Re: [PATCH] tests/tcg/s390x: Fix mvc, mvo and pack tests with Clang

2022-03-01 Thread Richard Henderson
On 2/28/22 23:39, Thomas Huth wrote: These instructions use addressing with a "base address", meaning that if register r0 is used, it is always treated as zero, no matter what value is stored in the register. So we have to make sure not to use register r0 for these instructions in our tests. Ther

Re: [PATCH] tests/tcg/s390x: Fix mvc, mvo and pack tests with Clang

2022-03-01 Thread David Hildenbrand
On 01.03.22 10:39, Thomas Huth wrote: > These instructions use addressing with a "base address", meaning > that if register r0 is used, it is always treated as zero, no matter > what value is stored in the register. So we have to make sure not > to use register r0 for these instructions in our test

[PATCH] tests/tcg/s390x: Fix mvc, mvo and pack tests with Clang

2022-03-01 Thread Thomas Huth
These instructions use addressing with a "base address", meaning that if register r0 is used, it is always treated as zero, no matter what value is stored in the register. So we have to make sure not to use register r0 for these instructions in our tests. There was no problem with GCC so far since