On Wed, Mar 22, 2017 at 11:54:48PM -0300, Marcos Paulo de Souza wrote:
> Hi Peter,
>
> first of all, thanks a lot for reading this patch so quickly and to
> point a lot of things to make this doc way better.
>
> See some notes below.
thanks for all the fixes, much appreciated.
just two comments
Hi Peter,
first of all, thanks a lot for reading this patch so quickly and to
point a lot of things to make this doc way better.
See some notes below.
On Wed, Mar 22, 2017 at 02:03:31PM +1000, Peter Hutterer wrote:
> Thanks for this, I'm getting enough questions about this, so it's nice to
>
On Thu, Mar 23, 2017 at 1:27 AM, Kees Cook wrote:
> This is a modified revert of commit 65fe935dd238 ("x86/KASLR, x86/power:
> Remove x86 hibernation restrictions"), since it appears that 32-bit
> hibernation still can't support KASLR. 64-bit is fine. Since people have
>
This is a modified revert of commit 65fe935dd238 ("x86/KASLR, x86/power:
Remove x86 hibernation restrictions"), since it appears that 32-bit
hibernation still can't support KASLR. 64-bit is fine. Since people have
been running with KASLR by default on 32-bit since v4.8, this disables
hibernation
>
> On Mon, Nov 07, 2016 at 07:27:38PM +, Winkler, Tomas wrote:
> > I value your opinion but I'm not responsible for inventing RPMB
> > and/or its implementation storage devices (eMMC, UFC, NVMe), it's pretty
> much done deal out there in the wild.
> > I'm just trying to provide common API
On Tue, Mar 21, 2017 at 03:38:26PM -0400, Sean Paul wrote:
> On Tue, Mar 21, 2017 at 04:52:28PM +0100, Daniel Vetter wrote:
> > The discussion pretty much concluded without objections, let's
> > document what we agreed on.
> >
> > Cc'ing linux-doc for the new tag in
This microbenchmark simulates how the use of different futex types
can affect the actual performanace of userspace rwlock locks. The
usage is:
perf bench futex rwlock
Two sets of simple rwlock lock and unlock functions are implemented
using the wait-wake futexes and glibc rwlock
The futex_pi_state structure will be overloaded in later patches to
store state information about non-PI futexes. So the structure name
itself is no longer a good description of its purpose. So its name
is changed to futex_state, a more generic name.
Some of the functions that process the futex
A new futex_setup_timer() helper function is added to consolidate all
the hrtimer_sleeper setup code.
Signed-off-by: Waiman Long
---
kernel/futex.c | 67 --
1 file changed, 37 insertions(+), 30 deletions(-)
diff --git
Two new helper functions cmpxchg_futex_value() and get_futex_value()
are added to access and change the futex value without the hash
bucket lock. As a result, page fault is enabled and the page will
be faulted in if not present yet.
Signed-off-by: Waiman Long
---
Currently, the futex state objects can only be located indirectly as
hash bucket => futex_q => futex state
Actually it can be beneficial in some cases to locate the futex state
object directly from the hash bucket without the futex_q middleman.
Therefore, a new list head to link the
This patch adds a new document file on how to use the TP futexes.
In term of locking performance, a futex locking microbenchmark
was written where separate userspace mutexes are implemented using
wait-wake (WW), PI and TP futexes respectively. This microbenchmark
was intructed to run 10s of
A new futex implementation called throughput-optimized (TP) futexes
is introduced. The goal of this new futex type is to maximize locking
throughput at the expense of fairness and deterministic latency. Its
throughput is higher than that of the wait-wake futexes especially
on systems with a large
As the futex_state structure will be overloaded in later patches
to be used by non-PI futexes, it is necessary to add a type field to
distinguish among different types of futexes.
Signed-off-by: Waiman Long
---
kernel/futex.c | 15 +++
1 file changed, 11
A performance disadvantage of TP futexes versus wait-wake futex is
the fact that locking was done in the kernel for TP futexes instead of
in the userspace like the WW futexes. The additional latency between
when the lock was taken in the kernel and when the task returned to
the userspace would be
By default, the TP futexes do not have preference for either readers
or writers. Most reader-writer locks allows users to decide if they
want to prefer readers or writers more.
This patch allows the setting of the prefer-reader mode in the val
argument of the futex system call. If that flag is
This patch extends the futex-mutex and futex-rwlock microbenchmarks to
test userspace mutexes and rwlocks built on top of the TP futexes. We
can then compare the relative performance of those userspace locks
based on different type of futexes.
Signed-off-by: Waiman Long
---
The current TP futexes has no guarantee that the top waiter
(serialization mutex owner) can get the lock within a finite time.
As a result, lock starvation can happen.
A lock handoff mechanism is added to the TP futexes to prevent lock
starvation from happening. The idea is that the top waiter
The TP futexes was designed to support userspace mutually exclusive
locks. They are now extended to support userspace reader/writer
locks as well.
Two new futex command codes are added:
1) FUTEX_LOCK_SHARED - to acquire a shared lock (reader-lock)
2) FUTEX_UNLOCK_SHARED - to release a shred
To better understand how the TP futexes are performing, it is useful to
return the internal status on the TP futexes. The FUTEX_LOCK futex(2)
syscall will now return a positive status code if no error happens. The
status code consists of the following 3 fields:
1) Bits 00-07: code on how the
Unlike other futexes, TP futexes do not accept the specification of
a timeout value. To align with the other futexes, timeout support is
now added. Like FUTEX_WAIT, the timeout value is relative to current
time, not an absolute time. However, the timeout isn't as precise as
the other futex types
The tp-futex.txt was updated to add description about shared locking
support.
Running the futex locking microbenchmark on a 2-socket 36-core E5-2699
v3 system (HT off) running on a 4.11 based kernel, the performance
of TP rwlock with 1:1 reader/writer ratio versus one based on the
wait-wake
Unlike wake_up_process(), wake_up_q() doesn't tell us how many
tasks have been woken up. This information can sometimes be useful
for tracking purpose. So wake_up_q() is now modified to return that
information.
Signed-off-by: Waiman Long
---
include/linux/sched/wake_q.h | 2
All the TP futex lock waiters are serialized in the kernel using a
kernel mutex which acts like a wait queue. The order at which the
waiters popped out from the wait queue will affect performance when
exclusive (writer) and shared (reader) lock waiters are mixed in the
queue. The worst case
v5->v6:
- Futex code changes:
1) Allow userspace locking of exclusive locks instead of just
kernel locking (patch 14). This is controlled by a flag passed
to the futex(2) syscall. This enables better throughput.
2) Add a reader-preferring mode to shared locking.
- Futex
This microbenchmark simulates how the use of different futex types
can affect the actual performanace of userspace mutex locks. The
usage is:
perf bench futex mutex
Three sets of simple mutex lock and unlock functions are implemented
using the wait-wake and PI futexes as well as the
Two new helper functions (task_pi_list_add & task_pi_list_del)
are created to consolidate all the pure pi_state_list addition and
insertion codes. The set_owner argument in task_pi_list_add() will
be needed in a later patch.
Signed-off-by: Waiman Long
---
kernel/futex.c | 52
On Tue, 2017-03-21 at 16:12 +0100, Neil Armstrong wrote:
> Some display pipelines can only provide non-RBG input pixels to the HDMI TX
> Controller, this patch takes the pixel format from the plat_data if provided.
>
> Signed-off-by: Neil Armstrong
On i.MX6 we could
On 03/22/2017 05:21 PM, Philipp Zabel wrote:
> On Tue, 2017-03-21 at 16:12 +0100, Neil Armstrong wrote:
>> Some display pipelines can only provide non-RBG input pixels to the HDMI TX
>> Controller, this patch takes the pixel format from the plat_data if provided.
>>
>> Signed-off-by: Neil
Hi Neil,
On 21-03-2017 15:12, Neil Armstrong wrote:
> Some display pipelines can only provide non-RBG input pixels to the HDMI TX
> Controller, this patch takes the pixel format from the plat_data if provided.
>
> Signed-off-by: Neil Armstrong
Looks fine, I'm assuming
On 03/22/2017 05:08 PM, Jose Abreu wrote:
> Hi Neil,
>
>
> On 21-03-2017 15:12, Neil Armstrong wrote:
>> Some display pipelines can only provide non-RBG input pixels to the HDMI TX
>> Controller, this patch takes the pixel format from the plat_data if provided.
>>
>> Signed-off-by: Neil
On Tuesday 21 March 2017 03:18 AM, Rob Herring wrote:
> On Mon, Mar 13, 2017 at 07:52:50PM +0530, Kishon Vijay Abraham I wrote:
>> According to errata i870, access to the PCIe slave port
>> that are not 32-bit aligned will result in incorrect mapping
>> to TLP Address and Byte enable fields.
>>
Hi Rob,
On Tuesday 21 March 2017 03:13 AM, Rob Herring wrote:
> On Mon, Mar 13, 2017 at 07:52:51PM +0530, Kishon Vijay Abraham I wrote:
>> Update device tree binding documentation of TI's dra7xx PCI
>> controller to include property for enabling unaligned mem access.
>>
>> Signed-off-by: Kishon
On Wed, Mar 22, 2017 at 10:10:05AM -0400, Gabriel L. Somlo wrote:
> On Wed, Mar 22, 2017 at 03:35:18PM +0200, Michael S. Tsirkin wrote:
> > On Tue, Mar 21, 2017 at 05:02:25PM -0700, Nadav Amit wrote:
> > >
> > > > On Mar 21, 2017, at 3:51 PM, Gabriel Somlo wrote:
> > > >
> > >
On Wed, Mar 22, 2017 at 03:35:18PM +0200, Michael S. Tsirkin wrote:
> On Tue, Mar 21, 2017 at 05:02:25PM -0700, Nadav Amit wrote:
> >
> > > On Mar 21, 2017, at 3:51 PM, Gabriel Somlo wrote:
> > >
> > > And I get the exact same results on the MacBookAir4,2 (which exhibits
> > >
On Wed, Mar 22, 2017 at 05:38:09AM +, Amit Kama IL wrote:
> Add initial support for boards based on Satixfy's SX3000b (Catniss) SoC.
> The SoC includes a MIPS interAptiv dual core 4 VPE processor and boots
> using device-tree.
>
> Signed-off-by: Amit Kama
>
> The
CC MST who touched this file in last commit on it.
On 03/22/2017 01:48 PM, Linas Vepstas wrote:
> Hi,
>
> On Tue, Mar 21, 2017 at 9:24 PM, Cao jin wrote:
>> Include whitespace shooting; correction; typo fix; superfluous word
>> dropping.
>
>>
>> diff --git
Add initial SX3000b platform related documentation to document tree:
- Vendor prefix
- Platform binding documentation
- Interrupt Controller Unit binding documentation.
Signed-off-by: Amit Kama
diff --git
38 matches
Mail list logo