Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1329782

[PATCH 3.2 15/87] powerpc: Make value-returning atomics fully ordered

From Ben Hutchings <ben@decadent.org.uk>
Newsgroups linux.kernel
Subject [PATCH 3.2 15/87] powerpc: Make value-returning atomics fully ordered
Date 2016-02-09 01:30 +0100
Message-ID <r04lQ-4r7-15@gated-at.bofh.it> (permalink)
References <r03SN-3V2-7@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw


3.2.77-rc1 review patch.  If anyone has any objections, please let me know.

------------------

From: Boqun Feng <boqun.feng@gmail.com>

commit 49e9cf3f0c04bf76ffa59242254110309554861d upstream.

According to memory-barriers.txt:

> Any atomic operation that modifies some state in memory and returns
> information about the state (old or new) implies an SMP-conditional
> general memory barrier (smp_mb()) on each side of the actual
> operation ...

Which mean these operations should be fully ordered. However on PPC,
PPC_ATOMIC_ENTRY_BARRIER is the barrier before the actual operation,
which is currently "lwsync" if SMP=y. The leading "lwsync" can not
guarantee fully ordered atomics, according to Paul Mckenney:

https://lkml.org/lkml/2015/10/14/970

To fix this, we define PPC_ATOMIC_ENTRY_BARRIER as "sync" to guarantee
the fully-ordered semantics.

This also makes futex atomics fully ordered, which can avoid possible
memory ordering problems if userspace code relies on futex system call
for fully ordered semantics.

Fixes: b97021f85517 ("powerpc: Fix atomic_xxx_return barrier semantics")
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
---
 arch/powerpc/include/asm/synch.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

--- a/arch/powerpc/include/asm/synch.h
+++ b/arch/powerpc/include/asm/synch.h
@@ -44,7 +44,7 @@ static inline void isync(void)
 	MAKE_LWSYNC_SECTION_ENTRY(97, __lwsync_fixup);
 #define PPC_ACQUIRE_BARRIER	 "\n" stringify_in_c(__PPC_ACQUIRE_BARRIER)
 #define PPC_RELEASE_BARRIER	 stringify_in_c(LWSYNC) "\n"
-#define PPC_ATOMIC_ENTRY_BARRIER "\n" stringify_in_c(LWSYNC) "\n"
+#define PPC_ATOMIC_ENTRY_BARRIER "\n" stringify_in_c(sync) "\n"
 #define PPC_ATOMIC_EXIT_BARRIER	 "\n" stringify_in_c(sync) "\n"
 #else
 #define PPC_ACQUIRE_BARRIER

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread


Thread

[PATCH 3.2 00/87] 3.2.77-rc1 review Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 79/87] crypto: af_alg - Forbid bind(2) when nokey  child sockets are present Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 17/87] asix: silence log message from oversize packet Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 73/87] crypto: algif_hash - Require setkey before  accept(2) Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 18/87] futex: Drop refcount if requeue_pi() acquired  the rtmutex Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 60/87] crypto: af_alg - Fix socket double-free when  accept fails Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 72/87] crypto: hash - Add crypto_ahash_has_setkey Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 70/87] crypto: af_alg - Add nokey compatibility path Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 57/87] ALSA: timer: Harden slave timer list handling Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 61/87] ALSA: hrtimer: Fix stall by hrtimer_cancel() Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 82/87] crypto: algif_skcipher - Load TX SG list after  waiting Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 11/87] rtlwifi: fix memory leak for USB device Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 64/87] ALSA: control: Avoid kernel warnings from tlv  ioctl with numid 0 Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 01/87] [media] gspca: ov534/topro: prevent a division by 0 Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:20 +0100
  [PATCH 3.2 08/87] mtd: nand: fix ONFI parameter page layout Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 80/87] crypto: algif_hash - Fix race condition in  hash_check_key Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 04/87] SCSI: initio: remove duplicate module device table Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 87/87] [media] usbvision: fix crash on detecting  device with invalid configuration Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 15/87] powerpc: Make value-returning atomics fully ordered Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 10/87] xhci: refuse loading if nousb is used Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 05/87] KVM: x86: expose MSR_TSC_AUX to userspace Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 86/87] [media] usbvision fix overflow of interfaces array Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 85/87] [media] usbvision: fix leak of usb_dev on  failure paths in  usbvision_probe() Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 07/87] ath9k_htc: check for underflow in  ath9k_htc_rx_msg() Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 13/87] wlcore/wl12xx: spi: fix oops on firmware load Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 14/87] EDAC: Robustify workqueues destruction Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 19/87] ALSA: fm801: propagate TUNER_ONLY bit when  autodetected Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 84/87] [media] usbvision-video: fix memory leak of  alt_max_pkt_size Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 12/87] wlcore: SPI - fix spi transfer_list Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 16/87] powerpc: Make {cmp}xchg* and their atomic_  versions fully ordered Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 78/87] crypto: algif_skcipher - Remove custom release  parent function Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  [PATCH 3.2 03/87] [media] rc: allow rc modules to be loaded if  rc-main is not a module Ben Hutchings <ben@decadent.org.uk> - 2016-02-09 01:30 +0100
  Re: [PATCH 3.2 00/87] 3.2.77-rc1 review Guenter Roeck <linux@roeck-us.net> - 2016-02-09 05:00 +0100
    Re: [PATCH 3.2 00/87] 3.2.77-rc1 review Ben Hutchings <ben@decadent.org.uk> - 2016-02-10 02:10 +0100

csiph-web