Hi,
Here we, LKP(Linux Kernel Performance, a project to do performance
benchmark testing for Linux kernel, and try to find and fix the
performance regressions), found a 10%-20% performance regression
of aim7 benchmark introduced by commit 5a50508: "mm/rmap: Convert
the struct anon_vma::mutex to an
On Tue, Jan 29, 2013 at 09:44:00AM +0100, Ingo Molnar wrote:
>
> * Yuanhan Liu wrote:
>
> > [...]
>
> Very nice measurements and analysis, thanks!
>
> > As stated above, anybody can have a chance to own the lock in
> > mutex once somebody release the lock.
1 1 1 1 2 2
cpu 15: 0 0 ... 2 0 0 1 0 1 1 1 1 1 2 2
Where you can see that CPU is much busier with this patch.
Suggested-by: Ingo Molnar
Cc: David Howells
Signed-off-by: Yuanha
On Tue, Jan 29, 2013 at 10:12:45AM +0100, Ingo Molnar wrote:
>
> * Yuanhan Liu wrote:
>
> > On Tue, Jan 29, 2013 at 09:44:00AM +0100, Ingo Molnar wrote:
> > >
> > > * Yuanhan Liu wrote:
> > >
> > > > [...]
> > >
> > &g
put get/get_uts() into CONFIG_PROC_SYSCTL code block as they are used
only when CONFIG_PROC_SYSCTL is enabled.
Signed-off-by: Yuanhan Liu
---
kernel/utsname_sysctl.c |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/kernel/utsname_sysctl.c b/kernel/utsname_sysctl.c
We can use user_ns, which is also assigned from task_cred_xxx(tsk,
user_ns), at the beginning of copy_namespaces().
Cc: Serge Hallyn
Signed-off-by: Yuanhan Liu
---
kernel/nsproxy.c |3 +--
1 files changed, 1 insertions(+), 2 deletions(-)
diff --git a/kernel/nsproxy.c b/kernel/nsproxy.c
Fix the wrong comment about the return value of clone_uts_ns()
Cc: Serge Hallyn
Signed-off-by: Yuanhan Liu
---
kernel/utsname.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/kernel/utsname.c b/kernel/utsname.c
index 08b197e..a47fc5d 100644
--- a/kernel/utsname.c
On Thu, Jan 31, 2013 at 10:39:31AM +0100, Ingo Molnar wrote:
>
> * Yuanhan Liu wrote:
>
> > We(Linux Kernel Performance project) found a regression introduced by
> > commit 5a50508, which just convert all mutex lock to rwsem write lock.
> > The semantics is same, bu
On Thu, Jan 31, 2013 at 11:45:41AM +0100, Ingo Molnar wrote:
>
> * Yuanhan Liu wrote:
>
> > > > output with this patch:
> > > > ---
> > > > cpu 00: 0 0 ... 1 1 2 1 1 1 2 1 1 1 1 3
> > > &
On Thu, Jan 31, 2013 at 03:57:51AM -0800, Michel Lespinasse wrote:
> On Wed, Jan 30, 2013 at 1:14 AM, Yuanhan Liu
> wrote:
> > We(Linux Kernel Performance project) found a regression introduced by
> > commit 5a50508, which just convert all mutex lock to rwsem write lock.
> &g
On Thu, Jan 31, 2013 at 02:12:28PM +0100, Ingo Molnar wrote:
>
> * Yuanhan Liu wrote:
>
> > BTW, mind to tell a nice test case for mmap_sem?
>
> this one was write-hitting on mmap_sem pretty hard, last I
> checked:
>
> http://people.redhat.com/mingo/threaded
On Thu, Jan 31, 2013 at 10:18:18PM +0100, Ingo Molnar wrote:
>
> * Yuanhan Liu wrote:
>
> > On Thu, Jan 31, 2013 at 02:12:28PM +0100, Ingo Molnar wrote:
> > >
> > > * Yuanhan Liu wrote:
> > >
> > > > BTW, mind to tell a nice test case fo
On Thu, Jan 31, 2013 at 12:22:52PM +0100, Ingo Molnar wrote:
>
> * Yuanhan Liu wrote:
>
> > > or whether the lock hold times could be reduced drastically
> >
> > I also found one, but it doesn't sound like the one will
> > reduce lock hold times
PU is much busier with this patch.
v2: make it stealable at __down_write_trylock as well, pointed by Michel
Reported-by: LKP project
Suggested-by: Ingo Molnar
Cc: David Howells
Cc: Michel Lespinasse
Signed-off-by: Yuanhan Liu
---
lib/rwsem-spinloc
It should be CONFIG_LOCK_STAT
Cc: Ingo Molnar
Signed-off-by: Yuanhan Liu
---
Documentation/lockstat.txt |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/Documentation/lockstat.txt b/Documentation/lockstat.txt
index cef00d4..dd2f7b2 100644
--- a/Documentation
Use spin_[un]lock instead of arch_spin_[un]lock in mutex-debug.h so
that we can collect the lock statistics of spin_lock_mutex from
/proc/lock_stat.
Cc: Ingo Molnar
Signed-off-by: Yuanhan Liu
---
kernel/mutex-debug.h |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a
On Thu, Jan 24, 2013 at 10:58:07AM +0100, Ingo Molnar wrote:
>
> * Yuanhan Liu wrote:
>
> > Use spin_[un]lock instead of arch_spin_[un]lock in mutex-debug.h so
> > that we can collect the lock statistics of spin_lock_mutex from
> > /proc/lock_stat.
> >
>
On Thu, Jan 24, 2013 at 11:14:50AM +0100, Ingo Molnar wrote:
>
> * Yuanhan Liu wrote:
>
> > On Thu, Jan 24, 2013 at 10:58:07AM +0100, Ingo Molnar wrote:
> > >
> > > * Yuanhan Liu wrote:
> > >
> > > > Use spin_[un]lock instead of arch
nction a bit and found it uses memcpy() to do the real _out_ things.
So, why we need do check here? memcpy can handle it no matter what
point type it is.
And please feel free to correct me if I'm wrong.
Thanks,
Yuanhan Liu
> This reason is the original
> code use the assignment to check
From: Yuanhan Liu
Firstly, this kind of type check doesn't work. It does something similay
like following:
void * __dummy = NULL;
__buf = __dummy;
__dummy is defined as void *. Thus it will not trigger warnings as
expected.
Second, we don't need that kind of check.
On Fri, Oct 26, 2012 at 07:38:31AM +0200, Stefani Seibold wrote:
> Am Freitag, den 26.10.2012, 09:46 +0800 schrieb Yuanhan Liu:
> > From: Yuanhan Liu
> >
> > Firstly, this kind of type check doesn't work. It does something similay
> > like following:
> >
On Fri, Oct 26, 2012 at 03:17:57PM +0800, Yuanhan Liu wrote:
> On Fri, Oct 26, 2012 at 08:51:06AM +0200, Stefani Seibold wrote:
> > Am Freitag, den 26.10.2012, 14:11 +0800 schrieb Yuanhan Liu:
> > > On Fri, Oct 26, 2012 at 07:38:31AM +0200, Stefani Seibold wrote:
> > > &
educe memory?
Thanks,
Yuanhan Liu
-
Cc: Stefani Seibold
Cc: Andrew Morton
Signed-off-by: Yuanhan Liu
---
kernel/kfifo.c |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/kernel/kfifo.c b/kernel/kfifo.c
index 59dcf5b..0f78378 100644
--- a/kernel/kfifo.c
+++ b/k
"not really happened yet" issue.
Cc: Stefani Seibold
Cc: Andrew Morton
Signed-off-by: Yuanhan Liu
---
kernel/kfifo.c | 22 ++
1 files changed, 10 insertions(+), 12 deletions(-)
diff --git a/kernel/kfifo.c b/kernel/kfifo.c
index 0f78378..e3a63c6 100644
--- a/kern
On Fri, Oct 26, 2012 at 08:51:06AM +0200, Stefani Seibold wrote:
> Am Freitag, den 26.10.2012, 14:11 +0800 schrieb Yuanhan Liu:
> > On Fri, Oct 26, 2012 at 07:38:31AM +0200, Stefani Seibold wrote:
> > > Am Freitag, den 26.10.2012, 09:46 +0800 schrieb Yuanhan Liu:
> &g
On Fri, Oct 26, 2012 at 05:52:44PM +0800, Richard Yang wrote:
> On Fri, Oct 26, 2012 at 02:11:45PM +0800, Yuanhan Liu wrote:
> >On Fri, Oct 26, 2012 at 07:38:31AM +0200, Stefani Seibold wrote:
> >> Am Freitag, den 26.10.2012, 09:46 +0800 schrieb Yuanhan Liu:
> &
On Fri, Oct 26, 2012 at 11:30:27AM +0200, Stefani Seibold wrote:
> Am Freitag, den 26.10.2012, 15:56 +0800 schrieb Yuanhan Liu:
> > Say, if we want to allocate a filo with size of 6 bytes, it would be safer
> > to allocate 8 bytes instead of 4 bytes.
> >
>
On Fri, Oct 26, 2012 at 11:26:31AM +0200, Stefani Seibold wrote:
> Am Freitag, den 26.10.2012, 15:17 +0800 schrieb Yuanhan Liu:
> > On Fri, Oct 26, 2012 at 08:51:06AM +0200, Stefani Seibold wrote:
> > > Am Freitag, den 26.10.2012, 14:11 +0800 schrieb Yuanhan Liu:
> > > &
On Fri, Oct 26, 2012 at 03:39:46PM +0200, Stefani Seibold wrote:
> Am Freitag, den 26.10.2012, 20:33 +0800 schrieb Yuanhan Liu:
> > On Fri, Oct 26, 2012 at 11:30:27AM +0200, Stefani Seibold wrote:
> > > Am Freitag, den 26.10.2012, 15:56 +0800 schrieb Yuanhan Liu:
> >
On Fri, Oct 26, 2012 at 03:42:44PM +0200, Stefani Seibold wrote:
> Am Freitag, den 26.10.2012, 21:04 +0800 schrieb Yuanhan Liu:
> > On Fri, Oct 26, 2012 at 11:26:31AM +0200, Stefani Seibold wrote:
> > > Am Freitag, den 26.10.2012, 15:17 +0800 schrieb Yuanhan Liu:
> > > &
ttps://lkml.org/lkml/2012/10/25/584
Cc: Andrew Morton
Cc: Wei Yang
Cc: Stefani Seibold
Cc: Fengguang Wu
Cc: Stephen Rothwell
Signed-off-by: Yuanhan Liu
---
include/linux/kfifo.h | 46 --
1 files changed, 12 insertions(+), 34 deletions(-)
diff
On Sat, Oct 27, 2012 at 09:55:58AM +0800, Richard Yang wrote:
> On Fri, Oct 26, 2012 at 08:31:38PM +0800, Yuanhan Liu wrote:
> >On Fri, Oct 26, 2012 at 05:52:44PM +0800, Richard Yang wrote:
> >> On Fri, Oct 26, 2012 at 02:11:45PM +0800, Yuanhan Liu wrote:
> >> >On F
Fix a coccinelle warning catched by Fengguang's 0-DAY system:
+ drivers/staging/dgrp/dgrp_net_ops.c:1061:11-27: ERROR: nd is NULL but
dereferenced.
Put the "done:" label a bit down would solve this issue.
Cc: Fengguang Wu
Cc: Julia Lawall
Cc: Greg Kroah-Hartman
Signed-off-
On Mon, Oct 29, 2012 at 01:59:35PM -0700, Andrew Morton wrote:
> On Fri, 26 Oct 2012 15:56:57 +0800
> Yuanhan Liu wrote:
>
> > Say, if we want to allocate a filo with size of 6 bytes, it would be safer
> > to allocate 8 bytes instead of 4 bytes.
> >
> > ...
>
On Wed, Oct 31, 2012 at 07:30:33AM +0100, Stefani Seibold wrote:
> Am Mittwoch, den 31.10.2012, 13:59 +0800 schrieb Yuanhan Liu:
> > On Mon, Oct 29, 2012 at 01:59:35PM -0700, Andrew Morton wrote:
> > > On Fri, 26 Oct 2012 15:56:57 +0800
> > > Yuanhan Liu wrote:
> &
aration of function
'__gpio_get_value'
include/asm-generic/gpio.h: In function 'gpio_set_value_cansleep':
nclude/asm-generic/gpio.h:226:2: error: implicit declaration of function
'__gpio_set_value'
Cc: Grant Likely
Cc: Linus Walleij
Cc: Fengguang Wu
Signed-off-by: Yuan
s without putting something like: #ifdef CONFIG_GPIOLIB
So, move those functions into proper section at
include/asm-generic/gpio.h would resolve 2 above issues.
Cc: Grant Likely
Cc: Linus Walleij
Cc: Fengguang Wu
Signed-off-by: Yuanhan Liu
---
include/asm-generic/gpio.h | 39 ++
On Wed, Oct 31, 2012 at 03:12:36PM +0800, Fengguang Wu wrote:
> On Wed, Oct 31, 2012 at 03:00:54PM +0800, Yuanhan Liu wrote:
> > There are 2 issues with current code:
> >
> > 1. redefinition of 'gpio_cansleep':
> > include/linux/gpio.h:60:19: error: redefiniti
On Fri, Sep 20, 2013 at 06:46:59AM -0700, tip-bot for Vladimir Davydov wrote:
> Commit-ID: 7e3115ef5149fc502e3a2e80719dba54a8e7409d
> Gitweb: http://git.kernel.org/tip/7e3115ef5149fc502e3a2e80719dba54a8e7409d
> Author: Vladimir Davydov
> AuthorDate: Sat, 14 Sep 2013 19:39:46 +0400
> Commi
Fix smatch warnings catched by Fengguang's 0-DAY system:
+ drivers/ptp/ptp_pch.c:641 pch_probe() warn: 'chip->mem_base' was not released
on error
+ drivers/ptp/ptp_pch.c:641 pch_probe() warn: 'chip->regs' was not released on
error
Cc: Richard Cochran
Cc: David S.
It maybe failed due to out of memory.
Cc: Matthew Wilcox
Cc: Keith Busch
Signed-off-by: Yuanhan Liu
---
drivers/block/nvme.c |4
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/drivers/block/nvme.c b/drivers/block/nvme.c
index 931769e..eb58978 100644
--- a/drivers
On Wed, Nov 14, 2012 at 08:03:49AM +0100, Stefani Seibold wrote:
> Am Freitag, den 09.11.2012, 10:32 +0800 schrieb Yuanhan Liu:
> > On Thu, Nov 08, 2012 at 01:37:15PM +0100, Stefani Seibold wrote:
> > > Am Donnerstag, den 08.11.2012, 20:24 +0800 schrieb Yuanhan Liu:
>
>
On Sun, Nov 04, 2012 at 06:47:12PM +0100, Linus Walleij wrote:
> On Wed, Oct 31, 2012 at 8:00 AM, Yuanhan Liu
> wrote:
>
> > Those functions are availabe only when CONFIG_GPIOLIB is set. So, we
> > should not call them under !CONFIG_GPIOLIB block.
> >
> > This
On Tue, Oct 30, 2012 at 11:52:10PM -0700, Andrew Morton wrote:
> On Wed, 31 Oct 2012 07:30:33 +0100 Stefani Seibold
> wrote:
>
> > > Yes, and I guess the same to give them a 64-element one.
> > >
> > > >
> > > > If there's absolutely no prospect that the kfifo code will ever support
> > > > 10
On Thu, Nov 08, 2012 at 01:37:15PM +0100, Stefani Seibold wrote:
> Am Donnerstag, den 08.11.2012, 20:24 +0800 schrieb Yuanhan Liu:
> > On Tue, Oct 30, 2012 at 11:52:10PM -0700, Andrew Morton wrote:
> > > On Wed, 31 Oct 2012 07:30:33 +0100 Stefani Seibold
> > > wrote:
] undefined!
ERROR: "adis_single_conversion" [drivers/staging/iio/gyro/adis16260.ko]
undefined!
Selecting IIO_ADIS_LIB at Kconfig file fixes this issue.
Reported-by: Fengguang Wu
Cc: Lars-Peter Clausen
Cc: Jonathan Cameron
Cc: Greg Kroah-Hartman
Signed-off-by: Yuanhan Liu
---
driver
On Mon, Dec 31, 2012 at 08:16:10AM +0100, Lars-Peter Clausen wrote:
> On 12/31/2012 06:36 AM, Yuanhan Liu wrote:
> > Commit 9d5e9fdf turned the code to use adis library, so it depends on
> > adis lib. But it forgot to select it at Kconfig file. Thus if adis is
> > not ena
scsi list
Cc: Stefani Seibold
Signed-off-by: Yuanhan Liu
---
drivers/scsi/libiscsi.c | 59 ---
include/scsi/libiscsi.h |4 +-
2 files changed, 32 insertions(+), 31 deletions(-)
diff --git a/drivers/scsi/libiscsi.c b/drivers/scsi/libiscsi.c
index 82c
;'
shows that this is one user of kfifo_init(2 in total).
Replace it with kfifo_alloc, as I propose to remove kfifo_init API
later.
Cc: James E.J. Bottomley
Cc: Stefani Seibold
Signed-off-by: Yuanhan Liu
---
drivers/scsi/libsrp.c | 22 ++
include/scsi/libsrp.h |
to following will make this issue gone:
if(address >= end && !spliting) {
...
}
Reported-by: LKP
CC: For 3.9+
Cc: H. Peter Anvin
Cc: Yinghai Lu
Bisected-by: "Xie, ChanglongX"
Signed-off-by: Yuanhan Liu
---
I reported this panic regression
On Sat, May 25, 2013 at 12:31:43AM -0700, Yinghai Lu wrote:
> On Fri, May 24, 2013 at 9:30 PM, Yuanhan Liu
> wrote:
> > Commit 8d57470d introduced a kernel panic while setting mem=2G at
> > boot time, and commit c9b3234a6 turns the the kernel panic to hang.
> >
> >
ar
Signed-off-by: Yuanhan Liu
---
kernel/timer.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/kernel/timer.c b/kernel/timer.c
index a61c093..0be4216 100644
--- a/kernel/timer.c
+++ b/kernel/timer.c
@@ -736,7 +736,7 @@ static struct tvec_base *lock_timer_base(struct
On Wed, Aug 22, 2012 at 04:00:25PM +1000, NeilBrown wrote:
> On Wed, 22 Aug 2012 11:57:02 +0800 Yuanhan Liu
> wrote:
>
> >
> > -#define NR_STRIPES 256
> > +#define NR_STRIPES 1024
>
> Changing one magic number into another magic number might he
On Mon, Aug 13, 2012 at 06:33:59PM +0800, Yuanhan Liu wrote:
> Embed block device name in struct block_device, thus we can replace
> the following code:
> char b[BDEVNAME_SIZE];
> printk(KERN_XXX "...%s..\n", bdevname(bdev, b);
>
> with a much simpler
No acutal case found. But logically, we should skip "OK" in case any
error met.
Signed-off-by: Yuanhan Liu
---
kernel/trace/trace_events.c |4 +++-
1 files changed, 3 insertions(+), 1 deletions(-)
diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index 29111d
akes the code a little cleaner and a bit smaller stack usage. If
this change is OK to you, I will file more patches to apply this change
to all other codes.
Cc: Alexander Viro
Cc: Jens Axboe
Cc: Andrew Morton
Signed-off-by: Yuanhan Liu
---
fs/buffer.c | 27 ---
fs
d_name);
This is a RFC version, which I just included one sample with this change
appiled. If this patch makes sense to you, I will file more patches to
apply this change.
Cc: Alexander Viro
Cc: Jens Axboe
Cc: Andrew Morton
Signed-off-by: Yuanhan Liu
---
fs/block_dev.c |2 ++
include/l
The old code would call __get_free_page() even though previous
allocation fail met. This is not needed.
Signed-off-by: Yuanhan Liu
Cc: Jeremy Fitzhardinge
Cc: Thomas Gleixner
Cc: Ingo Molnar
Cc: "H. Peter Anvin"
---
arch/x86/mm/pgtable.c | 18 +-
1 files
__generic_unplug_device() function is removed with commit
7eaceaccab5f40bbfda044629a6298616aeaed50, which forgot to
remove the declaration at meantime. Here remove it.
Signed-off-by: Yuanhan Liu
---
block/blk.h |1 -
1 files changed, 0 insertions(+), 1 deletions(-)
diff --git a/block/blk.h
Hi Ingo,
Ping...
On Fri, Feb 01, 2013 at 06:59:16PM +0800, Yuanhan Liu wrote:
> We(Linux Kernel Performance project) found a regression introduced by
> commit 5a50508, which just convert all mutex lock to rwsem write lock.
> The semantics is same, but the results is quite huge in s
acpi_no_s4_hw_signature is defined in #ifdef CONFIG_HIBERNATION block,
but the current code put the declare in #ifdef CONFIG_PM_SLEEP block.
Signed-off-by: Yuanhan Liu
---
include/linux/acpi.h |5 -
1 files changed, 4 insertions(+), 1 deletions(-)
diff --git a/include/linux/acpi.h b
On Tue, Oct 16, 2012 at 12:27:13PM +0800, Fengguang Wu wrote:
> The title could be made more descriptive:
>
> ACPI: move acpi_no_s4_hw_signature() declaration into #ifdef
> CONFIG_HIBERNATION
Yes, much better.
>
> On Tue, Oct 16, 2012 at 12:05:03PM +0800,
‘acpi_no_s4_hw_signature’ [-Werror=implicit-function-declaration]
v2: take better title and add build error message suggested by Fengguang
Signed-off-by: Yuanhan Liu
Reviewed-by: Fengguang Wu
---
include/linux/acpi.h |5 -
1 files changed, 4 insertions(+), 1 deletions(-)
diff --git a
On Tue, Oct 16, 2012 at 02:24:53PM +0800, Fengguang Wu wrote:
> On Tue, Oct 16, 2012 at 01:24:12PM +0800, Yuanhan Liu wrote:
> > acpi_no_s4_hw_signature is defined in #ifdef CONFIG_HIBERNATION block,
> > but the current code put the declaration in #ifdef CONFIG_PM_SLEEP block.
>
...@vger.kernel.org
Signed-off-by: Yuanhan Liu
Signed-off-by: Fengguang Wu
---
drivers/crypto/ux500/cryp/cryp_core.c |2 ++
drivers/crypto/ux500/hash/hash_core.c |2 ++
2 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/drivers/crypto/ux500/cryp/cryp_core.c
b/drivers/crypt
fn)
Only references the callbacks on CONFIG_PM_SLEEP (instead of CONFIG_PM).
Cc: John W. Linville
Cc: "Rafał Miłecki"
Cc: linux-wirel...@vger.kernel.org
Signed-off-by: Yuanhan Liu
Signed-off-by: Fengguang Wu
---
drivers/bcma/host_pci.c |6 +++---
1 files changed, 3 insertions(+)
fn)
Only references the callbacks on CONFIG_PM_SLEEP (instead of CONFIG_PM).
Cc: Jeff Garzik
Cc: Viresh Kumar
Cc: linux-...@vger.kernel.org
Signed-off-by: Yuanhan Liu
Signed-off-by: Fengguang Wu
---
drivers/ata/ahci_platform.c |2 +-
drivers/ata/pata_arasan_cf.c |2 +-
drivers/ata/sat
fn)
Only references the callbacks on CONFIG_PM_SLEEP (instead of CONFIG_PM).
Cc: Mattia Dongili
Cc: Arnd Bergmann
Cc: Greg Kroah-Hartman
Cc: platform-driver-...@vger.kernel.org
Signed-off-by: Yuanhan Liu
Signed-off-by: Fengguang Wu
---
drivers/char/sonypi.c |2 +-
1 files changed, 1 insert
This will fix warnings like following when CONFIG_PM_SLEEP is not set:
warning: 'xxx_suspend' defined but not used [-Wunused-function]
warning: 'xxx_resume' defined but not used [-Wunused-function]
Cc: "Rafael J. Wysocki"
Cc: Jens Axboe
Signed-off-
kfifo API
Thanks!
--yliu
Cc: Stefani Seibold
Cc: Andrew Morton
Cc: James E.J. Bottomley
Cc: iscsi list
Cc: Mike Christie
---
Yuanhan Liu (5):
kfifo: remove unnecessary type check
libsrp: replace kfifo_init with kfifo_alloc
libiscsi: replace kfifo_init with kfifo_alloc
kfifo: r
kfifo_init is buggy, we should never use that. Instead, we should use
kfifo_alloc.
Cc: Stefani Seibold
Cc: Andrew Morton
Signed-off-by: Yuanhan Liu
---
include/linux/kfifo.h | 27 ++-
kernel/kfifo.c| 23 ---
2 files changed, 2 insertions
ernel.org
Cc: linux-...@vger.kernel.org
Cc: linux...@kvack.org
Cc: d...@vger.kernel.org
Cc: linux-s...@vger.kernel.org
Signed-off-by: Yuanhan Liu
---
arch/arm/plat-omap/Kconfig |2 +-
arch/arm/plat-omap/mailbox.c|6 +++-
arch/powerpc/sysdev/fsl_rmu.c |
Signed-off-by: Yuanhan Liu
---
drivers/scsi/libsrp.c | 22 ++
include/scsi/libsrp.h |1 -
2 files changed, 10 insertions(+), 13 deletions(-)
diff --git a/drivers/scsi/libsrp.c b/drivers/scsi/libsrp.c
index 0707ecd..63db792 100644
--- a/drivers/scsi/libsrp.c
+++ b
libsrp.c: allocated items once by kcalloc.
Cc: Mike Christie
Cc: James E.J. Bottomley
Cc: iscsi list
Cc: Stefani Seibold
Cc: Andrew Morton
Signed-off-by: Yuanhan Liu
---
drivers/scsi/libiscsi.c | 59 ---
include/scsi/libiscsi.h |4 +-
2
ttps://lkml.org/lkml/2012/10/25/584
Cc: Stefani Seibold
Cc: Andrew Morton
Signed-off-by: Yuanhan Liu
---
include/linux/kfifo.h | 46 --
1 files changed, 12 insertions(+), 34 deletions(-)
diff --git a/include/linux/kfifo.h b/include/linux/kfifo.
On Tue, Jan 08, 2013 at 04:03:17PM +0100, Ulrich Windl wrote:
> >>> Yuanhan Liu schrieb am 08.01.2013 um 15:57
> >>> in
> Nachricht <1357657073-27352-1-git-send-email-yuanhan@linux.intel.com>:
>
> [...]
> > My proposal is to replace kfifo_init w
On Tue, Jan 08, 2013 at 10:51:04PM +0100, Stefani Seibold wrote:
> Am Dienstag, den 08.01.2013, 22:57 +0800 schrieb Yuanhan Liu:
> > Firstly, this kind of type check doesn't work. It does something similar
> > as following:
> > void * __dummy = NULL;
> > _
On Tue, Jan 08, 2013 at 10:16:46AM -0800, Dmitry Torokhov wrote:
> Hi Yuanhan,
>
> On Tue, Jan 08, 2013 at 10:57:53PM +0800, Yuanhan Liu wrote:
> > The current kfifo API take the kfifo size as input, while it rounds
> > _down_ the size to power of 2 at __kfifo_allo
On Wed, Jan 09, 2013 at 04:29:39PM +0100, Stefani Seibold wrote:
> Am Mittwoch, den 09.01.2013, 10:35 +0800 schrieb Yuanhan Liu:
> > On Tue, Jan 08, 2013 at 10:51:04PM +0100, Stefani Seibold wrote:
> > > Am Dienstag, den 08.01.2013, 22:57 +0800 schrieb Yuanhan Liu:
> > &
FYI, we found performance increasement, which is expected as commit patch says,
on `fsmark.files_per_sec' by c9dc4c6578502c2085705347375b82089aad18d0:
> commit c9dc4c6578502c2085705347375b82089aad18d0
> Author: Chris Mason
> AuthorDate: Sat Apr 4 17:14:42 2015 -0700
> Commit:
raid on those ramdisk
As you can see, though there are no much performance gain for hard disk
workload, the system time is dropped heavily, up to 97%. And as expected,
the performance increased a lot, up to 260%, for fast device(ram disk).
Signed-off-by: Yuanhan Liu
---
drivers/md/raid5.c
it wait_for_stripe, and here I introduce
wait_for_quiesce for the second usage. The name may not well taken, or
even taken wrongly. Feel free to correct me then.
This is also a prepare patch for next patch: make wait_for_stripe
exclusive.
Signed-off-by: Yuanhan Liu
---
drivers/md/raid5.c | 13 ++
On Mon, Apr 27, 2015 at 10:10:24AM +1000, NeilBrown wrote:
> On Fri, 24 Apr 2015 21:39:03 +0800 Yuanhan Liu
> wrote:
>
> > If I read code correctly, current wait_for_stripe actually has 2 usage:
> >
> > - wait for there is enough free stripe cache, triggered when
>
On Mon, Apr 27, 2015 at 10:24:05AM +1000, NeilBrown wrote:
> On Fri, 24 Apr 2015 21:39:04 +0800 Yuanhan Liu
> wrote:
>
> > I noticed heavy spin lock contention at get_active_stripe() with fsmark
> > multiple thread write workloads.
> >
> > Here is how this h
(active_stripes == 0)
Commit log refactor suggestion from Neil.
Signed-off-by: Yuanhan Liu
---
drivers/md/raid5.c | 15 +--
drivers/md/raid5.h | 1 +
2 files changed, 10 insertions(+), 6 deletions(-)
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 77dfd72..64d5bea 100644
That ends up introducing heavy lock contentions, and
hurts performance badly.
Here introduce wait_event_cmd_exclusive to relieve the lock contention
naturally by letting wake_up() just wake up one process.
Cc: Ingo Molnar
Cc: Peter Zijlstra
Signed-off-by: Yuanhan Liu
---
include/linux/w
o 97%. And as expected,
the performance increased a lot, up to 260%, for fast device(ram disk).
v2: use bits instead of array to note down wait queue need to wake up.
Signed-off-by: Yuanhan Liu
---
drivers/md/raid5.c | 27 +++
drivers/md/raid5.h | 2 +-
2 files changed
On Tue, Apr 28, 2015 at 04:13:15PM +0200, Peter Zijlstra wrote:
> On Mon, Apr 27, 2015 at 12:51:01PM +0800, Yuanhan Liu wrote:
> > It's just a variant of wait_event_cmd, with exclusive flag being set.
> >
> > For cases like RAID5, which puts many processes to sleep u
o 97%. And as expected,
the performance increased a lot, up to 260%, for fast device(ram disk).
v2: use bits instead of array to note down wait queue need to wake up.
Signed-off-by: Yuanhan Liu
---
drivers/md/raid5.c | 27 +++
drivers/md/raid5.h | 2 +-
2 files changed
e the same arguments - peterz
Signed-off-by: Yuanhan Liu
---
include/linux/wait.h | 13 +
1 file changed, 13 insertions(+)
diff --git a/include/linux/wait.h b/include/linux/wait.h
index 2db8334..db78c72 100644
--- a/include/linux/wait.h
+++ b/include/linux/wait.h
@@ -358,6 +35
(active_stripes == 0)
Commit log refactor suggestion from Neil.
Signed-off-by: Yuanhan Liu
---
drivers/md/raid5.c | 15 +--
drivers/md/raid5.h | 1 +
2 files changed, 10 insertions(+), 6 deletions(-)
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 77dfd72..64d5bea 100644
FYI, we noticed the below changes on
git://git.kernel.org/pub/scm/linux/kernel/git/mlin/linux.git block-generic-req
commit 5a19fe29ba7d052c0d8fa8a2bf461abc1e4d89bb ("block: make
generic_make_request handle arbitrarily sized bios")
testbox/testcase/testparams: vm-kbuild-1G/boot/1
v4.1-r
On Fri, Apr 24, 2015 at 12:15:59PM +1000, NeilBrown wrote:
> On Thu, 23 Apr 2015 14:55:59 +0800 Huang Ying wrote:
>
> > FYI, we noticed the below changes on
> >
> > git://neil.brown.name/md for-next
> > commit 878ee6792799e2f88bdcac329845efadb205252f ("RAID5: batch adjacent
> > full stripe writ
bion -> bios
Signed-off-by: Yuanhan Liu
---
drivers/md/raid5.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 697d77a..2651bda 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -2919,7 +291
Signed-off-by: Yuanhan Liu
---
drivers/md/raid5.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 2651bda..bae3e2c 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -5789,8 +5789,7 @@ static void raid5d(struct
On Thu, Apr 30, 2015 at 05:16:50PM +1000, NeilBrown wrote:
> On Thu, 30 Apr 2015 15:01:17 +0800 Yuanhan Liu
> wrote:
>
> > Signed-off-by: Yuanhan Liu
> > ---
> > drivers/md/raid5.c | 3 +--
> > 1 file changed, 1 insertion(+), 2 deletions(-)
> >
> >
On Thu, Apr 30, 2015 at 05:14:26PM +1000, NeilBrown wrote:
> On Thu, 30 Apr 2015 15:01:16 +0800 Yuanhan Liu
> wrote:
>
> > bion -> bios
> >
> > Signed-off-by: Yuanhan Liu
> > ---
> > drivers/md/raid5.c | 2 +-
> > 1 file changed, 1 insertion(+)
FYI, we noticed the below changes on
https://github.com/jiangliu/linux.git test/irq_common_data_v2
commit d5b2eacdbc280da7c6dfbe0f52bb293ef227d349 ("genirq: Introduce struct
irq_common_data to host shared irq data")
+-++---
FYI, we found changes on `fsmark.files_per_sec' by
78373b7319abdf15050af5b1632c4c8b8b398f33:
> commit 78373b7319abdf15050af5b1632c4c8b8b398f33
> Author: Jaegeuk Kim
> AuthorDate: Fri Mar 13 21:44:36 2015 -0700
> Commit: Jaegeuk Kim
> CommitDate: Fri Apr 10 15:08:45 2
On Fri, Nov 01, 2013 at 11:04:40AM -0700, Linus Torvalds wrote:
> On Fri, Nov 1, 2013 at 12:54 AM, Yuanhan Liu
> wrote:
> > @@ -67,19 +67,7 @@ static struct kmem_cache *anon_vma_chain_cachep;
> >
> > static inline struct anon_vma *anon_vma_alloc(void)
> > {
> &g
1 - 100 of 160 matches
Mail list logo
