On 9/2/2026 4:16 PM, Philippe Mathieu-Daudé wrote:
Hi Brian,
On 1/9/26 19:33, Brian Cain wrote:
GPCYCLELO/GPCYCLEHI read as zero unless the cycle counter is enabled
(SSR:CE), matching the hardware.
Reviewed-by: Pierrick Bouvier <[email protected]>
Signed-off-by: Brian Cain <[email protected]>
---
target/hexagon/cpu.c | 7 +++++--
target/hexagon/op_helper.c | 9 ++-------
2 files changed, 7 insertions(+), 9 deletions(-)
diff --git a/target/hexagon/op_helper.c b/target/hexagon/op_helper.c
index 23894ff3d28..2cea1927263 100644
--- a/target/hexagon/op_helper.c
+++ b/target/hexagon/op_helper.c
@@ -1912,13 +1912,8 @@ uint64_t
HELPER(greg_read_pair)(CPUHexagonState *env, uint32_t reg)
return (uint64_t)(env->greg[reg]) |
(((uint64_t)(env->greg[reg + 1])) << 32);
}
- switch (reg) {
- case HEX_GREG_GPCYCLELO:
- return hexagon_get_sys_pcycle_count(env);
- default:
- return (uint64_t)hexagon_greg_read(env, reg) |
- ((uint64_t)(hexagon_greg_read(env, reg + 1)) << 32);
- }
+ return (uint64_t)hexagon_greg_read(env, reg) |
+ ((uint64_t)(hexagon_greg_read(env, reg + 1)) << 32);
}
Just curious, does greg_read_pair() work with impair %reg?
Not sure if I understand "impair %reg"?
But - if this refers to the odd-value pairs, then no - they don't. The
guest reg encoding only has the "g{odd}:{even}" designated by the even
reg's value. Sent a follow up patch to assert this in greg_read_pair().
-Brian
