linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v3] ANDROID: binder: change down_write to down_read
@ 2018-03-29  2:37 Minchan Kim
  2018-03-29  2:42 ` Minchan Kim
  2018-03-29 23:42 ` kbuild test robot
  0 siblings, 2 replies; 6+ messages in thread
From: Minchan Kim @ 2018-03-29  2:37 UTC (permalink / raw)
  To: LKML
  Cc: Minchan Kim, Todd Kjos, Greg Kroah-Hartman,
	Arve Hjønnevåg, Martijn Coenen

binder_update_page_range needs down_write of mmap_sem because
vm_insert_page need to change vma->vm_flags to VM_MIXEDMAP unless
it is set. However, when I profile binder working, it seems
every binder buffers should be mapped in advance by binder_mmap.
It means we could set VM_MIXEDMAP in bider_mmap time which is
already hold a mmap_sem as down_write so binder_update_page_range
doesn't need to hold a mmap_sem as down_write.

Android suffers from mmap_sem contention so let's reduce mmap_sem
down_write.

Cc: Todd Kjos <tkjos@google.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Arve Hjønnevåg <arve@android.com>
Reviewed-by: Martijn Coenen <maco@android.com>
Signed-off-by: Minchan Kim <minchan@kernel.org>
---
>From v2:
  * Fix vma->flag setting - Arve

>From v1:
  * remove WARN_ON_ONCE - Greg
  * add reviewed-by - Martijn

Martijn, I took your LGTM of v1 as Reviewed-by. If you don't like it
or want to change it to acked-by, please, tell me.

 drivers/android/binder.c       | 3 ++-
 drivers/android/binder_alloc.c | 6 +++---
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/android/binder.c b/drivers/android/binder.c
index 764b63a5aade..875ab21fa2e0 100644
--- a/drivers/android/binder.c
+++ b/drivers/android/binder.c
@@ -4722,7 +4722,8 @@ static int binder_mmap(struct file *filp, struct vm_area_struct *vma)
 		failure_string = "bad vm_flags";
 		goto err_bad_arg;
 	}
-	vma->vm_flags = (vma->vm_flags | VM_DONTCOPY) & ~VM_MAYWRITE;
+	vma->vm_flags = (vma->flags | VM_DONTCOPY | VM_MIXEDMAP) &
+							~VM_MAYWRITE;
 	vma->vm_ops = &binder_vm_ops;
 	vma->vm_private_data = proc;
 
diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c
index 5a426c877dfb..4f382d51def1 100644
--- a/drivers/android/binder_alloc.c
+++ b/drivers/android/binder_alloc.c
@@ -219,7 +219,7 @@ static int binder_update_page_range(struct binder_alloc *alloc, int allocate,
 		mm = alloc->vma_vm_mm;
 
 	if (mm) {
-		down_write(&mm->mmap_sem);
+		down_read(&mm->mmap_sem);
 		vma = alloc->vma;
 	}
 
@@ -288,7 +288,7 @@ static int binder_update_page_range(struct binder_alloc *alloc, int allocate,
 		/* vm_insert_page does not seem to increment the refcount */
 	}
 	if (mm) {
-		up_write(&mm->mmap_sem);
+		up_read(&mm->mmap_sem);
 		mmput(mm);
 	}
 	return 0;
@@ -321,7 +321,7 @@ static int binder_update_page_range(struct binder_alloc *alloc, int allocate,
 	}
 err_no_vma:
 	if (mm) {
-		up_write(&mm->mmap_sem);
+		up_read(&mm->mmap_sem);
 		mmput(mm);
 	}
 	return vma ? -ENOMEM : -ESRCH;
-- 
2.17.0.rc1.321.gba9d0f2565-goog

^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [PATCH v3] ANDROID: binder: change down_write to down_read
  2018-03-29  2:37 [PATCH v3] ANDROID: binder: change down_write to down_read Minchan Kim
@ 2018-03-29  2:42 ` Minchan Kim
  2018-03-29  4:15   ` Greg Kroah-Hartman
  2018-03-29 23:42 ` kbuild test robot
  1 sibling, 1 reply; 6+ messages in thread
From: Minchan Kim @ 2018-03-29  2:42 UTC (permalink / raw)
  To: LKML
  Cc: Todd Kjos, Greg Kroah-Hartman, Arve Hjønnevåg, Martijn Coenen


On Thu, Mar 29, 2018 at 11:37:12AM +0900, Minchan Kim wrote:
> binder_update_page_range needs down_write of mmap_sem because
> vm_insert_page need to change vma->vm_flags to VM_MIXEDMAP unless
> it is set. However, when I profile binder working, it seems
> every binder buffers should be mapped in advance by binder_mmap.
> It means we could set VM_MIXEDMAP in bider_mmap time which is
> already hold a mmap_sem as down_write so binder_update_page_range
> doesn't need to hold a mmap_sem as down_write.
> 
> Android suffers from mmap_sem contention so let's reduce mmap_sem
> down_write.
> 
> Cc: Todd Kjos <tkjos@google.com>
> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Cc: Arve Hjønnevåg <arve@android.com>
> Reviewed-by: Martijn Coenen <maco@android.com>
> Signed-off-by: Minchan Kim <minchan@kernel.org>

Sent wrong version. Sorry for that. Please ignore this and take it.

>From 480e992d4a650fb98e1397114d75dea7af8e6d0c Mon Sep 17 00:00:00 2001
From: Minchan Kim <minchan@kernel.org>
Date: Wed, 28 Mar 2018 11:32:42 +0900
Subject: [PATCH v3] ANDROID: binder: change down_write to down_read
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

binder_update_page_range needs down_write of mmap_sem because
vm_insert_page need to change vma->vm_flags to VM_MIXEDMAP unless
it is set. However, when I profile binder working, it seems
every binder buffers should be mapped in advance by binder_mmap.
It means we could set VM_MIXEDMAP in bider_mmap time which is
already hold a mmap_sem as down_write so binder_update_page_range
doesn't need to hold a mmap_sem as down_write.

Android suffers from mmap_sem contention so let's reduce mmap_sem
down_write.

Cc: Arve Hjønnevåg <arve@android.com>
Cc: Todd Kjos <tkjos@google.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Martijn Coenen <maco@android.com>
Signed-off-by: Minchan Kim <minchan@kernel.org>
---

>From v2:
  * Fix vma->flag setting - Arve

>From v1:
  * remove WARN_ON_ONCE - Greg
  * add reviewed-by - Martijn

Martijn, I took your LGTM of v1 as Reviewed-by. If you don't like it
or want to change it to acked-by, please, tell me.

 drivers/android/binder.c       | 3 ++-
 drivers/android/binder_alloc.c | 6 +++---
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/android/binder.c b/drivers/android/binder.c
index 764b63a5aade..fe62be7d7113 100644
--- a/drivers/android/binder.c
+++ b/drivers/android/binder.c
@@ -4722,7 +4722,8 @@ static int binder_mmap(struct file *filp, struct vm_area_struct *vma)
 		failure_string = "bad vm_flags";
 		goto err_bad_arg;
 	}
-	vma->vm_flags = (vma->vm_flags | VM_DONTCOPY) & ~VM_MAYWRITE;
+	vma->vm_flags = (vma->vm_flags | VM_DONTCOPY | VM_MIXEDMAP) &
+							~VM_MAYWRITE;
 	vma->vm_ops = &binder_vm_ops;
 	vma->vm_private_data = proc;
 
diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c
index 5a426c877dfb..4f382d51def1 100644
--- a/drivers/android/binder_alloc.c
+++ b/drivers/android/binder_alloc.c
@@ -219,7 +219,7 @@ static int binder_update_page_range(struct binder_alloc *alloc, int allocate,
 		mm = alloc->vma_vm_mm;
 
 	if (mm) {
-		down_write(&mm->mmap_sem);
+		down_read(&mm->mmap_sem);
 		vma = alloc->vma;
 	}
 
@@ -288,7 +288,7 @@ static int binder_update_page_range(struct binder_alloc *alloc, int allocate,
 		/* vm_insert_page does not seem to increment the refcount */
 	}
 	if (mm) {
-		up_write(&mm->mmap_sem);
+		up_read(&mm->mmap_sem);
 		mmput(mm);
 	}
 	return 0;
@@ -321,7 +321,7 @@ static int binder_update_page_range(struct binder_alloc *alloc, int allocate,
 	}
 err_no_vma:
 	if (mm) {
-		up_write(&mm->mmap_sem);
+		up_read(&mm->mmap_sem);
 		mmput(mm);
 	}
 	return vma ? -ENOMEM : -ESRCH;
-- 
2.17.0.rc1.321.gba9d0f2565-goog

^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [PATCH v3] ANDROID: binder: change down_write to down_read
  2018-03-29  2:42 ` Minchan Kim
@ 2018-03-29  4:15   ` Greg Kroah-Hartman
  0 siblings, 0 replies; 6+ messages in thread
From: Greg Kroah-Hartman @ 2018-03-29  4:15 UTC (permalink / raw)
  To: Minchan Kim; +Cc: LKML, Todd Kjos, Arve Hjønnevåg, Martijn Coenen

On Thu, Mar 29, 2018 at 11:42:04AM +0900, Minchan Kim wrote:
> 
> On Thu, Mar 29, 2018 at 11:37:12AM +0900, Minchan Kim wrote:
> > binder_update_page_range needs down_write of mmap_sem because
> > vm_insert_page need to change vma->vm_flags to VM_MIXEDMAP unless
> > it is set. However, when I profile binder working, it seems
> > every binder buffers should be mapped in advance by binder_mmap.
> > It means we could set VM_MIXEDMAP in bider_mmap time which is
> > already hold a mmap_sem as down_write so binder_update_page_range
> > doesn't need to hold a mmap_sem as down_write.
> > 
> > Android suffers from mmap_sem contention so let's reduce mmap_sem
> > down_write.
> > 
> > Cc: Todd Kjos <tkjos@google.com>
> > Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> > Cc: Arve Hjønnevåg <arve@android.com>
> > Reviewed-by: Martijn Coenen <maco@android.com>
> > Signed-off-by: Minchan Kim <minchan@kernel.org>
> 
> Sent wrong version. Sorry for that. Please ignore this and take it.
> 
> >From 480e992d4a650fb98e1397114d75dea7af8e6d0c Mon Sep 17 00:00:00 2001
> From: Minchan Kim <minchan@kernel.org>
> Date: Wed, 28 Mar 2018 11:32:42 +0900
> Subject: [PATCH v3] ANDROID: binder: change down_write to down_read
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit

<snip>

I really don't want to hand-edit patches, please just send v4.

thanks,

greg k-h

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH v3] ANDROID: binder: change down_write to down_read
  2018-03-29  2:37 [PATCH v3] ANDROID: binder: change down_write to down_read Minchan Kim
  2018-03-29  2:42 ` Minchan Kim
@ 2018-03-29 23:42 ` kbuild test robot
  2018-03-30  0:37   ` Minchan Kim
  1 sibling, 1 reply; 6+ messages in thread
From: kbuild test robot @ 2018-03-29 23:42 UTC (permalink / raw)
  To: Minchan Kim
  Cc: kbuild-all, LKML, Minchan Kim, Todd Kjos, Greg Kroah-Hartman,
	Arve Hjønnevåg, Martijn Coenen

[-- Attachment #1: Type: text/plain, Size: 2606 bytes --]

Hi Minchan,

I love your patch! Yet something to improve:

[auto build test ERROR on staging/staging-testing]
[also build test ERROR on v4.16-rc7 next-20180329]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url:    https://github.com/0day-ci/linux/commits/Minchan-Kim/ANDROID-binder-change-down_write-to-down_read/20180330-043057
config: x86_64-randconfig-x014-201812 (attached as .config)
compiler: gcc-7 (Debian 7.3.0-1) 7.3.0
reproduce:
        # save the attached .config to linux build tree
        make ARCH=x86_64 

All errors (new ones prefixed by >>):

   drivers//android/binder.c: In function 'binder_mmap':
>> drivers//android/binder.c:4725:24: error: 'struct vm_area_struct' has no member named 'flags'; did you mean 'vm_flags'?
     vma->vm_flags = (vma->flags | VM_DONTCOPY | VM_MIXEDMAP) &
                           ^~~~~
                           vm_flags

vim +4725 drivers//android/binder.c

  4701	
  4702	static int binder_mmap(struct file *filp, struct vm_area_struct *vma)
  4703	{
  4704		int ret;
  4705		struct binder_proc *proc = filp->private_data;
  4706		const char *failure_string;
  4707	
  4708		if (proc->tsk != current->group_leader)
  4709			return -EINVAL;
  4710	
  4711		if ((vma->vm_end - vma->vm_start) > SZ_4M)
  4712			vma->vm_end = vma->vm_start + SZ_4M;
  4713	
  4714		binder_debug(BINDER_DEBUG_OPEN_CLOSE,
  4715			     "%s: %d %lx-%lx (%ld K) vma %lx pagep %lx\n",
  4716			     __func__, proc->pid, vma->vm_start, vma->vm_end,
  4717			     (vma->vm_end - vma->vm_start) / SZ_1K, vma->vm_flags,
  4718			     (unsigned long)pgprot_val(vma->vm_page_prot));
  4719	
  4720		if (vma->vm_flags & FORBIDDEN_MMAP_FLAGS) {
  4721			ret = -EPERM;
  4722			failure_string = "bad vm_flags";
  4723			goto err_bad_arg;
  4724		}
> 4725		vma->vm_flags = (vma->flags | VM_DONTCOPY | VM_MIXEDMAP) &
  4726								~VM_MAYWRITE;
  4727		vma->vm_ops = &binder_vm_ops;
  4728		vma->vm_private_data = proc;
  4729	
  4730		ret = binder_alloc_mmap_handler(&proc->alloc, vma);
  4731		if (ret)
  4732			return ret;
  4733		mutex_lock(&proc->files_lock);
  4734		proc->files = get_files_struct(current);
  4735		mutex_unlock(&proc->files_lock);
  4736		return 0;
  4737	
  4738	err_bad_arg:
  4739		pr_err("%s: %d %lx-%lx %s failed %d\n", __func__,
  4740		       proc->pid, vma->vm_start, vma->vm_end, failure_string, ret);
  4741		return ret;
  4742	}
  4743	

---
0-DAY kernel test infrastructure                Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all                   Intel Corporation

[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 30419 bytes --]

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH v3] ANDROID: binder: change down_write to down_read
  2018-03-29 23:42 ` kbuild test robot
@ 2018-03-30  0:37   ` Minchan Kim
  2018-03-30  2:15     ` Fengguang Wu
  0 siblings, 1 reply; 6+ messages in thread
From: Minchan Kim @ 2018-03-30  0:37 UTC (permalink / raw)
  To: kbuild test robot
  Cc: kbuild-all, LKML, Todd Kjos, Greg Kroah-Hartman,
	Arve Hjønnevåg, Martijn Coenen

On Fri, Mar 30, 2018 at 07:42:37AM +0800, kbuild test robot wrote:
> Hi Minchan,
> 
> I love your patch! Yet something to improve:

Glad to hear.
It's first time someone loves my patch. ;-)

> 
> [auto build test ERROR on staging/staging-testing]
> [also build test ERROR on v4.16-rc7 next-20180329]
> [if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
> 
> url:    https://github.com/0day-ci/linux/commits/Minchan-Kim/ANDROID-binder-change-down_write-to-down_read/20180330-043057
> config: x86_64-randconfig-x014-201812 (attached as .config)
> compiler: gcc-7 (Debian 7.3.0-1) 7.3.0
> reproduce:
>         # save the attached .config to linux build tree
>         make ARCH=x86_64 
> 
> All errors (new ones prefixed by >>):
> 
>    drivers//android/binder.c: In function 'binder_mmap':
> >> drivers//android/binder.c:4725:24: error: 'struct vm_area_struct' has no member named 'flags'; did you mean 'vm_flags'?
>      vma->vm_flags = (vma->flags | VM_DONTCOPY | VM_MIXEDMAP) &
>                            ^~~~~
>                            vm_flags

http://lkml.kernel.org/r/<20180329065424.203172-1-minchan@kernel.org>

This time, I was little bit fast. :)

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH v3] ANDROID: binder: change down_write to down_read
  2018-03-30  0:37   ` Minchan Kim
@ 2018-03-30  2:15     ` Fengguang Wu
  0 siblings, 0 replies; 6+ messages in thread
From: Fengguang Wu @ 2018-03-30  2:15 UTC (permalink / raw)
  To: Minchan Kim
  Cc: kbuild test robot, kbuild-all, LKML, Todd Kjos,
	Greg Kroah-Hartman, Arve Hjønnevåg, Martijn Coenen

On Fri, Mar 30, 2018 at 09:37:36AM +0900, Minchan Kim wrote:
>On Fri, Mar 30, 2018 at 07:42:37AM +0800, kbuild test robot wrote:
>> Hi Minchan,
>>
>> I love your patch! Yet something to improve:
>
>Glad to hear.
>It's first time someone loves my patch. ;-)

FYI, that message originates from Linus. :-)

>> [auto build test ERROR on staging/staging-testing]
>> [also build test ERROR on v4.16-rc7 next-20180329]
>> [if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
>>
>> url:    https://github.com/0day-ci/linux/commits/Minchan-Kim/ANDROID-binder-change-down_write-to-down_read/20180330-043057
>> config: x86_64-randconfig-x014-201812 (attached as .config)
>> compiler: gcc-7 (Debian 7.3.0-1) 7.3.0
>> reproduce:
>>         # save the attached .config to linux build tree
>>         make ARCH=x86_64
>>
>> All errors (new ones prefixed by >>):
>>
>>    drivers//android/binder.c: In function 'binder_mmap':
>> >> drivers//android/binder.c:4725:24: error: 'struct vm_area_struct' has no member named 'flags'; did you mean 'vm_flags'?
>>      vma->vm_flags = (vma->flags | VM_DONTCOPY | VM_MIXEDMAP) &
>>                            ^~~~~
>>                            vm_flags
>
>http://lkml.kernel.org/r/<20180329065424.203172-1-minchan@kernel.org>
>
>This time, I was little bit fast. :)

Hurrah, quick hands! :)

Cheers,
Fengguang

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2018-03-30  2:15 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-03-29  2:37 [PATCH v3] ANDROID: binder: change down_write to down_read Minchan Kim
2018-03-29  2:42 ` Minchan Kim
2018-03-29  4:15   ` Greg Kroah-Hartman
2018-03-29 23:42 ` kbuild test robot
2018-03-30  0:37   ` Minchan Kim
2018-03-30  2:15     ` Fengguang Wu

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).