Version 2 splits lazy do-it-all patch.
Patch 1 has an additional fix, so I dropped the r-b.

r~

Richard Henderson (15):
  linux-user/s390x: Fix sigframe types
  linux-user/s390x: Use uint16_t for signal retcode
  linux-user/s390x: Remove PSW_ADDR_AMODE
  linux-user/s390x: Remove restore_sigregs return value
  linux-user/s390x: Fix trace in restore_regs
  linux-user/s390x: Fix sigcontext sregs value
  linux-user/s390x: Use tswap_sigset in setup_rt_frame
  linux-user/s390x: Tidy save_sigregs
  linux-user/s390x: Clean up single-use gotos in signal.c
  linux-user/s390x: Set psw.mask properly for the signal handler
  linux-user/s390x: Add stub sigframe argument for last_break
  linux-user/s390x: Fix frame_addr corruption in setup_frame
  linux-user/s390x: Add build asserts for sigset sizes
  linux-user/s390x: Clean up signal.c
  linux-user/s390x: Handle vector regs in signal stack

 linux-user/s390x/signal.c | 280 +++++++++++++++++++++++---------------
 1 file changed, 170 insertions(+), 110 deletions(-)

-- 
2.25.1

On 4/28/21 12:33 PM, Richard Henderson wrote:
> Version 2 splits lazy do-it-all patch.
> Patch 1 has an additional fix, so I dropped the r-b.

... and I realized as I hit send that this depends on the 
target_restore_altstack cleanup that's part of

https://patchew.org/QEMU/20210426025334.1168495-1-richard.henderson@linaro.org/

For avoidance of doubt:
https://gitlab.com/rth7680/qemu/-/tree/fix-signals


r~

On Wed, 28 Apr 2021 12:33:53 -0700
Richard Henderson <richard.henderson@linaro.org> wrote:

> Version 2 splits lazy do-it-all patch.
> Patch 1 has an additional fix, so I dropped the r-b.
> 
> r~
> 
> Richard Henderson (15):
>   linux-user/s390x: Fix sigframe types
>   linux-user/s390x: Use uint16_t for signal retcode
>   linux-user/s390x: Remove PSW_ADDR_AMODE
>   linux-user/s390x: Remove restore_sigregs return value
>   linux-user/s390x: Fix trace in restore_regs
>   linux-user/s390x: Fix sigcontext sregs value
>   linux-user/s390x: Use tswap_sigset in setup_rt_frame
>   linux-user/s390x: Tidy save_sigregs
>   linux-user/s390x: Clean up single-use gotos in signal.c
>   linux-user/s390x: Set psw.mask properly for the signal handler
>   linux-user/s390x: Add stub sigframe argument for last_break
>   linux-user/s390x: Fix frame_addr corruption in setup_frame
>   linux-user/s390x: Add build asserts for sigset sizes
>   linux-user/s390x: Clean up signal.c
>   linux-user/s390x: Handle vector regs in signal stack
> 
>  linux-user/s390x/signal.c | 280 +++++++++++++++++++++++---------------
>  1 file changed, 170 insertions(+), 110 deletions(-)
> 

I assume the route-to-upstream for this is through the linux-user tree
and not the s390x tree, right?

Le 06/05/2021 à 13:54, Cornelia Huck a écrit :
> On Wed, 28 Apr 2021 12:33:53 -0700
> Richard Henderson <richard.henderson@linaro.org> wrote:
> 
>> Version 2 splits lazy do-it-all patch.
>> Patch 1 has an additional fix, so I dropped the r-b.
>>
>> r~
>>
>> Richard Henderson (15):
>>   linux-user/s390x: Fix sigframe types
>>   linux-user/s390x: Use uint16_t for signal retcode
>>   linux-user/s390x: Remove PSW_ADDR_AMODE
>>   linux-user/s390x: Remove restore_sigregs return value
>>   linux-user/s390x: Fix trace in restore_regs
>>   linux-user/s390x: Fix sigcontext sregs value
>>   linux-user/s390x: Use tswap_sigset in setup_rt_frame
>>   linux-user/s390x: Tidy save_sigregs
>>   linux-user/s390x: Clean up single-use gotos in signal.c
>>   linux-user/s390x: Set psw.mask properly for the signal handler
>>   linux-user/s390x: Add stub sigframe argument for last_break
>>   linux-user/s390x: Fix frame_addr corruption in setup_frame
>>   linux-user/s390x: Add build asserts for sigset sizes
>>   linux-user/s390x: Clean up signal.c
>>   linux-user/s390x: Handle vector regs in signal stack
>>
>>  linux-user/s390x/signal.c | 280 +++++++++++++++++++++++---------------
>>  1 file changed, 170 insertions(+), 110 deletions(-)
>>
> 
> I assume the route-to-upstream for this is through the linux-user tree
> and not the s390x tree, right?
> 

Yes, and I'll try to do a PR soon.

Thanks,
Laurent

Le 28/04/2021 à 21:33, Richard Henderson a écrit :
> Version 2 splits lazy do-it-all patch.
> Patch 1 has an additional fix, so I dropped the r-b.
> 
> r~
> 
> Richard Henderson (15):
>   linux-user/s390x: Fix sigframe types
>   linux-user/s390x: Use uint16_t for signal retcode
>   linux-user/s390x: Remove PSW_ADDR_AMODE
>   linux-user/s390x: Remove restore_sigregs return value
>   linux-user/s390x: Fix trace in restore_regs
>   linux-user/s390x: Fix sigcontext sregs value
>   linux-user/s390x: Use tswap_sigset in setup_rt_frame
>   linux-user/s390x: Tidy save_sigregs
>   linux-user/s390x: Clean up single-use gotos in signal.c
>   linux-user/s390x: Set psw.mask properly for the signal handler
>   linux-user/s390x: Add stub sigframe argument for last_break
>   linux-user/s390x: Fix frame_addr corruption in setup_frame
>   linux-user/s390x: Add build asserts for sigset sizes
>   linux-user/s390x: Clean up signal.c
>   linux-user/s390x: Handle vector regs in signal stack
> 
>  linux-user/s390x/signal.c | 280 +++++++++++++++++++++++---------------
>  1 file changed, 170 insertions(+), 110 deletions(-)
> 

Applied to my linux-user-for-6.1 branch.

Thanks,
Laurent

On 28.04.21 21:33, Richard Henderson wrote:
> Version 2 splits lazy do-it-all patch.

Yap, that helped a lot :)


-- 
Thanks,

David / dhildenb