Messages in this thread | | | From | David Laight <> | Subject | RE: [PATCH] x86: work around clang IAS bug referencing __force_order | Date | Sun, 23 Aug 2020 13:31:03 +0000 |
| |
From: Arvind Sankar > Sent: 22 August 2020 22:17 ... > Assuming we don't want to risk removing force_order, I'd suggest > - make it an input/output operand, so it enforces ordering fully. > - either restrict it to gcc < 8, or just provide a proper definition in > some file (maybe arch/x86/kernel/cpu/common.c)?
Is it possible to replace __force_order with a symbol that the linker scripts defines? Or just define __force_order in the linker script to some 'random' constant (eg 0).
ISTM that adding "m"(__force_order) to asm volatile can do no harm. Especially for accesses to CRn and MSRn (etc) which might have obscure side effects.
David
- Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
| |