* [PATCH 1/1] powerpc: Update page in counter for CMM
@ 2008-10-20 22:19 Brian King
2008-10-21 4:36 ` [1/1] " Milton Miller
0 siblings, 1 reply; 6+ messages in thread
From: Brian King @ 2008-10-20 22:19 UTC (permalink / raw)
To: paulus; +Cc: brking, linuxppc-dev
A new field has been added to the VPA as a method for
the client OS to communicate to firmware the number of
page ins it is performing when running collaborative
memory overcommit. The hypervisor will use this information
to better determine if a partition is experiencing memory
pressure and needs more memory allocated to it.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
---
arch/powerpc/include/asm/lppaca.h | 3 ++-
arch/powerpc/kernel/paca.c | 1 +
arch/powerpc/mm/fault.c | 8 ++++++--
3 files changed, 9 insertions(+), 3 deletions(-)
diff -puN arch/powerpc/mm/fault.c~powerpc_vrm_mm_pressure arch/powerpc/mm/fault.c
--- linux-2.6/arch/powerpc/mm/fault.c~powerpc_vrm_mm_pressure 2008-10-20 17:13:25.000000000 -0500
+++ linux-2.6-bjking1/arch/powerpc/mm/fault.c 2008-10-20 17:13:25.000000000 -0500
@@ -30,6 +30,8 @@
#include <linux/kprobes.h>
#include <linux/kdebug.h>
+#include <asm/atomic.h>
+#include <asm/firmware.h>
#include <asm/page.h>
#include <asm/pgtable.h>
#include <asm/mmu.h>
@@ -318,9 +320,11 @@ good_area:
goto do_sigbus;
BUG();
}
- if (ret & VM_FAULT_MAJOR)
+ if (ret & VM_FAULT_MAJOR) {
current->maj_flt++;
- else
+ if (firmware_has_feature(FW_FEATURE_CMO))
+ atomic_inc((atomic_t *)(&(get_lppaca()->page_ins)));
+ } else
current->min_flt++;
up_read(&mm->mmap_sem);
return 0;
diff -puN arch/powerpc/include/asm/lppaca.h~powerpc_vrm_mm_pressure arch/powerpc/include/asm/lppaca.h
--- linux-2.6/arch/powerpc/include/asm/lppaca.h~powerpc_vrm_mm_pressure 2008-10-20 17:13:25.000000000 -0500
+++ linux-2.6-bjking1/arch/powerpc/include/asm/lppaca.h 2008-10-20 17:13:25.000000000 -0500
@@ -133,7 +133,8 @@ struct lppaca {
//=============================================================================
// CACHE_LINE_4-5 0x0180 - 0x027F Contains PMC interrupt data
//=============================================================================
- u8 pmc_save_area[256]; // PMC interrupt Area x00-xFF
+ volatile u32 page_ins; // CMO Hint - # page ins by OS x00-x04
+ u8 pmc_save_area[252]; // PMC interrupt Area x04-xFF
} __attribute__((__aligned__(0x400)));
extern struct lppaca lppaca[];
diff -puN arch/powerpc/kernel/paca.c~powerpc_vrm_mm_pressure arch/powerpc/kernel/paca.c
--- linux-2.6/arch/powerpc/kernel/paca.c~powerpc_vrm_mm_pressure 2008-10-20 17:13:25.000000000 -0500
+++ linux-2.6-bjking1/arch/powerpc/kernel/paca.c 2008-10-20 17:13:25.000000000 -0500
@@ -37,6 +37,7 @@ struct lppaca lppaca[] = {
.end_of_quantum = 0xfffffffffffffffful,
.slb_count = 64,
.vmxregs_in_use = 0,
+ .page_ins = 0,
},
};
_
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [1/1] powerpc: Update page in counter for CMM
2008-10-20 22:19 [PATCH 1/1] powerpc: Update page in counter for CMM Brian King
@ 2008-10-21 4:36 ` Milton Miller
2008-10-21 20:24 ` Brian King
2008-10-21 20:27 ` [PATCHv2 1/1] " Brian King
0 siblings, 2 replies; 6+ messages in thread
From: Milton Miller @ 2008-10-21 4:36 UTC (permalink / raw)
To: Brian King; +Cc: linuxppc-dev
X-Patchwork-Id: 5144
On Mon Oct 20, 2008 near 12:19:21 GMT, Brian King wrote:
>
> A new field has been added to the VPA as a method for
> the client OS to communicate to firmware the number of
> page ins it is performing when running collaborative
> memory overcommit. The hypervisor will use this information
> to better determine if a partition is experiencing memory
> pressure and needs more memory allocated to it.
>
> Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
> ---
>
> arch/powerpc/include/asm/lppaca.h | 3 ++-
> arch/powerpc/kernel/paca.c | 1 +
> arch/powerpc/mm/fault.c | 8 ++++++--
> 3 files changed, 9 insertions(+), 3 deletions(-)
>
> diff -puN arch/powerpc/mm/fault.c~powerpc_vrm_mm_pressure arch/powerpc/mm/fault.c
> --- linux-2.6/arch/powerpc/mm/fault.c~powerpc_vrm_mm_pressure 2008-10-20 17:13:25.000000000 -0500
> +++ linux-2.6-bjking1/arch/powerpc/mm/fault.c 2008-10-20 17:13:25.000000000 -0500
..
> @@ -318,9 +320,11 @@ good_area:
> goto do_sigbus;
> BUG();
> }
> - if (ret & VM_FAULT_MAJOR)
> + if (ret & VM_FAULT_MAJOR) {
> current->maj_flt++;
> - else
> + if (firmware_has_feature(FW_FEATURE_CMO))
> + atomic_inc((atomic_t *)(&(get_lppaca()->page_ins)));
> + } else
> current->min_flt++;
> up_read(&mm->mmap_sem);
> return 0;
(1) why do we need atomic_inc and the hundreds or thousands of cycles to
do the atomic inc in a per-cpu area?
(2) assuming we make this a normal increment, should we keep the
feature test or just do it unconditionally (ie is the additional load
and branch worse that just doing the load and store of the counter --
the address was previously reserved, right? (no question if
it has to be atomic).
<Ramble things one might consider>
Ben asked if we need to worry about the hypervisor clearing the
count. If they treat it as a only-incrementing then we don't need
to worry. And since its only an indicator, then we may not care
about a big count by them interrupting between the load and store
If we are worried about linux preemption, then we need to disable
it to avoid crossing cpu variables or getting to this point multiple
times. (I have not looked at the context to see if we are already
disabled).
</Ramble>
milton
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [1/1] powerpc: Update page in counter for CMM
2008-10-21 4:36 ` [1/1] " Milton Miller
@ 2008-10-21 20:24 ` Brian King
2008-10-21 20:27 ` [PATCHv2 1/1] " Brian King
1 sibling, 0 replies; 6+ messages in thread
From: Brian King @ 2008-10-21 20:24 UTC (permalink / raw)
To: Milton Miller; +Cc: linuxppc-dev
Milton Miller wrote:
> X-Patchwork-Id: 5144
>> diff -puN arch/powerpc/mm/fault.c~powerpc_vrm_mm_pressure arch/powerpc/mm/fault.c
>> --- linux-2.6/arch/powerpc/mm/fault.c~powerpc_vrm_mm_pressure 2008-10-20 17:13:25.000000000 -0500
>> +++ linux-2.6-bjking1/arch/powerpc/mm/fault.c 2008-10-20 17:13:25.000000000 -0500
> ..
>> @@ -318,9 +320,11 @@ good_area:
>> goto do_sigbus;
>> BUG();
>> }
>> - if (ret & VM_FAULT_MAJOR)
>> + if (ret & VM_FAULT_MAJOR) {
>> current->maj_flt++;
>> - else
>> + if (firmware_has_feature(FW_FEATURE_CMO))
>> + atomic_inc((atomic_t *)(&(get_lppaca()->page_ins)));
>> + } else
>> current->min_flt++;
>> up_read(&mm->mmap_sem);
>> return 0;
>
> (1) why do we need atomic_inc and the hundreds or thousands of cycles to
> do the atomic inc in a per-cpu area?
We don't. I've now removed it.
> (2) assuming we make this a normal increment, should we keep the
> feature test or just do it unconditionally (ie is the additional load
> and branch worse that just doing the load and store of the counter --
> the address was previously reserved, right? (no question if
> it has to be atomic).
Simplified patch on the way...
> <Ramble things one might consider>
>
> Ben asked if we need to worry about the hypervisor clearing the
> count. If they treat it as a only-incrementing then we don't need
> to worry. And since its only an indicator, then we may not care
> about a big count by them interrupting between the load and store
This is a read only field from the hypervisor's perspective. They
shouldn't be clearing it.
> If we are worried about linux preemption, then we need to disable
> it to avoid crossing cpu variables or getting to this point multiple
> times. (I have not looked at the context to see if we are already
> disabled).
Looks to me like linux preemption is a possibility in this code, so
I've added the code to disable preemption around the increment.
-Brian
--
Brian King
Linux on Power Virtualization
IBM Linux Technology Center
^ permalink raw reply [flat|nested] 6+ messages in thread* [PATCHv2 1/1] powerpc: Update page in counter for CMM
2008-10-21 4:36 ` [1/1] " Milton Miller
2008-10-21 20:24 ` Brian King
@ 2008-10-21 20:27 ` Brian King
2008-10-22 0:11 ` Benjamin Herrenschmidt
1 sibling, 1 reply; 6+ messages in thread
From: Brian King @ 2008-10-21 20:27 UTC (permalink / raw)
To: Paul Mackerras; +Cc: linuxppc-dev, Milton Miller
A new field has been added to the VPA as a method for
the client OS to communicate to firmware the number of
page ins it is performing when running collaborative
memory overcommit. The hypervisor will use this information
to better determine if a partition is experiencing memory
pressure and needs more memory allocated to it.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
---
arch/powerpc/include/asm/lppaca.h | 3 ++-
arch/powerpc/kernel/paca.c | 1 +
arch/powerpc/mm/fault.c | 7 +++++--
3 files changed, 8 insertions(+), 3 deletions(-)
diff -puN arch/powerpc/mm/fault.c~powerpc_vrm_mm_pressure arch/powerpc/mm/fault.c
--- linux-2.6/arch/powerpc/mm/fault.c~powerpc_vrm_mm_pressure 2008-10-20 17:13:25.000000000 -0500
+++ linux-2.6-bjking1/arch/powerpc/mm/fault.c 2008-10-21 14:52:24.000000000 -0500
@@ -318,9 +318,12 @@ good_area:
goto do_sigbus;
BUG();
}
- if (ret & VM_FAULT_MAJOR)
+ if (ret & VM_FAULT_MAJOR) {
current->maj_flt++;
- else
+ preempt_disable();
+ get_lppaca()->page_ins++;
+ preempt_enable();
+ } else
current->min_flt++;
up_read(&mm->mmap_sem);
return 0;
diff -puN arch/powerpc/include/asm/lppaca.h~powerpc_vrm_mm_pressure arch/powerpc/include/asm/lppaca.h
--- linux-2.6/arch/powerpc/include/asm/lppaca.h~powerpc_vrm_mm_pressure 2008-10-20 17:13:25.000000000 -0500
+++ linux-2.6-bjking1/arch/powerpc/include/asm/lppaca.h 2008-10-21 13:46:45.000000000 -0500
@@ -133,7 +133,8 @@ struct lppaca {
//=============================================================================
// CACHE_LINE_4-5 0x0180 - 0x027F Contains PMC interrupt data
//=============================================================================
- u8 pmc_save_area[256]; // PMC interrupt Area x00-xFF
+ u32 page_ins; // CMO Hint - # page ins by OS x00-x04
+ u8 pmc_save_area[252]; // PMC interrupt Area x04-xFF
} __attribute__((__aligned__(0x400)));
extern struct lppaca lppaca[];
diff -puN arch/powerpc/kernel/paca.c~powerpc_vrm_mm_pressure arch/powerpc/kernel/paca.c
--- linux-2.6/arch/powerpc/kernel/paca.c~powerpc_vrm_mm_pressure 2008-10-20 17:13:25.000000000 -0500
+++ linux-2.6-bjking1/arch/powerpc/kernel/paca.c 2008-10-20 17:13:25.000000000 -0500
@@ -37,6 +37,7 @@ struct lppaca lppaca[] = {
.end_of_quantum = 0xfffffffffffffffful,
.slb_count = 64,
.vmxregs_in_use = 0,
+ .page_ins = 0,
},
};
_
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCHv2 1/1] powerpc: Update page in counter for CMM
2008-10-21 20:27 ` [PATCHv2 1/1] " Brian King
@ 2008-10-22 0:11 ` Benjamin Herrenschmidt
2008-10-22 15:53 ` [PATCHv3 " Brian King
0 siblings, 1 reply; 6+ messages in thread
From: Benjamin Herrenschmidt @ 2008-10-22 0:11 UTC (permalink / raw)
To: Brian King; +Cc: linuxppc-dev, Paul Mackerras, Milton Miller
> - if (ret & VM_FAULT_MAJOR)
> + if (ret & VM_FAULT_MAJOR) {
> current->maj_flt++;
> - else
> + preempt_disable();
> + get_lppaca()->page_ins++;
> + preempt_enable();
> + } else
That isn't going to work very well on 32-bit or non-pseries
platforms :-)
We need either something like a ppc_md. callback or at least a test for
firmware feature... (which probably still needs an ifdef to build on
32 bit).
Cheers,
Ben.
^ permalink raw reply [flat|nested] 6+ messages in thread* [PATCHv3 1/1] powerpc: Update page in counter for CMM
2008-10-22 0:11 ` Benjamin Herrenschmidt
@ 2008-10-22 15:53 ` Brian King
0 siblings, 0 replies; 6+ messages in thread
From: Brian King @ 2008-10-22 15:53 UTC (permalink / raw)
To: benh; +Cc: linuxppc-dev, Paul Mackerras, Milton Miller
A new field has been added to the VPA as a method for
the client OS to communicate to firmware the number of
page ins it is performing when running collaborative
memory overcommit. The hypervisor will use this information
to better determine if a partition is experiencing memory
pressure and needs more memory allocated to it.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
---
arch/powerpc/include/asm/lppaca.h | 3 ++-
arch/powerpc/kernel/paca.c | 1 +
arch/powerpc/mm/fault.c | 12 ++++++++++--
3 files changed, 13 insertions(+), 3 deletions(-)
diff -puN arch/powerpc/mm/fault.c~powerpc_vrm_mm_pressure arch/powerpc/mm/fault.c
--- linux-2.6/arch/powerpc/mm/fault.c~powerpc_vrm_mm_pressure 2008-10-20 17:13:25.000000000 -0500
+++ linux-2.6-bjking1/arch/powerpc/mm/fault.c 2008-10-22 08:53:13.000000000 -0500
@@ -30,6 +30,7 @@
#include <linux/kprobes.h>
#include <linux/kdebug.h>
+#include <asm/firmware.h>
#include <asm/page.h>
#include <asm/pgtable.h>
#include <asm/mmu.h>
@@ -318,9 +319,16 @@ good_area:
goto do_sigbus;
BUG();
}
- if (ret & VM_FAULT_MAJOR)
+ if (ret & VM_FAULT_MAJOR) {
current->maj_flt++;
- else
+#ifdef CONFIG_PPC_SMLPAR
+ if (firmware_has_feature(FW_FEATURE_CMO)) {
+ preempt_disable();
+ get_lppaca()->page_ins++;
+ preempt_enable();
+ }
+#endif
+ } else
current->min_flt++;
up_read(&mm->mmap_sem);
return 0;
diff -puN arch/powerpc/include/asm/lppaca.h~powerpc_vrm_mm_pressure arch/powerpc/include/asm/lppaca.h
--- linux-2.6/arch/powerpc/include/asm/lppaca.h~powerpc_vrm_mm_pressure 2008-10-20 17:13:25.000000000 -0500
+++ linux-2.6-bjking1/arch/powerpc/include/asm/lppaca.h 2008-10-21 13:46:45.000000000 -0500
@@ -133,7 +133,8 @@ struct lppaca {
//=============================================================================
// CACHE_LINE_4-5 0x0180 - 0x027F Contains PMC interrupt data
//=============================================================================
- u8 pmc_save_area[256]; // PMC interrupt Area x00-xFF
+ u32 page_ins; // CMO Hint - # page ins by OS x00-x04
+ u8 pmc_save_area[252]; // PMC interrupt Area x04-xFF
} __attribute__((__aligned__(0x400)));
extern struct lppaca lppaca[];
diff -puN arch/powerpc/kernel/paca.c~powerpc_vrm_mm_pressure arch/powerpc/kernel/paca.c
--- linux-2.6/arch/powerpc/kernel/paca.c~powerpc_vrm_mm_pressure 2008-10-20 17:13:25.000000000 -0500
+++ linux-2.6-bjking1/arch/powerpc/kernel/paca.c 2008-10-20 17:13:25.000000000 -0500
@@ -37,6 +37,7 @@ struct lppaca lppaca[] = {
.end_of_quantum = 0xfffffffffffffffful,
.slb_count = 64,
.vmxregs_in_use = 0,
+ .page_ins = 0,
},
};
_
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2008-10-22 15:54 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-10-20 22:19 [PATCH 1/1] powerpc: Update page in counter for CMM Brian King
2008-10-21 4:36 ` [1/1] " Milton Miller
2008-10-21 20:24 ` Brian King
2008-10-21 20:27 ` [PATCHv2 1/1] " Brian King
2008-10-22 0:11 ` Benjamin Herrenschmidt
2008-10-22 15:53 ` [PATCHv3 " Brian King
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).