This Commit 7747abf11487 misses the curly brackets. Follow Igor's and Eduardo's
suggestion,
Add a follow-up patch for it.
Signed-off-by: Dou Liyang
---
hw/core/machine.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/hw/core/machine.c b/hw/core/machi
Hi Eduardo,
At 07/10/2018 06:27 PM, Eduardo Habkost wrote:
On Tue, Jul 10, 2018 at 04:15:30PM +0800, Dou Liyang wrote:
[...]
+if (nb_numa_nodes)
machine_numa_finish_cpu_init(machine);
-}
missing curly brackets, should look like:
if (nb_numa_nodes
0] ACPI: SRAT: Node 2 PXM 2 [mem 0x1-0x13fff]
[0.00] ACPI: SRAT: Node 3 PXM 3 [mem 0x14000-0x33fff] hotplug
Signed-off-by: Dou Liyang
Reviewed-by: Igor Mammedov
---
Note to maintainer: update ACPI tables test blobs on commit.
---
hw/i386/acpi-build.c | 9 ++-
Now, QEmu adds a new check for memory-less NUMA nodes in build_srat().
It effects the ACPI test.
So, Update ACPI tables test blobs.
Signed-off-by: Dou Liyang
---
tests/acpi-test-data/pc/SRAT.numamem | Bin 224 -> 224 bytes
tests/acpi-test-data/q35/SRAT.numamem | Bin 224 -> 224 bytes
2
Hi Igor,
At 07/10/2018 03:52 PM, Igor Mammedov wrote:
On Thu, 5 Jul 2018 10:10:38 +0800
Dou Liyang wrote:
Currently, Qemu ACPI builder doesn't consider the non-memory NUMA nodes, eg:
s/non-memory/memory-less/ throughout subj/commit message
Yes, I will do it.
-m 4G,slots=4,m
Hi Igor,
At 07/10/2018 03:42 PM, Igor Mammedov wrote:
On Wed, 4 Jul 2018 21:22:39 +0800
Dou Liyang wrote:
Commit 7a3099fc9c5c("numa: postpone options post-processing till
machine_run_board_init()")
broke the commit 7b8be49d36fc("NUMA: Enable adding NUMA node imp
0] ACPI: SRAT: Node 2 PXM 2 [mem 0x1-0x13fff]
[0.00] ACPI: SRAT: Node 3 PXM 3 [mem 0x14000-0x33fff] hotplug
Signed-off-by: Dou Liyang
---
Have done a bootup test in Linux and window 10, 7
---
hw/i386/acpi-build.c | 9 ++---
tests/acpi-test-da
on need add a new node if memory hotplug
is enabled (slots > 0) even nb_numa_nodes=0.
So, Remove the check for numa_complete_configuration() to fix this.
Fixes 7a3099fc9c5c("numa: postpone options post-processing till
machine_run_board_init()")
Signed-off-by: Dou Liyang
---
v1 --&
At 07/04/2018 07:52 PM, Dou Liyang wrote:
Commit 7a3099fc9c5c("numa: postpone options post-processing till
machine_run_board_init()")
broke the commit 7b8be49d36fc("NUMA: Enable adding NUMA node implicitly").
The machine_run_board_init() doesn't do NUMA setup if nb_
l also check the value of nb_numa_nodes, remove the
check from machine_run_board_init() to fix ths bug.
Fixes 7a3099fc9c5c("numa: postpone options post-processing till
machine_run_board_init()")
Signed-off-by: Dou Liyang
---
hw/core/machine.c | 6 ++
1 file changed, 2 insertions(+), 4
At 01/12/2018 03:23 PM, Dou Liyang wrote:
[...]
BTW, I used "make TEST_ACPI_REBUILD_AML=1 check" to create the DSDT
file. Is it correct?
This sounds correct to me. Igor, can you confirm?
It's will rebuild/update reference tables.
I'm for merging it early and trackin
[...]
BTW, I used "make TEST_ACPI_REBUILD_AML=1 check" to create the DSDT
file. Is it correct?
This sounds correct to me. Igor, can you confirm?
It's will rebuild/update reference tables.
I'm for merging it early and tracking/fixing issue Peter saw during this dev
cycle.
Send an ack th
Hi Eduardo,
At 12/19/2017 06:09 AM, Eduardo Habkost wrote:
On Thu, Dec 14, 2017 at 12:08:53PM +0800, Dou Liyang wrote:
These are the patches left over from the pull request:
[Qemu-devel] [PULL 0/9] x86 and machine queue, 2017-10-05
because of some errors when tested by "make
enough.
Add a testcase for this situation to make sure the ACPI table is
correct for guest.
Suggested-by: Eduardo Habkost
Signed-off-by: Dou Liyang
---
tests/acpi-test-data/pc/DSDT.numamem | Bin 0 -> 5150 bytes
tests/acpi-test-data/pc/SRAT.numamem | Bin 0 -> 224 bytes
tests/acpi-tes
and retest this patch in v2.11.0
--Update the DSDT.numamem:
[1]: http://lists.gnu.org/archive/html/qemu-devel/2017-10/msg01334.html
Dou Liyang (2):
ACPI/unit-test: Add a testcase for RAM allocation in numa node
hw/acpi-build: Make next_base easy to follow
hw/i386/acpi-build.c
It may be hard to read the assignment statement of "next_base", so
S/next_base += (1ULL << 32) - pcms->below_4g_mem_size;
/next_base = mem_base + mem_len;
... for readability.
No functionality change.
Signed-off-by: Dou Liyang
---
hw/i386/acpi-build.c | 2 +-
1 file c
Hi, Igor,
At 11/28/2017 11:07 PM, Igor Mammedov wrote:
On Thu, 5 Oct 2017 17:36:30 -0300
Eduardo Habkost wrote:
From: Dou Liyang
It may be hard to read the assignment statement of "next_base", so
S/next_base += (1ULL << 32) - pcms->below_4g_mem_size;
/next_base =
Hi Igor,
At 11/28/2017 11:08 PM, Igor Mammedov wrote:
On Thu, 5 Oct 2017 17:36:31 -0300
Eduardo Habkost wrote:
From: Dou Liyang
As QEMU supports the memory-less node, it is possible that there is
no RAM in the first numa node(also be called as node0). eg:
... \
-m 128,slots=3,maxmem
Hi Igor,
[...]
+parse_numa_node(ms, &node, NULL);
I get build break here:
numa.c:451:13: error: too few arguments to function ‘parse_numa_node’
parse_numa_node(ms, &node, NULL);
In upstream tree, your commit
cc001888b780 ("numa: fixup parsed NumaNodeOptions earli
e memory hotplug
* Linux: switch to SWIOTLB DMA ops, to bounce DMA transfers to 32bit allocated
buffers that legacy drivers/hw can handle.
[Rewritten by Igor]
Reported-by: Thadeu Lima de Souza Cascardo
Suggested-by: Igor Mammedov
Signed-off-by: Dou Liyang
Cc: Paolo Bonzini
Cc: Richa
Hi Michael,
[...]
Seems to cause build failures:
/scm/qemu/numa.c:452:13: error: too many arguments to function ‘parse_numa_node’
parse_numa_node(ms, &node, NULL, NULL);
Yes, commit
cc001888b780 ("numa: fixup parsed NumaNodeOptions earlier")
removed a argument from the functi
e memory hotplug
* Linux: switch to SWIOTLB DMA ops, to bounce DMA transfers to 32bit allocated
buffers that legacy drivers/hw can handle.
[Rewritten by Igor]
Reported-by: Thadeu Lima de Souza Cascardo
Suggested-by: Igor Mammedov
Signed-off-by: Dou Liyang
Cc: Paolo Bonzini
Cc: Richa
Deer Eduardo,
At 10/26/2017 04:02 PM, Eduardo Habkost wrote:
Hi,
Sorry for taking so long to review it:
Not matter. It's my honor!
On Mon, Oct 23, 2017 at 09:33:42AM +0800, Dou Liyang wrote:
[...]
+ */
+if (ram_slots > 0 && QTAILQ_EMPTY(&numa_opts->h
Hi, Fam
At 10/23/2017 09:37 AM, no-re...@patchew.org wrote:
Hi,
This series failed build test on s390x host. Please find the details below.
Type: series
Message-id: 1508722422-3861-1-git-send-email-douly.f...@cn.fujitsu.com
Subject: [Qemu-devel] [PATCH v4] NUMA: Enable adding NUMA node implici
Hi Fam,
At 10/23/2017 09:37 AM, no-re...@patchew.org wrote:
Hi,
This series failed automatic build test. Please find the testing commands and
their output below. If you have docker installed, you can probably reproduce it
locally.
Type: series
Message-id: 1508722422-3861-1-git-send-email-douly
e memory hotplug
* Linux: switch to SWIOTLB DMA ops, to bounce DMA transfers to 32bit allocated
buffers that legacy drivers/hw can handle.
[Rewritten by Igor]
Reported-by: Thadeu Lima de Souza Cascardo
Suggested-by: Igor Mammedov
Signed-off-by: Dou Liyang
Cc: Paolo Bonzini
Cc: Richa
Hi Eduardo,
Thanks for your reply.
At 10/21/2017 03:15 AM, Eduardo Habkost wrote:
On Wed, Oct 18, 2017 at 11:50:32AM +0800, Dou Liyang wrote:
At 10/18/2017 11:48 AM, Dou Liyang wrote:
As QEMU supports the memory-less node, it is possible that there is
no RAM in the first numa node(also be
Hi Igor,
diff --git a/vl.c b/vl.c
index 9bb5058..d083b4d 100644
--- a/vl.c
+++ b/vl.c
@@ -4665,7 +4665,11 @@ int main(int argc, char **argv, char **envp)
default_drive(default_floppy, snapshot, IF_FLOPPY, 0, FD_OPTS);
default_drive(default_sdcard, snapshot, IF_SD, 0, SD_OPTS);
-pa
Hi all,
It seems that a month has passed
So, Ping...
Thanks,
dou.
At 09/21/2017 05:23 PM, Dou Liyang wrote:
Linux and Windows need ACPI SRAT table to make memory hotplug work properly,
however currently QEMU doesn't create SRAT table if numa options aren't present
on C
At 10/18/2017 11:48 AM, Dou Liyang wrote:
As QEMU supports the memory-less node, it is possible that there is
no RAM in the first numa node(also be called as node0). eg:
... \
-m 128,slots=3,maxmem=1G \
-numa node -numa node,mem=128M \
But, this makes it hard for QEMU to build a known-to
enough.
Add a testcase for this situation to make sure the ACPI table is
correct for guest.
Suggested-by: Eduardo Habkost
Signed-off-by: Dou Liyang
---
Changelog:
This patch can pass the 'make check' in Peter's machine,
But, in my own and Eduardo's machine, we can pass
Hi Peter,
At 10/10/2017 08:56 PM, Peter Maydell wrote:
On 10 October 2017 at 13:52, Eduardo Habkost wrote:
This shouldn't happen if CONFIG_IASL is not defined. Did you run
configure again after removing iasl?
I've neither added nor removed iasl from the build systems
in question. Pull reque
Hi Peter,
At 10/10/2017 05:40 PM, Peter Maydell wrote:
On 10 October 2017 at 08:41, Dou Liyang wrote:
If we want do the "bios-tables-test", we need iasl support.
otherwise, when I run "make V=1 check" without iasl, the test
will report error:
This works in master at
Hi Eduardo,
At 10/10/2017 08:52 PM, Eduardo Habkost wrote:
This shouldn't happen if CONFIG_IASL is not defined. Did you run
configure again after removing iasl?
Oops, Yes, the result above is in CONFIG_IASL = iasl. :-(
I have been aware of this problem.
I am waiting the result of the test w
Hi Eduardo,
I couldn't reproduce the failure too. But, I have some concern below.
At 10/06/2017 11:37 PM, Peter Maydell wrote:
On 6 October 2017 at 16:27, Eduardo Habkost wrote:
On Fri, Oct 06, 2017 at 01:17:42PM +0100, Peter Maydell wrote:
This fails tests, I'm afraid:
[...]
acpi-test: Wa
Hi Eduardo,
At 10/10/2017 10:29 AM, Eduardo Habkost wrote:
On Fri, Sep 01, 2017 at 10:10:03AM +0800, Dou Liyang wrote:
It may be hard to read the assignment statement of "next_base", so
S/next_base += (1ULL << 32) - pcms->below_4g_mem_size;
/next_base = mem_base
e memory hotplug
* Linux: switch to SWIOTLB DMA ops, to bounce DMA transfers to 32bit allocated
buffers that legacy drivers/hw can handle.
[Rewritten by Igor]
Reported-by: Thadeu Lima de Souza Cascardo
Suggested-by: Igor Mammedov
Signed-off-by: Dou Liyang
Cc: Paolo Bonzini
Cc: Richa
Hi Igor,
At 09/21/2017 03:54 PM, Igor Mammedov wrote:
On Thu, 21 Sep 2017 12:19:17 +0800
[...]
I've meant something like that:
parse_numa_opts() {
if (mc->auto_enable_numa_with_memhp == true) {
qemu_opts_parse_noisily(qemu_find_opts("numa"), "node", true);
}
}
Oops, Unde
Hi Igor,
I am sorry I missed some comments you gave to me.
my reply is below.
At 09/18/2017 05:24 PM, Dou Liyang wrote:
[...]
ranges where
*the guest will attempt to probe for a device that QEMU doesn't
*implement and a stub device is required.
+ * @numa_implicit_add_
to enable memory hotplug
* Linux: switch to SWIOTLB DMA ops, to bounce DMA transfers to 32bit allocated
buffers that legacy drivers/hw can handle.
[Rewritten by Igor]
Reported-by: Thadeu Lima de Souza Cascardo
Suggested-by: Igor Mammedov
Signed-off-by: Dou Liyang
Cc: Paolo Bonzini
Cc: Richa
Hi Igor,
At 09/18/2017 05:08 PM, Igor Mammedov wrote:
On Fri, 15 Sep 2017 16:33:18 +0800
Dou Liyang wrote:
In QEMU, if we enable NUMA and have nodes, QEMU will build ACPI SRAT table
for transfering NUMA configuration to the guest. So, the maximum memory in
SRAT can be used to determine
At 09/18/2017 03:40 PM, Igor Mammedov wrote:
On Mon, 18 Sep 2017 11:54:50 +0800
Dou Liyang wrote:
At 09/15/2017 06:05 PM, Dou Liyang wrote:
Hi Daniel,
At 09/15/2017 04:40 PM, Daniel P. Berrange wrote:
On Fri, Sep 15, 2017 at 04:33:18PM +0800, Dou Liyang wrote:
In QEMU, if we enable NUMA
At 09/15/2017 06:05 PM, Dou Liyang wrote:
Hi Daniel,
At 09/15/2017 04:40 PM, Daniel P. Berrange wrote:
On Fri, Sep 15, 2017 at 04:33:18PM +0800, Dou Liyang wrote:
In QEMU, if we enable NUMA and have nodes, QEMU will build ACPI SRAT
table
for transfering NUMA configuration to the guest. So
Hi Daniel,
At 09/15/2017 04:40 PM, Daniel P. Berrange wrote:
On Fri, Sep 15, 2017 at 04:33:18PM +0800, Dou Liyang wrote:
In QEMU, if we enable NUMA and have nodes, QEMU will build ACPI SRAT table
for transfering NUMA configuration to the guest. So, the maximum memory in
SRAT can be used to
Suggested-by: Igor Mammedov
Signed-off-by: Dou Liyang
Cc: Paolo Bonzini
Cc: Richard Henderson
Cc: Eduardo Habkost
Cc: "Michael S. Tsirkin"
Cc: Marcel Apfelbaum
Cc: Igor Mammedov
Cc: David Hildenbrand
Cc: Thomas Huth
Cc: Alistair Francis
Cc: f4...@amsat.org
Cc: Takao Indoh
Cc:
Hi Igor,
At 09/11/2017 06:58 PM, Igor Mammedov wrote:
> >
> > Igor's suggestion to enable NUMA implicitly sounds safer to me.
> >
>
> I agree with Igor too.
>
> Is anybody doing this? If not, may I make a patch to enable adding NUMA
> node implicitly first. let's see what it looks like.
As far
Hi Chen,
At 09/04/2017 02:14 PM, Zhang Chen wrote:
When network is busy, some tcp options(like sack) will unpredictable
occur in primary side or secondary side. it will make packet size
not same, but the two packet's payload is identical. colo just
care about packet payload, so we skip the optio
Hi Eduardo,
At 09/04/2017 09:08 PM, Eduardo Habkost wrote:
[...]
In my opinion, this may also add the hotpluggable memory, and see the
following commemts.
/*
* Entry is required for Windows to enable memory hotplug in OS
* and for Linux to enable SWIOTLB when booted with less than
Hi Igor,
At 09/04/2017 07:11 PM, Igor Mammedov wrote:
[...]
+if (mem_base <= HOLE_640K_START &&
+next_base > HOLE_640K_START) {
+mem_len -= next_base - HOLE_640K_START;
+if (mem_len > 0) {
+numamem = acpi_data_push(table_data, sizeof *n
At 09/04/2017 05:39 PM, Igor Mammedov wrote:
On Thu, 31 Aug 2017 20:04:26 +0800
Dou Liyang wrote:
From: Eduardo Habkost
Currently, Using the fisrt node without memory on the machine makes
QEMU unhappy. With this example command line:
... \
-m 1024M,slots=4,maxmem=32G \
-numa node
Hi Chen,
At 09/02/2017 12:02 AM, Zhang Chen wrote:
On 09/01/2017 05:35 PM, Dou Liyang wrote:
Hi chen,
At 08/21/2017 04:55 PM, Zhang Chen wrote:
When network is busy, some tcp options(like sack) will unpredictable
occur in primary side or secondary side. it will make packet size
not same
e_size,
pcms->numa_nodes - 1,
MEM_AFFINITY_HOTPLUGGABLE |
MEM_AFFINITY_ENABLED);
}
Thanks,
dou.
Once we apply the "Fix SRAT memory building in case of node 0
without RAM" patch from Dou Liyang, no memory affinity entries
will be generated if pc
Hi chen,
At 08/21/2017 04:55 PM, Zhang Chen wrote:
When network is busy, some tcp options(like sack) will unpredictable
occur in primary side or secondary side. it will make packet size
not same, but the two packet's payload is identical. colo just
care about packet payload, so we skip the optio
enough.
Add a testcase for this situation to make sure the ACPI table is
correct for guest.
Suggested-by: Eduardo Habkost
Signed-off-by: Dou Liyang
---
tests/acpi-test-data/pc/DSDT.numamem | Bin 0 -> 5104 bytes
tests/acpi-test-data/pc/SRAT.numamem | Bin 0 -> 224 bytes
tests/acpi-tes
eep code consistency and
reduce the loop times.
Signed-off-by: Dou Liyang
Reviewed-by: Igor Mammedov
Reviewed-by: Philippe Mathieu-Daudé
---
numa.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/numa.c b/numa.c
index e32af04..5f2916d 100644
--- a/numa.c
+++ b/numa.c
@@ -56
estcase.
This patchset fixs an ACPI building bug which caused by no RAM
in the first NUAM node. and also add a new testcase for the bug.
Dou Liyang (3):
hw/acpi-build: Make assignment statement of next_base easy to read
ACPI/unit-test: Add a new testcase for RAM allocation in numa node
em by cut out the 640K hole in the same way the PCI
4G hole does.
Signed-off-by: Eduardo Habkost
Signed-off-by: Dou Liyang
---
hw/i386/acpi-build.c | 28 ++--
1 file changed, 22 insertions(+), 6 deletions(-)
diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
It may be hard to read the assignment statement of "next_base", so
S/next_base += (1ULL << 32) - pcms->below_4g_mem_size;
/next_base = mem_base + mem_len;
... for readability.
No functionality change.
Signed-off-by: Dou Liyang
---
hw/i386/acpi-build.c | 2 +-
1 file c
Hi, Eduardo
At 09/01/2017 05:36 AM, Eduardo Habkost wrote:
On Thu, Aug 31, 2017 at 08:04:26PM +0800, Dou Liyang wrote:
From: Eduardo Habkost
Currently, Using the fisrt node without memory on the machine makes
QEMU unhappy. With this example command line:
... \
-m 1024M,slots=4,maxmem=32G
Hi Eduardo,
At 08/31/2017 06:38 PM, Dou Liyang wrote:
Hi Eduardo,
[...]
+continue;
}
Now the nodes will be out of order, if node 0 has no RAM. Why?
Because the code parsed the other node with RAM first, then parsed the
node 0.
Why not implement this in the same
m by cut out the 640K hole in the same way the PCI
4G hole does. Also do some cleanup.
Signed-off-by: Eduardo Habkost
Signed-off-by: Dou Liyang
---
hw/i386/acpi-build.c | 30 +++---
1 file changed, 23 insertions(+), 7 deletions(-)
diff --git a/hw/i386/acpi-build.c b
enough.
Add a testcase for this situation to make sure the ACPI table is
correct for guest.
Suggested-by: Eduardo Habkost
Signed-off-by: Dou Liyang
---
tests/acpi-test-data/pc/DSDT.numamem | Bin 0 -> 5104 bytes
tests/acpi-test-data/pc/SRAT.numamem | Bin 0 -> 224 bytes
tests/acpi-tes
RAM
in the first NUAM node. and also add a new testcase for the bug.
Dou Liyang (2):
ACPI/unit-test: Add a new testcase for RAM allocation in numa node
NUMA: Replace MAX_NODES with nb_numa_nodes in for loop
Eduardo Habkost (1):
hw/acpi-build: Fix SRAT memory building in case of node 0 witho
eep code consistency and
reduce the loop times.
Signed-off-by: Dou Liyang
Reviewed-by: Igor Mammedov
Reviewed-by: Philippe Mathieu-Daudé
---
numa.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/numa.c b/numa.c
index e32af04..5f2916d 100644
--- a/numa.c
+++ b/numa.c
@@ -56
Hi Eduardo,
[...]
+continue;
}
Now the nodes will be out of order, if node 0 has no RAM. Why?
Because the code parsed the other node with RAM first, then parsed the
node 0.
Why not implement this in the same way the PCI 4GB hole is
already implemented. e.g.:
In
enough.
Add a testcase for this situation to make sure the ACPI table is
correct for guest.
Suggested-by: Eduardo Habkost
Signed-off-by: Dou Liyang
---
v4 --> v5:
- rewrite the testcase.
- Drop the SLIT date
tests/acpi-test-data/pc/DSDT.numamem | Bin 0 -> 5104 bytes
tests/acpi-tes
Hi Igor,
At 08/24/2017 08:33 PM, Igor Mammedov wrote:
> test_acpi_one(" -numa node -numa node,mem=128", &data);
>
> but, the DSDT didn't match the default one. because, if we support
> NUMA, the DSDT will give us "_PXM" to map the CPU to node.
Ok, looks like you'll have to include your variant
Hi Igor,
At 08/24/2017 01:47 AM, Igor Mammedov wrote:
On Wed, 23 Aug 2017 21:35:29 +0800
Dou Liyang wrote:
Hi Igor,
At 08/23/2017 08:45 PM, Igor Mammedov wrote:
On Wed, 23 Aug 2017 20:12:51 +0800
Dou Liyang wrote:
Hi Igor,
At 08/23/2017 04:40 PM, Igor Mammedov wrote:
On Tue, 22 Aug
Hi Eduardo,
At 08/24/2017 01:25 AM, Eduardo Habkost wrote:
On Wed, Aug 23, 2017 at 09:35:29PM +0800, Dou Liyang wrote:
Hi Igor,
At 08/23/2017 08:45 PM, Igor Mammedov wrote:
On Wed, 23 Aug 2017 20:12:51 +0800
Dou Liyang wrote:
Hi Igor,
At 08/23/2017 04:40 PM, Igor Mammedov wrote:
On Tue
Hi Igor,
At 08/23/2017 08:45 PM, Igor Mammedov wrote:
On Wed, 23 Aug 2017 20:12:51 +0800
Dou Liyang wrote:
Hi Igor,
At 08/23/2017 04:40 PM, Igor Mammedov wrote:
On Tue, 22 Aug 2017 11:24:10 +0800
Dou Liyang wrote:
As QEMU supports the memory-less node, it is possible that there is
no
Hi Igor,
At 08/23/2017 04:40 PM, Igor Mammedov wrote:
On Tue, 22 Aug 2017 11:24:10 +0800
Dou Liyang wrote:
As QEMU supports the memory-less node, it is possible that there is
no RAM in the first numa node(also be called as node0). eg:
... \
-m 128,slots=3,maxmem=1G \
-numa node -numa
Hi Michael,
At 08/23/2017 10:48 AM, Michael S. Tsirkin wrote:
On Tue, Aug 22, 2017 at 11:24:08AM +0800, Dou Liyang wrote:
V3 --> v4:
-add a new testcase.
This patchset fixs an ACPI building bug which caused by no RAM
in the first NUAM node. and also add a new testcase for the bug.
tha
eep code consistency and
reduce the loop times.
Signed-off-by: Dou Liyang
---
numa.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/numa.c b/numa.c
index e32af04..5f2916d 100644
--- a/numa.c
+++ b/numa.c
@@ -567,7 +567,7 @@ void memory_region_allocate_system_memory(MemoryReg
ith the first node which has
memory on it. Add a new function for each node. Also do some cleanup.
Signed-off-by: Dou Liyang
---
hw/i386/acpi-build.c | 78 +---
1 file changed, 50 insertions(+), 28 deletions(-)
diff --git a/hw/i386/acpi-build.c b
V3 --> v4:
-add a new testcase.
This patchset fixs an ACPI building bug which caused by no RAM
in the first NUAM node. and also add a new testcase for the bug.
Dou Liyang (2):
hw/acpi-build: Fix SRAT memory building when there is no memory in
node0
ACPI/unit-test: Add a new testc
enough.
Add a testcase for this situation to make sure the ACPI table is
correct for guest.
Suggested-by: Eduardo Habkost
Signed-off-by: Dou Liyang
---
tests/acpi-test-data/pc/DSDT.numamem | Bin 0 -> 6463 bytes
tests/acpi-test-data/pc/SLIT.numamem | Bin 0 -> 48 bytes
tests/acpi-test-d
Hi, Igor
[...]
Ok, I will do it right now.
see commit 6b9c1dd2c for example of adding test case,
currently test has base SRAT tables that is reused by all tests.
When you create testcase that produces different SRAT table,
'make check' will print warnings that ables do not match
you can add "
Hi Eduardo,
At 08/19/2017 12:48 AM, Eduardo Habkost wrote:
On Thu, Aug 17, 2017 at 01:58:40PM +0800, Dou Liyang wrote:
Hi Igor,
I tested this patch with following guests:
1. RHEL 6.5 with Linux 2.6.32
2. RHEL 7.0 with Linux 3.10.0
3. Fedora 23 with Linux 4.13.0-rc5
4. window 2003 service
5
Hi Igor,
I tested this patch with following guests:
1. RHEL 6.5 with Linux 2.6.32
2. RHEL 7.0 with Linux 3.10.0
3. Fedora 23 with Linux 4.13.0-rc5
4. window 2003 service
5. window 7
6. window 10
Thanks,
dou.
At 08/16/2017 09:45 AM, Dou Liyang wrote:
Currently, Using the fisrt node
Hi Igor,
At 08/16/2017 05:18 PM, Igor Mammedov wrote:
On Wed, 16 Aug 2017 09:26:51 +0800
Dou Liyang wrote:
Currently, Using the fisrt node without memory on the machine makes
QEMU unhappy. With this example command line:
... \
-m 1024M,slots=4,maxmem=32G \
-numa node,nodeid=0 \
-numa
ith the first node which has
memory on it. Add a new function for each node. Also do some cleanup.
Signed-off-by: Dou Liyang
---
V3 --> V2
-Modify the title
V2 --> V1:
-Fix a coding style problem
Replace
for (node = 0;
node < pcms->numa_nodes && pcms->n
ith the first node which has
memory on it. Add a new function for each node. Also do some cleanup.
Signed-off-by: Dou Liyang
---
V2 --> V1:
-Fix a coding style problem
Replace
for (node = 0;
node < pcms->numa_nodes && pcms->node_mem[node] == 0;
node+
Hi Igor,
At 08/15/2017 04:04 PM, Igor Mammedov wrote:
On Tue, 15 Aug 2017 09:26:46 +0800
Dou Liyang wrote:
Hi Eduardo,
At 08/14/2017 08:44 PM, Eduardo Habkost wrote:
On Mon, Aug 14, 2017 at 06:11:11PM +0800, Dou Liyang wrote:
Currently, Using the fisrt node without memory on the machine
ith the first node which has
memory on it. Add a new function for each node. Also do some cleanup.
Signed-off-by: Dou Liyang
---
hw/i386/acpi-build.c | 76 +---
1 file changed, 48 insertions(+), 28 deletions(-)
diff --git a/hw/i386/acpi-build.c b
Hi Eduardo,
At 08/14/2017 08:44 PM, Eduardo Habkost wrote:
On Mon, Aug 14, 2017 at 06:11:11PM +0800, Dou Liyang wrote:
Currently, Using the fisrt node without memory on the machine makes
QEMU unhappy. With this example command line:
... \
-m 1024M,slots=4,maxmem=32G \
-numa node,nodeid=0
I'm sorry, forgot to cc Michael S. Tsirkin
At 08/14/2017 06:11 PM, Dou Liyang wrote:
Currently, Using the fisrt node without memory on the machine makes
QEMU unhappy. With this example command line:
... \
-m 1024M,slots=4,maxmem=32G \
-numa node,nodeid=0 \
-numa node,mem=1024M,nod
ation found" and the NUMA topology is
wrong.
This is because when QEMU builds ACPI SRAT, it regards node0 as the
default node to deal with the memory hole(640K-1M). this means the
node0 must have some memory(>1M) firstly.
Add a check in parse_numa_opts to avoid this situation.
Signed-off-b
Hi Igor,
At 06/05/2017 04:45 PM, Igor Mammedov wrote:
On Mon, 5 Jun 2017 15:53:02 +0800
Dou Liyang wrote:
Hi all,
I want to boot up a system with some removable memory.
So, i used '-object' and '-device' to cold-plug memory
in QEmu as following :
./x86_64-softmm
Hi all,
I want to boot up a system with some removable memory.
So, i used '-object' and '-device' to cold-plug memory
in QEmu as following :
./x86_64-softmmu/qemu-system-x86_64 \
...
-object memory-backend-ram,id=mem3,size=128M \
-device pc-dimm,id=dimm3,memdev=mem3 \
Hi, Max
At 02/01/2017 11:02 AM, Max Reitz wrote:
On 15.01.2017 09:01, Dou Liyang wrote:
In order to reduce the execution time, this patch optimize
the qmp_query_blockstats():
Remove the next_query_bds function.
Remove the bdrv_query_stats function.
Remove some judgement sentence.
The original
PUs may be ignored to synchronize.
The patch moves the cpu_synchronize_all_post_init func after generic
devices initialization to make sure that all the CPUs can be included.
Signed-off-by: Dou Liyang
---
Change log:
v2-> v3:
1. Rebase the patch.
2. Rewrite the log
v1-> v2:
1. Spl
Hi, Alex
Sorry for late reply.
At 01/27/2017 10:03 PM, Alex Bennée wrote:
Dou Liyang writes:
At the Qemu initialization, we call the cpu_synchronize_all_post_init()
to synchronize All CPU states to KVM in the ./vl.c::main().
Currently, it is called before we initialize the CPUs, which
Hi, Markus
Here is the non-RFC patch.
Any suggestions are welcome. :)
Thanks,
Liyang.
At 01/15/2017 04:01 PM, Dou Liyang wrote:
Change log v3 -> v4:
1. Develop these into the non-RFC patches.
2. Fix some comments.
3. do declarations first.
Change log v2 -> v3:
1. Remo
At 01/19/2017 01:13 AM, Igor Mammedov wrote:
possible_cpus could be initialized earlier then cpu objects,
s/then/than/
i.e. when -smp is parsed so move init code to possible_cpu_arch_ids()
[...]
Hi, Eduardo
At 01/19/2017 01:06 AM, Eduardo Habkost wrote:
On Wed, Jan 18, 2017 at 09:26:36PM +0800, Dou Liyang wrote:
Hi, All
**
ERROR:/tmp/qemu-test/src/tests/vhost-user-test.c:668:test_migrate: assertion failed:
(qdict_haskey(rsp, "return"))
GTester: last r
Hi, Igor,
At 01/18/2017 09:46 PM, Igor Mammedov wrote:
On Wed, 18 Jan 2017 20:40:04 +0800
Dou Liyang wrote:
As we fixed a bug(Bug 1) in below links, Named "Method-A":
https://lists.nongnu.org/archive/html/qemu-devel/2017-01/msg03354.html
Then, Eduardo gave us many suggests. T
moves the numa_post_machine_init func in the appropriate
location to make sure that all the CPUs has already been created
when it is called.
Signed-off-by: Dou Liyang
---
Change log v1-> v2:
1. Split it from
https://lists.nongnu.org/archive/html/qemu-devel/2017-01/msg03354.html
2. Rewrite the lo
Hi, All
**
ERROR:/tmp/qemu-test/src/tests/vhost-user-test.c:668:test_migrate: assertion failed:
(qdict_haskey(rsp, "return"))
GTester: last random seed: R02Sf52546c4daff8087416f43fa7c146db8
ftruncate: Permission denied
ftruncate: Permission denied
qemu-system-aarch64: /tmp/qemu-test/src/qom/cpu
Hi, Eduardo
At 01/18/2017 08:56 PM, Eduardo Habkost wrote:
On Wed, Jan 18, 2017 at 08:40:05PM +0800, Dou Liyang wrote:
Current default way of seting the CPUState::numa_node might be wrong
in case on cold/hot-plug CPUs. Making the users confused why the
NUMA info is different beetween the
Current function does some unnecessary operations, such as it makes
the assert() in the loop, and the loop was not stopped in time.
This patch moves the assert() out the loop and stops the loop in
time.
Signed-off-by: Dou Liyang
---
qom/cpu.c | 3 ++-
1 file changed, 2 insertions(+), 1
1 - 100 of 154 matches
Mail list logo
