The hotplug code contains only a single function, which is an SMP
function. Move that to platsmp.c where all other SMP runctions reside.
That allows removing hotplug.c and declaring the cpu_die function
static.
Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Avoid races and add synchronisation between the arch specific
kill and die routines.
The same synchronisation issue was fixed on IMX platform
by this commit:
"ARM: imx: fix sync issue between imx_cpu_die and imx_cpu_kill"
(sha1: 2f3edfd7e2)
Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
The core is not powered down, it is pointless to call the cpu_pm notifiers and
switch to the global timer.
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-and-tested-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
As there is no Power management unit on this board, it is not possible to power
down a core, just WFI is allowed. There is no point to invalidate the cache and
exit coherency.
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-and-tested-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
The DDR controller can detect idle periods and leverage low power
features clock stop. When new requests occur, the DDRC resumes
normal operation.
Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Add the DDR controller to the Zynq devicetree.
Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
There is an ISL9305 regulator on the Parallella board, add it to the DT
along with descriptions of all the supplies.
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
cifs provides two dummy functions 'sess_auth_lanman' and
'sess_auth_kerberos' for the case in which the respective
features are not defined. However, the caller is also under
an #ifdef, so we just get warnings about unused code:
fs/cifs/sess.c:1109:1: warning: 'sess_auth_kerberos' defined but not used [-Wunused-function]
sess_auth_kerberos(struct sess_data *sess_data)
Removing the dead functions gets rid of the warnings without
any downsides that I can see.
(Yalin Wang reported the identical problem and fix so added him)
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Yalin Wang <yalin.wang@sonymobile.com>
Signed-off-by: Steve French <smfrench@gmail.com>
This reverts commit 52a3624444.
Causes rmmod to fail for at least 7 seconds after unmount which
makes automated testing a little harder when reloading cifs.ko
between test runs.
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
CC: Jeff Layton <jlayton@primarydata.com>
Signed-off-by: Steve French <smfrench@gmail.com>
In the past when schedule scan was started while connected on one of
interfaces FW crashes were observed. In the newest FW this issue is
solved, so remove limitaion after examining corresponding TLV flag.
Signed-off-by: Alexander Bondar <alexander.bondar@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
When read, this new debugfs entry requests the current temperature
from the firmware and outputs it.
Signed-off-by: Luciano Coelho <luciano.coelho@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
The scheduler is a HW sub-block that directs the work of the Flow
Handler by issuing requests for frame transfers, specifying source
and destination. Its primary function is to allocate flows into the
TX FIFOs based upon a pre-determined mapping.
The driver has some responsibilities to the scheduler, namely
initialising and maintaining the hardware registers. This is
currently done by directly accessing them, which can cause races
with the firmware also accessing the registers.
To address this problem, change the driver to no longer directly
access the registers but go through the firmware for this if the
firmware has support for DQA and thus the new command.
Signed-off-by: Avri Altman <avri.altman@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Set RRM_ENABLED (bit 6) flag in scan commands. When this flag is set
the fw will add WFA vendor-specific TPC report and DS parameter set
IEs into probe requests.
This bit is set for SCAN_OFFLOAD_REQUEST_CMD(0x51).
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Reviewed-by: Assaf Krauss <assaf.krauss@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Add a ucode capability flag for support of Quiet Period (part of RRM).
If this flag is set, advertise device support.
Signed-off-by: Assaf Krauss <assaf.krauss@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Introduce new tx command flag (TX_CMD_FLG_WRITE_TX_POWER) which requests the
fw to update the tx power value at some constant offset.
Set this flag when transmitting either TPC report or Link measurement report
action frames.
In addition, introduce new tlv which indicates whether the fw supports power
insertion (IWL_UCODE_TLV_CAPA_TXPOWER_INSERTION_SUPPORT) and publish the
corresponding nl80211 feature.
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Since tx skbs are not modified by the driver, let mac80211 know the
driver can handle cloned skbs.
Signed-off-by: Ido Yariv <idox.yariv@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Hang is observed on virtual machines during CPU hotplug,
especially in big guests with many CPUs. (It reproducible
more often if host is over-committed).
It happens because master CPU gives up waiting on
secondary CPU and allows it to run wild. As result
AP causes locking or crashing system. For example
as described here:
https://lkml.org/lkml/2014/3/6/257
If master CPU have sent STARTUP IPI successfully,
and AP signalled to master CPU that it's ready
to start initialization, make master CPU wait
indefinitely till AP is onlined.
To ensure that AP won't ever run wild, make it
wait at early startup till master CPU confirms its
intention to wait for AP. If AP doesn't respond in 10
seconds, the master CPU will timeout and cancel
AP onlining.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Acked-by: Toshi Kani <toshi.kani@hp.com>
Tested-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: xen-devel@lists.xenproject.org
Link: http://lkml.kernel.org/r/1403266991-12233-1-git-send-email-imammedo@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
LTC2952 is an external power control chip, which signals
the OS to shut down. This patch documents the DT binding
for the chip.
Signed-off-by: René Moll <rene.moll@xsens.com>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
This adds a driver for the LTC2952, an external power control chip,
which signals the OS to shut down. Additionally this driver lets the
kernel power down the board.
Signed-off-by: René Moll <rene.moll@xsens.com>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
This piece of code was added so that we return -EPROBE_DEFER when no devices are
registered. But even if class_for_each_device() returns 0, we are going to
return -EPROBE_DEFER only.
And so this code isn't required at all. Remove it.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
Using 'goto' statements for freeing resources on failures is a good choice as it
makes code very clean, and reduces the chances of human errors.
Though in most cases compiler may take care of this. But adding unnecessary
'goto' statements wouldn't make anything better. Code becomes less readable
actually.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
In power_supply_show_property() routine, we call ->get_property() conditionally
and should check for failure in that case only. There is no point comparing
'ret' for errors when 'ret' is surely zero.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
At multiple places we are doing exactly what PTR_ERR_OR_ZERO() does. And so that
routine can be reused instead of increasing lines of code here.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
The 'if' statements in power_supply_changed_work() are mostly there for taking
care of races and normally they will always evaluate to true. Optimize them for
fast execution with 'likely' statements.
Also there is need to have better comments in code to mention about the races
clearly. Get them in place.
Cc: Zoran Markovic <zrn.markovic@gmail.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
There is no need to check the value of ret.intval before returning it, as we
will be returning zero explicitly when ret.intval is zero.
So essentially we will end up returning value of ret.intval as it is. Drop the
unnecessary 'if' statements.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
We need to stop 'class_for_each_device' loop when a supply matches with the
of-node. In order to achieve this we currently return -EINVAL from
__power_supply_populate_supplied_from() on successful match.
class_for_each_device() is free to return similar errors in other cases as well
and so the choice of return value here isn't particularly great.
This commit isn't removing the Hack but making it more elegant by returning '1'
instead.
Also power_supply_find_supply_from_node() can return errors other than
-EPROBE_DEFER now if class_for_each_device() fails.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
Callers of power_supply_find_supply_from_node(), i.e.
power_supply_check_supplies(), must propagate the errors returned by it instead
of returning their own.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
of_node_put() was called twice in power_supply_check_supplies() whereas a single
call will also work. Rearrange code a bit to make that feasible.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
In few routines, we need to end the do-while loop when no more "power-supplies"
are available. Currently we are doing 'continue' which will make the
'while(np)' conditional statement run again.
Skip this by doing a 'break' instead.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
Typecast from 'void *' to any other pointer type falls under implicit typecasts
category and so doesn't require explicit typecasts. Drop them.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
power_supply.h requires to forward declare few structures. One of them is done
at the top of the file and other one just before it is used. Declare them
together for better readability.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
'unnecessary' was wrongly spelled as 'unncessary', also it should have been
'unnecessarily'.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
If power-supply's DT node doesn't have a valid "power-supplies" entry, then
power_supply_check_supplies() should return early instead of trying to allocate
memory for "supplied_from" array.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
In routine power_supply_check_supplies(), 'cnt' is counting the number of
supplies passed in "power-supplies" field of a node. The value of 'cnt' will
always be one more than the number of supplies after the do-while loop ends. And
so we need to allocate memory for 'cnt - 1' char pointers. But we are allocating
memory for 'cnt' instead.
Fix this by not over-allocating memory.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sebastian Reichel <sre@kernel.org>
The options -mno-mmx and -mno-sse are unconditionally added to
KBUILD_CFLAGS in both branches of an ifeq and through a
$(cc-option) further down. We can safely remove the first
instances.
In fact, since the -mno-mmx and -mno-sse options were introduced
simultaneous with the other two options in the $(cc-option)
[according to http://www.gnu.org/software/gcc/gcc-3.1/changes.html],
and since the former were unconditionally used, one can deduce that
only gcc versions knowing about all four are supported. So also
eliminate the $(cc-option) wrap.
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Acked-by: Peter Foley <pefoley2@pefoley.com>
Link: http://lkml.kernel.org/r/1410365139-24440-1-git-send-email-linux@rasmusvillemoes.dk
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Revert PERF_EVENT_STATE_EXIT check on read syscall path.
It breaks standard way to read counter, which is to open
the counter, wait for the monitored process to die and
read the counter.
Reported-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Stephane Eranian <eranian@google.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: David Ahern <dsahern@gmail.com>
Link: http://lkml.kernel.org/r/20140908143107.GG17728@krava.brq.redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Add an appendix providing a simple self-contained code snippet
showing how SCHED_DEADLINE reservations can be created by
application developers.
Signed-off-by: Juri Lelli <juri.lelli@arm.com>
Reviewed-by: Henrik Austad <henrik@austad.us>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Dario Faggioli <raistlin@linux.it>
Cc: Juri Lelli <juri.lelli@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1410256636-26171-6-git-send-email-juri.lelli@arm.com
[ Fixed some whitespace inconsistencies. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Add an appendix briefly describing tools that can be used to
test SCHED_DEADLINE (and the scheduler in general). Links to
where source code of the tools is hosted are also provided.
Signed-off-by: Juri Lelli <juri.lelli@arm.com>
Reviewed-by: Henrik Austad <henrik@austad.us>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Dario Faggioli <raistlin@linux.it>
Cc: Juri Lelli <juri.lelli@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1410256636-26171-5-git-send-email-juri.lelli@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Admission control is of key importance for SCHED_DEADLINE, since
it guarantees system schedulability (or tells us something about
the degree of guarantees we can provide to the user).
This patch improves and clarifies bits and pieces regarding AC,
both for UP and SMP systems.
Signed-off-by: Luca Abeni <luca.abeni@unitn.it>
Signed-off-by: Juri Lelli <juri.lelli@arm.com>
Reviewed-by: Henrik Austad <henrik@austad.us>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Dario Faggioli <raistlin@linux.it>
Cc: Juri Lelli <juri.lelli@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1410256636-26171-4-git-send-email-juri.lelli@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Section 4 intro was still describing the old interface. Rewrite
it.
Signed-off-by: Juri Lelli <juri.lelli@arm.com>
Signed-off-by: Luca Abeni <luca.abeni@unitn.it>
Reviewed-by: Henrik Austad <henrik@austad.us>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Dario Faggioli <raistlin@linux.it>
Cc: Juri Lelli <juri.lelli@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1410256636-26171-3-git-send-email-juri.lelli@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Several small changes regarding SCHED_DEADLINE documentation
that fix terminology and improve clarity and readability:
- "current runtime" becomes "remaining runtime"
- readablity of an equation is improved by introducing more spacing
- clarify when admission control will certainly fail
- new URL for CBS technical report
- substitue "smallest" with "earliest"
Signed-off-by: Luca Abeni <luca.abeni@unitn.it>
Signed-off-by: Juri Lelli <juri.lelli@arm.com>
Reviewed-by: Henrik Austad <henrik@austad.us>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Dario Faggioli <raistlin@linux.it>
Cc: Juri Lelli <juri.lelli@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1410256636-26171-2-git-send-email-juri.lelli@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Currently we genarate a queueing route if we have matching policies
but can not resolve the states and the sysctl xfrm_larval_drop is
disabled. Here we assume that dst_output() is called to kill the
queued packets. Unfortunately this assumption is not true in all
cases, so it is possible that these packets leave the system unwanted.
We fix this by generating queueing routes only from the
route lookup functions, here we can guarantee a call to
dst_output() afterwards.
Fixes: a0073fe18e ("xfrm: Add a state resolution packet queue")
Reported-by: Konstantinos Kolelis <k.kolelis@sirrix.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Currently we genarate a blackhole route route whenever we have
matching policies but can not resolve the states. Here we assume
that dst_output() is called to kill the balckholed packets.
Unfortunately this assumption is not true in all cases, so
it is possible that these packets leave the system unwanted.
We fix this by generating blackhole routes only from the
route lookup functions, here we can guarantee a call to
dst_output() afterwards.
Fixes: 2774c131b1 ("xfrm: Handle blackhole route creation via afinfo.")
Reported-by: Konstantinos Kolelis <k.kolelis@sirrix.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
In arch/x86/kernel/setup.c::trim_bios_range(), the codes
introduced by 1b5576e6 (base on d8a9e6a5), it updates the first
4Kb of memory to be E820_RESERVED region. That's because it's a
BIOS owned area but generally not listed in the E820 table:
e820: BIOS-provided physical RAM map:
BIOS-e820: [mem 0x0000000000000000-0x0000000000096fff] usable
BIOS-e820: [mem 0x0000000000097000-0x0000000000097fff] reserved
...
e820: update [mem 0x00000000-0x00000fff] usable ==> reserved
e820: remove [mem 0x000a0000-0x000fffff] usable
But the region of first 4Kb didn't register to nosave memory:
PM: Registered nosave memory: [mem 0x00097000-0x00097fff]
PM: Registered nosave memory: [mem 0x000a0000-0x000fffff]
The code in e820_mark_nosave_regions() assumes the first e820
area to be RAM, so it causes the first 4Kb E820_RESERVED region
ignored when register to nosave. This patch removed assumption
of the first e820 area.
Signed-off-by: Lee, Chun-Yi <jlee@suse.com>
Acked-by: Pavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Len Brown <len.brown@intel.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Takashi Iwai <tiwai@suse.de>
Link: http://lkml.kernel.org/r/1410491038-17576-1-git-send-email-jlee@suse.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The setup_node_data() function allocates a pg_data_t object,
inserts it into the node_data[] array and initializes the
following fields: node_id, node_start_pfn and
node_spanned_pages.
However, a few function calls later during the kernel boot,
free_area_init_node() re-initializes those fields, possibly with
setup_node_data() is not used.
This causes a small glitch when running Linux as a hyperv numa
guest:
SRAT: PXM 0 -> APIC 0x00 -> Node 0
SRAT: PXM 0 -> APIC 0x01 -> Node 0
SRAT: PXM 1 -> APIC 0x02 -> Node 1
SRAT: PXM 1 -> APIC 0x03 -> Node 1
SRAT: Node 0 PXM 0 [mem 0x00000000-0x7fffffff]
SRAT: Node 1 PXM 1 [mem 0x80200000-0xf7ffffff]
SRAT: Node 1 PXM 1 [mem 0x100000000-0x1081fffff]
NUMA: Node 1 [mem 0x80200000-0xf7ffffff] + [mem 0x100000000-0x1081fffff] -> [mem 0x80200000-0x1081fffff]
Initmem setup node 0 [mem 0x00000000-0x7fffffff]
NODE_DATA [mem 0x7ffdc000-0x7ffeffff]
Initmem setup node 1 [mem 0x80800000-0x1081fffff]
NODE_DATA [mem 0x1081ea000-0x1081fdfff]
crashkernel: memory value expected
[ffffea0000000000-ffffea0001ffffff] PMD -> [ffff88007de00000-ffff88007fdfffff] on node 0
[ffffea0002000000-ffffea00043fffff] PMD -> [ffff880105600000-ffff8801077fffff] on node 1
Zone ranges:
DMA [mem 0x00001000-0x00ffffff]
DMA32 [mem 0x01000000-0xffffffff]
Normal [mem 0x100000000-0x1081fffff]
Movable zone start for each node
Early memory node ranges
node 0: [mem 0x00001000-0x0009efff]
node 0: [mem 0x00100000-0x7ffeffff]
node 1: [mem 0x80200000-0xf7ffffff]
node 1: [mem 0x100000000-0x1081fffff]
On node 0 totalpages: 524174
DMA zone: 64 pages used for memmap
DMA zone: 21 pages reserved
DMA zone: 3998 pages, LIFO batch:0
DMA32 zone: 8128 pages used for memmap
DMA32 zone: 520176 pages, LIFO batch:31
On node 1 totalpages: 524288
DMA32 zone: 7672 pages used for memmap
DMA32 zone: 491008 pages, LIFO batch:31
Normal zone: 520 pages used for memmap
Normal zone: 33280 pages, LIFO batch:7
In this dmesg, the SRAT table reports that the memory range for
node 1 starts at 0x80200000. However, the line starting with
"Initmem" reports that node 1 memory range starts at 0x80800000.
The "Initmem" line is reported by setup_node_data() and is
wrong, because the kernel ends up using the range as reported in
the SRAT table.
This commit drops all that dead code from setup_node_data(),
renames it to alloc_node_data() and adds a printk() to
free_area_init_node() so that we report a node's memory range
accurately.
Here's the same dmesg section with this patch applied:
SRAT: PXM 0 -> APIC 0x00 -> Node 0
SRAT: PXM 0 -> APIC 0x01 -> Node 0
SRAT: PXM 1 -> APIC 0x02 -> Node 1
SRAT: PXM 1 -> APIC 0x03 -> Node 1
SRAT: Node 0 PXM 0 [mem 0x00000000-0x7fffffff]
SRAT: Node 1 PXM 1 [mem 0x80200000-0xf7ffffff]
SRAT: Node 1 PXM 1 [mem 0x100000000-0x1081fffff]
NUMA: Node 1 [mem 0x80200000-0xf7ffffff] + [mem 0x100000000-0x1081fffff] -> [mem 0x80200000-0x1081fffff]
NODE_DATA(0) allocated [mem 0x7ffdc000-0x7ffeffff]
NODE_DATA(1) allocated [mem 0x1081ea000-0x1081fdfff]
crashkernel: memory value expected
[ffffea0000000000-ffffea0001ffffff] PMD -> [ffff88007de00000-ffff88007fdfffff] on node 0
[ffffea0002000000-ffffea00043fffff] PMD -> [ffff880105600000-ffff8801077fffff] on node 1
Zone ranges:
DMA [mem 0x00001000-0x00ffffff]
DMA32 [mem 0x01000000-0xffffffff]
Normal [mem 0x100000000-0x1081fffff]
Movable zone start for each node
Early memory node ranges
node 0: [mem 0x00001000-0x0009efff]
node 0: [mem 0x00100000-0x7ffeffff]
node 1: [mem 0x80200000-0xf7ffffff]
node 1: [mem 0x100000000-0x1081fffff]
Initmem setup node 0 [mem 0x00001000-0x7ffeffff]
On node 0 totalpages: 524174
DMA zone: 64 pages used for memmap
DMA zone: 21 pages reserved
DMA zone: 3998 pages, LIFO batch:0
DMA32 zone: 8128 pages used for memmap
DMA32 zone: 520176 pages, LIFO batch:31
Initmem setup node 1 [mem 0x80200000-0x1081fffff]
On node 1 totalpages: 524288
DMA32 zone: 7672 pages used for memmap
DMA32 zone: 491008 pages, LIFO batch:31
Normal zone: 520 pages used for memmap
Normal zone: 33280 pages, LIFO batch:7
This commit was tested on a two node bare-metal NUMA machine and
Linux as a numa guest on hyperv and qemu/kvm.
PS: The wrong memory range reported by setup_node_data() seems to be
harmless in the current kernel because it's just not used. However,
that bad range is used in kernel 2.6.32 to initialize the old boot
memory allocator, which causes a crash during boot.
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
