| Message ID | 20161109180706.GA8015@potion (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hi Radim,
[auto build test ERROR on kvm/linux-next]
[also build test ERROR on v4.9-rc4 next-20161109]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
url: https://github.com/0day-ci/linux/commits/Radim-Kr-m/KVM-x86-emulate-FXSAVE-and-FXRSTOR/20161110-021048
base: https://git.kernel.org/pub/scm/virt/kvm/kvm.git linux-next
config: x86_64-randconfig-x018-201645 (attached as .config)
compiler: gcc-6 (Debian 6.2.0-3) 6.2.0 20160901
reproduce:
# save the attached .config to linux build tree
make ARCH=x86_64
All errors (new ones prefixed by >>):
arch/x86/kvm/emulate.c: In function 'em_fxsave':
>> arch/x86/kvm/emulate.c:3910:7: error: implicit declaration of function 'asm_safe' [-Werror=implicit-function-declaration]
rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state));
^~~~~~~~
>> arch/x86/kvm/emulate.c:3910:32: error: expected expression before ',' token
rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state));
^
>> arch/x86/kvm/emulate.c:3910:35: error: 'fx' undeclared (first use in this function)
rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state));
^~
arch/x86/kvm/emulate.c:3910:35: note: each undeclared identifier is reported only once for each function it appears in
arch/x86/kvm/emulate.c: In function 'fxrstor_fixup':
arch/x86/kvm/emulate.c:3931:32: error: expected expression before ',' token
rc = asm_safe("fxsave %[fx]", , [fx] "+m"(old));
^
arch/x86/kvm/emulate.c:3931:35: error: 'fx' undeclared (first use in this function)
rc = asm_safe("fxsave %[fx]", , [fx] "+m"(old));
^~
arch/x86/kvm/emulate.c: In function 'em_fxrstor':
>> arch/x86/kvm/emulate.c:3977:34: error: expected expression before ':' token
rc = asm_safe("fxrstor %[fx]", : [fx] "m"(fx_state));
^
arch/x86/kvm/emulate.c: At top level:
>> arch/x86/kvm/emulate.c:4336:12: error: 'Aligned16' undeclared here (not in a function)
I(ModRM | Aligned16, em_fxsave),
^
arch/x86/kvm/emulate.c:4185:31: note: in definition of macro 'I'
#define I(_f, _e) { .flags = (_f), .u.execute = (_e) }
^~
cc1: some warnings being treated as errors
vim +/asm_safe +3910 arch/x86/kvm/emulate.c
3904 rc = check_fxsr(ctxt);
3905 if (rc != X86EMUL_CONTINUE)
3906 return rc;
3907
3908 ctxt->ops->get_fpu(ctxt);
3909
> 3910 rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state));
3911
3912 ctxt->ops->put_fpu(ctxt);
3913
3914 if (rc != X86EMUL_CONTINUE)
3915 return rc;
3916
3917 if (ctxt->ops->get_cr(ctxt, 4) & X86_CR4_OSFXSR)
3918 size = offsetof(struct fxregs_state, xmm_space[8 * 16/4]);
3919 else
3920 size = offsetof(struct fxregs_state, xmm_space[0]);
3921
3922 return segmented_write(ctxt, ctxt->memop.addr.mem, &fx_state, size);
3923 }
3924
3925 static int fxrstor_fixup(struct x86_emulate_ctxt *ctxt,
3926 struct fxregs_state *new)
3927 {
3928 int rc = X86EMUL_CONTINUE;
3929 struct fxregs_state old;
3930
> 3931 rc = asm_safe("fxsave %[fx]", , [fx] "+m"(old));
3932 if (rc != X86EMUL_CONTINUE)
3933 return rc;
3934
3935 /*
3936 * 64 bit host will restore XMM 8-15, which is not correct on non-64
3937 * bit guests. Load the current values in order to preserve 64 bit
3938 * XMMs after fxrstor.
3939 */
3940 #ifdef CONFIG_X86_64
3941 /* XXX: accessing XMM 8-15 very awkwardly */
3942 memcpy(&new->xmm_space[8 * 16/4], &old.xmm_space[8 * 16/4], 8 * 16);
3943 #endif
3944
3945 /*
3946 * Hardware doesn't save and restore XMM 0-7 without CR4.OSFXSR, but
3947 * does save and restore MXCSR.
3948 */
3949 if (!(ctxt->ops->get_cr(ctxt, 4) & X86_CR4_OSFXSR))
3950 memcpy(new->xmm_space, old.xmm_space, 8 * 16);
3951
3952 return rc;
3953 }
3954
3955 static int em_fxrstor(struct x86_emulate_ctxt *ctxt)
3956 {
3957 struct fxregs_state fx_state;
3958 int rc;
3959
3960 rc = check_fxsr(ctxt);
3961 if (rc != X86EMUL_CONTINUE)
3962 return rc;
3963
3964 rc = segmented_read(ctxt, ctxt->memop.addr.mem, &fx_state, 512);
3965 if (rc != X86EMUL_CONTINUE)
3966 return rc;
3967
3968 if (fx_state.mxcsr >> 16)
3969 return emulate_gp(ctxt, 0);
3970
3971 ctxt->ops->get_fpu(ctxt);
3972
3973 if (ctxt->mode < X86EMUL_MODE_PROT64)
3974 rc = fxrstor_fixup(ctxt, &fx_state);
3975
3976 if (rc == X86EMUL_CONTINUE)
> 3977 rc = asm_safe("fxrstor %[fx]", : [fx] "m"(fx_state));
3978
3979 ctxt->ops->put_fpu(ctxt);
3980
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
2016-11-10 02:42+0800, kbuild test robot: > Hi Radim, And I even replied to the series, so this would happen ... Btw. would the tool recognize it if the header was [PATCH v4 4/4]? > [auto build test ERROR on kvm/linux-next] > [also build test ERROR on v4.9-rc4 next-20161109] > [if your patch is applied to the wrong git tree, please drop us a note to help improve the system] > > url: https://github.com/0day-ci/linux/commits/Radim-Kr-m/KVM-x86-emulate-FXSAVE-and-FXRSTOR/20161110-021048 > base: https://git.kernel.org/pub/scm/virt/kvm/kvm.git linux-next > config: x86_64-randconfig-x018-201645 (attached as .config) > compiler: gcc-6 (Debian 6.2.0-3) 6.2.0 20160901 > reproduce: > # save the attached .config to linux build tree > make ARCH=x86_64 > > All errors (new ones prefixed by >>): > > arch/x86/kvm/emulate.c: In function 'em_fxsave': > >> arch/x86/kvm/emulate.c:3910:7: error: implicit declaration of function 'asm_safe' [-Werror=implicit-function-declaration] > rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state)); > ^~~~~~~~ > >> arch/x86/kvm/emulate.c:3910:32: error: expected expression before ',' token > rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state)); > ^ > >> arch/x86/kvm/emulate.c:3910:35: error: 'fx' undeclared (first use in this function) > rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state)); > ^~ > arch/x86/kvm/emulate.c:3910:35: note: each undeclared identifier is reported only once for each function it appears in > arch/x86/kvm/emulate.c: In function 'fxrstor_fixup': > arch/x86/kvm/emulate.c:3931:32: error: expected expression before ',' token > rc = asm_safe("fxsave %[fx]", , [fx] "+m"(old)); > ^ > arch/x86/kvm/emulate.c:3931:35: error: 'fx' undeclared (first use in this function) > rc = asm_safe("fxsave %[fx]", , [fx] "+m"(old)); > ^~ > arch/x86/kvm/emulate.c: In function 'em_fxrstor': > >> arch/x86/kvm/emulate.c:3977:34: error: expected expression before ':' token > rc = asm_safe("fxrstor %[fx]", : [fx] "m"(fx_state)); > ^ > arch/x86/kvm/emulate.c: At top level: > >> arch/x86/kvm/emulate.c:4336:12: error: 'Aligned16' undeclared here (not in a function) > I(ModRM | Aligned16, em_fxsave), > ^ > arch/x86/kvm/emulate.c:4185:31: note: in definition of macro 'I' > #define I(_f, _e) { .flags = (_f), .u.execute = (_e) } > ^~ > cc1: some warnings being treated as errors > > vim +/asm_safe +3910 arch/x86/kvm/emulate.c > > 3904 rc = check_fxsr(ctxt); > 3905 if (rc != X86EMUL_CONTINUE) > 3906 return rc; > 3907 > 3908 ctxt->ops->get_fpu(ctxt); > 3909 > > 3910 rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state)); > 3911 > 3912 ctxt->ops->put_fpu(ctxt); > 3913 > 3914 if (rc != X86EMUL_CONTINUE) > 3915 return rc; > 3916 > 3917 if (ctxt->ops->get_cr(ctxt, 4) & X86_CR4_OSFXSR) > 3918 size = offsetof(struct fxregs_state, xmm_space[8 * 16/4]); > 3919 else > 3920 size = offsetof(struct fxregs_state, xmm_space[0]); > 3921 > 3922 return segmented_write(ctxt, ctxt->memop.addr.mem, &fx_state, size); > 3923 } > 3924 > 3925 static int fxrstor_fixup(struct x86_emulate_ctxt *ctxt, > 3926 struct fxregs_state *new) > 3927 { > 3928 int rc = X86EMUL_CONTINUE; > 3929 struct fxregs_state old; > 3930 > > 3931 rc = asm_safe("fxsave %[fx]", , [fx] "+m"(old)); > 3932 if (rc != X86EMUL_CONTINUE) > 3933 return rc; > 3934 > 3935 /* > 3936 * 64 bit host will restore XMM 8-15, which is not correct on non-64 > 3937 * bit guests. Load the current values in order to preserve 64 bit > 3938 * XMMs after fxrstor. > 3939 */ > 3940 #ifdef CONFIG_X86_64 > 3941 /* XXX: accessing XMM 8-15 very awkwardly */ > 3942 memcpy(&new->xmm_space[8 * 16/4], &old.xmm_space[8 * 16/4], 8 * 16); > 3943 #endif > 3944 > 3945 /* > 3946 * Hardware doesn't save and restore XMM 0-7 without CR4.OSFXSR, but > 3947 * does save and restore MXCSR. > 3948 */ > 3949 if (!(ctxt->ops->get_cr(ctxt, 4) & X86_CR4_OSFXSR)) > 3950 memcpy(new->xmm_space, old.xmm_space, 8 * 16); > 3951 > 3952 return rc; > 3953 } > 3954 > 3955 static int em_fxrstor(struct x86_emulate_ctxt *ctxt) > 3956 { > 3957 struct fxregs_state fx_state; > 3958 int rc; > 3959 > 3960 rc = check_fxsr(ctxt); > 3961 if (rc != X86EMUL_CONTINUE) > 3962 return rc; > 3963 > 3964 rc = segmented_read(ctxt, ctxt->memop.addr.mem, &fx_state, 512); > 3965 if (rc != X86EMUL_CONTINUE) > 3966 return rc; > 3967 > 3968 if (fx_state.mxcsr >> 16) > 3969 return emulate_gp(ctxt, 0); > 3970 > 3971 ctxt->ops->get_fpu(ctxt); > 3972 > 3973 if (ctxt->mode < X86EMUL_MODE_PROT64) > 3974 rc = fxrstor_fixup(ctxt, &fx_state); > 3975 > 3976 if (rc == X86EMUL_CONTINUE) > > 3977 rc = asm_safe("fxrstor %[fx]", : [fx] "m"(fx_state)); > 3978 > 3979 ctxt->ops->put_fpu(ctxt); > 3980 > > --- > 0-DAY kernel test infrastructure Open Source Technology Center > https://lists.01.org/pipermail/kbuild-all Intel Corporation -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
CC Xiaolong. On Wed, Nov 09, 2016 at 07:46:28PM +0100, Radim Krčmář wrote: >2016-11-10 02:42+0800, kbuild test robot: >> Hi Radim, > >And I even replied to the series, so this would happen ... >Btw. would the tool recognize it if the header was [PATCH v4 4/4]? > >> [auto build test ERROR on kvm/linux-next] >> [also build test ERROR on v4.9-rc4 next-20161109] >> [if your patch is applied to the wrong git tree, please drop us a note to help improve the system] >> >> url: https://github.com/0day-ci/linux/commits/Radim-Kr-m/KVM-x86-emulate-FXSAVE-and-FXRSTOR/20161110-021048 >> base: https://git.kernel.org/pub/scm/virt/kvm/kvm.git linux-next >> config: x86_64-randconfig-x018-201645 (attached as .config) >> compiler: gcc-6 (Debian 6.2.0-3) 6.2.0 20160901 >> reproduce: >> # save the attached .config to linux build tree >> make ARCH=x86_64 >> >> All errors (new ones prefixed by >>): >> >> arch/x86/kvm/emulate.c: In function 'em_fxsave': >> >> arch/x86/kvm/emulate.c:3910:7: error: implicit declaration of function 'asm_safe' [-Werror=implicit-function-declaration] >> rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state)); >> ^~~~~~~~ >> >> arch/x86/kvm/emulate.c:3910:32: error: expected expression before ',' token >> rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state)); >> ^ >> >> arch/x86/kvm/emulate.c:3910:35: error: 'fx' undeclared (first use in this function) >> rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state)); >> ^~ >> arch/x86/kvm/emulate.c:3910:35: note: each undeclared identifier is reported only once for each function it appears in >> arch/x86/kvm/emulate.c: In function 'fxrstor_fixup': >> arch/x86/kvm/emulate.c:3931:32: error: expected expression before ',' token >> rc = asm_safe("fxsave %[fx]", , [fx] "+m"(old)); >> ^ >> arch/x86/kvm/emulate.c:3931:35: error: 'fx' undeclared (first use in this function) >> rc = asm_safe("fxsave %[fx]", , [fx] "+m"(old)); >> ^~ >> arch/x86/kvm/emulate.c: In function 'em_fxrstor': >> >> arch/x86/kvm/emulate.c:3977:34: error: expected expression before ':' token >> rc = asm_safe("fxrstor %[fx]", : [fx] "m"(fx_state)); >> ^ >> arch/x86/kvm/emulate.c: At top level: >> >> arch/x86/kvm/emulate.c:4336:12: error: 'Aligned16' undeclared here (not in a function) >> I(ModRM | Aligned16, em_fxsave), >> ^ >> arch/x86/kvm/emulate.c:4185:31: note: in definition of macro 'I' >> #define I(_f, _e) { .flags = (_f), .u.execute = (_e) } >> ^~ >> cc1: some warnings being treated as errors >> >> vim +/asm_safe +3910 arch/x86/kvm/emulate.c >> >> 3904 rc = check_fxsr(ctxt); >> 3905 if (rc != X86EMUL_CONTINUE) >> 3906 return rc; >> 3907 >> 3908 ctxt->ops->get_fpu(ctxt); >> 3909 >> > 3910 rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state)); >> 3911 >> 3912 ctxt->ops->put_fpu(ctxt); >> 3913 >> 3914 if (rc != X86EMUL_CONTINUE) >> 3915 return rc; >> 3916 >> 3917 if (ctxt->ops->get_cr(ctxt, 4) & X86_CR4_OSFXSR) >> 3918 size = offsetof(struct fxregs_state, xmm_space[8 * 16/4]); >> 3919 else >> 3920 size = offsetof(struct fxregs_state, xmm_space[0]); >> 3921 >> 3922 return segmented_write(ctxt, ctxt->memop.addr.mem, &fx_state, size); >> 3923 } >> 3924 >> 3925 static int fxrstor_fixup(struct x86_emulate_ctxt *ctxt, >> 3926 struct fxregs_state *new) >> 3927 { >> 3928 int rc = X86EMUL_CONTINUE; >> 3929 struct fxregs_state old; >> 3930 >> > 3931 rc = asm_safe("fxsave %[fx]", , [fx] "+m"(old)); >> 3932 if (rc != X86EMUL_CONTINUE) >> 3933 return rc; >> 3934 >> 3935 /* >> 3936 * 64 bit host will restore XMM 8-15, which is not correct on non-64 >> 3937 * bit guests. Load the current values in order to preserve 64 bit >> 3938 * XMMs after fxrstor. >> 3939 */ >> 3940 #ifdef CONFIG_X86_64 >> 3941 /* XXX: accessing XMM 8-15 very awkwardly */ >> 3942 memcpy(&new->xmm_space[8 * 16/4], &old.xmm_space[8 * 16/4], 8 * 16); >> 3943 #endif >> 3944 >> 3945 /* >> 3946 * Hardware doesn't save and restore XMM 0-7 without CR4.OSFXSR, but >> 3947 * does save and restore MXCSR. >> 3948 */ >> 3949 if (!(ctxt->ops->get_cr(ctxt, 4) & X86_CR4_OSFXSR)) >> 3950 memcpy(new->xmm_space, old.xmm_space, 8 * 16); >> 3951 >> 3952 return rc; >> 3953 } >> 3954 >> 3955 static int em_fxrstor(struct x86_emulate_ctxt *ctxt) >> 3956 { >> 3957 struct fxregs_state fx_state; >> 3958 int rc; >> 3959 >> 3960 rc = check_fxsr(ctxt); >> 3961 if (rc != X86EMUL_CONTINUE) >> 3962 return rc; >> 3963 >> 3964 rc = segmented_read(ctxt, ctxt->memop.addr.mem, &fx_state, 512); >> 3965 if (rc != X86EMUL_CONTINUE) >> 3966 return rc; >> 3967 >> 3968 if (fx_state.mxcsr >> 16) >> 3969 return emulate_gp(ctxt, 0); >> 3970 >> 3971 ctxt->ops->get_fpu(ctxt); >> 3972 >> 3973 if (ctxt->mode < X86EMUL_MODE_PROT64) >> 3974 rc = fxrstor_fixup(ctxt, &fx_state); >> 3975 >> 3976 if (rc == X86EMUL_CONTINUE) >> > 3977 rc = asm_safe("fxrstor %[fx]", : [fx] "m"(fx_state)); >> 3978 >> 3979 ctxt->ops->put_fpu(ctxt); >> 3980 >> >> --- >> 0-DAY kernel test infrastructure Open Source Technology Center >> https://lists.01.org/pipermail/kbuild-all Intel Corporation -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c index 6af3cac6ec89..7d4f9b7f06ee 100644 --- a/arch/x86/kvm/emulate.c +++ b/arch/x86/kvm/emulate.c @@ -3883,6 +3883,131 @@ static int em_movsxd(struct x86_emulate_ctxt *ctxt) return X86EMUL_CONTINUE; } +static int check_fxsr(struct x86_emulate_ctxt *ctxt) +{ + u32 eax = 1, ebx, ecx = 0, edx; + + ctxt->ops->get_cpuid(ctxt, &eax, &ebx, &ecx, &edx); + if (!(edx & FFL(FXSR))) + return emulate_ud(ctxt); + + if (ctxt->ops->get_cr(ctxt, 0) & (X86_CR0_TS | X86_CR0_EM)) + return emulate_nm(ctxt); + + /* + * Don't emulate a case that should never be hit, instead of working + * around a lack of fxsave64/fxrstor64 on old compilers. + */ + if (ctxt->mode >= X86EMUL_MODE_PROT64) + return X86EMUL_UNHANDLEABLE; + + return X86EMUL_CONTINUE; +} + +/* + * FXSAVE and FXRSTOR have 4 different formats depending on execution mode, + * 1) 16 bit mode + * 2) 32 bit mode + * - like (1), but FIP and FDP (foo) are only 16 bit. At least Intel CPUs + * preserve whole 32 bit values, though, so (1) and (2) are the same wrt. + * save and restore + * 3) 64-bit mode with REX.W prefix + * - like (2), but XMM 8-15 are being saved and restored + * 4) 64-bit mode without REX.W prefix + * - like (3), but FIP and FDP are 64 bit + * + * Emulation uses (3) for (1) and (2) and preserves XMM 8-15 to reach the + * desired result. (4) is not emulated. + * + * Note: Guest and host CPUID.(EAX=07H,ECX=0H):EBX[bit 13] (deprecate FPU CS + * and FPU DS) should match. + */ +static int em_fxsave(struct x86_emulate_ctxt *ctxt) +{ + struct fxregs_state fx_state; + size_t size; + int rc; + + rc = check_fxsr(ctxt); + if (rc != X86EMUL_CONTINUE) + return rc; + + ctxt->ops->get_fpu(ctxt); + + rc = asm_safe("fxsave %[fx]", , [fx] "+m"(fx_state)); + + ctxt->ops->put_fpu(ctxt); + + if (rc != X86EMUL_CONTINUE) + return rc; + + if (ctxt->ops->get_cr(ctxt, 4) & X86_CR4_OSFXSR) + size = offsetof(struct fxregs_state, xmm_space[8 * 16/4]); + else + size = offsetof(struct fxregs_state, xmm_space[0]); + + return segmented_write(ctxt, ctxt->memop.addr.mem, &fx_state, size); +} + +static int fxrstor_fixup(struct x86_emulate_ctxt *ctxt, + struct fxregs_state *new) +{ + int rc = X86EMUL_CONTINUE; + struct fxregs_state old; + + rc = asm_safe("fxsave %[fx]", , [fx] "+m"(old)); + if (rc != X86EMUL_CONTINUE) + return rc; + + /* + * 64 bit host will restore XMM 8-15, which is not correct on non-64 + * bit guests. Load the current values in order to preserve 64 bit + * XMMs after fxrstor. + */ +#ifdef CONFIG_X86_64 + /* XXX: accessing XMM 8-15 very awkwardly */ + memcpy(&new->xmm_space[8 * 16/4], &old.xmm_space[8 * 16/4], 8 * 16); +#endif + + /* + * Hardware doesn't save and restore XMM 0-7 without CR4.OSFXSR, but + * does save and restore MXCSR. + */ + if (!(ctxt->ops->get_cr(ctxt, 4) & X86_CR4_OSFXSR)) + memcpy(new->xmm_space, old.xmm_space, 8 * 16); + + return rc; +} + +static int em_fxrstor(struct x86_emulate_ctxt *ctxt) +{ + struct fxregs_state fx_state; + int rc; + + rc = check_fxsr(ctxt); + if (rc != X86EMUL_CONTINUE) + return rc; + + rc = segmented_read(ctxt, ctxt->memop.addr.mem, &fx_state, 512); + if (rc != X86EMUL_CONTINUE) + return rc; + + if (fx_state.mxcsr >> 16) + return emulate_gp(ctxt, 0); + + ctxt->ops->get_fpu(ctxt); + + if (ctxt->mode < X86EMUL_MODE_PROT64) + rc = fxrstor_fixup(ctxt, &fx_state); + + if (rc == X86EMUL_CONTINUE) + rc = asm_safe("fxrstor %[fx]", : [fx] "m"(fx_state)); + + ctxt->ops->put_fpu(ctxt); + + return rc; +} + static bool valid_cr(int nr) { switch (nr) { @@ -4235,7 +4360,9 @@ static const struct gprefix pfx_0f_ae_7 = { }; static const struct group_dual group15 = { { - N, N, N, N, N, N, N, GP(0, &pfx_0f_ae_7), + I(ModRM | Aligned16, em_fxsave), + I(ModRM | Aligned16, em_fxrstor), + N, N, N, N, N, GP(0, &pfx_0f_ae_7), }, { N, N, N, N, N, N, N, N, } };
Internal errors were reported on 16 bit fxsave and fxrstor with ipxe. Old Intels don't have unrestricted_guest, so we have to emulate them. The patch takes advantage of the hardware implementation. AMD and Intel differ in saving and restoring other fields in first 32 bytes. A test wrote 0xff to the fxsave area, 0 to upper bits of MCSXR in the fxsave area, executed fxrstor, rewrote the fxsave area to 0xee, and executed fxsave: Intel (Nehalem): 7f 1f 7f 7f ff 00 ff 07 ff ff ff ff ff ff 00 00 ff ff ff ff ff ff 00 00 ff ff 00 00 ff ff 00 00 Intel (Haswell -- deprecated FPU CS and FPU DS): 7f 1f 7f 7f ff 00 ff 07 ff ff ff ff 00 00 00 00 ff ff ff ff 00 00 00 00 ff ff 00 00 ff ff 00 00 AMD (Opteron 2300-series): 7f 1f 7f 7f ff 00 ee ee ee ee ee ee ee ee ee ee ee ee ee ee ee ee ee ee ff ff 00 00 ff ff 02 00 fxsave/fxrstor will only be emulated on early Intels, so KVM can't do much to improve the situation. Signed-off-by: Radim Krčmář <rkrcmar@redhat.com> --- v4: - add support for CR4.OSFXSR [Paolo] I tested hardware implementation on AMD and Intel: neither of them save XMM 0-7, but both save and restore MCSXR without CR4.OSFXSR, so KVM behaves like that as well. v3: - remove fxsave64 and extra colons at the end of asm to make old GCCs happy (fxsave64 could have been implemented using other nmemonics, but there is no point when it won't be used + removing it makes the code nicer.) v2: - throws #GP to the guest when reserved MXCSR are set [Nadav] - returns internal emulation error if an exception is hit during execution - preserves XMM 8-15 --- arch/x86/kvm/emulate.c | 129 ++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 128 insertions(+), 1 deletion(-)