Hi Julien,
On Mon, Jul 24, 2017 at 4:54 PM, Julien Grall wrote:
> Hi Vijay,
>
>
> On 18/07/17 12:41, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> For each cpu, update cpu_to_node[] with node id from
>> the numa-node-id DT property. Also, initialize cpu_to_node[]
>> with node 0.
Hi Julien,
On Thu, Jul 20, 2017 at 4:56 PM, Julien Grall wrote:
>
>
> On 19/07/17 19:39, Julien Grall wrote:
>>>
>>> cell = (const __be32 *)prop->data;
>>> banks = fdt32_to_cpu(prop->len) / (reg_cells * sizeof (u32));
>>>
>>> -for ( i = 0; i < banks && bootinfo.mem.nr_banks < NR_MEM
On Thu, Jul 20, 2017 at 5:39 PM, Julien Grall wrote:
>
>
> On 20/07/17 13:05, Vijay Kilari wrote:
>>
>> On Thu, Jul 20, 2017 at 4:30 PM, Julien Grall
>> wrote:
>>>
>>> Hi Vijay,
>>>
>>>
>>> On 20/07/17 08:00, Vijay Ki
On Thu, Jul 20, 2017 at 4:30 PM, Julien Grall wrote:
> Hi Vijay,
>
>
> On 20/07/17 08:00, Vijay Kilari wrote:
>>
>> On Wed, Jul 19, 2017 at 9:53 PM, Julien Grall
>> wrote:
>>>
>>> Hi Vijay,
>>>
>>> On 18/07/17 12:41,
Hi Julien,
On Thu, Jul 20, 2017 at 4:47 PM, Julien Grall wrote:
> Hi Vijay,
>
> On 18/07/17 12:41, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Move code from xen/arch/x86/srat.c to xen/common/numa.c
>> so that it can be used by other archs.
>>
>> Apart from moving the code the
On Thu, Jul 20, 2017 at 12:09 AM, Julien Grall wrote:
> Hi Vijay,
>
> On 18/07/17 12:41, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Parse memory node and fetch numa-node-id information.
>> For each memory range, store in node_memblk_range[]
>> along with node id.
>>
>> When bo
On Wed, Jul 19, 2017 at 11:28 PM, Julien Grall wrote:
> Hi Vijay,
>
> On 18/07/17 12:41, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> The common code allows numa initialization only when
>> ACPI_NUMA config is enabled. Allow initialization when
>> NUMA config is enabled for DT.
On Tue, Jul 18, 2017 at 11:36 PM, Julien Grall wrote:
> Hi Vijay,
>
> On 18/07/17 12:41, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Right now CONFIG_NUMA is not enabled for ARM and
>> existing code in asm-arm/numa.h is for !CONFIG_NUMA.
>> Hence put this code under #ifndef CON
On Wed, Jul 19, 2017 at 11:56 PM, Julien Grall wrote:
> Hi,
>
>
> On 18/07/17 12:41, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Parse CPU node and fetch numa-node-id information.
>> For each node-id found, update nodemask_t mask.
>> Refer to Documentation/devicetree/bindings/n
On Wed, Jul 19, 2017 at 11:11 PM, Julien Grall wrote:
> Hi Vijay,
>
> On 18/07/17 12:41, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Move code from xen/arch/x86/numa.c to xen/common/numa.c
>> so that it can be used by other archs.
>>
>> The following changes are done:
>> - Few
On Wed, Jul 19, 2017 at 10:48 PM, Julien Grall wrote:
>
>
> On 19/07/17 07:40, Vijay Kilari wrote:
>>
>> On Tue, Jul 18, 2017 at 8:59 PM, Wei Liu wrote:
>>>
>>> On Tue, Jul 18, 2017 at 05:11:27PM +0530, vijay.kil...@gmail.com wrote:
>>>>
>
On Wed, Jul 19, 2017 at 9:25 PM, Julien Grall wrote:
> Hi Vijay,
>
>
> On 19/07/2017 08:00, Vijay Kilari wrote:
>>
>> On Tue, Jul 18, 2017 at 11:25 PM, Julien Grall
>> wrote:
>>>
>>> Hi,
>>>
>>>
>>> On 18/07/
On Wed, Jul 19, 2017 at 9:53 PM, Julien Grall wrote:
> Hi Vijay,
>
> On 18/07/17 12:41, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Fix coding style, trailing spaces, tabs in NUMA code.
>> Also drop unused macros and functions.
>> There is no functional change.
>>
>> Signed-off
On Wed, Jul 19, 2017 at 10:42 PM, Julien Grall wrote:
> Hi Vijay,
>
> On 18/07/17 12:41, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> memnode_shift variable is changed from int to unsigned int.
>> With this change, compute_memnode_shift() returns error value
>> instead of return
On Tue, Jul 18, 2017 at 11:25 PM, Julien Grall wrote:
> Hi,
>
>
> On 18/07/17 12:41, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Introduce NR_NODES config option to specify number
>> of NUMA nodes supported. By default value is set at
>> 64 for x86 and 8 for arm. Dropped NODES_
On Tue, Jul 18, 2017 at 11:46 PM, Julien Grall wrote:
> Hi,
>
>
> On 18/07/17 16:29, Wei Liu wrote:
>>
>> On Tue, Jul 18, 2017 at 05:11:30PM +0530, vijay.kil...@gmail.com wrote:
>> [...]
>>>
>>> diff --git a/xen/common/numa.c b/xen/common/numa.c
>>> new file mode 100644
>>> index 000..0381f1b
On Tue, Jul 18, 2017 at 8:59 PM, Wei Liu wrote:
> On Tue, Jul 18, 2017 at 05:11:27PM +0530, vijay.kil...@gmail.com wrote:
>> From: Vijaya Kumar K
>>
>> Add accessors for nodes[] and other static variables and
>> use those accessors. These variables are later accessed
>> outside the file when the
On Tue, Jul 18, 2017 at 9:06 PM, Wei Liu wrote:
> On Tue, Jul 18, 2017 at 05:11:40PM +0530, vijay.kil...@gmail.com wrote:
>> From: Vijaya Kumar K
>>
>> Move SRAT handling code which is common across
>> architectures is moved to new file xen/drivers/acpi/srat.c
>> from xen/arch/x86/srat.c file. Ne
On Tue, Jul 18, 2017 at 9:48 PM, Julien Grall wrote:
> Hi,
>
> On 18/07/17 12:41, vijay.kil...@gmail.com wrote:
>>
>> This patch is tested on Thunderx platform.
>> No changes are made to x86 implementation only code is sanitized
>> and refactored. Hence only compilation tested for x86.
>>
>> This
From: Vijaya Kumar K
CONFIG_NUMA is defined in xen/drivers/acpi/Kconfig.
Move to common/Kconfig and enabled by default.
Also, NUMA feature uses PDX for physical address to
memory node mapping. Hence make HAS_PDX dependent
for NUMA.
Signed-off-by: Vijaya Kumar K
---
xen/common/Kconfig | 4
From: Vijaya Kumar K
Parse distance-matrix and fetch node distance information.
Store distance information in node_distance[].
Register dt_node_distance() function pointer with
the ARM numa code. This approach can be later used for
ACPI.
Signed-off-by: Vijaya Kumar K
---
v3: - Moved __node_dis
From: Vijaya Kumar K
Add CONFIG_ACPI_NUMA to xen/drivers/acpi/Kconfig and
drop CONFIG_ACPI_NUMA set in asm-x86/config.h.
Signed-off-by: Vijaya Kumar K
---
xen/drivers/acpi/Kconfig | 4
xen/include/asm-x86/config.h | 1 -
2 files changed, 4 insertions(+), 1 deletion(-)
diff --git a/xe
From: Vijaya Kumar K
Parse CPU node and fetch numa-node-id information.
For each node-id found, update nodemask_t mask.
Refer to Documentation/devicetree/bindings/numa.txt
in linux kernel.
Signed-off-by: Vijaya Kumar K
---
v3: - Parse cpu nodes under path /cpus
- Move changes to bootfdt.c a
From: Vijaya Kumar K
Move code from xen/arch/x86/srat.c to xen/common/numa.c
so that it can be used by other archs.
Apart from moving the code the following changes are done
- Coding style of code moved to numa.c is changed to xen style
- {memory,processor}_nodes_parsed are made global and mov
From: Vijaya Kumar K
Implement arch_sanitize_nodes_memory() which looks at all banks
in bootinfo.mem, update nodes[] with corresponding nodeid.
Call numa_scan_nodes() generic function with ram start and
end address, which takes care of further computing memnodeshift
and populating memnodemap[] us
From: Vijaya Kumar K
Register SRAT entry handler for type
ACPI_SRAT_TYPE_GICC_AFFINITY to parse SRAT table
and extract proximity for all CPU IDs.
Signed-off-by: Vijaya Kumar
---
xen/arch/arm/acpi/boot.c | 2 +
xen/arch/arm/numa/acpi_numa.c | 124 +
From: Vijaya Kumar K
SRAT's X2APIC_CPU_AFFINITY and CPU_AFFINITY types are not used
by ARM. Hence move handling of this SRAT types to arch specific
file and handle them under arch_table_parse_srat().
Signed-off-by: Vijaya Kumar K
---
xen/arch/arm/numa/acpi_numa.c | 5 +
xen/arch/x86/srat.
From: Vijaya Kumar K
Call ACPI NUMA initialization under CONFIG_ACPI_NUMA.
Signed-off-by: Vijaya Kumar
---
xen/arch/arm/numa/acpi_numa.c | 27 ++-
xen/arch/arm/numa/numa.c | 15 +--
xen/common/numa.c | 14 ++
xen/include/asm-arm/
From: Vijaya Kumar K
Parse MADT table and extract MPIDR for all
CPU IDs in MADT ACPI_MADT_TYPE_GENERIC_INTERRUPT entries
and store in cpuid_to_hwid_map[]
This mapping is used by SRAT table parsing to extract MPIDR
of the CPU ID.
MADT table is also parsed in arm/acpi/boot.c during smp boot.
Howe
From: Vijaya Kumar K
For each cpu, update cpu_to_node[] with node id from
the numa-node-id DT property. Also, initialize cpu_to_node[]
with node 0.
Add macros to access cpu_to_node[] information.
Signed-off-by: Vijaya Kumar K
---
v3: - Dropped numa_add_cpu declaration from asm-arm/numa.h
-
From: Vijaya Kumar K
Delete numa-node-id and distance map from DOM0 DT
so that NUMA information is not exposed to DOM0.
This helps particularly to boot Node 1 devices
as if booting on Node0.
However this approach has limitation where memory allocation
for the devices should be local.
Also, do n
From: Vijaya Kumar K
Right now CONFIG_NUMA is not enabled for ARM and
existing code in asm-arm/numa.h is for !CONFIG_NUMA.
Hence put this code under #ifndef CONFIG_NUMA.
This help to make this changes work when CONFIG_NUMA
is not enabled. Though CONFIG_NUMA is enabled by default,
manually disabl
From: Vijaya Kumar K
Move SRAT handling code which is common across
architectures is moved to new file xen/drivers/acpi/srat.c
from xen/arch/x86/srat.c file. New header file srat.h is
introduced.
Other major changes are:
- Coding style of code moved is changed.
- Moved struct pxm2node from srat.
From: Vijaya Kumar K
Introduce new api device_tree_type_matches() to check for
device type. Also export device_tree_get_u32() and
device_tree_node_compatible()
These functions are later used for parsing NUMA information.
Signed-off-by: Vijaya Kumar K
---
v3: Export device_tree_node_compatible(
From: Vijaya Kumar K
Parse memory node and fetch numa-node-id information.
For each memory range, store in node_memblk_range[]
along with node id.
When booting in UEFI mode, UEFI passes memory information
to Dom0 using EFI memory descriptor table and deletes the
memory nodes from the host DT. Ho
From: Vijaya Kumar K
Introduce NR_NODES config option to specify number
of NUMA nodes supported. By default value is set at
64 for x86 and 8 for arm. Dropped NODES_SHIFT macro.
Also move NR_NODE_MEMBLKS from asm-x86/acpi.h to xen/numa.h
Signed-off-by: Vijaya Kumar K
---
xen/arch/Kconfig
From: Vijaya Kumar K
With this RFC patch series, NUMA support is added for ARM platform.
Both DT and ACPI based NUMA support is added.
Only Xen is made aware of NUMA platform. NUMA awareness to DOM0 is not
added.
As part of this series, the code under x86 architecture is
reused by moving into co
From: Vijaya Kumar K
The common code allows numa initialization only when
ACPI_NUMA config is enabled. Allow initialization when
NUMA config is enabled for DT.
In this patch, along with acpi_numa, check for acpi_disabled
is added.
Signed-off-by: Vijaya Kumar K
---
xen/common/numa.c | 4 +---
From: Vijaya Kumar K
Change u{8,32,64} to uint{8,32,64}_t, u64 to paddr_t
wherever applicable.
Fix attributes coding styles.
Also changed
- Some variables from int to unsigned int
- Used pfn_to_paddr/paddr_to_pfn whereever required.
- Alloc memnodemap[] of size BITS_PER_LONG.
Signed-off-by
From: Vijaya Kumar K
memnode_shift variable is changed from int to unsigned int.
With this change, compute_memnode_shift() returns error value
instead of returning shift value. The memnode_shift is updated inside
compute_memnode_shift().
Also, following changes are made
- Rename compute_hash_s
From: Vijaya Kumar K
Add accessors for nodes[] and other static variables and
use those accessors. These variables are later accessed
outside the file when the code made generic in later
patches. However the coding style is not changed.
Signed-off-by: Vijaya Kumar K
---
v3: - Changed accessors
From: Vijaya Kumar K
Move code from xen/arch/x86/numa.c to xen/common/numa.c
so that it can be used by other archs.
The following changes are done:
- Few generic static functions in x86/numa.c is made
non-static common/numa.c
- The generic contents of header file asm-x86/numa.h
are moved to
From: Vijaya Kumar K
Fix coding style, trailing spaces, tabs in NUMA code.
Also drop unused macros and functions.
There is no functional change.
Signed-off-by: Vijaya Kumar K
Reviewed-by: Wei Liu
---
v3: - Change commit message
- Changed VIRTUAL_BUG_ON to ASSERT
- Dropped useless inner
From: Vijaya Kumar K
Rename some function in ACPI code as follow
- Rename setup_node to acpi_setup_node
- Rename bad_srat to numa_failed
- Rename nodes_cover_memory to arch_sanitize_nodes_memory
and changed return type to bool
- Rename acpi_scan_nodes to numa_scan_nodes
Also introduce res
Hi Jan,
Sorry for late reply.
On Fri, Jun 30, 2017 at 7:35 PM, Jan Beulich wrote:
03/28/17 5:54 PM >>>
>> --- a/xen/arch/x86/numa.c
>> +++ b/xen/arch/x86/numa.c
>> @@ -53,15 +53,15 @@ int srat_disabled(void)
>> /*
>> * Given a shift value, try to populate memnodemap[]
>> * Returns :
On Wed, May 31, 2017 at 4:07 PM, Jan Beulich wrote:
On 31.05.17 at 12:18, wrote:
>> On 31/05/17 11:04, Jan Beulich wrote:
>> On 28.03.17 at 17:53, wrote:
--- a/xen/common/Kconfig
+++ b/xen/common/Kconfig
@@ -41,6 +41,10 @@ config HAS_GDBSX
config HAS_IOPORTS
and
On Mon, May 8, 2017 at 10:36 PM, Julien Grall wrote:
> Hi Vijay,
>
>
> On 28/03/17 16:53, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Move code from xen/arch/x86/srat.c to xen/common/numa.c
>> so that it can be used by other archs.
>> Few generic static functions in x86/s
On Mon, May 8, 2017 at 11:01 PM, Julien Grall wrote:
> Hi Vijay,
>
> The title likely needs to have the work device-tree/DT in it.
>
> On 28/03/17 16:53, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Parse CPU node and fetch numa-node-id information.
>> For each node-id found, up
On Mon, May 8, 2017 at 10:21 PM, Julien Grall wrote:
> On 28/03/17 16:53, vijay.kil...@gmail.com wrote:
>>
>> diff --git a/xen/common/numa.c b/xen/common/numa.c
>> new file mode 100644
>> index 000..207ebd8
>> --- /dev/null
>> +++ b/xen/common/numa.c
>> @@ -0,0 +1,488 @@
>> +/*
>> + * Common N
On Mon, May 8, 2017 at 10:11 PM, Julien Grall wrote:
> Hi Vijay,
>
>
> On 28/03/17 16:53, vijay.kil...@gmail.com wrote:
>>
>> diff --git a/xen/arch/x86/numa.c b/xen/arch/x86/numa.c
>> index 3bdab9a..33c6806 100644
>> --- a/xen/arch/x86/numa.c
>> +++ b/xen/arch/x86/numa.c
>> @@ -10,286 +10,13 @@
>>
On Mon, May 8, 2017 at 9:28 PM, Julien Grall wrote:
> Hi Vijay,
>
> On 28/03/17 16:53, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Right now CONFIG_NUMA is not enabled for ARM and
>> existing code in asm-arm/numa.h is for !CONFIG_NUMA.
>> Hence put this code under #ifndef CONFI
On Mon, May 8, 2017 at 8:09 PM, Julien Grall wrote:
> Hi Vijay,
>
> On 28/03/17 16:53, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Add accessor for nodes[] and other static variables and
>
>
> s/accessor/accessors/
>
>> used those accessors.
>
>
> Also, I am not sure to underst
On Tue, Apr 25, 2017 at 9:13 PM, Jan Beulich wrote:
>>>> On 25.04.17 at 17:14, wrote:
>> On 25/04/17 15:54, Vijay Kilari wrote:
>>> On Tue, Apr 25, 2017 at 5:58 PM, Julien Grall wrote:
>>>>>>>
>>>>>>> By setting 1, we
On Tue, Apr 25, 2017 at 5:58 PM, Julien Grall wrote:
>
>
> On 25/04/17 13:20, Vijay Kilari wrote:
>>
>> On Tue, Apr 25, 2017 at 5:34 PM, Julien Grall
>> wrote:
>>>
>>> Hello Vijay,
>>>
>>> On 25/04/17 07:54, Vijay Kilari wrot
On Tue, Apr 25, 2017 at 5:34 PM, Julien Grall wrote:
> Hello Vijay,
>
> On 25/04/17 07:54, Vijay Kilari wrote:
>>
>> On Thu, Apr 20, 2017 at 9:29 PM, Julien Grall
>> wrote:
>>>
>>> Hi Vijay,
>>>
>>> On 28/03/17 16:53, vi
On Thu, Apr 20, 2017 at 9:42 PM, Julien Grall wrote:
> Hi Vijay,
>
>
> On 28/03/17 16:53, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Split numa_initmem_init() so that the numa fallback code is moved
>> as separate function which is generic.
>>
>> Signed-off-by: Vijaya Kumar K
On Thu, Apr 20, 2017 at 9:29 PM, Julien Grall wrote:
> Hi Vijay,
>
> On 28/03/17 16:53, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> Add accessor functions for acpi_numa and numa_off static
>> variables. Init value of acpi_numa is set 1 instead of 0.
>
>
> Please explain why you
Hi Andre,
On Sat, Apr 8, 2017 at 3:37 AM, Andre Przywara wrote:
> Hi,
>
> an only slightly modified repost of the last version.
> I added the Reviewed-by: and Acked-by: tags from Stefano and Julien
> and rebased on top of the latest staging tree:
> commit 89216c7999eb5b8558bfac7d61ae0d5ab844ce3f
From: Vijaya Kumar K
On ARM platforms with NUMA, while initializing second memory node,
panic is triggered from init_node_heap() when virt_to_mfn()
is called for DIRECTMAP_VIRT region address because DIRECTMAP_VIRT
region is not mapped to any virtual address.
The check virt_to_mfn() here is used
Hi Julien,
On Mon, Apr 3, 2017 at 3:31 PM, Julien Grall wrote:
> Hi Vijay,
>
> On 28/03/17 13:35, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> On ARM64, virt_to_mfn uses the hardware for address
>> translation. So if the virtual address is not mapped translation
>> fault is rai
On Fri, Mar 31, 2017 at 11:35 PM, Andre Przywara wrote:
> The MAPTI commands associates a DeviceID/EventID pair with a LPI/CPU
> pair and actually instantiates LPI interrupts.
> We connect the already allocated host LPI to this virtual LPI, so that
> any triggering IRQ on the host can be quickly f
Hi Andre,
On Fri, Mar 31, 2017 at 11:35 PM, Andre Przywara wrote:
> The ITS uses device IDs to map LPIs to a device. Dom0 will later use
> those IDs, which we directly pass on to the host.
> For this we have to map each device that Dom0 may request to a host
> ITS device with the same identifier.
Hi Andre,
On Thu, Mar 16, 2017 at 4:50 PM, Andre Przywara wrote:
> The ITS uses device IDs to map LPIs to a device. Dom0 will later use
> those IDs, which we directly pass on to the host.
> For this we have to map each device that Dom0 may request to a host
> ITS device with the same identifier.
From: Vijaya Kumar K
Call ACPI NUMA initialization under CONFIG_ACPI_NUMA.
Signed-off-by: Vijaya Kumar
---
xen/arch/arm/numa/acpi_numa.c | 28 +++-
xen/arch/arm/numa/numa.c | 6 ++
xen/common/numa.c | 14 ++
xen/include/asm-arm/numa.h
From: Vijaya Kumar K
Parse distance-matrix and fetch node distance information.
Store distance information in node_distance[].
Register dt_node_distance() function pointer with
the ARM numa code. This approach can be later used for
ACPI.
Signed-off-by: Vijaya Kumar K
---
xen/arch/arm/bootfdt.
From: Vijaya Kumar K
Register SRAT entry handler for type
ACPI_SRAT_TYPE_GICC_AFFINITY to parse SRAT table
and extract proximity for all CPU IDs.
Signed-off-by: Vijaya Kumar
---
xen/arch/arm/acpi/boot.c | 2 +
xen/arch/arm/numa/acpi_numa.c | 126 +
From: Vijaya Kumar K
Parse MADT table and extract MPIDR for all
CPU IDs in MADT ACPI_MADT_TYPE_GENERIC_INTERRUPT entries
and store in cpuid_to_hwid_map[]
This mapping is used by SRAT table parsing to
extract MPIDR of the CPU ID.
Signed-off-by: Vijaya Kumar
---
xen/arch/arm/numa/Makefile|
From: Vijaya Kumar K
SRAT's X2APIC_CPU_AFFINITY and CPU_AFFINITY types are not used
by ARM. Hence move handling of this SRAT types to arch specific
file and handle them under arch_table_parse_srat().
Signed-off-by: Vijaya Kumar K
---
xen/arch/arm/numa/acpi_numa.c | 5 +
xen/arch/x86/srat.
From: Vijaya Kumar K
On NUMA initialization failure, reset all the
NUMA structures to emulate as single node.
Signed-off-by: Vijaya Kumar
---
xen/arch/arm/numa/numa.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/xen/arch/arm/numa/numa.c b/xen/arch/arm/numa/numa.c
index 7583a40..89
From: Vijaya Kumar K
Add CONFIG_ACPI_NUMA to xen/drivers/acpi/Kconfig and
drop CONFIG_ACPI_NUMA set in asm-x86/config.h.
Signed-off-by: Vijaya Kumar K
---
xen/drivers/acpi/Kconfig | 4
xen/include/asm-x86/config.h | 1 -
2 files changed, 4 insertions(+), 1 deletion(-)
diff --git a/xe
From: Vijaya Kumar K
CONFIG_NUMA is defined in xen/drivers/acpi/Kconfig.
Move to common/Kconfig and enabled by default.
Also, NUMA feature uses PDX for physical address to
memory node mapping. Hence make HAS_PDX dependent
for NUMA.
Signed-off-by: Vijaya Kumar K
---
xen/common/Kconfig | 4
From: Vijaya Kumar K
Move SRAT handling code which is common across
architecture is moved to new file xen/drivers/acpi/srat.c
from xen/arch/x86/srat.c file. New header file srat.h is
introduced.
Signed-off-by: Vijaya Kumar K
---
xen/arch/x86/dom0_build.c | 1 +
xen/arch/x86/mm.c
From: Vijaya Kumar K
Delete numa-node-id and distance map from DOM0 DT
so that NUMA information is not exposed to DOM0.
This helps particularly to boot Node 1 devices
as if booting on Node0.
However this approach has limitation where memory allocation
for the devices should be local.
Also, do n
From: Vijaya Kumar K
Parse CPU node and fetch numa-node-id information.
For each node-id found, update nodemask_t mask.
Refer to /Documentation/devicetree/bindings/numa.txt.
Signed-off-by: Vijaya Kumar K
---
xen/arch/arm/Makefile | 1 +
xen/arch/arm/bootfdt.c | 16 --
xen/a
From: Vijaya Kumar K
For each cpu, update cpu_to_node[] with node id from
the numa-node-id DT property. Also, initialize cpu_to_node[]
with node 0.
Add macros to access cpu_to_node[] information.
Signed-off-by: Vijaya Kumar K
---
xen/arch/arm/numa/numa.c | 21 +
xen/arch
From: Vijaya Kumar K
Implement arch_sanitize_nodes_memory() which looks at all banks
in bootinfo.mem, update nodes[] with corresponding nodeid.
Call numa_scan_nodes() generic function with ram start and
end address, which takes care of further computing memnodeshift
and populating memnodemap[] us
From: Vijaya Kumar K
Parse memory node and fetch numa-node-id information.
For each memory range, store in node_memblk_range[]
along with node id.
When booting in UEFI mode, UEFI passes memory information
to Dom0 using EFI memory descriptor table and deletes the
memory nodes from the host DT. Ho
From: Vijaya Kumar K
Move code from xen/arch/x86/numa.c to xen/common/numa.c
so that it can be used by other archs.
Few generic static functions in x86/numa.c is made
non-static common/numa.c
The generic contents of header file asm-x86/numa.h
are moved to xen/numa.h.
Signed-off-by: Vijaya Kumar
From: Vijaya Kumar K
Move code from xen/arch/x86/srat.c to xen/common/numa.c
so that it can be used by other archs.
Few generic static functions in x86/srat.c are made
non-static common/numa.c
Signed-off-by: Vijaya Kumar K
---
xen/arch/x86/srat.c| 152 ++
From: Vijaya Kumar K
Introduce acpi_node_distance() and call from __node_distance().
This helps to implement arch specific __node_distance().
Also introduce LOCAL_DISTANCE & REMOTE DISTANCE macros.
Signed-off-by: Vijaya Kumar K
---
xen/arch/x86/srat.c| 13 +
xen/include/xen/num
From: Vijaya Kumar K
Right now CONFIG_NUMA is not enabled for ARM and
existing code in asm-arm/numa.h is for !CONFIG_NUMA.
Hence put this code under #ifndef CONFIG_NUMA.
This help to make this changes work when CONFIG_NUMA
is not enabled.
Also define NODES_SHIFT macro for ARM to value 2.
This l
From: Vijaya Kumar K
Rename some function in ACPI code as follow
- Rename setup_node to acpi_setup_node
- Rename bad_srat to numa_failed
- Rename nodes_cover_memory to arch_sanitize_nodes_memory
- Rename acpi_scan_nodes to numa_scan_nodes
Also introduce reset_pxm2node() to reset pxm2node var
From: Vijaya Kumar K
Add accessor for nodes[] and other static variables and
used those accessors.
Signed-off-by: Vijaya Kumar K
---
xen/arch/x86/srat.c | 108 +++-
1 file changed, 82 insertions(+), 26 deletions(-)
diff --git a/xen/arch/x86/srat
From: Vijaya Kumar K
Change u{8,32,64} to uint{8,32,64}_t and bool_t to bool.
Fix attributes coding styles.
Also change memnodeshift to unsigned int.
Signed-off-by: Vijaya Kumar K
---
xen/arch/x86/numa.c| 40 +--
xen/arch/x86/srat.c| 52 +
From: Vijaya Kumar K
With this RFC patch series, NUMA support is added for ARM platform.
Both DT and ACPI based NUMA support is added.
Only Xen is made aware of NUMA platform. NUMA awareness to DOM0 is not
added.
As part of this series, the code under x86 architecture is
reused by moving into co
From: Vijaya Kumar K
Following changes are made
- Rename compute_hash_shift to compute_memnode_shift
and return error values instead of shift.
- Changes prototypes of populate_memnodemap()
and extract_lsb_from_nodes()
Signed-off-by: Vijaya Kumar K
---
xen/arch/x86/numa.c| 47 ++
From: Vijaya Kumar K
Split numa_initmem_init() so that the numa fallback code is moved
as separate function which is generic.
Signed-off-by: Vijaya Kumar K
---
xen/arch/x86/numa.c | 29 +
1 file changed, 17 insertions(+), 12 deletions(-)
diff --git a/xen/arch/x86/n
From: Vijaya Kumar K
Add accessor functions for acpi_numa and numa_off static
variables. Init value of acpi_numa is set 1 instead of 0.
Also return value of srat_disabled is changed to bool.
Signed-off-by: Vijaya Kumar K
---
xen/arch/x86/numa.c| 43 +++--
From: Vijaya Kumar K
Fix coding style, trailing spaces, tabs in NUMA code.
Also drop unused macros and functions.
Signed-off-by: Vijaya Kumar K
---
xen/arch/x86/numa.c| 47 +-
xen/arch/x86/srat.c| 2 +-
xen/include/asm-x86/numa.h | 4
From: Vijaya Kumar K
On ARM64, virt_to_mfn uses the hardware for address
translation. So if the virtual address is not mapped translation
fault is raised. On ARM64, DIRECTMAP_VIRT region is direct mapped.
On ARM platforms with NUMA, While initializing second memory node,
panic is triggered from
I don't have any specific requirement. I am also
looking with ACPI support.
Please share your RFC patches so that I can test on our platform.
> Thanks,
> Sameer
>
> On 3/20/2017 11:58 PM, Vijay Kilari wrote:
>> Hi,
>>
>> Is there any effort put by anyone to get
Hi Jan,
Thanks for the review.
On Mon, Mar 27, 2017 at 12:50 PM, Jan Beulich wrote:
On 27.03.17 at 09:10, wrote:
>> @@ -254,7 +262,6 @@ static inline int gvirt_to_maddr(vaddr_t va, paddr_t
>> *pa, unsigned int flags)
>> #define virt_to_mfn(va) (virt_to_maddr(va) >> PAGE_SHIFT)
>> #d
From: Vijaya Kumar K
On ARM64, virt_to_mfn uses the hardware for address
translation. So if the virtual address is not mapped translation
fault is raised. On ARM64, DIRECTMAP_VIRT region is direct mapped.
On ARM platforms with NUMA, While initializing second memory node,
panic is triggered from
Hi,
Is there any effort put by anyone to get SMMUv3 support in Xen for ARM64?.
Would be glad to know.
Regards
Vijay
___
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel
On Thu, Mar 16, 2017 at 9:55 PM, Shanker Donthineni
wrote:
> Hi Andre,
>
>
> On 03/16/2017 06:20 AM, Andre Przywara wrote:
>> Create a new file to hold the emulation code for the ITS widget.
>> For now we emulate the memory mapped ITS registers and provide a stub
>> to introduce the ITS command ha
On Tue, Mar 14, 2017 at 9:02 PM, Julien Grall wrote:
> Hello Vijay,
>
> On 13/03/17 11:43, vijay.kil...@gmail.com wrote:
>>
>> From: Vijaya Kumar K
>>
>> On ARM, virt_to_mfn uses the hardware for address
>> translation. So if the virtual address is not mapped translation
>> fault is raised.On ARM
On Mon, Mar 13, 2017 at 7:12 PM, Jan Beulich wrote:
On 13.03.17 at 12:43, wrote:
>> --- a/xen/include/asm-arm/mm.h
>> +++ b/xen/include/asm-arm/mm.h
>> @@ -260,6 +260,13 @@ static inline int gvirt_to_maddr(vaddr_t va, paddr_t
>> *pa, unsigned int flags)
>> #define virt_to_mfn(va) (virt_t
From: Vijaya Kumar K
On ARM, virt_to_mfn uses the hardware for address
translation. So if the virtual address is not mapped translation
fault is raised.On ARM, DIRECTMAP_VIRT region is direct mapped.
On ARM with NUMA, While initializing second memory node,
panic is triggered from init_node_heap(
On Fri, Mar 3, 2017 at 8:52 PM, Jan Beulich wrote:
On 03.03.17 at 16:16, wrote:
>> On Fri, Mar 3, 2017 at 8:22 PM, Julien Grall wrote:
>>> int __init acpi_numa_init(void)
>>> {
>>> if (!acpi_parse_table()) {
>>> acpi_table_parse_srat(TYPE_CPU_AFFINITY);
>>
>> Thi
On Fri, Mar 10, 2017 at 3:09 PM, Julien Grall wrote:
> Hello,
>
> It is not the first time I am saying this. Please CC *all* the maintainers
> of the code you modify. Give a look at scripts/get_maintainers.pl.
I got below maintainers when I ran the script on complete patch as below.
ubuntu@ubunt
1 - 100 of 530 matches
Mail list logo
