diff mbox series

[v7,07/27] arm64/sve: Check SVE virtualisability

Message ID 1553864452-15080-8-git-send-email-Dave.Martin@arm.com (mailing list archive)
State New, archived
Headers show
Series KVM: arm64: SVE guest support | expand

Commit Message

Dave Martin March 29, 2019, 1 p.m. UTC
Due to the way the effective SVE vector length is controlled and
trapped at different exception levels, certain mismatches in the
sets of vector lengths supported by different physical CPUs in the
system may prevent straightforward virtualisation of SVE at parity
with the host.

This patch analyses the extent to which SVE can be virtualised
safely without interfering with migration of vcpus between physical
CPUs, and rejects late secondary CPUs that would erode the
situation further.

It is left up to KVM to decide what to do with this information.

Signed-off-by: Dave Martin <Dave.Martin@arm.com>
Reviewed-by: Julien Thierry <julien.thierry@arm.com>
Tested-by: zhang.lei <zhang.lei@jp.fujitsu.com>

---

QUESTION: The final structure of this code makes it quite natural to
clamp the vector length for KVM guests to the maximum value supportable
across all CPUs; such a value is guaranteed to exist, but may be
surprisingly small on a given hardware platform.

Should we be stricter and actually refuse to support KVM at all on such
hardware?  This may help to force people to fix Linux (or the
architecture) if/when this issue comes up.

For now, I stick with pr_warn() and make do with a limited SVE vector
length for guests.

Changes since v5:

 * pr_info() about the presence of unvirtualisable vector lengths in
   sve_setup() upgrade to pr_warn(), for consistency with
   sve_verify_vq_map().
---
 arch/arm64/include/asm/fpsimd.h |  1 +
 arch/arm64/kernel/cpufeature.c  |  2 +-
 arch/arm64/kernel/fpsimd.c      | 86 ++++++++++++++++++++++++++++++++++-------
 3 files changed, 73 insertions(+), 16 deletions(-)

Comments

Andrew Jones April 4, 2019, 9:21 p.m. UTC | #1
On Fri, Mar 29, 2019 at 01:00:32PM +0000, Dave Martin wrote:
> Due to the way the effective SVE vector length is controlled and
> trapped at different exception levels, certain mismatches in the
> sets of vector lengths supported by different physical CPUs in the
> system may prevent straightforward virtualisation of SVE at parity
> with the host.
> 
> This patch analyses the extent to which SVE can be virtualised
> safely without interfering with migration of vcpus between physical
> CPUs, and rejects late secondary CPUs that would erode the
> situation further.
> 
> It is left up to KVM to decide what to do with this information.
> 
> Signed-off-by: Dave Martin <Dave.Martin@arm.com>
> Reviewed-by: Julien Thierry <julien.thierry@arm.com>
> Tested-by: zhang.lei <zhang.lei@jp.fujitsu.com>
> 
> ---
> 
> QUESTION: The final structure of this code makes it quite natural to
> clamp the vector length for KVM guests to the maximum value supportable
> across all CPUs; such a value is guaranteed to exist, but may be
> surprisingly small on a given hardware platform.
> 
> Should we be stricter and actually refuse to support KVM at all on such
> hardware?  This may help to force people to fix Linux (or the
> architecture) if/when this issue comes up.

Blocking KVM would be too harsh, since the users of the host may not
care about guests with SVE, but still care about guests.

> 
> For now, I stick with pr_warn() and make do with a limited SVE vector
> length for guests.

I think that's probably the best we can do.

> 
> Changes since v5:
> 
>  * pr_info() about the presence of unvirtualisable vector lengths in
>    sve_setup() upgrade to pr_warn(), for consistency with
>    sve_verify_vq_map().
> ---
>  arch/arm64/include/asm/fpsimd.h |  1 +
>  arch/arm64/kernel/cpufeature.c  |  2 +-
>  arch/arm64/kernel/fpsimd.c      | 86 ++++++++++++++++++++++++++++++++++-------
>  3 files changed, 73 insertions(+), 16 deletions(-)
> 
> diff --git a/arch/arm64/include/asm/fpsimd.h b/arch/arm64/include/asm/fpsimd.h
> index dd1ad39..964adc9 100644
> --- a/arch/arm64/include/asm/fpsimd.h
> +++ b/arch/arm64/include/asm/fpsimd.h
> @@ -87,6 +87,7 @@ extern void sve_kernel_enable(const struct arm64_cpu_capabilities *__unused);
>  extern u64 read_zcr_features(void);
>  
>  extern int __ro_after_init sve_max_vl;
> +extern int __ro_after_init sve_max_virtualisable_vl;
>  
>  #ifdef CONFIG_ARM64_SVE
>  
> diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
> index 4061de1..7f8cc51 100644
> --- a/arch/arm64/kernel/cpufeature.c
> +++ b/arch/arm64/kernel/cpufeature.c
> @@ -1863,7 +1863,7 @@ static void verify_sve_features(void)
>  	unsigned int len = zcr & ZCR_ELx_LEN_MASK;
>  
>  	if (len < safe_len || sve_verify_vq_map()) {
> -		pr_crit("CPU%d: SVE: required vector length(s) missing\n",
> +		pr_crit("CPU%d: SVE: vector length support mismatch\n",
>  			smp_processor_id());
>  		cpu_die_early();
>  	}
> diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
> index f59ea67..b219796a 100644
> --- a/arch/arm64/kernel/fpsimd.c
> +++ b/arch/arm64/kernel/fpsimd.c
> @@ -18,6 +18,7 @@
>   */
>  
>  #include <linux/bitmap.h>
> +#include <linux/bitops.h>
>  #include <linux/bottom_half.h>
>  #include <linux/bug.h>
>  #include <linux/cache.h>
> @@ -48,6 +49,7 @@
>  #include <asm/sigcontext.h>
>  #include <asm/sysreg.h>
>  #include <asm/traps.h>
> +#include <asm/virt.h>
>  
>  #define FPEXC_IOF	(1 << 0)
>  #define FPEXC_DZF	(1 << 1)
> @@ -130,14 +132,18 @@ static int sve_default_vl = -1;
>  
>  /* Maximum supported vector length across all CPUs (initially poisoned) */
>  int __ro_after_init sve_max_vl = SVE_VL_MIN;
> +int __ro_after_init sve_max_virtualisable_vl = SVE_VL_MIN;
>  /* Set of available vector lengths, as vq_to_bit(vq): */
>  static __ro_after_init DECLARE_BITMAP(sve_vq_map, SVE_VQ_MAX);
> +/* Set of vector lengths present on at least one cpu: */
> +static __ro_after_init DECLARE_BITMAP(sve_vq_partial_map, SVE_VQ_MAX);
>  static void __percpu *efi_sve_state;
>  
>  #else /* ! CONFIG_ARM64_SVE */
>  
>  /* Dummy declaration for code that will be optimised out: */
>  extern __ro_after_init DECLARE_BITMAP(sve_vq_map, SVE_VQ_MAX);
> +extern __ro_after_init DECLARE_BITMAP(sve_vq_partial_map, SVE_VQ_MAX);
>  extern void __percpu *efi_sve_state;
>  
>  #endif /* ! CONFIG_ARM64_SVE */
> @@ -623,12 +629,6 @@ int sve_get_current_vl(void)
>  	return sve_prctl_status(0);
>  }
>  
> -/*
> - * Bitmap for temporary storage of the per-CPU set of supported vector lengths
> - * during secondary boot.
> - */
> -static DECLARE_BITMAP(sve_secondary_vq_map, SVE_VQ_MAX);
> -
>  static void sve_probe_vqs(DECLARE_BITMAP(map, SVE_VQ_MAX))
>  {
>  	unsigned int vq, vl;
> @@ -654,6 +654,7 @@ static void sve_probe_vqs(DECLARE_BITMAP(map, SVE_VQ_MAX))
>  void __init sve_init_vq_map(void)
>  {
>  	sve_probe_vqs(sve_vq_map);
> +	bitmap_copy(sve_vq_partial_map, sve_vq_map, SVE_VQ_MAX);
>  }
>  
>  /*
> @@ -663,8 +664,11 @@ void __init sve_init_vq_map(void)
>   */
>  void sve_update_vq_map(void)
>  {
> -	sve_probe_vqs(sve_secondary_vq_map);
> -	bitmap_and(sve_vq_map, sve_vq_map, sve_secondary_vq_map, SVE_VQ_MAX);
> +	DECLARE_BITMAP(tmp_map, SVE_VQ_MAX);
> +
> +	sve_probe_vqs(tmp_map);
> +	bitmap_and(sve_vq_map, sve_vq_map, tmp_map, SVE_VQ_MAX);
> +	bitmap_or(sve_vq_partial_map, sve_vq_partial_map, tmp_map, SVE_VQ_MAX);
>  }
>  
>  /*
> @@ -673,18 +677,48 @@ void sve_update_vq_map(void)
>   */
>  int sve_verify_vq_map(void)
>  {
> -	int ret = 0;
> +	DECLARE_BITMAP(tmp_map, SVE_VQ_MAX);
> +	unsigned long b;
>  
> -	sve_probe_vqs(sve_secondary_vq_map);
> -	bitmap_andnot(sve_secondary_vq_map, sve_vq_map, sve_secondary_vq_map,
> -		      SVE_VQ_MAX);
> -	if (!bitmap_empty(sve_secondary_vq_map, SVE_VQ_MAX)) {
> +	sve_probe_vqs(tmp_map);
> +
> +	bitmap_complement(tmp_map, tmp_map, SVE_VQ_MAX);
> +	if (bitmap_intersects(tmp_map, sve_vq_map, SVE_VQ_MAX)) {
>  		pr_warn("SVE: cpu%d: Required vector length(s) missing\n",
>  			smp_processor_id());
> -		ret = -EINVAL;
> +		return -EINVAL;
>  	}
>  
> -	return ret;
> +	if (!IS_ENABLED(CONFIG_KVM) || !is_hyp_mode_available())
> +		return 0;
> +
> +	/*
> +	 * For KVM, it is necessary to ensure that this CPU doesn't
> +	 * support any vector length that guests may have probed as
> +	 * unsupported.
> +	 */
> +
> +	/* Recover the set of supported VQs: */
> +	bitmap_complement(tmp_map, tmp_map, SVE_VQ_MAX);
> +	/* Find VQs supported that are not globally supported: */
> +	bitmap_andnot(tmp_map, tmp_map, sve_vq_map, SVE_VQ_MAX);
> +
> +	/* Find the lowest such VQ, if any: */
> +	b = find_last_bit(tmp_map, SVE_VQ_MAX);
> +	if (b >= SVE_VQ_MAX)
> +		return 0; /* no mismatches */
> +
> +	/*
> +	 * Mismatches above sve_max_virtualisable_vl are fine, since
> +	 * no guest is allowed to configure ZCR_EL2.LEN to exceed this:
> +	 */
> +	if (sve_vl_from_vq(bit_to_vq(b)) <= sve_max_virtualisable_vl) {
> +		pr_warn("SVE: cpu%d: Unsupported vector length(s) present\n",
> +			smp_processor_id());
> +		return -EINVAL;
> +	}
> +
> +	return 0;
>  }
>  
>  static void __init sve_efi_setup(void)
> @@ -751,6 +785,8 @@ u64 read_zcr_features(void)
>  void __init sve_setup(void)
>  {
>  	u64 zcr;
> +	DECLARE_BITMAP(tmp_map, SVE_VQ_MAX);
> +	unsigned long b;
>  
>  	if (!system_supports_sve())
>  		return;
> @@ -779,11 +815,31 @@ void __init sve_setup(void)
>  	 */
>  	sve_default_vl = find_supported_vector_length(64);
>  
> +	bitmap_andnot(tmp_map, sve_vq_partial_map, sve_vq_map,
> +		      SVE_VQ_MAX);
> +
> +	b = find_last_bit(tmp_map, SVE_VQ_MAX);
> +	if (b >= SVE_VQ_MAX)
> +		/* No non-virtualisable VLs found */
> +		sve_max_virtualisable_vl = SVE_VQ_MAX;
> +	else if (WARN_ON(b == SVE_VQ_MAX - 1))
> +		/* No virtualisable VLs?  This is architecturally forbidden. */
> +		sve_max_virtualisable_vl = SVE_VQ_MIN;
> +	else /* b + 1 < SVE_VQ_MAX */
> +		sve_max_virtualisable_vl = sve_vl_from_vq(bit_to_vq(b + 1));
> +
> +	if (sve_max_virtualisable_vl > sve_max_vl)
> +		sve_max_virtualisable_vl = sve_max_vl;
> +
>  	pr_info("SVE: maximum available vector length %u bytes per vector\n",
>  		sve_max_vl);
>  	pr_info("SVE: default vector length %u bytes per vector\n",
>  		sve_default_vl);
>  
> +	/* KVM decides whether to support mismatched systems. Just warn here: */
> +	if (sve_max_virtualisable_vl < sve_max_vl)
> +		pr_warn("SVE: unvirtualisable vector lengths present\n");
> +
>  	sve_efi_setup();
>  }
>  
> -- 
> 2.1.4
>

Reviewed-by: Andrew Jones <drjones@redhat.com>
Dave Martin April 5, 2019, 9:35 a.m. UTC | #2
On Thu, Apr 04, 2019 at 11:21:04PM +0200, Andrew Jones wrote:
> On Fri, Mar 29, 2019 at 01:00:32PM +0000, Dave Martin wrote:
> > Due to the way the effective SVE vector length is controlled and
> > trapped at different exception levels, certain mismatches in the
> > sets of vector lengths supported by different physical CPUs in the
> > system may prevent straightforward virtualisation of SVE at parity
> > with the host.
> > 
> > This patch analyses the extent to which SVE can be virtualised
> > safely without interfering with migration of vcpus between physical
> > CPUs, and rejects late secondary CPUs that would erode the
> > situation further.
> > 
> > It is left up to KVM to decide what to do with this information.
> > 
> > Signed-off-by: Dave Martin <Dave.Martin@arm.com>
> > Reviewed-by: Julien Thierry <julien.thierry@arm.com>
> > Tested-by: zhang.lei <zhang.lei@jp.fujitsu.com>
> > 
> > ---
> > 
> > QUESTION: The final structure of this code makes it quite natural to
> > clamp the vector length for KVM guests to the maximum value supportable
> > across all CPUs; such a value is guaranteed to exist, but may be
> > surprisingly small on a given hardware platform.
> > 
> > Should we be stricter and actually refuse to support KVM at all on such
> > hardware?  This may help to force people to fix Linux (or the
> > architecture) if/when this issue comes up.
> 
> Blocking KVM would be too harsh, since the users of the host may not
> care about guests with SVE, but still care about guests.
> 
> > 
> > For now, I stick with pr_warn() and make do with a limited SVE vector
> > length for guests.
> 
> I think that's probably the best we can do.

Agreed.  Since it fell out quite nicely this way in the code, this was
my preferred option.

[...]

> Reviewed-by: Andrew Jones <drjones@redhat.com>

Thanks

---Dave
diff mbox series

Patch

diff --git a/arch/arm64/include/asm/fpsimd.h b/arch/arm64/include/asm/fpsimd.h
index dd1ad39..964adc9 100644
--- a/arch/arm64/include/asm/fpsimd.h
+++ b/arch/arm64/include/asm/fpsimd.h
@@ -87,6 +87,7 @@  extern void sve_kernel_enable(const struct arm64_cpu_capabilities *__unused);
 extern u64 read_zcr_features(void);
 
 extern int __ro_after_init sve_max_vl;
+extern int __ro_after_init sve_max_virtualisable_vl;
 
 #ifdef CONFIG_ARM64_SVE
 
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 4061de1..7f8cc51 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -1863,7 +1863,7 @@  static void verify_sve_features(void)
 	unsigned int len = zcr & ZCR_ELx_LEN_MASK;
 
 	if (len < safe_len || sve_verify_vq_map()) {
-		pr_crit("CPU%d: SVE: required vector length(s) missing\n",
+		pr_crit("CPU%d: SVE: vector length support mismatch\n",
 			smp_processor_id());
 		cpu_die_early();
 	}
diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
index f59ea67..b219796a 100644
--- a/arch/arm64/kernel/fpsimd.c
+++ b/arch/arm64/kernel/fpsimd.c
@@ -18,6 +18,7 @@ 
  */
 
 #include <linux/bitmap.h>
+#include <linux/bitops.h>
 #include <linux/bottom_half.h>
 #include <linux/bug.h>
 #include <linux/cache.h>
@@ -48,6 +49,7 @@ 
 #include <asm/sigcontext.h>
 #include <asm/sysreg.h>
 #include <asm/traps.h>
+#include <asm/virt.h>
 
 #define FPEXC_IOF	(1 << 0)
 #define FPEXC_DZF	(1 << 1)
@@ -130,14 +132,18 @@  static int sve_default_vl = -1;
 
 /* Maximum supported vector length across all CPUs (initially poisoned) */
 int __ro_after_init sve_max_vl = SVE_VL_MIN;
+int __ro_after_init sve_max_virtualisable_vl = SVE_VL_MIN;
 /* Set of available vector lengths, as vq_to_bit(vq): */
 static __ro_after_init DECLARE_BITMAP(sve_vq_map, SVE_VQ_MAX);
+/* Set of vector lengths present on at least one cpu: */
+static __ro_after_init DECLARE_BITMAP(sve_vq_partial_map, SVE_VQ_MAX);
 static void __percpu *efi_sve_state;
 
 #else /* ! CONFIG_ARM64_SVE */
 
 /* Dummy declaration for code that will be optimised out: */
 extern __ro_after_init DECLARE_BITMAP(sve_vq_map, SVE_VQ_MAX);
+extern __ro_after_init DECLARE_BITMAP(sve_vq_partial_map, SVE_VQ_MAX);
 extern void __percpu *efi_sve_state;
 
 #endif /* ! CONFIG_ARM64_SVE */
@@ -623,12 +629,6 @@  int sve_get_current_vl(void)
 	return sve_prctl_status(0);
 }
 
-/*
- * Bitmap for temporary storage of the per-CPU set of supported vector lengths
- * during secondary boot.
- */
-static DECLARE_BITMAP(sve_secondary_vq_map, SVE_VQ_MAX);
-
 static void sve_probe_vqs(DECLARE_BITMAP(map, SVE_VQ_MAX))
 {
 	unsigned int vq, vl;
@@ -654,6 +654,7 @@  static void sve_probe_vqs(DECLARE_BITMAP(map, SVE_VQ_MAX))
 void __init sve_init_vq_map(void)
 {
 	sve_probe_vqs(sve_vq_map);
+	bitmap_copy(sve_vq_partial_map, sve_vq_map, SVE_VQ_MAX);
 }
 
 /*
@@ -663,8 +664,11 @@  void __init sve_init_vq_map(void)
  */
 void sve_update_vq_map(void)
 {
-	sve_probe_vqs(sve_secondary_vq_map);
-	bitmap_and(sve_vq_map, sve_vq_map, sve_secondary_vq_map, SVE_VQ_MAX);
+	DECLARE_BITMAP(tmp_map, SVE_VQ_MAX);
+
+	sve_probe_vqs(tmp_map);
+	bitmap_and(sve_vq_map, sve_vq_map, tmp_map, SVE_VQ_MAX);
+	bitmap_or(sve_vq_partial_map, sve_vq_partial_map, tmp_map, SVE_VQ_MAX);
 }
 
 /*
@@ -673,18 +677,48 @@  void sve_update_vq_map(void)
  */
 int sve_verify_vq_map(void)
 {
-	int ret = 0;
+	DECLARE_BITMAP(tmp_map, SVE_VQ_MAX);
+	unsigned long b;
 
-	sve_probe_vqs(sve_secondary_vq_map);
-	bitmap_andnot(sve_secondary_vq_map, sve_vq_map, sve_secondary_vq_map,
-		      SVE_VQ_MAX);
-	if (!bitmap_empty(sve_secondary_vq_map, SVE_VQ_MAX)) {
+	sve_probe_vqs(tmp_map);
+
+	bitmap_complement(tmp_map, tmp_map, SVE_VQ_MAX);
+	if (bitmap_intersects(tmp_map, sve_vq_map, SVE_VQ_MAX)) {
 		pr_warn("SVE: cpu%d: Required vector length(s) missing\n",
 			smp_processor_id());
-		ret = -EINVAL;
+		return -EINVAL;
 	}
 
-	return ret;
+	if (!IS_ENABLED(CONFIG_KVM) || !is_hyp_mode_available())
+		return 0;
+
+	/*
+	 * For KVM, it is necessary to ensure that this CPU doesn't
+	 * support any vector length that guests may have probed as
+	 * unsupported.
+	 */
+
+	/* Recover the set of supported VQs: */
+	bitmap_complement(tmp_map, tmp_map, SVE_VQ_MAX);
+	/* Find VQs supported that are not globally supported: */
+	bitmap_andnot(tmp_map, tmp_map, sve_vq_map, SVE_VQ_MAX);
+
+	/* Find the lowest such VQ, if any: */
+	b = find_last_bit(tmp_map, SVE_VQ_MAX);
+	if (b >= SVE_VQ_MAX)
+		return 0; /* no mismatches */
+
+	/*
+	 * Mismatches above sve_max_virtualisable_vl are fine, since
+	 * no guest is allowed to configure ZCR_EL2.LEN to exceed this:
+	 */
+	if (sve_vl_from_vq(bit_to_vq(b)) <= sve_max_virtualisable_vl) {
+		pr_warn("SVE: cpu%d: Unsupported vector length(s) present\n",
+			smp_processor_id());
+		return -EINVAL;
+	}
+
+	return 0;
 }
 
 static void __init sve_efi_setup(void)
@@ -751,6 +785,8 @@  u64 read_zcr_features(void)
 void __init sve_setup(void)
 {
 	u64 zcr;
+	DECLARE_BITMAP(tmp_map, SVE_VQ_MAX);
+	unsigned long b;
 
 	if (!system_supports_sve())
 		return;
@@ -779,11 +815,31 @@  void __init sve_setup(void)
 	 */
 	sve_default_vl = find_supported_vector_length(64);
 
+	bitmap_andnot(tmp_map, sve_vq_partial_map, sve_vq_map,
+		      SVE_VQ_MAX);
+
+	b = find_last_bit(tmp_map, SVE_VQ_MAX);
+	if (b >= SVE_VQ_MAX)
+		/* No non-virtualisable VLs found */
+		sve_max_virtualisable_vl = SVE_VQ_MAX;
+	else if (WARN_ON(b == SVE_VQ_MAX - 1))
+		/* No virtualisable VLs?  This is architecturally forbidden. */
+		sve_max_virtualisable_vl = SVE_VQ_MIN;
+	else /* b + 1 < SVE_VQ_MAX */
+		sve_max_virtualisable_vl = sve_vl_from_vq(bit_to_vq(b + 1));
+
+	if (sve_max_virtualisable_vl > sve_max_vl)
+		sve_max_virtualisable_vl = sve_max_vl;
+
 	pr_info("SVE: maximum available vector length %u bytes per vector\n",
 		sve_max_vl);
 	pr_info("SVE: default vector length %u bytes per vector\n",
 		sve_default_vl);
 
+	/* KVM decides whether to support mismatched systems. Just warn here: */
+	if (sve_max_virtualisable_vl < sve_max_vl)
+		pr_warn("SVE: unvirtualisable vector lengths present\n");
+
 	sve_efi_setup();
 }