Re: [PATCH for-8.2?] target/i386: Fix 32-bit wrapping of pc/eip computation

[Richard Henderson <richard.henderson@linaro.org>]

On 12/25/23 07:49, Michael Tokarev wrote:
> 12.12.2023 20:25, Richard Henderson:
>> In 32-bit mode, pc = eip + cs_base is also 32-bit, and must wrap.
>> Failure to do so results in incorrect memory exceptions to the guest.
>> Before 732d548732ed, this was implicitly done via truncation to
>> target_ulong but only in qemu-system-i386, not qemu-system-x86_64.
>>
>> To fix this, we must add conditional zero-extensions.
>> Since we have to test for 32 vs 64-bit anyway, note that cs_base
>> is always zero in 64-bit mode.
>>
>> Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2022
>> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
>> ---
>>
>> This may be too late for 8.2; if not, then 8.2.1 and 8.1.next.
>> I think I have found all forms of pc <-> eip, but another set
>> of eyes would be appreciated.
> 
> This change breaks trivial 4M edk2 boot - both in 8.2.0 and in
> 8.1.4 (which also has this commit now).
> 
>   qemu-system-x86_64 -machine q35 -no-user-config -nodefaults -display none \
>    -serial stdio \
>    -drive file=/usr/share/OVMF/OVMF_CODE_4M.secboot.fd,if=pflash,format=raw,readonly=on \
>    -drive file=/usr/share/OVMF/OVMF_VARS_4M.ms.fd,if=pflash,format=raw,snapshot=on
> 
> After this change, nothing is printed on the serial console anymore
> (or in vga, whatever). Before that commit, usual edk2 boot sequence
> is seen.
> 
> Nothing has changed with the 2M variant though.

Ack.  Looking at it....


r~