[<prev] [next>] [day] [month] [year] [list]
Message-ID: <20260809010303.601628-1-alex@alexrp.com>
Date: Sun, 9 Aug 2026 03:03:03 +0200
From: Alex Rønne Petersen <alex@...xrp.com>
To: musl@...ts.openwall.com
Cc: Alex Rønne Petersen <alex@...xrp.com>
Subject: [PATCH] powerpc: set up a proper stack frame in the parent thread in clone()
The ABI requires a stack frame to, at minimum, consist of the backchain slot and
the LR save slot at sp+0 and sp+4 respectively. The old code spilled r30/r31
into those slots, meaning that a backchain-based unwinder would see a nonsense
value as the backchain pointer and go on a wild goose chase.
It's admittedly a very small window where this is possible -- a thread that's
stopped in the middle of the clone() parent body -- but fixing it just requires
shifting the r30/r31 spill slots down by 8 bytes and storing the old sp in the
backchain slot, so seems reasonable to do.
---
src/thread/powerpc/clone.s | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/src/thread/powerpc/clone.s b/src/thread/powerpc/clone.s
index da13f446..55163585 100644
--- a/src/thread/powerpc/clone.s
+++ b/src/thread/powerpc/clone.s
@@ -16,8 +16,9 @@ __clone:
# store non-volatile regs r30, r31 on stack in order to put our
# start func and its arg there
-stwu 30, -16(1)
-stw 31, 4(1)
+stwu 1, -16(1)
+stw 30, 8(1)
+stw 31, 12(1)
# save r3 (func) into r30, and r6(arg) into r31
mr 30, 3
@@ -65,8 +66,8 @@ sc
2:
# restore stack
-lwz 30, 0(1)
-lwz 31, 4(1)
+lwz 30, 8(1)
+lwz 31, 12(1)
addi 1, 1, 16
blr
--
2.53.0
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
