[kernel/f16] Add patch to fix epoll lockdep warnings

Josh Boyer jwboyer at fedoraproject.org
Tue Aug 2 19:31:34 UTC 2011


commit ec0811cedbdac76b50ffd4ee48fd796c64b4bd57
Author: Josh Boyer <jwboyer at redhat.com>
Date:   Tue Aug 2 14:33:26 2011 -0400

    Add patch to fix epoll lockdep warnings
    
    make debug as we're still really early in the release

 config-generic                            |    8 +-
 config-nodebug                            |  102 +++++++++++++-------------
 config-x86_64-generic                     |    2 +-
 epoll-fix-spurious-lockdep-warnings.patch |  116 +++++++++++++++++++++++++++++
 kernel.spec                               |   10 ++-
 5 files changed, 180 insertions(+), 58 deletions(-)
---
diff --git a/config-generic b/config-generic
index 40ec3cf..a1ebab7 100644
--- a/config-generic
+++ b/config-generic
@@ -1472,12 +1472,12 @@ CONFIG_ATMEL=m
 CONFIG_B43=m
 CONFIG_B43_PCMCIA=y
 CONFIG_B43_SDIO=y
-# CONFIG_B43_DEBUG is not set
+CONFIG_B43_DEBUG=y
 CONFIG_B43_PHY_LP=y
 CONFIG_B43_PHY_N=y
 # CONFIG_B43_FORCE_PIO is not set
 CONFIG_B43LEGACY=m
-# CONFIG_B43LEGACY_DEBUG is not set
+CONFIG_B43LEGACY_DEBUG=y
 CONFIG_B43LEGACY_DMA=y
 CONFIG_B43LEGACY_PIO=y
 CONFIG_B43LEGACY_DMA_AND_PIO_MODE=y
@@ -4004,7 +4004,7 @@ CONFIG_IBMASR=m
 CONFIG_PM_DEBUG=y
 CONFIG_PM_TRACE=y
 # CONFIG_PM_VERBOSE is not set
-# CONFIG_PM_TEST_SUSPEND is not set
+CONFIG_PM_TEST_SUSPEND=y
 CONFIG_PM_RUNTIME=y
 
 CONFIG_CPU_FREQ=y
@@ -4483,7 +4483,7 @@ CONFIG_USB_ATMEL=m
 # CONFIG_RCU_CPU_STALL_DETECTOR is not set
 # CONFIG_DEBUG_BLOCK_EXT_DEVT is not set
 # CONFIG_FUNCTION_GRAPH_TRACER is not set
-# CONFIG_BOOT_TRACER is not set
+CONFIG_BOOT_TRACER=y
 CONFIG_EARLY_PRINTK_DBGP=y
 
 CONFIG_SECURITYFS=y
diff --git a/config-nodebug b/config-nodebug
index 62650aa..5730cd1 100644
--- a/config-nodebug
+++ b/config-nodebug
@@ -2,99 +2,99 @@ CONFIG_SND_VERBOSE_PRINTK=y
 CONFIG_SND_DEBUG=y
 CONFIG_SND_PCM_XRUN_DEBUG=y
 
-# CONFIG_DEBUG_MUTEXES is not set
-# CONFIG_DEBUG_RT_MUTEXES is not set
-# CONFIG_DEBUG_LOCK_ALLOC is not set
-# CONFIG_PROVE_LOCKING is not set
-# CONFIG_DEBUG_VM is not set
-# CONFIG_DEBUG_SPINLOCK is not set
-# CONFIG_PROVE_RCU is not set
+CONFIG_DEBUG_MUTEXES=y
+CONFIG_DEBUG_RT_MUTEXES=y
+CONFIG_DEBUG_LOCK_ALLOC=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_DEBUG_VM=y
+CONFIG_DEBUG_SPINLOCK=y
+CONFIG_PROVE_RCU=y
 # CONFIG_PROVE_RCU_REPEATEDLY is not set
-# CONFIG_DEBUG_PER_CPU_MAPS is not set
+CONFIG_DEBUG_PER_CPU_MAPS=y
 CONFIG_CPUMASK_OFFSTACK=y
 
-# CONFIG_CPU_NOTIFIER_ERROR_INJECT is not set
+CONFIG_CPU_NOTIFIER_ERROR_INJECT=m
 
-# CONFIG_FAULT_INJECTION is not set
-# CONFIG_FAILSLAB is not set
-# CONFIG_FAIL_PAGE_ALLOC is not set
-# CONFIG_FAIL_MAKE_REQUEST is not set
-# CONFIG_FAULT_INJECTION_DEBUG_FS is not set
-# CONFIG_FAULT_INJECTION_STACKTRACE_FILTER is not set
-# CONFIG_FAIL_IO_TIMEOUT is not set
+CONFIG_FAULT_INJECTION=y
+CONFIG_FAILSLAB=y
+CONFIG_FAIL_PAGE_ALLOC=y
+CONFIG_FAIL_MAKE_REQUEST=y
+CONFIG_FAULT_INJECTION_DEBUG_FS=y
+CONFIG_FAULT_INJECTION_STACKTRACE_FILTER=y
+CONFIG_FAIL_IO_TIMEOUT=y
 
-# CONFIG_SLUB_DEBUG_ON is not set
+CONFIG_SLUB_DEBUG_ON=y
 
-# CONFIG_LOCK_STAT is not set
+CONFIG_LOCK_STAT=y
 
-# CONFIG_DEBUG_STACK_USAGE is not set
+CONFIG_DEBUG_STACK_USAGE=y
 
-# CONFIG_ACPI_DEBUG is not set
+CONFIG_ACPI_DEBUG=y
 # CONFIG_ACPI_DEBUG_FUNC_TRACE is not set
 
-# CONFIG_DEBUG_SG is not set
+CONFIG_DEBUG_SG=y
 
 # CONFIG_DEBUG_PAGEALLOC is not set
 
-# CONFIG_DEBUG_WRITECOUNT is not set
-# CONFIG_DEBUG_OBJECTS is not set
+CONFIG_DEBUG_WRITECOUNT=y
+CONFIG_DEBUG_OBJECTS=y
 # CONFIG_DEBUG_OBJECTS_SELFTEST is not set
-# CONFIG_DEBUG_OBJECTS_FREE is not set
-# CONFIG_DEBUG_OBJECTS_TIMERS is not set
-# CONFIG_DEBUG_OBJECTS_RCU_HEAD is not set
+CONFIG_DEBUG_OBJECTS_FREE=y
+CONFIG_DEBUG_OBJECTS_TIMERS=y
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=y
 CONFIG_DEBUG_OBJECTS_ENABLE_DEFAULT=1
 
-# CONFIG_X86_PTDUMP is not set
+CONFIG_X86_PTDUMP=y
 
-# CONFIG_CAN_DEBUG_DEVICES is not set
+CONFIG_CAN_DEBUG_DEVICES=y
 
-# CONFIG_MODULE_FORCE_UNLOAD is not set
+CONFIG_MODULE_FORCE_UNLOAD=y
 
-# CONFIG_SYSCTL_SYSCALL_CHECK is not set
+CONFIG_SYSCTL_SYSCALL_CHECK=y
 
-# CONFIG_DEBUG_NOTIFIERS is not set
+CONFIG_DEBUG_NOTIFIERS=y
 
-# CONFIG_DMA_API_DEBUG is not set
+CONFIG_DMA_API_DEBUG=y
 
-# CONFIG_MMIOTRACE is not set
+CONFIG_MMIOTRACE=y
 
-# CONFIG_DEBUG_CREDENTIALS is not set
+CONFIG_DEBUG_CREDENTIALS=y
 
 # off in both production debug and nodebug builds,
 #  on in rawhide nodebug builds
-# CONFIG_DEBUG_FORCE_WEAK_PER_CPU is not set
+CONFIG_DEBUG_FORCE_WEAK_PER_CPU=y
 
-# CONFIG_EXT4_DEBUG is not set
+CONFIG_EXT4_DEBUG=y
 
-# CONFIG_DEBUG_PERF_USE_VMALLOC is not set
+CONFIG_DEBUG_PERF_USE_VMALLOC=y
 
-# CONFIG_JBD2_DEBUG is not set
+CONFIG_JBD2_DEBUG=y
 
-# CONFIG_DEBUG_CFQ_IOSCHED is not set
+CONFIG_DEBUG_CFQ_IOSCHED=y
 
-# CONFIG_DRBD_FAULT_INJECTION is not set
+CONFIG_DRBD_FAULT_INJECTION=y
 
-# CONFIG_ATH_DEBUG is not set
-# CONFIG_CARL9170_DEBUGFS is not set
-# CONFIG_IWLWIFI_DEVICE_TRACING is not set
+CONFIG_ATH_DEBUG=y
+CONFIG_CARL9170_DEBUGFS=y
+CONFIG_IWLWIFI_DEVICE_TRACING=y
 
-# CONFIG_DEBUG_OBJECTS_WORK is not set
+CONFIG_DEBUG_OBJECTS_WORK=y
 # CONFIG_DEBUG_STRICT_USER_COPY_CHECKS is not set
 
-# CONFIG_DMADEVICES_DEBUG is not set
-# CONFIG_DMADEVICES_VDEBUG is not set
+CONFIG_DMADEVICES_DEBUG=y
+CONFIG_DMADEVICES_VDEBUG=y
 
 CONFIG_PM_ADVANCED_DEBUG=y
 
-# CONFIG_CEPH_LIB_PRETTYDEBUG is not set
-# CONFIG_QUOTA_DEBUG is not set
+CONFIG_CEPH_LIB_PRETTYDEBUG=y
+CONFIG_QUOTA_DEBUG=y
 
 CONFIG_PCI_DEFAULT_USE_CRS=y
 
 CONFIG_KGDB_KDB=y
 CONFIG_KDB_KEYBOARD=y
 
-# CONFIG_DEBUG_OBJECTS_PERCPU_COUNTER is not set
-# CONFIG_TEST_LIST_SORT is not set
+CONFIG_DEBUG_OBJECTS_PERCPU_COUNTER=y
+CONFIG_TEST_LIST_SORT=y
 
-# CONFIG_DEBUG_SET_MODULE_RONX is not set
+CONFIG_DEBUG_SET_MODULE_RONX=y
diff --git a/config-x86_64-generic b/config-x86_64-generic
index 51879a7..f4c3c30 100644
--- a/config-x86_64-generic
+++ b/config-x86_64-generic
@@ -16,7 +16,7 @@ CONFIG_K8_NUMA=y
 CONFIG_AMD_NUMA=y
 CONFIG_X86_64_ACPI_NUMA=y
 # CONFIG_NUMA_EMU is not set
-CONFIG_NR_CPUS=256
+CONFIG_NR_CPUS=512
 CONFIG_X86_POWERNOW_K8=m
 CONFIG_X86_P4_CLOCKMOD=m
 CONFIG_IA32_EMULATION=y
diff --git a/epoll-fix-spurious-lockdep-warnings.patch b/epoll-fix-spurious-lockdep-warnings.patch
new file mode 100644
index 0000000..131796a
--- /dev/null
+++ b/epoll-fix-spurious-lockdep-warnings.patch
@@ -0,0 +1,116 @@
+epoll can acquire multiple ep->mutex on multiple "struct eventpoll"s
+at once in the case where one epoll fd is monitoring another epoll
+fd. This is perfectly OK, since we're careful about the lock ordering,
+but causes spurious lockdep warnings. Annotate the recursion using
+mutex_lock_nested, and add a comment explaining the nesting rules for
+good measure.
+
+Reported-by: Paul Bolle <pebolle at tiscali.nl>
+Signed-off-by: Nelson Elhage <nelhage at nelhage.com>
+---
+ I've tested this on a synthetic epoll test case, that just adds e1 to
+ e2 and then does an epoll_wait(). I verified that it caused lockdep
+ problems on 3.0 and that this patch fixed it, but I haven't done more
+ extensive testing. Paul, are you able to test systemd against this?
+
+ fs/eventpoll.c |   25 ++++++++++++++++++-------
+ 1 files changed, 18 insertions(+), 7 deletions(-)
+
+diff --git a/fs/eventpoll.c b/fs/eventpoll.c
+index f9cfd16..0cb7bc6 100644
+--- a/fs/eventpoll.c
++++ b/fs/eventpoll.c
+@@ -76,6 +76,15 @@
+  * Events that require holding "epmutex" are very rare, while for
+  * normal operations the epoll private "ep->mtx" will guarantee
+  * a better scalability.
++ * It is possible to acquire multiple "ep->mtx"es at once in the case
++ * when one epoll fd is added to another. In this case, we always
++ * acquire the locks in the order of nesting (i.e. after epoll_ctl(e1,
++ * EPOLL_CTL_ADD, e2), e1->mtx will always be acquired before
++ * e2->mtx). Since we disallow cycles of epoll file descriptors, this
++ * ensures that the mutexes are well-ordered. In order to communicate
++ * this nesting to lockdep, when walking a tree of epoll file
++ * descriptors, we use the current recursion depth as the lockdep
++ * subkey.
+  */
+ 
+ /* Epoll private bits inside the event mask */
+@@ -464,13 +473,15 @@ static void ep_unregister_pollwait(struct eventpoll *ep, struct epitem *epi)
+  * @ep: Pointer to the epoll private data structure.
+  * @sproc: Pointer to the scan callback.
+  * @priv: Private opaque data passed to the @sproc callback.
++ * @depth: The current depth of recursive f_op->poll calls.
+  *
+  * Returns: The same integer error code returned by the @sproc callback.
+  */
+ static int ep_scan_ready_list(struct eventpoll *ep,
+ 			      int (*sproc)(struct eventpoll *,
+ 					   struct list_head *, void *),
+-			      void *priv)
++			      void *priv,
++			      int depth)
+ {
+ 	int error, pwake = 0;
+ 	unsigned long flags;
+@@ -481,7 +492,7 @@ static int ep_scan_ready_list(struct eventpoll *ep,
+ 	 * We need to lock this because we could be hit by
+ 	 * eventpoll_release_file() and epoll_ctl().
+ 	 */
+-	mutex_lock(&ep->mtx);
++	mutex_lock_nested(&ep->mtx, depth);
+ 
+ 	/*
+ 	 * Steal the ready list, and re-init the original one to the
+@@ -670,7 +681,7 @@ static int ep_read_events_proc(struct eventpoll *ep, struct list_head *head,
+ 
+ static int ep_poll_readyevents_proc(void *priv, void *cookie, int call_nests)
+ {
+-	return ep_scan_ready_list(priv, ep_read_events_proc, NULL);
++	return ep_scan_ready_list(priv, ep_read_events_proc, NULL, call_nests + 1);
+ }
+ 
+ static unsigned int ep_eventpoll_poll(struct file *file, poll_table *wait)
+@@ -737,7 +748,7 @@ void eventpoll_release_file(struct file *file)
+ 
+ 		ep = epi->ep;
+ 		list_del_init(&epi->fllink);
+-		mutex_lock(&ep->mtx);
++		mutex_lock_nested(&ep->mtx, 0);
+ 		ep_remove(ep, epi);
+ 		mutex_unlock(&ep->mtx);
+ 	}
+@@ -1134,7 +1145,7 @@ static int ep_send_events(struct eventpoll *ep,
+ 	esed.maxevents = maxevents;
+ 	esed.events = events;
+ 
+-	return ep_scan_ready_list(ep, ep_send_events_proc, &esed);
++	return ep_scan_ready_list(ep, ep_send_events_proc, &esed, 0);
+ }
+ 
+ static inline struct timespec ep_set_mstimeout(long ms)
+@@ -1267,7 +1278,7 @@ static int ep_loop_check_proc(void *priv, void *cookie, int call_nests)
+ 	struct rb_node *rbp;
+ 	struct epitem *epi;
+ 
+-	mutex_lock(&ep->mtx);
++	mutex_lock_nested(&ep->mtx, call_nests + 1);
+ 	for (rbp = rb_first(&ep->rbr); rbp; rbp = rb_next(rbp)) {
+ 		epi = rb_entry(rbp, struct epitem, rbn);
+ 		if (unlikely(is_file_epoll(epi->ffd.file))) {
+@@ -1409,7 +1420,7 @@ SYSCALL_DEFINE4(epoll_ctl, int, epfd, int, op, int, fd,
+ 	}
+ 
+ 
+-	mutex_lock(&ep->mtx);
++	mutex_lock_nested(&ep->mtx, 0);
+ 
+ 	/*
+ 	 * Try to lookup the file inside our RB tree, Since we grabbed "mtx"
+-- 
+1.7.4.1
+
+--
+To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
+the body of a message to majordomo at vger.kernel.org
+More majordomo info at  http://vger.kernel.org/majordomo-info.html
\ No newline at end of file
diff --git a/kernel.spec b/kernel.spec
index be59935..7e8210d 100644
--- a/kernel.spec
+++ b/kernel.spec
@@ -131,7 +131,7 @@ Summary: The Linux kernel
 %define doc_build_fail true
 %endif
 
-%define rawhide_skip_docs 0
+%define rawhide_skip_docs 1
 %if 0%{?rawhide_skip_docs}
 %define with_doc 0
 %define doc_build_fail true
@@ -154,7 +154,7 @@ Summary: The Linux kernel
 # Set debugbuildsenabled to 1 for production (build separate debug kernels)
 #  and 0 for rawhide (all kernels are debug kernels).
 # See also 'make debug' and 'make release'.
-%define debugbuildsenabled 1
+%define debugbuildsenabled 0
 
 # Want to build a vanilla kernel build without any non-upstream patches?
 %define with_vanilla %{?_with_vanilla: 1} %{?!_with_vanilla: 0}
@@ -696,6 +696,7 @@ Patch12205: runtime_pm_fixups.patch
 Patch12303: dmar-disable-when-ricoh-multifunction.patch
 
 Patch13000: fix-scsi_dispatch_cmd.patch
+Patch13001: epoll-fix-spurious-lockdep-warnings.patch
 
 Patch20000: utrace.patch
 
@@ -1280,6 +1281,7 @@ ApplyPatch udlfb-bind-framebuffer-to-interface.patch
 ApplyPatch dmar-disable-when-ricoh-multifunction.patch
 
 ApplyPatch fix-scsi_dispatch_cmd.patch
+ApplyPatch epoll-fix-spurious-lockdep-warnings.patch
 
 # utrace.
 ApplyPatch utrace.patch
@@ -1894,6 +1896,10 @@ fi
 
 %changelog
 * Tue Aug 02 2011 Josh Boyer <jwboyer at redhat.com>
+- Fix epoll recursive lockdep warnings (rhbz 722472)
+- Turn debug builds back on
+
+* Tue Aug 02 2011 Josh Boyer <jwboyer at redhat.com>
 - Add change from Yanko Kaneti to get the rt2x00 drivers in modules.networking
   (rhbz 708314)
 


More information about the scm-commits mailing list