On 29/12/2018 09:52, Pu Wen wrote: > On 2018/12/28 20:40, Andrew Cooper wrote: >> AMD hardware before Zen doesn't safe/restore the FPU error pointers >> unless an unmasked FPU exception is pending. Zen processors have a >> feature bit indicating that this (mis)behaviour no longer exists. >> >> Express the common logic in terms of cpu_bug_fpu_err_ptr as Hygon >> processors (being Zen derivatives) won't inherit this behaviour. > ... >> diff --git a/xen/arch/x86/xstate.c b/xen/arch/x86/xstate.c >> index 15edd5d..7ca5684 100644 >> --- a/xen/arch/x86/xstate.c >> +++ b/xen/arch/x86/xstate.c >> @@ -369,15 +369,13 @@ void xrstor(struct vcpu *v, uint64_t mask) >> unsigned int faults, prev_faults; >> >> /* >> - * AMD CPUs don't save/restore FDP/FIP/FOP unless an exception >> + * Some CPUs don't save/restore FDP/FIP/FOP unless an exception >> * is pending. Clear the x87 state here by setting it to fixed >> * values. The hypervisor data segment can be sometimes 0 and >> * sometimes new user value. Both should be ok. Use the FPU saved >> * data block as a safe address because it should be in L1. >> */ >> - if ( (mask & ptr->xsave_hdr.xstate_bv & X86_XCR0_FP) && >> - !(ptr->fpu_sse.fsw & ~ptr->fpu_sse.fcw & 0x003f) && >> - boot_cpu_data.x86_vendor == X86_VENDOR_AMD ) >> + if ( cpu_bug_fpu_err_ptr ) >> asm volatile ( "fnclex\n\t" /* clear exceptions */ >> "ffree %%st(7)\n\t" /* clear stack tag */ >> "fildl %0" /* load to clear state */ > It seems that the functions fpu_fxrstor and fpu_fxsave have the same > problem, even though they are not called on Hygon Dhyana platform. > Should them be fixed in the same way?
Oops yes - they should be fixed. They will get used if the user passes "xsave=0" on the command line. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
