From: Sasha Levin <sashal@kernel.org>
To: patches@lists.linux.dev, stable@vger.kernel.org
Cc: Arnd Bergmann <arnd@arndb.de>,
"David S . Miller" <davem@davemloft.net>,
Aleksandr Loktionov <aleksandr.loktionov@intel.com>,
Alexander Lobakin <aleksander.lobakin@intel.com>,
Sunitha Mekala <sunithax.d.mekala@intel.com>,
Tony Nguyen <anthony.l.nguyen@intel.com>,
Sasha Levin <sashal@kernel.org>,
przemyslaw.kitszel@intel.com, intel-wired-lan@lists.osuosl.org
Subject: [PATCH AUTOSEL 6.1 1/3] ethernet: intel: fix building with large NR_CPUS
Date: Tue, 22 Jul 2025 20:58:55 -0400 [thread overview]
Message-ID: <20250723005857.1023488-1-sashal@kernel.org> (raw)
From: Arnd Bergmann <arnd@arndb.de>
[ 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.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: David S. Miller <davem@davemloft.net>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Tested-by: Sunitha Mekala <sunithax.d.mekala@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
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(-)
diff --git a/drivers/net/ethernet/intel/fm10k/fm10k.h b/drivers/net/ethernet/intel/fm10k/fm10k.h
index 6119a41088381..65a2816142d96 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k.h
+++ b/drivers/net/ethernet/intel/fm10k/fm10k.h
@@ -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;
diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index 22ac8c48ca340..61590e92f3abc 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -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;
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe.h b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
index 2bf387e52e202..f49b99b175ef4 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
@@ -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 */
--
2.39.5
next reply other threads:[~2025-07-23 0:59 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-07-23 0:58 Sasha Levin [this message]
2025-07-23 0:58 ` [PATCH AUTOSEL 6.1 2/3] ASoC: amd: yc: Add DMI entries to support HP 15-fb1xxx Sasha Levin
2025-07-23 0:58 ` [PATCH AUTOSEL 6.1 3/3] ASoC: Intel: fix SND_SOC_SOF dependencies Sasha Levin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250723005857.1023488-1-sashal@kernel.org \
--to=sashal@kernel.org \
--cc=aleksander.lobakin@intel.com \
--cc=aleksandr.loktionov@intel.com \
--cc=anthony.l.nguyen@intel.com \
--cc=arnd@arndb.de \
--cc=davem@davemloft.net \
--cc=intel-wired-lan@lists.osuosl.org \
--cc=patches@lists.linux.dev \
--cc=przemyslaw.kitszel@intel.com \
--cc=stable@vger.kernel.org \
--cc=sunithax.d.mekala@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).