Quoting Janosch Frank (2022-08-02 09:56:12) > On 8/2/22 09:10, Nico Boehr wrote: > > Quoting Janosch Frank (2022-07-29 10:26:28) > > [...] > >> diff --git a/s390x/snippets/asm/snippet-pv-diag-500.S b/s390x/snippets/asm/snippet-pv-diag-500.S > >> index 8dd66bd9..f4d75388 100644 > >> --- a/s390x/snippets/asm/snippet-pv-diag-500.S > >> +++ b/s390x/snippets/asm/snippet-pv-diag-500.S > >> @@ -8,6 +8,7 @@ > >> * Janosch Frank <frankja@xxxxxxxxxxxxx> > >> */ > >> #include <asm/asm-offsets.h> > >> +#include "macros.S" > >> .section .text > >> > >> /* Clean and pre-load registers that are used for diag 500 */ > >> @@ -21,10 +22,7 @@ lghi %r3, 3 > >> lghi %r4, 4 > >> > >> /* Let's jump to the next label on a PGM */ > >> -xgr %r5, %r5 > >> -stg %r5, GEN_LC_PGM_NEW_PSW > > > > So previously the PSW mask was zero and hence we had 24-bit addressing, no? Now, we have bits 31 and 32 one and hence 64 bit addressing. > > Yes > Also the linker script patch will exchange the mask for an invalid one > so we need to replace both the mask and the address. > > > > > I guess 24-bit addressing is not appropriate here (or at least doesn't matter too much), so I guess this is intended, isn't it? Allrighty, then: Reviewed-by: Nico Boehr <nrb@xxxxxxxxxxxxx>
