[PATCH] x86/HPET: don't needlessly set up channels for broadcast

[PATCH] x86/HPET: don't needlessly set up channels for broadcast

[Jan Beulich]

[-- Attachment #1: Type: text/plain, Size: 1141 bytes --]

When there are more FSB delivery capable HPET channels than CPU cores
(or threads), we can simply use a dedicated channel per CPU. This
avoids wasting the resources to handle the excess channels (including
the pointless triggering of the respective interrupt on each
wraparound) as well as the ping-pong of the interrupts' affinities
(when getting assigned to different CPUs).

Signed-off-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/arch/x86/hpet.c
+++ b/xen/arch/x86/hpet.c
@@ -369,7 +369,7 @@ static void __init hpet_fsb_cap_lookup(v
     if ( !hpet_events )
         return;
 
-    for ( i = 0; i < num_chs; i++ )
+    for ( i = 0; i < num_chs && num_hpets_used < nr_cpu_ids; i++ )
     {
         struct hpet_event_channel *ch = &hpet_events[num_hpets_used];
         u32 cfg = hpet_read32(HPET_Tn_CFG(i));
@@ -408,6 +408,9 @@ static struct hpet_event_channel *hpet_g
     if ( num_hpets_used == 0 )
         return hpet_events;
 
+    if ( num_hpets_used >= nr_cpu_ids )
+        return &hpet_events[cpu];
+
     do {
         next = next_channel;
         if ( (i = next + 1) == num_hpets_used )




[-- Attachment #2: x86-HPET-reduce.patch --]
[-- Type: text/plain, Size: 1195 bytes --]

x86/HPET: don't needlessly set up channels for broadcast

When there are more FSB delivery capable HPET channels than CPU cores
(or threads), we can simply use a dedicated channel per CPU. This
avoids wasting the resources to handle the excess channels (including
the pointless triggering of the respective interrupt on each
wraparound) as well as the ping-pong of the interrupts' affinities
(when getting assigned to different CPUs).

Signed-off-by: Jan Beulich <jbeulich@suse.com>

--- a/xen/arch/x86/hpet.c
+++ b/xen/arch/x86/hpet.c
@@ -369,7 +369,7 @@ static void __init hpet_fsb_cap_lookup(v
     if ( !hpet_events )
         return;
 
-    for ( i = 0; i < num_chs; i++ )
+    for ( i = 0; i < num_chs && num_hpets_used < nr_cpu_ids; i++ )
     {
         struct hpet_event_channel *ch = &hpet_events[num_hpets_used];
         u32 cfg = hpet_read32(HPET_Tn_CFG(i));
@@ -408,6 +408,9 @@ static struct hpet_event_channel *hpet_g
     if ( num_hpets_used == 0 )
         return hpet_events;
 
+    if ( num_hpets_used >= nr_cpu_ids )
+        return &hpet_events[cpu];
+
     do {
         next = next_channel;
         if ( (i = next + 1) == num_hpets_used )

[-- Attachment #3: Type: text/plain, Size: 126 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

Re: [PATCH] x86/HPET: don't needlessly set up channels for broadcast

[Keir Fraser]

On 27/09/2012 15:55, "Jan Beulich" <JBeulich@suse.com> wrote:

> When there are more FSB delivery capable HPET channels than CPU cores
> (or threads), we can simply use a dedicated channel per CPU. This
> avoids wasting the resources to handle the excess channels (including
> the pointless triggering of the respective interrupt on each
> wraparound) as well as the ping-pong of the interrupts' affinities
> (when getting assigned to different CPUs).
> 
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

Acked-by: Keir Fraser <keir@xen.org>

> --- a/xen/arch/x86/hpet.c
> +++ b/xen/arch/x86/hpet.c
> @@ -369,7 +369,7 @@ static void __init hpet_fsb_cap_lookup(v
>      if ( !hpet_events )
>          return;
>  
> -    for ( i = 0; i < num_chs; i++ )
> +    for ( i = 0; i < num_chs && num_hpets_used < nr_cpu_ids; i++ )
>      {
>          struct hpet_event_channel *ch = &hpet_events[num_hpets_used];
>          u32 cfg = hpet_read32(HPET_Tn_CFG(i));
> @@ -408,6 +408,9 @@ static struct hpet_event_channel *hpet_g
>      if ( num_hpets_used == 0 )
>          return hpet_events;
>  
> +    if ( num_hpets_used >= nr_cpu_ids )
> +        return &hpet_events[cpu];
> +
>      do {
>          next = next_channel;
>          if ( (i = next + 1) == num_hpets_used )
> 
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel