eal/x86: fix some CPU extended features
Checks
Commit Message
Caught while checking CPUID related stuff in OVS.
According to [1], for Structured Extended Feature Flags Enumeration Leaf
(EAX = 0x07H, ECX = 0):
- BMI1 is associated to EBX, bit 3 (was incorrectly 2),
- SMEP is associated to EBX, bit 7 (was incorrectly 6),
- BMI2 is associated to EBX, bit 8 (was incorrectly 7),
- ERMS is associated to EBX, bit 9 (was incorrectly 8),
This patch then sorts the rest of the extended features (leaf 0) for
readability.
1: https://software.intel.com/sites/default/files/managed/c5/15/architecture-instruction-set-extensions-programming-reference.pdf
Fixes: af75078fece3 ("first public release")
Cc: stable@dpdk.org
Signed-off-by: David Marchand <david.marchand@redhat.com>
---
lib/eal/x86/rte_cpuflags.c | 46 +++++++++++++++++++-------------------
1 file changed, 23 insertions(+), 23 deletions(-)
Comments
On Wed, Oct 06, 2021 at 09:20:08PM +0200, David Marchand wrote:
> Caught while checking CPUID related stuff in OVS.
>
> According to [1], for Structured Extended Feature Flags Enumeration Leaf
> (EAX = 0x07H, ECX = 0):
>
> - BMI1 is associated to EBX, bit 3 (was incorrectly 2),
> - SMEP is associated to EBX, bit 7 (was incorrectly 6),
> - BMI2 is associated to EBX, bit 8 (was incorrectly 7),
> - ERMS is associated to EBX, bit 9 (was incorrectly 8),
>
> This patch then sorts the rest of the extended features (leaf 0) for
> readability.
>
> 1: https://software.intel.com/sites/default/files/managed/c5/15/architecture-instruction-set-extensions-programming-reference.pdf
>
> Fixes: af75078fece3 ("first public release")
Wow, I'm surprised at it being incorrect this long, but checking the
original commit of code, it does indeed seem true that this is the correct
fixes line!
> Cc: stable@dpdk.org
>
> Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Ideally, I think the fixes should go in a separate patch from the
reordering, but I don't feel strongly about it so this is ok as either one
or two patches IMHO.
> ---
> lib/eal/x86/rte_cpuflags.c | 46 +++++++++++++++++++-------------------
> 1 file changed, 23 insertions(+), 23 deletions(-)
>
> diff --git a/lib/eal/x86/rte_cpuflags.c b/lib/eal/x86/rte_cpuflags.c
> index d339734a8c..d6b518251b 100644
> --- a/lib/eal/x86/rte_cpuflags.c
> +++ b/lib/eal/x86/rte_cpuflags.c
> @@ -100,18 +100,36 @@ const struct feature_entry rte_cpu_feature_table[] = {
> FEAT_DEF(ENERGY_EFF, 0x00000006, 0, RTE_REG_ECX, 3)
>
> FEAT_DEF(FSGSBASE, 0x00000007, 0, RTE_REG_EBX, 0)
> - FEAT_DEF(BMI1, 0x00000007, 0, RTE_REG_EBX, 2)
> + FEAT_DEF(BMI1, 0x00000007, 0, RTE_REG_EBX, 3)
> FEAT_DEF(HLE, 0x00000007, 0, RTE_REG_EBX, 4)
> FEAT_DEF(AVX2, 0x00000007, 0, RTE_REG_EBX, 5)
> - FEAT_DEF(SMEP, 0x00000007, 0, RTE_REG_EBX, 6)
> - FEAT_DEF(BMI2, 0x00000007, 0, RTE_REG_EBX, 7)
> - FEAT_DEF(ERMS, 0x00000007, 0, RTE_REG_EBX, 8)
> + FEAT_DEF(SMEP, 0x00000007, 0, RTE_REG_EBX, 7)
> + FEAT_DEF(BMI2, 0x00000007, 0, RTE_REG_EBX, 8)
> + FEAT_DEF(ERMS, 0x00000007, 0, RTE_REG_EBX, 9)
> FEAT_DEF(INVPCID, 0x00000007, 0, RTE_REG_EBX, 10)
> FEAT_DEF(RTM, 0x00000007, 0, RTE_REG_EBX, 11)
> FEAT_DEF(AVX512F, 0x00000007, 0, RTE_REG_EBX, 16)
> + FEAT_DEF(AVX512DQ, 0x00000007, 0, RTE_REG_EBX, 17)
> FEAT_DEF(RDSEED, 0x00000007, 0, RTE_REG_EBX, 18)
> + FEAT_DEF(AVX512IFMA, 0x00000007, 0, RTE_REG_EBX, 21)
> + FEAT_DEF(AVX512CD, 0x00000007, 0, RTE_REG_EBX, 28)
> + FEAT_DEF(AVX512BW, 0x00000007, 0, RTE_REG_EBX, 30)
> + FEAT_DEF(AVX512VL, 0x00000007, 0, RTE_REG_EBX, 31)
> +
> + FEAT_DEF(AVX512VBMI, 0x00000007, 0, RTE_REG_ECX, 1)
> + FEAT_DEF(WAITPKG, 0x00000007, 0, RTE_REG_ECX, 5)
> + FEAT_DEF(AVX512VBMI2, 0x00000007, 0, RTE_REG_ECX, 6)
> + FEAT_DEF(GFNI, 0x00000007, 0, RTE_REG_ECX, 8)
> + FEAT_DEF(VAES, 0x00000007, 0, RTE_REG_ECX, 9)
> + FEAT_DEF(VPCLMULQDQ, 0x00000007, 0, RTE_REG_ECX, 10)
> + FEAT_DEF(AVX512VNNI, 0x00000007, 0, RTE_REG_ECX, 11)
> + FEAT_DEF(AVX512BITALG, 0x00000007, 0, RTE_REG_ECX, 12)
> + FEAT_DEF(AVX512VPOPCNTDQ, 0x00000007, 0, RTE_REG_ECX, 14)
> + FEAT_DEF(CLDEMOTE, 0x00000007, 0, RTE_REG_ECX, 25)
> + FEAT_DEF(MOVDIRI, 0x00000007, 0, RTE_REG_ECX, 27)
> + FEAT_DEF(MOVDIR64B, 0x00000007, 0, RTE_REG_ECX, 28)
>
> - FEAT_DEF(WAITPKG, 0x00000007, 0, RTE_REG_ECX, 5)
> + FEAT_DEF(AVX512VP2INTERSECT, 0x00000007, 0, RTE_REG_EDX, 8)
>
> FEAT_DEF(LAHF_SAHF, 0x80000001, 0, RTE_REG_ECX, 0)
> FEAT_DEF(LZCNT, 0x80000001, 0, RTE_REG_ECX, 4)
> @@ -123,24 +141,6 @@ const struct feature_entry rte_cpu_feature_table[] = {
> FEAT_DEF(EM64T, 0x80000001, 0, RTE_REG_EDX, 29)
>
> FEAT_DEF(INVTSC, 0x80000007, 0, RTE_REG_EDX, 8)
> -
> - FEAT_DEF(AVX512DQ, 0x00000007, 0, RTE_REG_EBX, 17)
> - FEAT_DEF(AVX512IFMA, 0x00000007, 0, RTE_REG_EBX, 21)
> - FEAT_DEF(AVX512CD, 0x00000007, 0, RTE_REG_EBX, 28)
> - FEAT_DEF(AVX512BW, 0x00000007, 0, RTE_REG_EBX, 30)
> - FEAT_DEF(AVX512VL, 0x00000007, 0, RTE_REG_EBX, 31)
> - FEAT_DEF(AVX512VBMI, 0x00000007, 0, RTE_REG_ECX, 1)
> - FEAT_DEF(AVX512VBMI2, 0x00000007, 0, RTE_REG_ECX, 6)
> - FEAT_DEF(GFNI, 0x00000007, 0, RTE_REG_ECX, 8)
> - FEAT_DEF(VAES, 0x00000007, 0, RTE_REG_ECX, 9)
> - FEAT_DEF(VPCLMULQDQ, 0x00000007, 0, RTE_REG_ECX, 10)
> - FEAT_DEF(AVX512VNNI, 0x00000007, 0, RTE_REG_ECX, 11)
> - FEAT_DEF(AVX512BITALG, 0x00000007, 0, RTE_REG_ECX, 12)
> - FEAT_DEF(AVX512VPOPCNTDQ, 0x00000007, 0, RTE_REG_ECX, 14)
> - FEAT_DEF(CLDEMOTE, 0x00000007, 0, RTE_REG_ECX, 25)
> - FEAT_DEF(MOVDIRI, 0x00000007, 0, RTE_REG_ECX, 27)
> - FEAT_DEF(MOVDIR64B, 0x00000007, 0, RTE_REG_ECX, 28)
> - FEAT_DEF(AVX512VP2INTERSECT, 0x00000007, 0, RTE_REG_EDX, 8)
> };
>
> int
> --
> 2.23.0
>
On Fri, Oct 8, 2021 at 12:15 PM Bruce Richardson
<bruce.richardson@intel.com> wrote:
>
> On Wed, Oct 06, 2021 at 09:20:08PM +0200, David Marchand wrote:
> > Caught while checking CPUID related stuff in OVS.
> >
> > According to [1], for Structured Extended Feature Flags Enumeration Leaf
> > (EAX = 0x07H, ECX = 0):
> >
> > - BMI1 is associated to EBX, bit 3 (was incorrectly 2),
> > - SMEP is associated to EBX, bit 7 (was incorrectly 6),
> > - BMI2 is associated to EBX, bit 8 (was incorrectly 7),
> > - ERMS is associated to EBX, bit 9 (was incorrectly 8),
> >
> > This patch then sorts the rest of the extended features (leaf 0) for
> > readability.
> >
> > 1: https://software.intel.com/sites/default/files/managed/c5/15/architecture-instruction-set-extensions-programming-reference.pdf
> >
> > Fixes: af75078fece3 ("first public release")
>
> Wow, I'm surprised at it being incorrect this long, but checking the
> original commit of code, it does indeed seem true that this is the correct
> fixes line!
>
> > Cc: stable@dpdk.org
> >
> > Signed-off-by: David Marchand <david.marchand@redhat.com>
>
> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
>
> Ideally, I think the fixes should go in a separate patch from the
> reordering, but I don't feel strongly about it so this is ok as either one
> or two patches IMHO.
I was reading the datasheet and comparing the defines one by one, so
easier with sorted defines.
Now that I identified the wrong entries, I am fine with splitting
since it eases backport on 19.11.
I'll send v2 with your ack on both.
@@ -100,18 +100,36 @@ const struct feature_entry rte_cpu_feature_table[] = {
FEAT_DEF(ENERGY_EFF, 0x00000006, 0, RTE_REG_ECX, 3)
FEAT_DEF(FSGSBASE, 0x00000007, 0, RTE_REG_EBX, 0)
- FEAT_DEF(BMI1, 0x00000007, 0, RTE_REG_EBX, 2)
+ FEAT_DEF(BMI1, 0x00000007, 0, RTE_REG_EBX, 3)
FEAT_DEF(HLE, 0x00000007, 0, RTE_REG_EBX, 4)
FEAT_DEF(AVX2, 0x00000007, 0, RTE_REG_EBX, 5)
- FEAT_DEF(SMEP, 0x00000007, 0, RTE_REG_EBX, 6)
- FEAT_DEF(BMI2, 0x00000007, 0, RTE_REG_EBX, 7)
- FEAT_DEF(ERMS, 0x00000007, 0, RTE_REG_EBX, 8)
+ FEAT_DEF(SMEP, 0x00000007, 0, RTE_REG_EBX, 7)
+ FEAT_DEF(BMI2, 0x00000007, 0, RTE_REG_EBX, 8)
+ FEAT_DEF(ERMS, 0x00000007, 0, RTE_REG_EBX, 9)
FEAT_DEF(INVPCID, 0x00000007, 0, RTE_REG_EBX, 10)
FEAT_DEF(RTM, 0x00000007, 0, RTE_REG_EBX, 11)
FEAT_DEF(AVX512F, 0x00000007, 0, RTE_REG_EBX, 16)
+ FEAT_DEF(AVX512DQ, 0x00000007, 0, RTE_REG_EBX, 17)
FEAT_DEF(RDSEED, 0x00000007, 0, RTE_REG_EBX, 18)
+ FEAT_DEF(AVX512IFMA, 0x00000007, 0, RTE_REG_EBX, 21)
+ FEAT_DEF(AVX512CD, 0x00000007, 0, RTE_REG_EBX, 28)
+ FEAT_DEF(AVX512BW, 0x00000007, 0, RTE_REG_EBX, 30)
+ FEAT_DEF(AVX512VL, 0x00000007, 0, RTE_REG_EBX, 31)
+
+ FEAT_DEF(AVX512VBMI, 0x00000007, 0, RTE_REG_ECX, 1)
+ FEAT_DEF(WAITPKG, 0x00000007, 0, RTE_REG_ECX, 5)
+ FEAT_DEF(AVX512VBMI2, 0x00000007, 0, RTE_REG_ECX, 6)
+ FEAT_DEF(GFNI, 0x00000007, 0, RTE_REG_ECX, 8)
+ FEAT_DEF(VAES, 0x00000007, 0, RTE_REG_ECX, 9)
+ FEAT_DEF(VPCLMULQDQ, 0x00000007, 0, RTE_REG_ECX, 10)
+ FEAT_DEF(AVX512VNNI, 0x00000007, 0, RTE_REG_ECX, 11)
+ FEAT_DEF(AVX512BITALG, 0x00000007, 0, RTE_REG_ECX, 12)
+ FEAT_DEF(AVX512VPOPCNTDQ, 0x00000007, 0, RTE_REG_ECX, 14)
+ FEAT_DEF(CLDEMOTE, 0x00000007, 0, RTE_REG_ECX, 25)
+ FEAT_DEF(MOVDIRI, 0x00000007, 0, RTE_REG_ECX, 27)
+ FEAT_DEF(MOVDIR64B, 0x00000007, 0, RTE_REG_ECX, 28)
- FEAT_DEF(WAITPKG, 0x00000007, 0, RTE_REG_ECX, 5)
+ FEAT_DEF(AVX512VP2INTERSECT, 0x00000007, 0, RTE_REG_EDX, 8)
FEAT_DEF(LAHF_SAHF, 0x80000001, 0, RTE_REG_ECX, 0)
FEAT_DEF(LZCNT, 0x80000001, 0, RTE_REG_ECX, 4)
@@ -123,24 +141,6 @@ const struct feature_entry rte_cpu_feature_table[] = {
FEAT_DEF(EM64T, 0x80000001, 0, RTE_REG_EDX, 29)
FEAT_DEF(INVTSC, 0x80000007, 0, RTE_REG_EDX, 8)
-
- FEAT_DEF(AVX512DQ, 0x00000007, 0, RTE_REG_EBX, 17)
- FEAT_DEF(AVX512IFMA, 0x00000007, 0, RTE_REG_EBX, 21)
- FEAT_DEF(AVX512CD, 0x00000007, 0, RTE_REG_EBX, 28)
- FEAT_DEF(AVX512BW, 0x00000007, 0, RTE_REG_EBX, 30)
- FEAT_DEF(AVX512VL, 0x00000007, 0, RTE_REG_EBX, 31)
- FEAT_DEF(AVX512VBMI, 0x00000007, 0, RTE_REG_ECX, 1)
- FEAT_DEF(AVX512VBMI2, 0x00000007, 0, RTE_REG_ECX, 6)
- FEAT_DEF(GFNI, 0x00000007, 0, RTE_REG_ECX, 8)
- FEAT_DEF(VAES, 0x00000007, 0, RTE_REG_ECX, 9)
- FEAT_DEF(VPCLMULQDQ, 0x00000007, 0, RTE_REG_ECX, 10)
- FEAT_DEF(AVX512VNNI, 0x00000007, 0, RTE_REG_ECX, 11)
- FEAT_DEF(AVX512BITALG, 0x00000007, 0, RTE_REG_ECX, 12)
- FEAT_DEF(AVX512VPOPCNTDQ, 0x00000007, 0, RTE_REG_ECX, 14)
- FEAT_DEF(CLDEMOTE, 0x00000007, 0, RTE_REG_ECX, 25)
- FEAT_DEF(MOVDIRI, 0x00000007, 0, RTE_REG_ECX, 27)
- FEAT_DEF(MOVDIR64B, 0x00000007, 0, RTE_REG_ECX, 28)
- FEAT_DEF(AVX512VP2INTERSECT, 0x00000007, 0, RTE_REG_EDX, 8)
};
int