Hi,
On 19/05/2023 16:13, Luca Fancellu wrote:
On 19/05/2023 15:51, Luca Fancellu wrote:
/* Control Registers */
/*
* CPTR_EL2 needs to be written before calling vfp_restore_state, a
* synchronization instruction is expected after the write (isb)
*/
WRITE_SYSREG(n->arch.cptr_el2, CPTR_EL2);
WRITE_SYSREG(n->arch.cpacr, CPACR_EL1);
/*
* This write to sysreg CONTEXTIDR_EL1 ensures we don't hit erratum
* #852523 (Cortex-A57) or #853709 (Cortex-A72).
* I.e DACR32_EL2 is not correctly synchronized.
*/
WRITE_SYSREG(n->arch.contextidr, CONTEXTIDR_EL1);
WRITE_SYSREG(n->arch.tpidr_el0, TPIDR_EL0);
WRITE_SYSREG(n->arch.tpidrro_el0, TPIDRRO_EL0);
WRITE_SYSREG(n->arch.tpidr_el1, TPIDR_EL1);
if ( is_32bit_domain(n->domain) && cpu_has_thumbee )
{
WRITE_SYSREG(n->arch.teecr, TEECR32_EL1);
WRITE_SYSREG(n->arch.teehbr, TEEHBR32_EL1);
}
#ifdef CONFIG_ARM_32
WRITE_CP32(n->arch.joscr, JOSCR);
WRITE_CP32(n->arch.jmcr, JMCR);
#endif
isb();
/* VFP - call vfp_restore_state after writing on CPTR_EL2 + isb */
vfp_restore_state(n);
Maybe I misunderstood your preference, do you want me to put the write to
CPTR_EL2
right before the isb() that precedes vfp_restore_state?
Yes please. Unless there is a reason to keep it "far away". The comments look
good to me.
Ok, a question regarding README.LinuxPrimitives, is it some file taken from an
automated tool?
I am not aware of any automated tools using it. All the re-syncs I have
seen recently were manual.
Because I see there is some kind of structure, how can I know if my syntax is
correct?
There are some commands to help syncing when the file is very similar to
Linux. In your case, I would follow what we did for the atomics as only
some functions are the same.
Cheers,
--
Julien Grall
