At memory hot-remove time we can retrieve an LMB's nid from its
corresponding memory_block. There is no need to store the nid
in multiple locations.
Signed-off-by: Scott Cheloha
---
The linear search in powerpc's memory_add_physaddr_to_nid() has become a
bottleneck at boot on systems
Hi Michal,
On Thu, Mar 12, 2020 at 06:02:37AM +0100, Michal Suchánek wrote:
> On Wed, Mar 11, 2020 at 06:08:15PM -0500, Scott Cheloha wrote:
> > At memory hot-remove time we can retrieve an LMB's nid from its
> > corresponding memory_block. There is no need to store the
256 LMB(s) at index
8100
[...]
[ 103.934092] pseries-hotplug-mem: Memory at 3fff000 (drc index 80003fff)
was hot-added
Signed-off-by: Scott Cheloha
---
These linear searches become a serious bottleneck as the machine
approaches 64TB. There are just too many LMBs to use a linear
search
On Tue, Jan 28, 2020 at 05:56:55PM -0600, Nathan Lynch wrote:
> Scott Cheloha writes:
> > LMB lookup is currently an O(n) linear search. This scales poorly when
> > there are many LMBs.
> >
> > If we cache each LMB by both its base address and its DRC index
> > i
On Thu, Jan 30, 2020 at 10:09:32AM -0600, Fontenot, Nathan wrote:
> On 1/29/2020 12:10 PM, Scott Cheloha wrote:
> > On Tue, Jan 28, 2020 at 05:56:55PM -0600, Nathan Lynch wrote:
> >> Scott Cheloha writes:
> >>> LMB lookup is currently an O(n) linear search. This
On Tue, Jan 28, 2020 at 05:56:55PM -0600, Nathan Lynch wrote:
> Scott Cheloha writes:
> > LMB lookup is currently an O(n) linear search. This scales poorly when
> > there are many LMBs.
> >
> > If we cache each LMB by both its base address and its DRC index
> > i
d() so this patch will also
improve memory hot-add speeds on big machines.
Signed-off-by: Scott Cheloha
---
arch/powerpc/include/asm/drmem.h | 1 +
arch/powerpc/mm/drmem.c | 24
arch/powerpc/mm/numa.c | 29 ++---
3 files chang
This patch series introduces two xarrays of LMBs in an effort to speed
up the drmem code and simplify the hotplug code on pseries machines.
The first patch introduces an xarray of LMBs indexed by physical
address. xa_load() is then used to accelerate LMB lookup during
memory_add_physaddr_to_nid()
ht to run
faster, too: several linear searches have been replaced with xa_load(),
which runs in sub-linear time.
The array of LMBs, drmem_info.lmbs[], is kept to preserve the ordering
of LMBs read from the firmware in drmem_init() during firmware writes
in drmem_update_dt().
Signed-off-by: Scott Ch
> On Nov 28, 2023, at 9:21 AM, Nathan Lynch wrote:
>
> Nick Child writes:
>> Hi Nathan,
>> Patches 1 and 3 LGTM
>
> thanks.
>
>> Regarding this patch, dlpar_memory_remove_by_count() calls
>> dlpar_add_lmb() and does not free drc on add error.
>> dlpar_add_lmb() is called here in error recover
the drconf range are considered
first. On an otherwise idle LPAR with so many LMBs we would expect to
find the LMBs we need near the start of the drconf range, hence the
smaller speedup.
Signed-off-by: Scott Cheloha
---
arch/powerpc/mm/numa.c | 2 +-
arch/powerpc/platf
On Fri, Aug 21, 2020 at 10:33:10AM +0200, Laurent Dufour wrote:
> Le 11/08/2020 à 03:51, Scott Cheloha a écrit :
> >
> > [...]
> >
> > @@ -631,7 +638,7 @@ static int dlpar_memory_remove_by_ic(u32
> > lmbs_to_remove, u32 drc_index)
> > static
the drconf range are considered
first. On an otherwise idle LPAR with so many LMBs we would expect to
find the LMBs we need near the start of the drconf range, hence the
smaller speedup.
Signed-off-by: Scott Cheloha
---
Changelog:
v1:
https://lore.kernel.org/linuxppc-dev/202009101
the drconf range are considered
first. On an otherwise idle LPAR with so many LMBs we would expect to
find the LMBs we need near the start of the drconf range, hence the
smaller speedup.
Signed-off-by: Scott Cheloha
---
Changelog:
v1:
https://lore.kernel.org/linuxppc-dev/202009101
On Wed, Sep 16, 2020 at 09:39:53AM +0200, David Hildenbrand wrote:
> On 15.09.20 21:46, Scott Cheloha wrote:
> > During memory hot-add, dlpar_add_lmb() calls memory_add_physaddr_to_nid()
> > to determine which node id (nid) to use when later calling __add_memory().
> >
>
the drconf range are considered
first. On an otherwise idle LPAR with so many LMBs we would expect to
find the LMBs we need near the start of the drconf range, hence the
smaller speedup.
Signed-off-by: Scott Cheloha
---
Changelog:
v1:
https://lore.kernel.org/linuxppc-dev/202009101
Add a non-NUMA definition for of_drconf_to_nid_single() to topology.h
so we have one even if powerpc/mm/numa.c is not compiled. On a non-NUMA
kernel the appropriate node id is always first_online_node.
Signed-off-by: Scott Cheloha
Reported-by: kernel test robot
Fixes: 72cdd117c449 ("ps
Add a non-NUMA definition for of_drconf_to_nid_single() to topology.h
so we have one even if powerpc/mm/numa.c is not compiled. On a non-NUMA
kernel the appropriate node id is always first_online_node.
Signed-off-by: Scott Cheloha
Reported-by: kernel test robot
Fixes: 72cdd117c449 ("ps
PAPR v2.12 defines a new hypercall, H_WATCHDOG. The hypercall permits
guest control of one or more virtual watchdog timers.
Add the opcode for the H_WATCHDOG hypercall to hvcall.h. While here,
add a definition for H_NOOP, a possible return code for H_WATCHDOG.
Signed-off-by: Scott Cheloha
ned-off-by: Scott Cheloha
---
arch/powerpc/include/asm/firmware.h | 3 ++-
arch/powerpc/platforms/pseries/firmware.c | 1 +
2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/include/asm/firmware.h
b/arch/powerpc/include/asm/firmware.h
index 834b8ecf..398e0b5e4
workaround we represent them as platform devices.
This patch registers a single platform device, "pseries-wdt", with the
platform bus if the FW_FEATURE_WATCHDOG flag is set.
A driver for this device, "pseries-wdt", will be introduced in a
subsequent patch.
Signed-off-by: Scott Cheloh
es_wdt_start(wd);
+ return 0;
+}
+
+static const struct platform_device_id pseries_wdt_id[] = {
+ { .name = "pseries-wdt" },
+ {}
+};
+MODULE_DEVICE_TABLE(platform, pseries_wdt_id);
+
+static struct platform_driver pseries_wdt_driver = {
+ .driver = {
+ .name = DRV_NAME,
+ .owner = THIS_MODULE,
+ },
+ .id_table = pseries_wdt_id,
+ .probe = pseries_wdt_probe,
+ .resume = pseries_wdt_resume,
+ .suspend = pseries_wdt_suspend,
+};
+module_platform_driver(pseries_wdt_driver);
+
+MODULE_AUTHOR("Alexey Kardashevskiy ");
+MODULE_AUTHOR("Scott Cheloha ");
+MODULE_DESCRIPTION("POWER Architecture Platform Watchdog Driver");
+MODULE_LICENSE("GPL");
--
2.27.0
This series is preceded by two RFCs:
RFC v1:
https://lore.kernel.org/linux-watchdog/20220413165104.179144-1-chel...@linux.ibm.com/
RFC v2:
https://lore.kernel.org/linux-watchdog/20220509174357.5448-1-chel...@linux.ibm.com/
Changes of note from RFC v2:
- Separate platform device registration in
PAPR v2.12 defines a new hypercall, H_WATCHDOG. The hypercall permits
guest control of one or more virtual watchdog timers.
Add the opcode for the H_WATCHDOG hypercall to hvcall.h. While here,
add a definition for H_NOOP, a possible return code for H_WATCHDOG.
Signed-off-by: Scott Cheloha
ned-off-by: Scott Cheloha
---
arch/powerpc/include/asm/firmware.h | 3 ++-
arch/powerpc/platforms/pseries/firmware.c | 1 +
2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/include/asm/firmware.h
b/arch/powerpc/include/asm/firmware.h
index 834b8ecf..398e0b5e4
workaround we represent them as platform devices.
This patch registers a single platform device, "pseries-wdt", with the
platform bus if the FW_FEATURE_WATCHDOG flag is set.
A driver for this device, "pseries-wdt", will be introduced in a
subsequent patch.
Signed-off-by: Scott Cheloh
es_wdt_start(wd);
+ return 0;
+}
+
+static const struct platform_device_id pseries_wdt_id[] = {
+ { .name = "pseries-wdt" },
+ {}
+};
+MODULE_DEVICE_TABLE(platform, pseries_wdt_id);
+
+static struct platform_driver pseries_wdt_driver = {
+ .driver = {
+ .name = DRV_NAME,
+ .owner = THIS_MODULE,
+ },
+ .id_table = pseries_wdt_id,
+ .probe = pseries_wdt_probe,
+ .resume = pseries_wdt_resume,
+ .suspend = pseries_wdt_suspend,
+};
+module_platform_driver(pseries_wdt_driver);
+
+MODULE_AUTHOR("Alexey Kardashevskiy ");
+MODULE_AUTHOR("Scott Cheloha ");
+MODULE_DESCRIPTION("POWER Architecture Platform Watchdog Driver");
+MODULE_LICENSE("GPL");
--
2.27.0
The H_GetPerformanceCounterInfo (GPCI) hypercall input/output structs are
useful to modules outside of perf/, so move them into asm/hvcall.h to live
alongside the other powerpc hypercall structs.
Leave the perf-specific GPCI stuff in perf/hv-gpci.h.
Signed-off-by: Scott Cheloha
---
arch
A score of 0 indicates the worst possible affinity while a
score of 100 indicates perfect affinity. The score can be used to
reason about performance.
This patch adds the score for the local LPAR to the lparcfg procfile
under a new 'partition_affinity_score' key.
Signed-off-by: Scot
we're no longer seeing the
soft lockup. drmem_init() now completes in negligible time, even when
the LMB count is large.
Signed-off-by: Scott Cheloha
---
arch/powerpc/include/asm/drmem.h | 21 ---
arch/powerpc/mm/drmem.c | 6 +-
.../plat
we're no longer seeing the
soft lockup. drmem_init() now completes in negligible time, even when
the LMB count is large.
Signed-off-by: Scott Cheloha
---
v1:
- RFC
v2:
- Adjusted commit message.
- Miscellaneous cleanup.
v3:
- Correct issue found by Laurent Dufour :
- Add missing
terInfo hypercall is already used elsewhere in
the kernel, in powerpc/perf/hv-gpci.c. Refactoring that code and this
code into a more general API might be worthwhile if additional modules
require the hypercall in the future.
Signed-off-by: Scott Cheloha
---
arch/powerpc/platforms/ps
On Wed, May 25, 2022 at 04:35:11PM +1000, Alexey Kardashevskiy wrote:
>
> On 5/21/22 04:35, Scott Cheloha wrote:
> > PAPR v2.12 defines a new hypercall, H_WATCHDOG. The hypercall permits
> > guest control of one or more virtual watchdog timers. The timers have
> > milli
On Wed, May 25, 2022 at 12:52:09AM -0700, Guenter Roeck wrote:
> On 5/24/22 23:35, Alexey Kardashevskiy wrote:
> >
> > On 5/21/22 04:35, Scott Cheloha wrote:
> > > PAPR v2.12 defines a new hypercall, H_WATCHDOG. The hypercall permits
> > > guest control of on
On Wed, Jun 01, 2022 at 08:45:03AM -0700, Guenter Roeck wrote:
> On 6/1/22 08:07, Scott Cheloha wrote:
> [ ... ]
> > > > > +static unsigned long action = PSERIES_WDTF_ACTION_HARD_RESTART;
> > > > > +
> > > > > +static int
PAPR v2.12 defines a new hypercall, H_WATCHDOG. This patch series
adds support for this hypercall to powerpc/pseries kernels and
introduces a new watchdog driver, "pseries-wdt", for the virtual
timers exposed by the hypercall.
This series is preceded by the following:
RFC v1:
https://lore.kerne
PAPR v2.12 defines a new hypercall, H_WATCHDOG. The hypercall permits
guest control of one or more virtual watchdog timers.
Add the opcode for the H_WATCHDOG hypercall to hvcall.h. While here,
add a definition for H_NOOP, a possible return code for H_WATCHDOG.
Signed-off-by: Scott Cheloha
ned-off-by: Scott Cheloha
---
arch/powerpc/include/asm/firmware.h | 3 ++-
arch/powerpc/platforms/pseries/firmware.c | 1 +
2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/include/asm/firmware.h
b/arch/powerpc/include/asm/firmware.h
index 834b8ecf..398e0b5e4
dt" },
+ {}
+};
+MODULE_DEVICE_TABLE(platform, pseries_wdt_id);
+
+static struct platform_driver pseries_wdt_driver = {
+ .driver = {
+ .name = DRV_NAME,
+ .owner = THIS_MODULE,
+ },
+ .id_table = pseries_wdt_id,
+ .probe = pseries_wdt_probe,
+ .resume = pseries_wdt_resume,
+ .suspend = pseries_wdt_suspend,
+};
+module_platform_driver(pseries_wdt_driver);
+
+MODULE_AUTHOR("Alexey Kardashevskiy ");
+MODULE_AUTHOR("Scott Cheloha ");
+MODULE_DESCRIPTION("POWER Architecture Platform Watchdog Driver");
+MODULE_LICENSE("GPL");
--
2.27.0
workaround we represent them as platform devices.
This patch registers a single platform device, "pseries-wdt", with the
platform bus if the FW_FEATURE_WATCHDOG flag is set.
A driver for this device, "pseries-wdt", will be introduced in a
subsequent patch.
Signed-off-by: Scott Cheloh
On Fri, Jun 24, 2022 at 11:51:01PM +1000, Michael Ellerman wrote:
> Scott Cheloha writes:
> ...
> > +
> > +static struct platform_driver pseries_wdt_driver = {
> > + .driver = {
> > + .name = DRV_NAME,
> > + .owner = THIS_MODULE,
>
On Fri, Jun 24, 2022 at 11:27:36PM +1000, Michael Ellerman wrote:
> Nathan Lynch writes:
> > Scott Cheloha writes:
> >> PAPR v2.12 defines a new hypercall, H_WATCHDOG. The hypercall permits
> >> guest control of one or more virtual watchdog timers.
> ...
&g
On Fri, Jun 24, 2022 at 11:27:24PM +1000, Michael Ellerman wrote:
> Hi Scott,
>
> A few comments below ...
>
> Scott Cheloha writes:
> >
> > [...]
> >
> > diff --git a/Documentation/watchdog/watchdog-parameters.rst
> > b/Documentation/watchdog/w
PAPR v2.12 defines a new hypercall, H_WATCHDOG. This patch series
adds support for this hypercall to powerpc/pseries kernels and
introduces a new watchdog driver, "pseries-wdt", for the virtual
timers exposed by the hypercall.
This series is preceded by the following:
RFC v1:
https://lore.kerne
PAPR v2.12 defines a new hypercall, H_WATCHDOG. The hypercall permits
guest control of one or more virtual watchdog timers.
Add the opcode for the H_WATCHDOG hypercall to hvcall.h. While here,
add a definition for H_NOOP, a possible return code for H_WATCHDOG.
Signed-off-by: Scott Cheloha
ned-off-by: Scott Cheloha
---
arch/powerpc/include/asm/firmware.h | 3 ++-
arch/powerpc/platforms/pseries/firmware.c | 1 +
2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/arch/powerpc/include/asm/firmware.h
b/arch/powerpc/include/asm/firmware.h
index 834b8ecf..398e0b5e4
workaround we represent them as platform devices.
This patch registers a single platform device, "pseries-wdt", with the
platform bus if the FW_FEATURE_WATCHDOG flag is set.
A driver for this device, "pseries-wdt", will be introduced in a
subsequent patch.
Signed-off-by: Scott Cheloh
{}
+};
+MODULE_DEVICE_TABLE(platform, pseries_wdt_id);
+
+static struct platform_driver pseries_wdt_driver = {
+ .driver = {
+ .name = DRV_NAME,
+ },
+ .id_table = pseries_wdt_id,
+ .probe = pseries_wdt_probe,
+ .resume = pseries_wdt_resume,
+ .suspend = pseries_wdt_suspend,
+};
+module_platform_driver(pseries_wdt_driver);
+
+MODULE_AUTHOR("Alexey Kardashevskiy");
+MODULE_AUTHOR("Scott Cheloha");
+MODULE_DESCRIPTION("POWER Architecture Platform Watchdog Driver");
+MODULE_LICENSE("GPL");
--
2.27.0
On Wed, Jul 13, 2022 at 01:50:14PM -0700, Guenter Roeck wrote:
> On 7/13/22 13:23, Scott Cheloha wrote:
> > PAPR v2.12 defines a new hypercall, H_WATCHDOG. The hypercall permits
> > guest control of one or more virtual watchdog timers. The timers have
> > millisecond granu
49 matches
Mail list logo
