Re: [kvm-unit-tests PATCH] x86/emulator: Test POP-SS blocking

[Michal Luczaj <mhal@xxxxxxx>]

On 8/24/22 20:32, Sean Christopherson wrote:
> Eh, let's completely skip usermode for code #DBs and not tweak __run_single_step_db_test().
> It's easier to just have a standalone function.

Something like this?

static void test_pop_ss_code_db(bool fep_available)
{
	write_ss(KERNEL_DS);

	write_dr7(DR7_FIXED_1 |
		  DR7_ENABLE_DRx(0) |
		  DR7_EXECUTE_DRx(0) |
		  DR7_LEN_1_DRx(0));

#define POPSS_DB(desc, fep1, fep2)				\
({								\
	unsigned int r;						\
								\
	n = 0;							\
	asm volatile(/* jump to 32-bit code segment */		\
		     "ljmp *1f\n\t"				\
		     "1:\n\t"					\
		     "	.long 2f\n\t"				\
		     "	.word " xstr(KERNEL_CS32) "\n\t"	\
		     /* exercise POP SS blocking */		\
		     ".code32\n\t"				\
		     "2: lea 3f, %0\n\t"			\
		     "mov %0, %%dr0\n\t"			\
		     "push %%ss\n\t"				\
		     fep1 "pop %%ss\n\t"			\
		     fep2 "3: xor %0, %0\n\t"			\
		     /* back to long mode */			\
		     "ljmp %[cs64], $4f\n\t"			\
		     ".code64\n\t"				\
		     "4:"					\
		     : "=R" (r)					\
		     : [cs64] "i" (KERNEL_CS64)			\
		     : "memory");				\
								\
	report(!n && !r, desc ": #DB suppressed after POP SS"); \
})

	POPSS_DB("no fep", "", "");
	if (fep_available) {
		POPSS_DB("fep POP-SS", KVM_FEP, "");
		POPSS_DB("fep XOR", "", KVM_FEP);
		POPSS_DB("fep POP-SS/fep XOR", KVM_FEP, KVM_FEP);
	}

	write_dr7(DR7_FIXED_1);
}

Results:

em_pop_sreg unpatched
PASS: no fep: #DB suppressed after POP SS
FAIL: fep POP-SS: #DB suppressed after POP SS
PASS: fep XOR: #DB suppressed after POP SS
PASS: fep POP-SS/fep XOR: #DB suppressed after POP SS

em_pop_sreg patched
PASS: no fep: #DB suppressed after POP SS
PASS: fep POP-SS: #DB suppressed after POP SS
PASS: fep XOR: #DB suppressed after POP SS
PASS: fep POP-SS/fep XOR: #DB suppressed after POP SS

thanks,
Michal