HWPOISON updates for 2.6.32-rc5 for review

HWPOISON updates for 2.6.32-rc5 for review

[Andi Kleen]

Here are the hwpoison bug fixes queued up for 2.6.32-rc5. In particularly
they fix the page flags races that were discussed earlier and a couple of
random bug fixes. 

Please review.

-Andi

[PATCH] HWPOISON: Clean up PR_MCE_KILL interface

[Andi Kleen]

From: Andi Kleen <ak@linux.intel.com>

While writing the manpage I noticed some shortcomings in the
current interface.

- Define symbolic names for all the different values
- Boundary check the kill mode values
- For symmetry add a get interface too. This allows library
code to get/set the current state.
- For consistency define a PR_MCE_KILL_DEFAULT value

Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
 include/linux/prctl.h |   12 ++++++++++++
 kernel/sys.c          |   23 ++++++++++++++++++-----
 2 files changed, 30 insertions(+), 5 deletions(-)

diff --git a/include/linux/prctl.h b/include/linux/prctl.h
index 9311505..a3baeb2 100644
--- a/include/linux/prctl.h
+++ b/include/linux/prctl.h
@@ -88,6 +88,18 @@
 #define PR_TASK_PERF_EVENTS_DISABLE		31
 #define PR_TASK_PERF_EVENTS_ENABLE		32
 
+/*
+ * Set early/late kill mode for hwpoison memory corruption.
+ * This influences when the process gets killed on a memory corruption.
+ */
 #define PR_MCE_KILL	33
+# define PR_MCE_KILL_CLEAR   0
+# define PR_MCE_KILL_SET     1
+
+# define PR_MCE_KILL_LATE    0
+# define PR_MCE_KILL_EARLY   1
+# define PR_MCE_KILL_DEFAULT 2
+
+#define PR_MCE_KILL_GET 34
 
 #endif /* _LINUX_PRCTL_H */
diff --git a/kernel/sys.c b/kernel/sys.c
index 255475d..f6afe07 100644
--- a/kernel/sys.c
+++ b/kernel/sys.c
@@ -1546,24 +1546,37 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3,
 			if (arg4 | arg5)
 				return -EINVAL;
 			switch (arg2) {
-			case 0:
+			case PR_MCE_KILL_CLEAR:
 				if (arg3 != 0)
 					return -EINVAL;
 				current->flags &= ~PF_MCE_PROCESS;
 				break;
-			case 1:
+			case PR_MCE_KILL_SET:
 				current->flags |= PF_MCE_PROCESS;
-				if (arg3 != 0)
+				if (arg3 == PR_MCE_KILL_EARLY)
 					current->flags |= PF_MCE_EARLY;
-				else
+				else if (arg3 == PR_MCE_KILL_LATE)
 					current->flags &= ~PF_MCE_EARLY;
+				else if (arg3 == PR_MCE_KILL_DEFAULT)
+					current->flags &=
+						~(PF_MCE_EARLY|PF_MCE_PROCESS);
+				else
+					return -EINVAL;
 				break;
 			default:
 				return -EINVAL;
 			}
 			error = 0;
 			break;
-
+		case PR_MCE_KILL_GET:
+			if (arg2 | arg3 | arg4 | arg5)
+				return -EINVAL;
+			if (current->flags & PF_MCE_PROCESS)
+				error = (current->flags & PF_MCE_EARLY) ?
+					PR_MCE_KILL_EARLY : PR_MCE_KILL_LATE;
+			else
+				error = PR_MCE_KILL_DEFAULT;
+			break;
 		default:
 			error = -EINVAL;
 			break;
-- 
1.6.0.2

[PATCH] HWPOISON: Add brief hwpoison description to Documentation

[Andi Kleen]

From: Andi Kleen <ak@linux.intel.com>

Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
 Documentation/vm/hwpoison.txt |  136 +++++++++++++++++++++++++++++++++++++++++
 1 files changed, 136 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/vm/hwpoison.txt

diff --git a/Documentation/vm/hwpoison.txt b/Documentation/vm/hwpoison.txt
new file mode 100644
index 0000000..3ffadf8
--- /dev/null
+++ b/Documentation/vm/hwpoison.txt
@@ -0,0 +1,136 @@
+What is hwpoison?
+
+Upcoming Intel CPUs have support for recovering from some memory errors
+(``MCA recovery''). This requires the OS to declare a page "poisoned",
+kill the processes associated with it and avoid using it in the future.
+
+This patchkit implements the necessary infrastructure in the VM.
+
+To quote the overview comment:
+
+ * High level machine check handler. Handles pages reported by the
+ * hardware as being corrupted usually due to a 2bit ECC memory or cache
+ * failure.
+ *
+ * This focusses on pages detected as corrupted in the background.
+ * When the current CPU tries to consume corruption the currently
+ * running process can just be killed directly instead. This implies
+ * that if the error cannot be handled for some reason it's safe to
+ * just ignore it because no corruption has been consumed yet. Instead
+ * when that happens another machine check will happen.
+ *
+ * Handles page cache pages in various states. The tricky part
+ * here is that we can access any page asynchronous to other VM
+ * users, because memory failures could happen anytime and anywhere,
+ * possibly violating some of their assumptions. This is why this code
+ * has to be extremely careful. Generally it tries to use normal locking
+ * rules, as in get the standard locks, even if that means the
+ * error handling takes potentially a long time.
+ *
+ * Some of the operations here are somewhat inefficient and have non
+ * linear algorithmic complexity, because the data structures have not
+ * been optimized for this case. This is in particular the case
+ * for the mapping from a vma to a process. Since this case is expected
+ * to be rare we hope we can get away with this.
+
+The code consists of a the high level handler in mm/memory-failure.c,
+a new page poison bit and various checks in the VM to handle poisoned
+pages.
+
+The main target right now is KVM guests, but it works for all kinds
+of applications. KVM support requires a recent qemu-kvm release.
+
+For the KVM use there was need for a new signal type so that
+KVM can inject the machine check into the guest with the proper
+address. This in theory allows other applications to handle
+memory failures too. The expection is that near all applications
+won't do that, but some very specialized ones might.
+
+---
+
+There are two (actually three) modi memory failure recovery can be in:
+
+vm.memory_failure_recovery sysctl set to zero:
+	All memory failures cause a panic. Do not attempt recovery.
+	(on x86 this can be also affected by the tolerant level of the
+	MCE subsystem)
+
+early kill
+	(can be controlled globally and per process)
+	Send SIGBUS to the application as soon as the error is detected
+	This allows applications who can process memory errors in a gentle
+	way (e.g. drop affected object)
+	This is the mode used by KVM qemu.
+
+late kill
+	Send SIGBUS when the application runs into the corrupted page.
+	This is best for memory error unaware applications and default
+	Note some pages are always handled as late kill.
+
+---
+
+User control:
+
+vm.memory_failure_recovery
+	See sysctl.txt
+
+vm.memory_failure_early_kill
+	Enable early kill mode globally
+
+PR_MCE_KILL
+	Set early/late kill mode/revert to system default
+	arg1: PR_MCE_KILL_CLEAR: Revert to system default
+	arg1: PR_MCE_KILL_SET: arg2 defines thread specific mode
+		PR_MCE_KILL_EARLY: Early kill
+		PR_MCE_KILL_LATE:  Late kill
+		PR_MCE_KILL_DEFAULT: Use system global default
+PR_MCE_KILL_GET
+	return current mode
+
+
+---
+
+Testing:
+
+madvise(MADV_POISON, ....)
+	(as root)
+	Poison a page in the process for testing
+
+
+hwpoison-inject module through debugfs
+	/sys/debug/hwpoison/corrupt-pfn
+
+Inject hwpoison fault at PFN echoed into this file
+
+
+Architecture specific MCE injector
+
+x86 has mce-inject, mce-test
+
+Some portable hwpoison test programs in mce-test, see blow.
+
+---
+
+References:
+
+http://halobates.de/mce-lc09-2.pdf
+	Overview presentation from LinuxCon 09
+
+git://git.kernel.org/pub/scm/utils/cpu/mce/mce-test.git
+	Test suite (hwpoison specific portable tests in tsrc)
+
+git://git.kernel.org/pub/scm/utils/cpu/mce/mce-inject.git
+	x86 specific injector
+
+
+---
+
+Limitations:
+
+- Not all page types are supported and never will. Most kernel internal
+objects cannot be recovered, only LRU pages for now.
+- Right now hugepage support is missing.
+
+---
+Andi Kleen, Oct 2009
+
-- 
1.6.0.2

[PATCH] HWPOISON: return early on non-LRU pages

[Andi Kleen]

From: Wu Fengguang <fengguang.wu@intel.com>

Right now we have some trouble with non atomic access
to page flags when locking the page. To plug this hole
for now, limit error recovery to LRU pages for now.

This could be better fixed by defining a suitable protocol,
but let's go this simple way for now

This avoids unnecessary races with __set_page_locked() and
__SetPageSlab*() and maybe more non-atomic page flag operations.

This loses isolated pages which are currently in page reclaim, but these
are relatively limited compared to the total memory.

Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
[AK: new description, bug fixes, cleanups]
---
 mm/memory-failure.c |   49 ++++++++++++++++++++++++-------------------------
 1 files changed, 24 insertions(+), 25 deletions(-)

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 729d4b1..e17ec3f 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -370,9 +370,6 @@ static int me_pagecache_clean(struct page *p, unsigned long pfn)
 	int ret = FAILED;
 	struct address_space *mapping;
 
-	if (!isolate_lru_page(p))
-		page_cache_release(p);
-
 	/*
 	 * For anonymous pages we're done the only reference left
 	 * should be the one m_f() holds.
@@ -498,30 +495,18 @@ static int me_pagecache_dirty(struct page *p, unsigned long pfn)
  */
 static int me_swapcache_dirty(struct page *p, unsigned long pfn)
 {
-	int ret = FAILED;
-
 	ClearPageDirty(p);
 	/* Trigger EIO in shmem: */
 	ClearPageUptodate(p);
 
-	if (!isolate_lru_page(p)) {
-		page_cache_release(p);
-		ret = DELAYED;
-	}
-
-	return ret;
+	return DELAYED;
 }
 
 static int me_swapcache_clean(struct page *p, unsigned long pfn)
 {
-	int ret = FAILED;
-
-	if (!isolate_lru_page(p)) {
-		page_cache_release(p);
-		ret = RECOVERED;
-	}
 	delete_from_swap_cache(p);
-	return ret;
+
+	return RECOVERED;
 }
 
 /*
@@ -611,8 +596,6 @@ static struct page_state {
 	{ 0,		0,		"unknown page state",	me_unknown },
 };
 
-#undef lru
-
 static void action_result(unsigned long pfn, char *msg, int result)
 {
 	struct page *page = NULL;
@@ -664,9 +647,6 @@ static void hwpoison_user_mappings(struct page *p, unsigned long pfn,
 	if (PageReserved(p) || PageCompound(p) || PageSlab(p))
 		return;
 
-	if (!PageLRU(p))
-		lru_add_drain_all();
-
 	/*
 	 * This check implies we don't kill processes if their pages
 	 * are in the swap cache early. Those are always late kills.
@@ -738,6 +718,7 @@ static void hwpoison_user_mappings(struct page *p, unsigned long pfn,
 
 int __memory_failure(unsigned long pfn, int trapno, int ref)
 {
+	unsigned long lru_flag;
 	struct page_state *ps;
 	struct page *p;
 	int res;
@@ -775,6 +756,24 @@ int __memory_failure(unsigned long pfn, int trapno, int ref)
 	}
 
 	/*
+	 * We ignore non-LRU pages for good reasons.
+	 * - PG_locked is only well defined for LRU pages and a few others
+	 * - to avoid races with __set_page_locked()
+	 * - to avoid races with __SetPageSlab*() (and more non-atomic ops)
+	 * The check (unnecessarily) ignores LRU pages being isolated and
+	 * walked by the page reclaim code, however that's not a big loss.
+	 */
+	if (!PageLRU(p))
+		lru_add_drain_all();
+	lru_flag = p->flags & lru;
+	if (isolate_lru_page(p)) {
+		action_result(pfn, "non LRU", IGNORED);
+		put_page(p);
+		return -EBUSY;
+	}
+	page_cache_release(p);
+
+	/*
 	 * Lock the page and wait for writeback to finish.
 	 * It's very difficult to mess with pages currently under IO
 	 * and in many cases impossible, so we just avoid it here.
@@ -790,7 +789,7 @@ int __memory_failure(unsigned long pfn, int trapno, int ref)
 	/*
 	 * Torn down by someone else?
 	 */
-	if (PageLRU(p) && !PageSwapCache(p) && p->mapping == NULL) {
+	if ((lru_flag & lru) && !PageSwapCache(p) && p->mapping == NULL) {
 		action_result(pfn, "already truncated LRU", IGNORED);
 		res = 0;
 		goto out;
@@ -798,7 +797,7 @@ int __memory_failure(unsigned long pfn, int trapno, int ref)
 
 	res = -EBUSY;
 	for (ps = error_states;; ps++) {
-		if ((p->flags & ps->mask) == ps->res) {
+		if (((p->flags | lru_flag)& ps->mask) == ps->res) {
 			res = page_action(ps, p, pfn, ref);
 			break;
 		}
-- 
1.6.0.2

[PATCH] HWPOISON: Fix page count leak in hwpoison late kill in do_swap_page

[Andi Kleen]

From: Andi Kleen <ak@linux.intel.com>

When returning due to a poisoned page drop the page count.

It wasn't a fatal problem because noone cares about the page count
on a poisoned page (except when it wraps), but it's cleaner to fix it.

Pointed out by Linus.

Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
 mm/memory.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/mm/memory.c b/mm/memory.c
index 7e91b5f..7a3b0ad 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -2539,7 +2539,7 @@ static int do_swap_page(struct mm_struct *mm, struct vm_area_struct *vma,
 	} else if (PageHWPoison(page)) {
 		ret = VM_FAULT_HWPOISON;
 		delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
-		goto out;
+		goto out_release;
 	}
 
 	lock_page(page);
@@ -2611,6 +2611,7 @@ out_nomap:
 	pte_unmap_unlock(page_table, ptl);
 out_page:
 	unlock_page(page);
+out_release:
 	page_cache_release(page);
 	return ret;
 }
-- 
1.6.0.2

[PATCH] HWPOISON: fix oops on ksm pages

[Andi Kleen]

From: Hugh Dickins <hugh.dickins@tiscali.co.uk>

Memory failure on a KSM page currently oopses on its NULL anon_vma in
page_lock_anon_vma(): that may not be much worse than the consequence
of ignoring it, but it is better to be consistent with how ZERO_PAGE
and hugetlb pages and other awkward cases are treated.  Just skip it.

Signed-off-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
 mm/memory-failure.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index e17ec3f..e354b9f 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -35,6 +35,7 @@
 #include <linux/mm.h>
 #include <linux/page-flags.h>
 #include <linux/sched.h>
+#include <linux/ksm.h>
 #include <linux/rmap.h>
 #include <linux/pagemap.h>
 #include <linux/swap.h>
@@ -644,7 +645,7 @@ static void hwpoison_user_mappings(struct page *p, unsigned long pfn,
 	int i;
 	int kill = 1;
 
-	if (PageReserved(p) || PageCompound(p) || PageSlab(p))
+	if (PageReserved(p) || PageCompound(p) || PageSlab(p) || PageKsm(p))
 		return;
 
 	/*
-- 
1.6.0.2

[PATCH] HWPOISON: fix/proc/meminfo alignment

[Andi Kleen]

From: Hugh Dickins <hugh.dickins@tiscali.co.uk>

Given such a long name, the kB count in /proc/meminfo's HardwareCorrupted
line is being shown too far right (it does align with x86_64's VmallocChunk
above, but I hope nobody will ever have that much corrupted!).  Align it.

Signed-off-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
 fs/proc/meminfo.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
index c7bff4f..a65239c 100644
--- a/fs/proc/meminfo.c
+++ b/fs/proc/meminfo.c
@@ -99,7 +99,7 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
 		"VmallocUsed:    %8lu kB\n"
 		"VmallocChunk:   %8lu kB\n"
 #ifdef CONFIG_MEMORY_FAILURE
-		"HardwareCorrupted: %8lu kB\n"
+		"HardwareCorrupted: %5lu kB\n"
 #endif
 		,
 		K(i.totalram),
-- 
1.6.0.2

[PATCH] HWPOISON: Allow schedule_on_each_cpu() from keventd

[Andi Kleen]

From: Andi Kleen <ak@linux.intel.com>

Right now when calling schedule_on_each_cpu() from keventd there
is a deadlock because it tries to schedule a work item on the current CPU
too. This happens via lru_add_drain_all() in hwpoison.

Just call the function for the current CPU in this case. This is actually
faster too.

Debugging with Fengguang Wu & Max Asbock

Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
 kernel/workqueue.c |   21 +++++++++++++++++++--
 1 files changed, 19 insertions(+), 2 deletions(-)

diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index addfe2d..f61a2fe 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -667,21 +667,38 @@ EXPORT_SYMBOL(schedule_delayed_work_on);
 int schedule_on_each_cpu(work_func_t func)
 {
 	int cpu;
+	int orig = -1;
 	struct work_struct *works;
 
 	works = alloc_percpu(struct work_struct);
 	if (!works)
 		return -ENOMEM;
 
+	/*
+	 * when running in keventd don't schedule a work item on itself.
+	 * Can just call directly because the work queue is already bound.
+	 * This also is faster.
+	 * Make this a generic parameter for other workqueues?
+	 */
+	if (current_is_keventd()) {
+		orig = raw_smp_processor_id();
+		INIT_WORK(per_cpu_ptr(works, orig), func);
+		func(per_cpu_ptr(works, orig));
+	}
+
 	get_online_cpus();
 	for_each_online_cpu(cpu) {
 		struct work_struct *work = per_cpu_ptr(works, cpu);
 
+		if (cpu == orig)
+			continue;
 		INIT_WORK(work, func);
 		schedule_work_on(cpu, work);
 	}
-	for_each_online_cpu(cpu)
-		flush_work(per_cpu_ptr(works, cpu));
+	for_each_online_cpu(cpu) {
+		if (cpu != orig)
+			flush_work(per_cpu_ptr(works, cpu));
+	}
 	put_online_cpus();
 	free_percpu(works);
 	return 0;
-- 
1.6.0.2

[PATCH] HWPOISON: fix invalid page count in printk output

[Andi Kleen]

From: Wu Fengguang <fengguang.wu@intel.com>

The madvise injector already holds a reference when passing in a page
to the memory-failure code. The code corrects for this additional reference
for its checks, but the final printk output didn't. Fix that.

Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
---
 mm/memory-failure.c |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index e354b9f..dacc641 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -613,13 +613,16 @@ static int page_action(struct page_state *ps, struct page *p,
 			unsigned long pfn, int ref)
 {
 	int result;
+	int count;
 
 	result = ps->action(p, pfn);
 	action_result(pfn, ps->msg, result);
-	if (page_count(p) != 1 + ref)
+
+	count = page_count(p) - 1 - ref;
+	if (count != 0)
 		printk(KERN_ERR
 		       "MCE %#lx: %s page still referenced by %d users\n",
-		       pfn, ps->msg, page_count(p) - 1);
+		       pfn, ps->msg, count);
 
 	/* Could do more checks here if page looks ok */
 	/*
-- 
1.6.0.2