diff mbox series

[net-next,04/24] net: Use nested-BH locking for napi_alloc_cache.

Message ID 20231215171020.687342-5-bigeasy@linutronix.de
State New
Headers
Series locking: Introduce nested-BH locking. |

Commit Message

Sebastian Andrzej Siewior Dec. 15, 2023, 5:07 p.m. UTC 
  napi_alloc_cache is a per-CPU variable and relies on disabled BH for its
locking. Without per-CPU locking in local_bh_disable() on PREEMPT_RT
this data structure requires explicit locking.

Add a local_lock_t to the data structure and use local_lock_nested_bh()
for locking. This change adds only lockdep coverage and does not alter
the functional behaviour for !PREEMPT_RT.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 net/core/skbuff.c | 57 +++++++++++++++++++++++++++--------------------
 1 file changed, 33 insertions(+), 24 deletions(-)

Comments

kernel test robot Dec. 16, 2023, 4:43 a.m. UTC | #1 
Hi Sebastian,

kernel test robot noticed the following build warnings:

[auto build test WARNING on net-next/main]

url:    https://github.com/intel-lab-lkp/linux/commits/Sebastian-Andrzej-Siewior/locking-local_lock-Introduce-guard-definition-for-local_lock/20231216-011911
base:   net-next/main
patch link:    https://lore.kernel.org/r/20231215171020.687342-5-bigeasy%40linutronix.de
patch subject: [PATCH net-next 04/24] net: Use nested-BH locking for napi_alloc_cache.
config: x86_64-randconfig-121-20231216 (https://download.01.org/0day-ci/archive/20231216/202312161210.q8xdLxsl-lkp@intel.com/config)
compiler: clang version 16.0.4 (https://github.com/llvm/llvm-project.git ae42196bc493ffe877a7e3dff8be32035dea4d07)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20231216/202312161210.q8xdLxsl-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202312161210.q8xdLxsl-lkp@intel.com/

sparse warnings: (new ones prefixed by >>)
>> net/core/skbuff.c:302:38: sparse: sparse: incorrect type in argument 1 (different address spaces) @@     expected struct local_lock_t [usertype] *l @@     got struct local_lock_t [noderef] __percpu * @@
   net/core/skbuff.c:302:38: sparse:     expected struct local_lock_t [usertype] *l
   net/core/skbuff.c:302:38: sparse:     got struct local_lock_t [noderef] __percpu *
   net/core/skbuff.c:331:38: sparse: sparse: incorrect type in argument 1 (different address spaces) @@     expected struct local_lock_t [usertype] *l @@     got struct local_lock_t [noderef] __percpu * @@
   net/core/skbuff.c:331:38: sparse:     expected struct local_lock_t [usertype] *l
   net/core/skbuff.c:331:38: sparse:     got struct local_lock_t [noderef] __percpu *
   net/core/skbuff.c:734:17: sparse: sparse: incorrect type in argument 1 (different address spaces) @@     expected struct local_lock_t [usertype] *l @@     got struct local_lock_t [noderef] __percpu * @@
   net/core/skbuff.c:734:17: sparse:     expected struct local_lock_t [usertype] *l
   net/core/skbuff.c:734:17: sparse:     got struct local_lock_t [noderef] __percpu *
   net/core/skbuff.c:806:9: sparse: sparse: incorrect type in argument 1 (different address spaces) @@     expected struct local_lock_t [usertype] *l @@     got struct local_lock_t [noderef] __percpu * @@
   net/core/skbuff.c:806:9: sparse:     expected struct local_lock_t [usertype] *l
   net/core/skbuff.c:806:9: sparse:     got struct local_lock_t [noderef] __percpu *
   net/core/skbuff.c:1317:38: sparse: sparse: incorrect type in argument 1 (different address spaces) @@     expected struct local_lock_t [usertype] *l @@     got struct local_lock_t [noderef] __percpu * @@
   net/core/skbuff.c:1317:38: sparse:     expected struct local_lock_t [usertype] *l
   net/core/skbuff.c:1317:38: sparse:     got struct local_lock_t [noderef] __percpu *
   net/core/skbuff.c: note: in included file (through include/linux/mmzone.h, include/linux/gfp.h, include/linux/umh.h, include/linux/kmod.h, ...):
   include/linux/page-flags.h:242:46: sparse: sparse: self-comparison always evaluates to false
   net/core/skbuff.c: note: in included file (through include/linux/mmzone.h, include/linux/gfp.h, include/linux/umh.h, include/linux/kmod.h, ...):
>> include/linux/local_lock.h:71:1: sparse: sparse: incorrect type in initializer (different address spaces) @@     expected void const [noderef] __percpu *__vpp_verify @@     got struct local_lock_t [usertype] * @@
   include/linux/local_lock.h:71:1: sparse:     expected void const [noderef] __percpu *__vpp_verify
   include/linux/local_lock.h:71:1: sparse:     got struct local_lock_t [usertype] *
>> include/linux/local_lock.h:71:1: sparse: sparse: incorrect type in initializer (different address spaces) @@     expected void const [noderef] __percpu *__vpp_verify @@     got struct local_lock_t [usertype] * @@
   include/linux/local_lock.h:71:1: sparse:     expected void const [noderef] __percpu *__vpp_verify
   include/linux/local_lock.h:71:1: sparse:     got struct local_lock_t [usertype] *
>> include/linux/local_lock.h:71:1: sparse: sparse: incorrect type in initializer (different address spaces) @@     expected void const [noderef] __percpu *__vpp_verify @@     got struct local_lock_t [usertype] * @@
   include/linux/local_lock.h:71:1: sparse:     expected void const [noderef] __percpu *__vpp_verify
   include/linux/local_lock.h:71:1: sparse:     got struct local_lock_t [usertype] *
   net/core/skbuff.c: note: in included file (through include/net/net_namespace.h, include/linux/inet.h):
   include/linux/skbuff.h:2715:28: sparse: sparse: self-comparison always evaluates to false
   net/core/skbuff.c: note: in included file (through include/linux/skbuff.h, include/net/net_namespace.h, include/linux/inet.h):
   include/net/checksum.h:33:39: sparse: sparse: incorrect type in argument 3 (different base types) @@     expected restricted __wsum [usertype] sum @@     got unsigned int @@
   include/net/checksum.h:33:39: sparse:     expected restricted __wsum [usertype] sum
   include/net/checksum.h:33:39: sparse:     got unsigned int

vim +302 net/core/skbuff.c

   296	
   297	void *__napi_alloc_frag_align(unsigned int fragsz, unsigned int align_mask)
   298	{
   299		struct napi_alloc_cache *nc = this_cpu_ptr(&napi_alloc_cache);
   300	
   301		fragsz = SKB_DATA_ALIGN(fragsz);
 > 302		guard(local_lock_nested_bh)(&napi_alloc_cache.bh_lock);
   303	
   304		return page_frag_alloc_align(&nc->page, fragsz, GFP_ATOMIC, align_mask);
   305	}
   306	EXPORT_SYMBOL(__napi_alloc_frag_align);
   307
Sebastian Andrzej Siewior Jan. 12, 2024, 10:58 a.m. UTC | #2 
On 2023-12-16 12:43:23 [+0800], kernel test robot wrote:
> Hi Sebastian,
Hi,

> >> net/core/skbuff.c:302:38: sparse: sparse: incorrect type in argument 1 (different address spaces) @@     expected struct local_lock_t [usertype] *l @@     got struct local_lock_t [noderef] __percpu * @@

updated the guard definition for that.
Thanks for the report.

Sebastian
diff mbox series

Patch

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index de9397e45718a..9c3073dcc80f1 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -265,6 +265,7 @@  static void *page_frag_alloc_1k(struct page_frag_1k *nc, gfp_t gfp_mask)
 #endif
 
 struct napi_alloc_cache {
+	local_lock_t bh_lock;
 	struct page_frag_cache page;
 	struct page_frag_1k page_small;
 	unsigned int skb_count;
@@ -272,7 +273,9 @@  struct napi_alloc_cache {
 };
 
 static DEFINE_PER_CPU(struct page_frag_cache, netdev_alloc_cache);
-static DEFINE_PER_CPU(struct napi_alloc_cache, napi_alloc_cache);
+static DEFINE_PER_CPU(struct napi_alloc_cache, napi_alloc_cache) = {
+	.bh_lock = INIT_LOCAL_LOCK(bh_lock),
+};
 
 /* Double check that napi_get_frags() allocates skbs with
  * skb->head being backed by slab, not a page fragment.
@@ -296,6 +299,7 @@  void *__napi_alloc_frag_align(unsigned int fragsz, unsigned int align_mask)
 	struct napi_alloc_cache *nc = this_cpu_ptr(&napi_alloc_cache);
 
 	fragsz = SKB_DATA_ALIGN(fragsz);
+	guard(local_lock_nested_bh)(&napi_alloc_cache.bh_lock);
 
 	return page_frag_alloc_align(&nc->page, fragsz, GFP_ATOMIC, align_mask);
 }
@@ -324,6 +328,7 @@  static struct sk_buff *napi_skb_cache_get(void)
 	struct napi_alloc_cache *nc = this_cpu_ptr(&napi_alloc_cache);
 	struct sk_buff *skb;
 
+	guard(local_lock_nested_bh)(&napi_alloc_cache.bh_lock);
 	if (unlikely(!nc->skb_count)) {
 		nc->skb_count = kmem_cache_alloc_bulk(skbuff_cache,
 						      GFP_ATOMIC,
@@ -726,9 +731,11 @@  struct sk_buff *__netdev_alloc_skb(struct net_device *dev, unsigned int len,
 		pfmemalloc = nc->pfmemalloc;
 	} else {
 		local_bh_disable();
-		nc = this_cpu_ptr(&napi_alloc_cache.page);
-		data = page_frag_alloc(nc, len, gfp_mask);
-		pfmemalloc = nc->pfmemalloc;
+		scoped_guard(local_lock_nested_bh, &napi_alloc_cache.bh_lock) {
+			nc = this_cpu_ptr(&napi_alloc_cache.page);
+			data = page_frag_alloc(nc, len, gfp_mask);
+			pfmemalloc = nc->pfmemalloc;
+		}
 		local_bh_enable();
 	}
 
@@ -793,31 +800,32 @@  struct sk_buff *__napi_alloc_skb(struct napi_struct *napi, unsigned int len,
 		goto skb_success;
 	}
 
-	nc = this_cpu_ptr(&napi_alloc_cache);
-
 	if (sk_memalloc_socks())
 		gfp_mask |= __GFP_MEMALLOC;
 
-	if (NAPI_HAS_SMALL_PAGE_FRAG && len <= SKB_WITH_OVERHEAD(1024)) {
-		/* we are artificially inflating the allocation size, but
-		 * that is not as bad as it may look like, as:
-		 * - 'len' less than GRO_MAX_HEAD makes little sense
-		 * - On most systems, larger 'len' values lead to fragment
-		 *   size above 512 bytes
-		 * - kmalloc would use the kmalloc-1k slab for such values
-		 * - Builds with smaller GRO_MAX_HEAD will very likely do
-		 *   little networking, as that implies no WiFi and no
-		 *   tunnels support, and 32 bits arches.
-		 */
-		len = SZ_1K;
+	scoped_guard(local_lock_nested_bh, &napi_alloc_cache.bh_lock) {
+		nc = this_cpu_ptr(&napi_alloc_cache);
+		if (NAPI_HAS_SMALL_PAGE_FRAG && len <= SKB_WITH_OVERHEAD(1024)) {
+			/* we are artificially inflating the allocation size, but
+			 * that is not as bad as it may look like, as:
+			 * - 'len' less than GRO_MAX_HEAD makes little sense
+			 * - On most systems, larger 'len' values lead to fragment
+			 *   size above 512 bytes
+			 * - kmalloc would use the kmalloc-1k slab for such values
+			 * - Builds with smaller GRO_MAX_HEAD will very likely do
+			 *   little networking, as that implies no WiFi and no
+			 *   tunnels support, and 32 bits arches.
+			 */
+			len = SZ_1K;
 
-		data = page_frag_alloc_1k(&nc->page_small, gfp_mask);
-		pfmemalloc = NAPI_SMALL_PAGE_PFMEMALLOC(nc->page_small);
-	} else {
-		len = SKB_HEAD_ALIGN(len);
+			data = page_frag_alloc_1k(&nc->page_small, gfp_mask);
+			pfmemalloc = NAPI_SMALL_PAGE_PFMEMALLOC(nc->page_small);
+		} else {
+			len = SKB_HEAD_ALIGN(len);
 
-		data = page_frag_alloc(&nc->page, len, gfp_mask);
-		pfmemalloc = nc->page.pfmemalloc;
+			data = page_frag_alloc(&nc->page, len, gfp_mask);
+			pfmemalloc = nc->page.pfmemalloc;
+		}
 	}
 
 	if (unlikely(!data))
@@ -1306,6 +1314,7 @@  static void napi_skb_cache_put(struct sk_buff *skb)
 	struct napi_alloc_cache *nc = this_cpu_ptr(&napi_alloc_cache);
 	u32 i;
 
+	guard(local_lock_nested_bh)(&napi_alloc_cache.bh_lock);
 	kasan_poison_object_data(skbuff_cache, skb);
 	nc->skb_cache[nc->skb_count++] = skb;