
 	{
 		.driver_data = &acp6x_card,
 		.matches = {
Signature · 1 line
2.39.5
author maintainer
#2 lore raw

PATCH 3/3 ASoC: Intel: fix SND_SOC_SOF dependencies

In reply to Sasha Levin

From: Arnd Bergmann <[email protected]>

[ Upstream commit e837b59f8b411b5baf5e3de7a5aea10b1c545a63 ]

It is currently possible to configure a kernel with all Intel SoC configs as loadable modules, but the board config as built-in. This causes a link failure in the reference to the snd_soc_sof.ko module:

x86_64-linux-ld: sound/soc/intel/boards/sof_rt5682.o: in function `sof_rt5682_hw_params':
sof_rt5682.c:(.text+0x1f9): undefined reference to `sof_dai_get_mclk'
x86_64-linux-ld: sof_rt5682.c:(.text+0x234): undefined reference to `sof_dai_get_bclk'
x86_64-linux-ld: sound/soc/intel/boards/sof_rt5682.o: in function `sof_rt5682_codec_init':
sof_rt5682.c:(.text+0x3e0): undefined reference to `sof_dai_get_mclk'
x86_64-linux-ld: sound/soc/intel/boards/sof_cs42l42.o: in function `sof_cs42l42_hw_params':
sof_cs42l42.c:(.text+0x2a): undefined reference to `sof_dai_get_bclk'
x86_64-linux-ld: sound/soc/intel/boards/sof_nau8825.o: in function `sof_nau8825_hw_params':
sof_nau8825.c:(.text+0x7f): undefined reference to `sof_dai_get_bclk'
x86_64-linux-ld: sound/soc/intel/boards/sof_da7219.o: in function `da7219_codec_init':
sof_da7219.c:(.text+0xbf): undefined reference to `sof_dai_get_mclk'
x86_64-linux-ld: sound/soc/intel/boards/sof_maxim_common.o: in function `max_98373_hw_params':
sof_maxim_common.c:(.text+0x6f9): undefined reference to `sof_dai_get_tdm_slots'
x86_64-linux-ld: sound/soc/intel/boards/sof_realtek_common.o: in function `rt1015_hw_params':
sof_realtek_common.c:(.text+0x54c): undefined reference to `sof_dai_get_bclk'
x86_64-linux-ld: sound/soc/intel/boards/sof_realtek_common.o: in function `rt1308_hw_params':
sof_realtek_common.c:(.text+0x702): undefined reference to `sof_dai_get_mclk'
x86_64-linux-ld: sound/soc/intel/boards/sof_cirrus_common.o: in function `cs35l41_hw_params':
sof_cirrus_common.c:(.text+0x2f): undefined reference to `sof_dai_get_bclk'

Add an optional dependency on SND_SOC_SOF_INTEL_COMMON, to ensure that whenever the SOF support is in a loadable module, none of the board code can be built-in.

This may be be a little heavy-handed, but I also don't see a reason why one would want the boards to be built-in but not the SoC, so it shouldn't actually cause any usability problems.

---

LLM Generated explanations, may be completely bogus:

Based on my analysis of the commit and its context, here is my assessment:

**YES** - This commit should be backported to stable kernel trees.

## Extensive Explanation:

### 1. **The commit fixes a critical build failure**
The commit addresses link errors that occur when Intel ASoC board drivers are built-in (=y) while the SOF core module is built as a loadable module (=m). The error messages show undefined references to SOF DAI functions:
- sof_dai_get_mclk
- sof_dai_get_bclk
- sof_dai_get_tdm_slots

These functions are exported from the snd-sof module (defined in sound/soc/sof/sof-audio.c), but when board drivers are built-in, they cannot reference symbols from loadable modules.

### 2. **The fix is minimal and contained**
The change is a single-line modification to the Kconfig dependency:

```diff
-if SND_SOC_INTEL_MACH
+if SND_SOC_INTEL_MACH && (SND_SOC_SOF_INTEL_COMMON ||
!SND_SOC_SOF_INTEL_COMMON)
```

This clever dependency ensures that when SND_SOC_SOF_INTEL_COMMON is built as a module (=m), the Intel machine drivers cannot be built-in (=y). The || !SND_SOC_SOF_INTEL_COMMON part allows the machine drivers to be enabled when SOF support is completely disabled.

### 3. **Similar to previous backported fixes**
Looking at the historical commits:
- Commit 8d0872f6239f (Similar Commit #1) fixed a similar link failure for sof-nau8825 and was backported (YES)
- Commit f4df4e4042b0 (Similar Commit #4) fixed a similar dependency issue with SOF IMX and was backported (YES)

Both of these commits addressed build/link failures caused by incorrect Kconfig dependencies, just like this commit.

### 4. **No risk of regression**
The change only affects build configuration and doesn't modify any runtime code. It prevents invalid configurations that would fail to build anyway. As the commit message states: "it shouldn't actually cause any usability problems" because there's no practical reason to have board drivers built-in while the SOF core is a module.

### 5. **Meets stable tree criteria**
- Fixes a real bug (build failure)
- Small and contained change
- No new features or architectural changes
- Similar fixes have been backported before
- Low risk of introducing new issues

The commit follows the pattern of other Kconfig dependency fixes in the Intel ASoC subsystem that have been successfully backported to stable trees.

 sound/soc/intel/boards/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
sound/soc/intel/boards/Kconfig +1 -1
index ca49cc49c378c..109efda305bfa 100644
@@ -11,7 +11,7 @@ menuconfig SND_SOC_INTEL_MACH
 	 kernel: saying N will just cause the configurator to skip all
 	 the questions about Intel ASoC machine drivers.
 
-if SND_SOC_INTEL_MACH
+if SND_SOC_INTEL_MACH && (SND_SOC_SOF_INTEL_COMMON || !SND_SOC_SOF_INTEL_COMMON)
 
 config SND_SOC_INTEL_USER_FRIENDLY_LONG_NAMES
 	bool "Use more user friendly long card names"
Signature · 1 line
2.39.5