@subsection Atomic instructions
-Atomic exchange-and-add instructions are provided in two flavors: one
-for swapping 64-bit quantities and another for 32-bit quantities.
+Atomic exchange instructions are provided in two flavors: one for
+compare-and-swap, one for unconditional exchange.
+
+@table @code
+@item acmp [rd + offset16], rs
+@itemx r0 = cmpxchg_64 (rd + offset16, r0, rs)
+Atomic compare-and-swap. Compares value in @code{r0} to value
+addressed by @code{rd + offset16}. On match, the value addressed by
+@code{rd + offset16} is replaced with the value in @code{rs}.
+Regardless, the value that was at @code{rd + offset16} is
+zero-extended and loaded into @code{r0}.
+
+@item axchg [rd + offset16], rs
+@itemx rs = xchg_64 (rd + offset16, rs)
+Atomic exchange. Atomically exchanges the value in @code{rs} with
+the value addressed by @code{rd + offset16}.
+@end table
+
+@noindent
+The following instructions provide atomic arithmetic operations.
@table @code
@item aadd [rd + offset16], rs
@subsection 32-bit atomic instructions
-Atomic exchange-and-add instructions are provided in two flavors: one
-for swapping 32-bit quantities and another for 32-bit quantities.
+32-bit atomic exchange instructions are provided in two flavors: one
+for compare-and-swap, one for unconditional exchange.
+
+@table @code
+@item acmp32 [rd + offset16], rs
+@itemx w0 = cmpxchg32_32 (rd + offset16, w0, ws)
+Atomic compare-and-swap. Compares value in @code{w0} to value
+addressed by @code{rd + offset16}. On match, the value addressed by
+@code{rd + offset16} is replaced with the value in @code{ws}.
+Regardless, the value that was at @code{rd + offset16} is
+zero-extended and loaded into @code{w0}.
+
+@item axchg [rd + offset16], rs
+@itemx ws = xchg32_32 (rd + offset16, ws)
+Atomic exchange. Atomically exchanges the value in @code{ws} with
+the value addressed by @code{rd + offset16}.
+@end table
+
+@noindent
+The following instructions provide 32-bit atomic arithmetic operations.
@table @code
@item aadd32 [rd + offset16], rs
78: c3 12 1e ef 00 00 00 41 w2=atomic_fetch_or\(\(u32\*\)\(r1\+0x1eef\),w2\)
80: db 12 1e ef 00 00 00 a1 r2=atomic_fetch_xor\(\(u64\*\)\(r1\+0x1eef\),r2\)
88: c3 12 1e ef 00 00 00 a1 w2=atomic_fetch_xor\(\(u32\*\)\(r1\+0x1eef\),w2\)
+ 90: db 12 00 04 00 00 00 f1 r0=cmpxchg_64\(r1\+0x4,r0,r2\)
+ 98: c3 23 00 04 00 00 00 f1 w0=cmpxchg32_32\(r2\+0x4,w0,w3\)
+ a0: db 12 00 08 00 00 00 e1 r2=xchg_64\(r1\+0x8,r2\)
+ a8: c3 13 00 08 00 00 00 e1 w3=xchg32_32\(r1\+0x8,w3\)
68: c3 12 1e ef 00 00 00 41 afor32 \[%r1\+0x1eef\],%r2
70: db 12 1e ef 00 00 00 a1 afxor \[%r1\+0x1eef\],%r2
78: c3 12 1e ef 00 00 00 a1 afxor32 \[%r1\+0x1eef\],%r2
+ 80: db 12 00 04 00 00 00 f1 acmp \[%r1\+0x4\],%r2
+ 88: c3 23 00 04 00 00 00 f1 acmp32 \[%r2\+0x4\],%r3
+ 90: db 12 00 08 00 00 00 e1 axchg \[%r1\+0x8\],%r2
+ 98: c3 13 00 08 00 00 00 e1 axchg32 \[%r1\+0x8\],%r3
78: c3 21 ef 1e 41 00 00 00 w2=atomic_fetch_or\(\(u32\*\)\(r1\+0x1eef\),w2\)
80: db 21 ef 1e a1 00 00 00 r2=atomic_fetch_xor\(\(u64\*\)\(r1\+0x1eef\),r2\)
88: c3 21 ef 1e a1 00 00 00 w2=atomic_fetch_xor\(\(u32\*\)\(r1\+0x1eef\),w2\)
+ 90: db 21 04 00 f1 00 00 00 r0=cmpxchg_64\(r1\+0x4,r0,r2\)
+ 98: c3 32 04 00 f1 00 00 00 w0=cmpxchg32_32\(r2\+0x4,w0,w3\)
+ a0: db 21 08 00 e1 00 00 00 r2=xchg_64\(r1\+0x8,r2\)
+ a8: c3 31 08 00 e1 00 00 00 w3=xchg32_32\(r1\+0x8,w3\)
w2 = atomic_fetch_or((u32*)(r1+0x1eef),w2)
r2 = atomic_fetch_xor((u64*)(r1+0x1eef),r2)
w2 = atomic_fetch_xor((u32*)(r1+0x1eef),w2)
+ r0 = cmpxchg_64(r1+0x4,r0,r2)
+ w0 = cmpxchg32_32(r2+0x4,w0,w3)
+ r2 = xchg_64(r1+0x8,r2)
+ w3 = xchg32_32(r1+0x8,w3)
68: c3 21 ef 1e 41 00 00 00 afor32 \[%r1\+0x1eef\],%r2
70: db 21 ef 1e a1 00 00 00 afxor \[%r1\+0x1eef\],%r2
78: c3 21 ef 1e a1 00 00 00 afxor32 \[%r1\+0x1eef\],%r2
+ 80: db 21 04 00 f1 00 00 00 acmp \[%r1\+0x4\],%r2
+ 88: c3 32 04 00 f1 00 00 00 acmp32 \[%r2\+0x4\],%r3
+ 90: db 21 08 00 e1 00 00 00 axchg \[%r1\+0x8\],%r2
+ 98: c3 31 08 00 e1 00 00 00 axchg32 \[%r1\+0x8\],%r3
afor32 [%r1+0x1eef], %r2
afxor [%r1+0x1eef], %r2
afxor32 [%r1+0x1eef], %r2
+
+ acmp [%r1+4], %r2
+ acmp32 [%r2+4], %r3
+ axchg [%r1+8], %r2
+ axchg32 [%r1+8], %r3
#define BPF_IMM32_AFAND ((uint64_t)0x00000051)
#define BPF_IMM32_AFXOR ((uint64_t)0x000000a1)
#define BPF_IMM32_AXCHG ((uint64_t)0x000000e1)
-#define BPF_IMM32_ACMP ((uint64_t)b0x000000f1)
+#define BPF_IMM32_ACMP ((uint64_t)0x000000f1)
/* Unique identifiers for BPF instructions. */
BPF_INSN_AADD32, BPF_INSN_AOR32, BPF_INSN_AAND32, BPF_INSN_AXOR32,
/* Atomic instructions with fetching (32-bit.) */
BPF_INSN_AFADD32, BPF_INSN_AFOR32, BPF_INSN_AFAND32, BPF_INSN_AFXOR32,
+ /* Atomic compare-and-swap, atomic exchange. */
+ BPF_INSN_ACMP, BPF_INSN_AXCHG,
+ /* Atomic compare-and-swap, atomic exchange (32-bit). */
+ BPF_INSN_ACMP32, BPF_INSN_AXCHG32,
/* GNU simulator specific instruction. */
BPF_INSN_BRKPT,
};
{BPF_INSN_AFXOR32, "afxor32%W[ %dr %o16 ] , %sr", "%sw = atomic_fetch_xor ( ( u32 * ) ( %dr %o16 ) , %sw )",
BPF_V3, BPF_CODE|BPF_IMM32, BPF_CLASS_STX|BPF_SIZE_W|BPF_MODE_ATOMIC|BPF_IMM32_AFXOR},
+ /* Atomic compare-and-swap, atomic exchange. */
+ {BPF_INSN_ACMP, "acmp%W[ %dr %o16 ] , %sr", "r0 = cmpxchg_64 ( %dr %o16 , r0 , %sr )",
+ BPF_V3, BPF_CODE|BPF_IMM32, BPF_CLASS_STX|BPF_SIZE_DW|BPF_MODE_ATOMIC|BPF_IMM32_ACMP},
+ {BPF_INSN_AXCHG, "axchg%W[ %dr %o16 ] , %sr", "%sr = xchg_64 ( %dr %o16 , %sr )",
+ BPF_V3, BPF_CODE|BPF_IMM32, BPF_CLASS_STX|BPF_SIZE_DW|BPF_MODE_ATOMIC|BPF_IMM32_AXCHG},
+
+ /* Atomic compare-and-swap, atomic exchange (32-bit). */
+ {BPF_INSN_ACMP32, "acmp32%W[ %dr %o16 ], %sr", "w0 = cmpxchg32_32 ( %dr %o16 , w0 , %sw )",
+ BPF_V3, BPF_CODE|BPF_IMM32, BPF_CLASS_STX|BPF_SIZE_W|BPF_MODE_ATOMIC|BPF_IMM32_ACMP},
+ {BPF_INSN_AXCHG32, "axchg32%W[ %dr %o16 ], %sr", "%sw = xchg32_32 ( %dr %o16 , %sw )",
+ BPF_V3, BPF_CODE|BPF_IMM32, BPF_CLASS_STX|BPF_SIZE_W|BPF_MODE_ATOMIC|BPF_IMM32_AXCHG},
+
/* Old versions of aadd and aadd32. */
{BPF_INSN_AADD, "xadddw%W[ %dr %o16 ] , %sr", "* ( u64 * ) ( %dr %o16 ) += %sr",
BPF_V1, BPF_CODE|BPF_IMM32, BPF_CLASS_STX|BPF_SIZE_DW|BPF_MODE_ATOMIC|BPF_IMM32_AADD},