Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1467186 > unrolled thread

[PATCH 0/2] IB/core: Fine-tuning for ib_is_udata_cleared()

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  [PATCH 0/2] IB/core: Fine-tuning for ib_is_udata_cleared() SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-21 19:40 +0200
    [PATCH 2/2] IB/core: Delete an unnecessary initialisation in  ib_is_udata_cleared() SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-21 19:50 +0200
    [PATCH 1/2] IB/core: Use memdup_user() rather than duplicating its  implementation SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-21 19:50 +0200
    Re: [PATCH 0/2] IB/core: Fine-tuning for ib_is_udata_cleared() Joe Perches <joe@perches.com> - 2016-08-21 20:10 +0200
      Re: IB/core: Fine-tuning for ib_is_udata_cleared() Joe Perches <joe@perches.com> - 2016-08-21 22:00 +0200
        Re: IB/core: Fine-tuning for ib_is_udata_cleared() SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-21 22:20 +0200
          Re: IB/core: Fine-tuning for ib_is_udata_cleared() Yann Droneaud <ydroneaud@opteya.com> - 2016-08-22 11:50 +0200
            [PATCH v2] IB/core: Use memdup_user() rather than duplicating its  implementation SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-22 18:40 +0200
              Re: [PATCH v2] IB/core: Use memdup_user() rather than duplicating its  implementation Doug Ledford <dledford@redhat.com> - 2016-08-23 18:50 +0200
      Re: IB/core: Fine-tuning for ib_is_udata_cleared() SF Markus Elfring <elfring@users.sourceforge.net> - 2016-08-21 22:00 +0200

#1467186 — [PATCH 0/2] IB/core: Fine-tuning for ib_is_udata_cleared()

From: Markus Elfring <elfring@users.sourceforge.net>
Date: Sun, 21 Aug 2016 19:34:12 +0200

A few update suggestions were taken into account
from static source code analysis.

Markus Elfring (2):
  Use memdup_user()
  Delete an unnecessary initialisation

 include/rdma/ib_verbs.h | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)

-- 
2.9.3

[toc] | [next] | [standalone]

#1467187 — [PATCH 2/2] IB/core: Delete an unnecessary initialisation in ib_is_udata_cleared()

From: Markus Elfring <elfring@users.sourceforge.net>
Date: Sun, 21 Aug 2016 19:23:12 +0200

The local variable "ret" will be set to an appropriate value a bit later.
Thus omit the explicit initialisation at the beginning.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
---
 include/rdma/ib_verbs.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 033610b..0d8100f 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -2199,7 +2199,7 @@ static inline bool ib_is_udata_cleared(struct ib_udata *udata,
 				       size_t len)
 {
 	const void __user *p = udata->inbuf + offset;
-	bool ret = false;
+	bool ret;
 	u8 *buf;
 
 	if (len > USHRT_MAX)
-- 
2.9.3

[toc] | [prev] | [next] | [standalone]

#1467188 — [PATCH 1/2] IB/core: Use memdup_user() rather than duplicating its implementation

From: Markus Elfring <elfring@users.sourceforge.net>
Date: Sun, 21 Aug 2016 18:45:22 +0200

Reuse existing functionality from memdup_user() instead of keeping
duplicate source code.

This issue was detected by using the Coccinelle software.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
---
 include/rdma/ib_verbs.h | 9 ++-------
 1 file changed, 2 insertions(+), 7 deletions(-)

diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 6f667dd..033610b 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -2205,16 +2205,11 @@ static inline bool ib_is_udata_cleared(struct ib_udata *udata,
 	if (len > USHRT_MAX)
 		return false;
 
-	buf = kmalloc(len, GFP_KERNEL);
-	if (!buf)
+	buf = memdup_user(p, len);
+	if (IS_ERR(buf))
 		return false;
 
-	if (copy_from_user(buf, p, len))
-		goto free;
-
 	ret = !memchr_inv(buf, 0, len);
-
-free:
 	kfree(buf);
 	return ret;
 }
-- 
2.9.3

[toc] | [prev] | [next] | [standalone]

#1467192

On Sun, 2016-08-21 at 19:39 +0200, SF Markus Elfring wrote:
> From: Markus Elfring <elfring@users.sourceforge.net>
> Date: Sun, 21 Aug 2016 19:34:12 +0200
> 
> A few update suggestions were taken into account
> from static source code analysis.

Don't introduce a defect in patch 1 and correct
that introduced defect in patch 2.

This should be a single patch.

[toc] | [prev] | [next] | [standalone]

#1467228 — Re: IB/core: Fine-tuning for ib_is_udata_cleared()

On Sun, 2016-08-21 at 21:51 +0200, SF Markus Elfring wrote:

> > Don't introduce a defect in patch 1 and correct
> > that introduced defect in patch 2.
> Which detail do you not like here?

See above.

[toc] | [prev] | [next] | [standalone]

#1467238 — Re: IB/core: Fine-tuning for ib_is_udata_cleared()

>>> Don't introduce a defect in patch 1 and correct
>>> that introduced defect in patch 2.
>> Which detail do you not like here?
> 
> See above.

This feedback is not clearer.

I find that the two update steps should work in principle,
shouldn't they?

I guess that we have got different preferences for the shown
patch granularity. Another update variant can follow a bit later
with the changes squashed together.

Regards,
Markus

[toc] | [prev] | [next] | [standalone]

#1467513 — Re: IB/core: Fine-tuning for ib_is_udata_cleared()

Hi,

Le dimanche 21 août 2016 à 22:15 +0200, SF Markus Elfring a écrit :
> > 
> > > 
> > > > 
> > > > Don't introduce a defect in patch 1 and correct
> > > > that introduced defect in patch 2.
> > > Which detail do you not like here?
> > 
> > See above.
> 
> This feedback is not clearer.
> 

It's clear enough: your second patch fixes an issue you introduced in
your first patch by removing the code which made use of the ret
initialization value:

-       if (copy_from_user(buf, p, len))
-               goto free;

> I find that the two update steps should work in principle,
> shouldn't they?
> 

It would be better to squash them here.

Regards.

-- 
Yann Droneaud
OPTEYA

[toc] | [prev] | [next] | [standalone]

#1467778 — [PATCH v2] IB/core: Use memdup_user() rather than duplicating its implementation

From: Markus Elfring <elfring@users.sourceforge.net>
Date: Mon, 22 Aug 2016 18:23:24 +0200

* Reuse existing functionality from memdup_user() instead of keeping
  duplicate source code.

  This issue was detected by using the Coccinelle software.

* The local variable "ret" will be set to an appropriate value a bit later.
  Thus omit the explicit initialisation at the beginning.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
---

v2: The desired changes were put into a single patch instead of
    distributing them over two update steps.

 include/rdma/ib_verbs.h | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)

diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 6f667dd..0d8100f 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -2199,22 +2199,17 @@ static inline bool ib_is_udata_cleared(struct ib_udata *udata,
 				       size_t len)
 {
 	const void __user *p = udata->inbuf + offset;
-	bool ret = false;
+	bool ret;
 	u8 *buf;
 
 	if (len > USHRT_MAX)
 		return false;
 
-	buf = kmalloc(len, GFP_KERNEL);
-	if (!buf)
+	buf = memdup_user(p, len);
+	if (IS_ERR(buf))
 		return false;
 
-	if (copy_from_user(buf, p, len))
-		goto free;
-
 	ret = !memchr_inv(buf, 0, len);
-
-free:
 	kfree(buf);
 	return ret;
 }
-- 
2.9.3

[toc] | [prev] | [next] | [standalone]

#1468707 — Re: [PATCH v2] IB/core: Use memdup_user() rather than duplicating its implementation

[Multipart message — attachments visible in raw view] — view raw

On 8/22/2016 12:30 PM, SF Markus Elfring wrote:
> From: Markus Elfring <elfring@users.sourceforge.net>
> Date: Mon, 22 Aug 2016 18:23:24 +0200
> 
> * Reuse existing functionality from memdup_user() instead of keeping
>   duplicate source code.
> 
>   This issue was detected by using the Coccinelle software.
> 
> * The local variable "ret" will be set to an appropriate value a bit later.
>   Thus omit the explicit initialisation at the beginning.
> 
> Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>

Applied, thanks.


-- 
Doug Ledford <dledford@redhat.com>
    GPG Key ID: 0E572FDD

[toc] | [prev] | [next] | [standalone]

#1467230 — Re: IB/core: Fine-tuning for ib_is_udata_cleared()

>> A few update suggestions were taken into account
>> from static source code analysis.
> 
> Don't introduce a defect in patch 1 and correct
> that introduced defect in patch 2.

Which detail do you not like here?


> This should be a single patch.

Do any more software software developers would like to get
the proposed changes in a single update step?

Regards,
Markus

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web