From: Laurent Dufour <ldufour@linux.vnet.ibm.com>
To: paulmck@linux.vnet.ibm.com, peterz@infradead.org,
        akpm@linux-foundation.org, kirill@shutemov.name, ak@linux.intel.com,
        mhocko@kernel.org, dave@stgolabs.net, jack@suse.cz
Cc: linux-kernel@vger.kernel.org, linux-mm@kvack.org,
        haren@linux.vnet.ibm.com, khandual@linux.vnet.ibm.com,
        npiggin@gmail.com, bsingharora@gmail.com
Subject: [RFC v3 08/17] mm/spf: Try spin lock in speculative path
Date: Thu, 27 Apr 2017 17:52:47 +0200
In-Reply-To: <1493308376-23851-1-git-send-email-ldufour@linux.vnet.ibm.com>
References: <1493308376-23851-1-git-send-email-ldufour@linux.vnet.ibm.com>
Message-Id: <1493308376-23851-9-git-send-email-ldufour@linux.vnet.ibm.com>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2690
Lines: 84

There is a deadlock when a CPU is doing a speculative page fault and
another one is calling do_unmap().

The deadlock occurred because the speculative path try to spinlock the
pte while the interrupt are disabled. When the other CPU in the
unmap's path has locked the pte then is waiting for all the CPU to
invalidate the TLB. As the CPU doing the speculative fault have the
interrupt disable it can't invalidate the TLB, and can't get the lock.

Since we are in a speculative path, we can race with other mm action.
So let assume that the lock may not get acquired and fail the
speculative page fault.

Here are the stack captured during the deadlock:

CPU 0
native_flush_tlb_others+0x7c/0x260
flush_tlb_mm_range+0x6a/0x220
tlb_flush_mmu_tlbonly+0x63/0xc0
unmap_page_range+0x897/0x9d0
? unmap_single_vma+0x7d/0xe0
? release_pages+0x2b3/0x360
unmap_single_vma+0x7d/0xe0
unmap_vmas+0x51/0xa0
unmap_region+0xbd/0x130
do_munmap+0x279/0x460
SyS_munmap+0x53/0x70

CPU 1
do_raw_spin_lock+0x14e/0x160
_raw_spin_lock+0x5d/0x80
? pte_map_lock+0x169/0x1b0
pte_map_lock+0x169/0x1b0
handle_pte_fault+0xbf2/0xd80
? trace_hardirqs_on+0xd/0x10
handle_speculative_fault+0x272/0x280
handle_speculative_fault+0x5/0x280
__do_page_fault+0x187/0x580
trace_do_page_fault+0x52/0x260
do_async_page_fault+0x19/0x70
async_page_fault+0x28/0x30

Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
---
 mm/memory.c | 17 ++++++++++++++---
 1 file changed, 14 insertions(+), 3 deletions(-)

diff --git a/mm/memory.c b/mm/memory.c
index fd3a0dc122c5..458f579feb6f 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -2118,7 +2118,8 @@ static bool pte_spinlock(struct vm_fault *vmf)
 		goto out;
 
 	vmf->ptl = pte_lockptr(vmf->vma->vm_mm, vmf->pmd);
-	spin_lock(vmf->ptl);
+	if (unlikely(!spin_trylock(vmf->ptl)))
+		goto out;
 
 	if (vma_is_dead(vmf->vma, vmf->sequence)) {
 		spin_unlock(vmf->ptl);
@@ -2152,8 +2153,18 @@ static bool pte_map_lock(struct vm_fault *vmf)
 	if (vma_is_dead(vmf->vma, vmf->sequence))
 		goto out;
 
-	vmf->pte = pte_offset_map_lock(vmf->vma->vm_mm, vmf->pmd,
-				       vmf->address, &vmf->ptl);
+	/* Same as pte_offset_map_lock() except that we call
+	 * spin_trylock() in place of spin_lock() to avoid race with
+	 * unmap path which may have the lock and wait for this CPU
+	 * to invalidate TLB but this CPU has irq disabled.
+	 * Since we are in a speculative patch, accept it could fail
+	 */
+	vmf->ptl = pte_lockptr(vmf->vma->vm_mm, vmf->pmd);
+	vmf->pte = pte_offset_map(vmf->pmd, vmf->address);
+	if (unlikely(!spin_trylock(vmf->ptl))) {
+		pte_unmap(vmf->pte);
+		goto out;
+	}
 
 	if (vma_is_dead(vmf->vma, vmf->sequence)) {
 		pte_unmap_unlock(vmf->pte, vmf->ptl);
-- 
2.7.4