Commit 89b4d774 authored by Noah Misch's avatar Noah Misch

For PowerPC instruction "addi", use constraint "b".

Without "b", a variant of the tas() code miscompiles on macOS 10.4.
This may also fix a compilation failure involving macOS 10.1.  Today's
compilers have been allocating acceptable registers with or without this
change, but this future-proofs the code by precisely conveying the
acceptable registers.  Back-patch to 9.4 (all supported versions).

Reviewed by Tom Lane.

Discussion: https://postgr.es/m/20191009063900.GA4066266@rfd.leadboat.com
parent f25968c4
...@@ -452,6 +452,11 @@ typedef unsigned int slock_t; ...@@ -452,6 +452,11 @@ typedef unsigned int slock_t;
#define TAS_SPIN(lock) (*(lock) ? 1 : TAS(lock)) #define TAS_SPIN(lock) (*(lock) ? 1 : TAS(lock))
/* /*
* The second operand of addi can hold a constant zero or a register number,
* hence constraint "=&b" to avoid allocating r0. "b" stands for "address
* base register"; most operands having this register-or-zero property are
* address bases, e.g. the second operand of lwax.
*
* NOTE: per the Enhanced PowerPC Architecture manual, v1.0 dated 7-May-2002, * NOTE: per the Enhanced PowerPC Architecture manual, v1.0 dated 7-May-2002,
* an isync is a sufficient synchronization barrier after a lwarx/stwcx loop. * an isync is a sufficient synchronization barrier after a lwarx/stwcx loop.
* On newer machines, we can use lwsync instead for better performance. * On newer machines, we can use lwsync instead for better performance.
...@@ -488,7 +493,7 @@ tas(volatile slock_t *lock) ...@@ -488,7 +493,7 @@ tas(volatile slock_t *lock)
#endif #endif
" li %1,0 \n" " li %1,0 \n"
: "=&r"(_t), "=r"(_res), "+m"(*lock) : "=&b"(_t), "=r"(_res), "+m"(*lock)
: "r"(lock) : "r"(lock)
: "memory", "cc"); : "memory", "cc");
return _res; return _res;
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment