Date: Thu, 22 Mar 2012 14:01:59 -0400 From: Jeremiah Lott <jlott@averesystems.com> To: freebsd-amd64@freebsd.org Cc: alc@freebsd.org, kib@freebsd.org Subject: page fault after wiring page Message-ID: <DB3A5662-87CD-459E-9DC6-9E41EF58ACF0@averesystems.com>
next in thread | raw e-mail | index | archive | help
We've been seeing some panics and deadlocks that appear to be related to =
getting a page fault when accessing a page after it has been wired (on =
amd64). All the ones we have seen are related to sysctl handlers that =
call sysctl_wire_old_buffer, then lock a mutex, then call SYSCTL_OUT. =
When it does the copyout, it gets a page fault even though the page has =
been wired, sometimes causing it to sleep while holding a mutex or =
recurse on non-recursable mutexes. Here are the two panics that are =
most easy to follow:
Sleeping thread (tid 100458, pid 2737) owns a non-sleepable lock
sched_switch() at 0xffffffff80603bf5 =3D sched_switch+0x146
mi_switch() at 0xffffffff805e8e15 =3D mi_switch+0x183
sleepq_switch() at 0xffffffff8061e6e7 =3D sleepq_switch+0xb1
sleepq_wait() at 0xffffffff8061f0ea =3D sleepq_wait+0x3d
_sx_slock_hard() at 0xffffffff805e7ca7 =3D _sx_slock_hard+0x41d
_sx_slock() at 0xffffffff805e7e32 =3D _sx_slock+0x3d
vm_map_lookup() at 0xffffffff807909e4 =3D vm_map_lookup+0x54
vm_fault() at 0xffffffff80786c20 =3D vm_fault+0x11c
trap_pfault() at 0xffffffff80844dd0 =3D trap_pfault+0xe1
trap() at 0xffffffff80845286 =3D trap+0x337
calltrap() at 0xffffffff80827f28 =3D calltrap+0x8
--- trap 0xc, rip =3D 0xffffffff8084296b, rsp =3D 0xffffff811391e7e0, =
rbp =3D 0xffffff811391e810 ---
copyout() at 0xffffffff8084296b =3D copyout+0x3b
sysctl_rtsock() at 0xffffffff806a5ef7 =3D sysctl_rtsock+0x499
sysctl_root() at 0xffffffff805eab9e =3D sysctl_root+0xea
userland_sysctl() at 0xffffffff805eae6e =3D userland_sysctl+0x14f
sysctl() at 0xffffffff805eb258 =3D sysctl+0x9a
amd64_syscall() at 0xffffffff80844065 =3D amd64_syscall+0x145
Xfast_syscall() at 0xffffffff8082821c =3D Xfast_syscall+0xfc
login: panic: _mtx_lock_sleep: recursed on non-recursive mutex process =
lock @ ../../../amd64/amd64/trap.c:731
cpuid =3D 0
KDB: stack backtrace:
gdb_trace_self_wrapper() at 0xffffffff8057e7ea =3D =
gdb_trace_self_wrapper+0x2a
kdb_backtrace() at 0xffffffff8062ffdc =3D kdb_backtrace+0x37
panic() at 0xffffffff805f89ca =3D panic+0x2ad
_mtx_lock_flags() at 0xffffffff805e9376 =3D _mtx_lock_flags
_mtx_lock_flags() at 0xffffffff805e9417 =3D _mtx_lock_flags+0xa1
trap_pfault() at 0xffffffff80880450 =3D trap_pfault+0xa1
trap() at 0xffffffff80880ac7 =3D trap+0x4b8
calltrap() at 0xffffffff80861af8 =3D calltrap+0x8
--- trap 0xc, rip =3D 0xffffffff8087de8b, rsp =3D 0xffffff807b7e9410, =
rbp =3D 0xffffff807b7e9440 ---
copyout() at 0xffffffff8087de8b =3D copyout+0x3b
sysctl_out_proc() at 0xffffffff805ed305 =3D sysctl_out_proc+0x16c
sysctl_root() at 0xffffffff80606141 =3D sysctl_root+0x13a
userland_sysctl() at 0xffffffff8060640a =3D userland_sysctl+0x14f
sysctl() at 0xffffffff806067f8 =3D sysctl+0x9a
amd64_syscall() at 0xffffffff8087f635 =3D amd64_syscall+0x145
Xfast_syscall() at 0xffffffff80861dec =3D Xfast_syscall+0xfc
--- syscall (202, FreeBSD ELF64, sysctl), rip =3D 0x801c12b0c, rsp =3D =
0x7fffffffb768, rbp =3D 0x7fffffffb7b0 ---
--- curthread 0xffffff000465b000, tid 100142
After doing some instrumentation, I think I've figured out what is =
causing this. It seems that when I am wiring the page, in some =
situations the page table entry is being changed from read-only -> =
read-write as well as being wired. I haven't figured out the exact =
scenario that causes this, but I can definitely see it in my added =
trace. Here is an example page table entry transition I am seeing in =
pmap_enter that is called as a result of the wire:
pmap_enter: origpte: 80000000ad201425 newpte: 80000000ad201607
This means that we are setting PG_W (wired) and PG_RW (read/write) in =
this pmap_enter operation. Everytime I saw a page-fault after wiring it =
was immediately preceded by a transition like this (in the cases that =
did not page fault, the page table entry already had PG_RW set). This =
made me suspect that a read-only version of the page table entry was =
cached in the TLB. I noticed we invalidate in some situations in =
pmap_enter, but this transition is not one of them. I was able to =
eliminate the panics by making this change:
diff --git a/src/sys/amd64/amd64/pmap.c b/src/sys/amd64/amd64/pmap.c
--- a/src/sys/amd64/amd64/pmap.c
+++ b/src/sys/amd64/amd64/pmap.c
@@ -3251,6 +3251,11 @@ validate:
if (opa !=3D VM_PAGE_TO_PHYS(m) || =
((origpte &
PG_NX) =3D=3D 0 && (newpte & =
PG_NX)))
invlva =3D TRUE;
+ if ((newpte & PG_W) &&
+ ((origpte & PG_RW) =3D=3D 0) &&
+ (newpte & PG_RW)) {
+ invlva =3D TRUE;
+ }
}
if ((origpte & (PG_M | PG_RW)) =3D=3D (PG_M | =
PG_RW)) {
if ((origpte & PG_MANAGED) !=3D 0)
I wanted to see if anyone has seen issues in this area, and if this fix =
seems appropriate. I'm running 8.2, but I didn't see any obvious =
changes to pmap stuff in head which would change this behavior. Thanks =
for any feedback,
Jeremiah Lott
Avere Systems=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?DB3A5662-87CD-459E-9DC6-9E41EF58ACF0>