[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH 12/26] target/s390x: Move masking of psw.addr to cpu_get_tb_c
Re: [PATCH 12/26] target/s390x: Move masking of psw.addr to cpu_get_tb_cpu_state
Sat, 5 Nov 2022 09:27:07 +1100
Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101 Thunderbird/102.2.2
On 11/4/22 00:42, Ilya Leoshkevich wrote:
On Wed, Oct 05, 2022 at 08:44:07PM -0700, Richard Henderson wrote:
Masking after the fact in s390x_tr_init_disas_context
provides incorrect information to tb_lookup.
Signed-off-by: Richard Henderson <email@example.com>
target/s390x/cpu.h | 13 +++++++------
target/s390x/tcg/translate.c | 6 ------
2 files changed, 7 insertions(+), 12 deletions(-)
How can we end up in a situation where this matters? E.g. if we are in
64-bit mode and execute
we will get a specification exception, and cpu_get_tb_cpu_state() will
not run. And for valid 31-bit addresses masking should be a no-op.
Ah, true. I was mislead by the presence of the code in s390x_tr_init_disas_context.
Perhaps a tcg_debug_assert or just a comment?