diff options
| author | Iain Sandoe <iain@sandoe.co.uk> | 2021-04-03 10:07:48 +0100 |
|---|---|---|
| committer | Iain Sandoe <iain@sandoe.co.uk> | 2021-04-03 21:28:15 +0100 |
| commit | 89bc1d4e7cdd0b2d012050134ad1d464ec357f0b (patch) | |
| tree | 687fe8d93b89cace4e5a7c98fdf1cf819b8a73f1 /gcc | |
| parent | ad7c6298b6f1cf4eaa70a53af3479fab14018896 (diff) | |
| download | gcc-89bc1d4e7cdd0b2d012050134ad1d464ec357f0b.zip gcc-89bc1d4e7cdd0b2d012050134ad1d464ec357f0b.tar.gz gcc-89bc1d4e7cdd0b2d012050134ad1d464ec357f0b.tar.bz2 | |
Darwin : Fix out-of-bounds access to df_regs_ever_live.
During changes made for LRA (or, perhaps, even before) we omitted
a check that the current register we are working on is a hard reg
before we tried to note its liveness.
A stage 1 built with fsanitize=address catches this, as does any
attempt to build master with clang and -std=c++11.
gcc/ChangeLog:
* config/darwin.c (machopic_legitimize_pic_address): Check
that the current pic register is one of the hard reg set
before setting liveness.
Diffstat (limited to 'gcc')
| -rw-r--r-- | gcc/config/darwin.c | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/gcc/config/darwin.c b/gcc/config/darwin.c index c1086a0..5d17391 100644 --- a/gcc/config/darwin.c +++ b/gcc/config/darwin.c @@ -907,7 +907,7 @@ machopic_legitimize_pic_address (rtx orig, machine_mode mode, rtx reg) pic = reg; } - if (lra_in_progress) + if (lra_in_progress && HARD_REGISTER_P (pic)) df_set_regs_ever_live (REGNO (pic), true); pic_ref = gen_rtx_PLUS (Pmode, pic, machopic_gen_offset (XEXP (orig, 0))); @@ -974,7 +974,8 @@ machopic_legitimize_pic_address (rtx orig, machine_mode mode, rtx reg) emit_move_insn (reg, pic); pic = reg; } - if (lra_in_progress) + + if (lra_in_progress && HARD_REGISTER_P (pic)) df_set_regs_ever_live (REGNO (pic), true); pic_ref = gen_rtx_PLUS (Pmode, pic, |