diff mbox

[v1,04/13] x86: implement data structure and CPU init flow for MBA

Message ID 1502264512-4648-5-git-send-email-yi.y.sun@linux.intel.com (mailing list archive)
State New, archived
Headers show

Commit Message

Yi Sun Aug. 9, 2017, 7:41 a.m. UTC 
This patch implements main data structures of MBA.

Like CAT features, MBA HW info has cos_max which means the max cos
registers number, and thrtl_max which means the max throttle value
(delay value). It also has a flag to represent if the throttle
value is linear or not.

One COS register of MBA stores a throttle value for one or more
domains. The throttle value means the transaction time between L2
cache and next level memory to be delayed.

This patch also implements init flow for MBA and register stub
callback functions.

Signed-off-by: Yi Sun <yi.y.sun@linux.intel.com>
---
v1:
    - rebase codes onto L2 CAT v15.
    - use '(1u << X)'.
      (suggested by Wei Liu)
    - move comment to appropriate place.
      (suggested by Chao Peng)
    - implement 'mba_init_feature' and keep 'cat_init_feature'.
      (suggested by Chao Peng)
    - keep 'regs.b' into a local variable to avoid reading CPUID every time.
      (suggested by Chao Peng)
---
 xen/arch/x86/psr.c              | 144 ++++++++++++++++++++++++++++++++++------
 xen/include/asm-x86/msr-index.h |   1 +
 xen/include/asm-x86/psr.h       |   2 +
 3 files changed, 126 insertions(+), 21 deletions(-)

Comments

Wei Liu Aug. 15, 2017, 10:50 a.m. UTC | #1 
On Wed, Aug 09, 2017 at 03:41:43PM +0800, Yi Sun wrote:
> This patch implements main data structures of MBA.
> 
> Like CAT features, MBA HW info has cos_max which means the max cos
> registers number, and thrtl_max which means the max throttle value
> (delay value). It also has a flag to represent if the throttle
> value is linear or not.
> 
> One COS register of MBA stores a throttle value for one or more
> domains. The throttle value means the transaction time between L2
> cache and next level memory to be delayed.
> 
> This patch also implements init flow for MBA and register stub
> callback functions.
> 
> Signed-off-by: Yi Sun <yi.y.sun@linux.intel.com>
> ---
> v1:
>     - rebase codes onto L2 CAT v15.
>     - use '(1u << X)'.
>       (suggested by Wei Liu)
>     - move comment to appropriate place.
>       (suggested by Chao Peng)
>     - implement 'mba_init_feature' and keep 'cat_init_feature'.
>       (suggested by Chao Peng)
>     - keep 'regs.b' into a local variable to avoid reading CPUID every time.
>       (suggested by Chao Peng)
> ---
>  xen/arch/x86/psr.c              | 144 ++++++++++++++++++++++++++++++++++------
>  xen/include/asm-x86/msr-index.h |   1 +
>  xen/include/asm-x86/psr.h       |   2 +
>  3 files changed, 126 insertions(+), 21 deletions(-)
> 
> diff --git a/xen/arch/x86/psr.c b/xen/arch/x86/psr.c
> index 5ec00a9..d94a5b1 100644
> --- a/xen/arch/x86/psr.c
> +++ b/xen/arch/x86/psr.c
> @@ -27,13 +27,16 @@
>   * - CMT         Cache Monitoring Technology
>   * - COS/CLOS    Class of Service. Also mean COS registers.
>   * - COS_MAX     Max number of COS for the feature (minus 1)
> + * - MBA         Memory Bandwidth Allocation
>   * - MSRs        Machine Specific Registers
>   * - PSR         Intel Platform Shared Resource
> + * - THRTL_MAX   Max throttle value (delay value) of MBA
>   */
>  
> -#define PSR_CMT        (1<<0)
> -#define PSR_CAT        (1<<1)
> -#define PSR_CDP        (1<<2)
> +#define PSR_CMT        (1u << 0)
> +#define PSR_CAT        (1u << 1)
> +#define PSR_CDP        (1u << 2)
> +#define PSR_MBA        (1u << 3)

I would split this part out to a separate patch so that it can be
applied immediately.

>  
>  #define CAT_CBM_LEN_MASK 0x1f
>  #define CAT_COS_MAX_MASK 0xffff
> @@ -60,10 +63,14 @@
>   */
>  #define MAX_COS_NUM 2
>  
> +#define MBA_LINEAR         (1u << 2)
> +#define MBA_THRTL_MAX_MASK 0xfff
> +
>  enum psr_feat_type {
>      FEAT_TYPE_L3_CAT,
>      FEAT_TYPE_L3_CDP,
>      FEAT_TYPE_L2_CAT,
> +    FEAT_TYPE_MBA,
>      FEAT_TYPE_NUM,
>      FEAT_TYPE_UNKNOWN,
>  };
> @@ -71,7 +78,6 @@ enum psr_feat_type {
>  /*
>   * This structure represents one feature.
>   * cos_max     - The max COS registers number got through CPUID.
> - * cbm_len     - The length of CBM got through CPUID.
>   * cos_reg_val - Array to store the values of COS registers. One entry stores
>   *               the value of one COS register.
>   *               For L3 CAT and L2 CAT, one entry corresponds to one COS_ID.
> @@ -80,9 +86,23 @@ enum psr_feat_type {
>   *               cos_reg_val[1] (Code).
>   */
>  struct feat_node {
> -    /* cos_max and cbm_len are common values for all features so far. */
> +    /* cos_max is common values for all features so far. */
>      unsigned int cos_max;
> -    unsigned int cbm_len;
> +
> +    /* Feature specific HW info. */
> +    union {
> +        struct {
> +            /* The length of CBM got through CPUID. */
> +            unsigned int cbm_len;
> +        } cat_info;
> +
> +        struct {
> +            /* The max throttling value got through CPUID. */
> +            unsigned int thrtl_max;
> +            unsigned int linear;
> +        } mba_info;
> +    };
> +

I suggest you add a tag to specify which struct is in effect in the
union and ASSERT accordingly in their respective type specific
functions.
Chao Peng Aug. 16, 2017, 3:14 a.m. UTC | #2 
On Wed, 2017-08-09 at 15:41 +0800, Yi Sun wrote:
> This patch implements main data structures of MBA.
> 
> Like CAT features, MBA HW info has cos_max which means the max cos
> registers number, and thrtl_max which means the max throttle value

Similarly, there is no existence of 'cos register', 'cos number' is even
better or anything else.

> (delay value). It also has a flag to represent if the throttle
> value is linear or not.
> 
> One COS register of MBA stores a throttle value for one or more

Ditto.
                                                         ^s
...

> -/* CAT common functions implementation. */
> +/* Implementation of allocation features' functions. */
>  static int cat_init_feature(const struct cpuid_leaf *regs,
>                              struct feat_node *feat,
>                              struct psr_socket_info *info,
> @@ -289,7 +310,6 @@ static int cat_init_feature(const struct
> cpuid_leaf *regs,
>      if ( !regs->a || !regs->d )
>          return -ENOENT;
>  
> -    feat->cbm_len = (regs->a & CAT_CBM_LEN_MASK) + 1;

Does this have to be moved?

...

> +
> +    feat->cos_reg_val[0] = 0;
> +    wrmsrl(MSR_IA32_PSR_MBA_MASK(0), 0);

It's better to have a comment here to explain what the defaul value '0'
stands for in both linear mode and non-linear mode.

Chao
Yi Sun Aug. 16, 2017, 7:18 a.m. UTC | #3 
On 17-08-15 11:50:15, Wei Liu wrote:
> On Wed, Aug 09, 2017 at 03:41:43PM +0800, Yi Sun wrote:
> > -#define PSR_CAT        (1<<1)
> > -#define PSR_CDP        (1<<2)
> > +#define PSR_CMT        (1u << 0)
> > +#define PSR_CAT        (1u << 1)
> > +#define PSR_CDP        (1u << 2)
> > +#define PSR_MBA        (1u << 3)
> 
> I would split this part out to a separate patch so that it can be
> applied immediately.
> 
A fix patch has been sent out.

> >  struct feat_node {
> > -    /* cos_max and cbm_len are common values for all features so far. */
> > +    /* cos_max is common values for all features so far. */
> >      unsigned int cos_max;
> > -    unsigned int cbm_len;
> > +
> > +    /* Feature specific HW info. */
> > +    union {
> > +        struct {
> > +            /* The length of CBM got through CPUID. */
> > +            unsigned int cbm_len;
> > +        } cat_info;
> > +
> > +        struct {
> > +            /* The max throttling value got through CPUID. */
> > +            unsigned int thrtl_max;
> > +            unsigned int linear;
> > +        } mba_info;
> > +    };
> > +
> 
> I suggest you add a tag to specify which struct is in effect in the
> union and ASSERT accordingly in their respective type specific
> functions.

Before using struct in this union at all places, there is check to make sure the
correct feature node is used. So, is it necessary to add an additional flag to
check? Thanks!
Wei Liu Aug. 17, 2017, 9:49 a.m. UTC | #4 
On Wed, Aug 16, 2017 at 03:18:13PM +0800, Yi Sun wrote:
> On 17-08-15 11:50:15, Wei Liu wrote:
> > >  struct feat_node {
> > > -    /* cos_max and cbm_len are common values for all features so far. */
> > > +    /* cos_max is common values for all features so far. */
> > >      unsigned int cos_max;
> > > -    unsigned int cbm_len;
> > > +
> > > +    /* Feature specific HW info. */
> > > +    union {
> > > +        struct {
> > > +            /* The length of CBM got through CPUID. */
> > > +            unsigned int cbm_len;
> > > +        } cat_info;
> > > +
> > > +        struct {
> > > +            /* The max throttling value got through CPUID. */
> > > +            unsigned int thrtl_max;
> > > +            unsigned int linear;
> > > +        } mba_info;
> > > +    };
> > > +
> > 
> > I suggest you add a tag to specify which struct is in effect in the
> > union and ASSERT accordingly in their respective type specific
> > functions.
> 
> Before using struct in this union at all places, there is check to make sure the
> correct feature node is used. So, is it necessary to add an additional flag to
> check? Thanks!

OK, in that case I think it is fine to not add a tag.
diff mbox

Patch

diff --git a/xen/arch/x86/psr.c b/xen/arch/x86/psr.c
index 5ec00a9..d94a5b1 100644
--- a/xen/arch/x86/psr.c
+++ b/xen/arch/x86/psr.c
@@ -27,13 +27,16 @@ 
  * - CMT         Cache Monitoring Technology
  * - COS/CLOS    Class of Service. Also mean COS registers.
  * - COS_MAX     Max number of COS for the feature (minus 1)
+ * - MBA         Memory Bandwidth Allocation
  * - MSRs        Machine Specific Registers
  * - PSR         Intel Platform Shared Resource
+ * - THRTL_MAX   Max throttle value (delay value) of MBA
  */
 
-#define PSR_CMT        (1<<0)
-#define PSR_CAT        (1<<1)
-#define PSR_CDP        (1<<2)
+#define PSR_CMT        (1u << 0)
+#define PSR_CAT        (1u << 1)
+#define PSR_CDP        (1u << 2)
+#define PSR_MBA        (1u << 3)
 
 #define CAT_CBM_LEN_MASK 0x1f
 #define CAT_COS_MAX_MASK 0xffff
@@ -60,10 +63,14 @@ 
  */
 #define MAX_COS_NUM 2
 
+#define MBA_LINEAR         (1u << 2)
+#define MBA_THRTL_MAX_MASK 0xfff
+
 enum psr_feat_type {
     FEAT_TYPE_L3_CAT,
     FEAT_TYPE_L3_CDP,
     FEAT_TYPE_L2_CAT,
+    FEAT_TYPE_MBA,
     FEAT_TYPE_NUM,
     FEAT_TYPE_UNKNOWN,
 };
@@ -71,7 +78,6 @@  enum psr_feat_type {
 /*
  * This structure represents one feature.
  * cos_max     - The max COS registers number got through CPUID.
- * cbm_len     - The length of CBM got through CPUID.
  * cos_reg_val - Array to store the values of COS registers. One entry stores
  *               the value of one COS register.
  *               For L3 CAT and L2 CAT, one entry corresponds to one COS_ID.
@@ -80,9 +86,23 @@  enum psr_feat_type {
  *               cos_reg_val[1] (Code).
  */
 struct feat_node {
-    /* cos_max and cbm_len are common values for all features so far. */
+    /* cos_max is common values for all features so far. */
     unsigned int cos_max;
-    unsigned int cbm_len;
+
+    /* Feature specific HW info. */
+    union {
+        struct {
+            /* The length of CBM got through CPUID. */
+            unsigned int cbm_len;
+        } cat_info;
+
+        struct {
+            /* The max throttling value got through CPUID. */
+            unsigned int thrtl_max;
+            unsigned int linear;
+        } mba_info;
+    };
+
     uint32_t cos_reg_val[MAX_COS_REG_CNT];
 };
 
@@ -161,6 +181,7 @@  static DEFINE_PER_CPU(struct psr_assoc, psr_assoc);
  */
 static struct feat_node *feat_l3;
 static struct feat_node *feat_l2_cat;
+static struct feat_node *feat_mba;
 
 /* Common functions */
 #define cat_default_val(len) (0xffffffff >> (32 - (len)))
@@ -273,7 +294,7 @@  static bool psr_check_cbm(unsigned int cbm_len, unsigned long cbm)
     return true;
 }
 
-/* CAT common functions implementation. */
+/* Implementation of allocation features' functions. */
 static int cat_init_feature(const struct cpuid_leaf *regs,
                             struct feat_node *feat,
                             struct psr_socket_info *info,
@@ -289,7 +310,6 @@  static int cat_init_feature(const struct cpuid_leaf *regs,
     if ( !regs->a || !regs->d )
         return -ENOENT;
 
-    feat->cbm_len = (regs->a & CAT_CBM_LEN_MASK) + 1;
     feat->cos_max = min(opt_cos_max, regs->d & CAT_COS_MAX_MASK);
 
     switch ( type )
@@ -299,13 +319,15 @@  static int cat_init_feature(const struct cpuid_leaf *regs,
         if ( feat->cos_max < 1 )
             return -ENOENT;
 
+        feat->cat_info.cbm_len = (regs->a & CAT_CBM_LEN_MASK) + 1;
+
         /* We reserve cos=0 as default cbm (all bits within cbm_len are 1). */
-        feat->cos_reg_val[0] = cat_default_val(feat->cbm_len);
+        feat->cos_reg_val[0] = cat_default_val(feat->cat_info.cbm_len);
 
         wrmsrl((type == FEAT_TYPE_L3_CAT ?
                 MSR_IA32_PSR_L3_MASK(0) :
                 MSR_IA32_PSR_L2_MASK(0)),
-               cat_default_val(feat->cbm_len));
+               cat_default_val(feat->cat_info.cbm_len));
 
         break;
 
@@ -316,15 +338,19 @@  static int cat_init_feature(const struct cpuid_leaf *regs,
         if ( feat->cos_max < 3 )
             return -ENOENT;
 
+        feat->cat_info.cbm_len = (regs->a & CAT_CBM_LEN_MASK) + 1;
+
         /* Cut half of cos_max when CDP is enabled. */
         feat->cos_max = (feat->cos_max - 1) >> 1;
 
         /* We reserve cos=0 as default cbm (all bits within cbm_len are 1). */
-        get_cdp_code(feat, 0) = cat_default_val(feat->cbm_len);
-        get_cdp_data(feat, 0) = cat_default_val(feat->cbm_len);
+        get_cdp_code(feat, 0) = cat_default_val(feat->cat_info.cbm_len);
+        get_cdp_data(feat, 0) = cat_default_val(feat->cat_info.cbm_len);
 
-        wrmsrl(MSR_IA32_PSR_L3_MASK(0), cat_default_val(feat->cbm_len));
-        wrmsrl(MSR_IA32_PSR_L3_MASK(1), cat_default_val(feat->cbm_len));
+        wrmsrl(MSR_IA32_PSR_L3_MASK(0),
+               cat_default_val(feat->cat_info.cbm_len));
+        wrmsrl(MSR_IA32_PSR_L3_MASK(1),
+               cat_default_val(feat->cat_info.cbm_len));
         rdmsrl(MSR_IA32_PSR_L3_QOS_CFG, val);
         wrmsrl(MSR_IA32_PSR_L3_QOS_CFG,
                val | (1ull << PSR_L3_QOS_CDP_ENABLE_BIT));
@@ -344,7 +370,45 @@  static int cat_init_feature(const struct cpuid_leaf *regs,
 
     printk(XENLOG_INFO "%s: enabled on socket %u, cos_max:%u, cbm_len:%u\n",
            cat_feat_name[type], cpu_to_socket(smp_processor_id()),
-           feat->cos_max, feat->cbm_len);
+           feat->cos_max, feat->cat_info.cbm_len);
+
+    return 0;
+}
+
+static int mba_init_feature(const struct cpuid_leaf *regs,
+                            struct feat_node *feat,
+                            struct psr_socket_info *info,
+                            enum psr_feat_type type)
+{
+    /* No valid value so do not enable feature. */
+    if ( !regs->a || !regs->d )
+        return -ENOENT;
+
+    if ( type != FEAT_TYPE_MBA )
+        return -ENOENT;
+
+    feat->cos_max = min(opt_cos_max, regs->d & CAT_COS_MAX_MASK);
+    if ( feat->cos_max < 1 )
+        return -ENOENT;
+
+    feat->mba_info.thrtl_max = (regs->a & MBA_THRTL_MAX_MASK) + 1;
+
+    if ( regs->c & MBA_LINEAR )
+        feat->mba_info.linear = 1;
+
+    feat->cos_reg_val[0] = 0;
+    wrmsrl(MSR_IA32_PSR_MBA_MASK(0), 0);
+
+    /* Add this feature into array. */
+    info->features[type] = feat;
+
+    if ( !opt_cpu_info )
+        return 0;
+
+    printk(XENLOG_INFO "MBA: enabled on socket %u, cos_max:%u,"
+           "thrtl_max:%u, linear:%u.\n",
+           cpu_to_socket(smp_processor_id()),
+           feat->cos_max, feat->mba_info.thrtl_max, feat->mba_info.linear);
 
     return 0;
 }
@@ -356,7 +420,7 @@  static bool cat_get_feat_info(const struct feat_node *feat,
         return false;
 
     data[PSR_INFO_IDX_COS_MAX] = feat->cos_max;
-    data[PSR_INFO_IDX_CAT_CBM_LEN] = feat->cbm_len;
+    data[PSR_INFO_IDX_CAT_CBM_LEN] = feat->cat_info.cbm_len;
     data[PSR_INFO_IDX_CAT_FLAG] = 0;
 
     return true;
@@ -422,6 +486,26 @@  static const struct feat_props l2_cat_props = {
     .write_msr = l2_cat_write_msr,
 };
 
+/* MBA props */
+static bool mba_get_feat_info(const struct feat_node *feat,
+                              uint32_t data[], unsigned int array_len)
+{
+    return false;
+}
+
+static void mba_write_msr(unsigned int cos, uint32_t val,
+                          enum psr_val_type type)
+{
+}
+
+static const struct feat_props mba_props = {
+    .cos_num = 1,
+    .type[0] = PSR_VAL_TYPE_MBA,
+    .alt_type = PSR_VAL_TYPE_UNKNOWN,
+    .get_feat_info = mba_get_feat_info,
+    .write_msr = mba_write_msr,
+};
+
 static void __init parse_psr_bool(char *s, char *value, char *feature,
                                   unsigned int mask)
 {
@@ -457,6 +541,7 @@  static void __init parse_psr_param(char *s)
         parse_psr_bool(s, val_str, "cmt", PSR_CMT);
         parse_psr_bool(s, val_str, "cat", PSR_CAT);
         parse_psr_bool(s, val_str, "cdp", PSR_CDP);
+        parse_psr_bool(s, val_str, "mba", PSR_MBA);
 
         if ( val_str && !strcmp(s, "rmid_max") )
             opt_rmid_max = simple_strtoul(val_str, NULL, 0);
@@ -863,7 +948,7 @@  static int insert_val_into_array(uint32_t val[],
     if ( array_len < props->cos_num )
         return -ENOSPC;
 
-    if ( !psr_check_cbm(feat->cbm_len, new_val) )
+    if ( !psr_check_cbm(feat->cat_info.cbm_len, new_val) )
         return -EINVAL;
 
     /*
@@ -1380,6 +1465,10 @@  static int psr_cpu_prepare(void)
          (feat_l2_cat = xzalloc(struct feat_node)) == NULL )
         return -ENOMEM;
 
+    if ( feat_mba == NULL &&
+         (feat_mba = xzalloc(struct feat_node)) == NULL )
+        return -ENOMEM;
+
     return 0;
 }
 
@@ -1389,6 +1478,7 @@  static void psr_cpu_init(void)
     unsigned int socket, cpu = smp_processor_id();
     struct feat_node *feat;
     struct cpuid_leaf regs;
+    uint32_t reg_b;
 
     if ( !psr_alloc_feat_enabled() || !boot_cpu_has(X86_FEATURE_PQE) )
         goto assoc_init;
@@ -1407,7 +1497,8 @@  static void psr_cpu_init(void)
     spin_lock_init(&info->ref_lock);
 
     cpuid_count_leaf(PSR_CPUID_LEVEL_CAT, 0, &regs);
-    if ( regs.b & PSR_RESOURCE_TYPE_L3 )
+    reg_b = regs.b;
+    if ( reg_b & PSR_RESOURCE_TYPE_L3 )
     {
         cpuid_count_leaf(PSR_CPUID_LEVEL_CAT, 1, &regs);
 
@@ -1430,8 +1521,7 @@  static void psr_cpu_init(void)
         }
     }
 
-    cpuid_count_leaf(PSR_CPUID_LEVEL_CAT, 0, &regs);
-    if ( regs.b & PSR_RESOURCE_TYPE_L2 )
+    if ( reg_b & PSR_RESOURCE_TYPE_L2 )
     {
         cpuid_count_leaf(PSR_CPUID_LEVEL_CAT, 2, &regs);
 
@@ -1443,6 +1533,18 @@  static void psr_cpu_init(void)
             feat_l2_cat = feat;
     }
 
+    if ( reg_b & PSR_RESOURCE_TYPE_MBA )
+    {
+        cpuid_count_leaf(PSR_CPUID_LEVEL_CAT, 3, &regs);
+
+        feat = feat_mba;
+        feat_mba = NULL;
+        if ( !mba_init_feature(&regs, feat, info, FEAT_TYPE_MBA) )
+            feat_props[FEAT_TYPE_MBA] = &mba_props;
+        else
+            feat_mba = feat;
+    }
+
     info->feat_init = true;
 
  assoc_init:
@@ -1502,7 +1604,7 @@  static int __init psr_presmp_init(void)
     if ( (opt_psr & PSR_CMT) && opt_rmid_max )
         init_psr_cmt(opt_rmid_max);
 
-    if ( opt_psr & (PSR_CAT | PSR_CDP) )
+    if ( opt_psr & (PSR_CAT | PSR_CDP | PSR_MBA) )
         init_psr();
 
     if ( psr_cpu_prepare() )
diff --git a/xen/include/asm-x86/msr-index.h b/xen/include/asm-x86/msr-index.h
index 4e08de6..41f1677 100644
--- a/xen/include/asm-x86/msr-index.h
+++ b/xen/include/asm-x86/msr-index.h
@@ -348,6 +348,7 @@ 
 #define MSR_IA32_PSR_L3_MASK_CODE(n)	(0x00000c90 + (n) * 2 + 1)
 #define MSR_IA32_PSR_L3_MASK_DATA(n)	(0x00000c90 + (n) * 2)
 #define MSR_IA32_PSR_L2_MASK(n)		(0x00000d10 + (n))
+#define MSR_IA32_PSR_MBA_MASK(n)	(0x00000d50 + (n))
 
 /* Intel Model 6 */
 #define MSR_P6_PERFCTR(n)		(0x000000c1 + (n))
diff --git a/xen/include/asm-x86/psr.h b/xen/include/asm-x86/psr.h
index 1b6f22f..551ccf3 100644
--- a/xen/include/asm-x86/psr.h
+++ b/xen/include/asm-x86/psr.h
@@ -24,6 +24,7 @@ 
 /* Resource Type Enumeration */
 #define PSR_RESOURCE_TYPE_L3            0x2
 #define PSR_RESOURCE_TYPE_L2            0x4
+#define PSR_RESOURCE_TYPE_MBA           0x8
 
 /* L3 Monitoring Features */
 #define PSR_CMT_L3_OCCUPANCY            0x1
@@ -58,6 +59,7 @@  enum psr_val_type {
     PSR_VAL_TYPE_L3_CODE,
     PSR_VAL_TYPE_L3_DATA,
     PSR_VAL_TYPE_L2,
+    PSR_VAL_TYPE_MBA,
     PSR_VAL_TYPE_UNKNOWN,
 };