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

Groups > linux.kernel > #1230294 > unrolled thread

[PATCH v2 2/2] EDAC: Fix sysfs dimm_label store operation

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 v2 2/2] EDAC: Fix sysfs dimm_label store operation Toshi Kani <toshi.kani@hpe.com> - 2015-09-22 17:10 +0200
    Re: [PATCH v2 2/2] EDAC: Fix sysfs dimm_label store operation Borislav Petkov <bp@alien8.de> - 2015-09-24 18:50 +0200
      Re: [PATCH v2 2/2] EDAC: Fix sysfs dimm_label store operation Toshi Kani <toshi.kani@hpe.com> - 2015-09-24 21:10 +0200
        Re: [PATCH v2 2/2] EDAC: Fix sysfs dimm_label store operation Borislav Petkov <bp@alien8.de> - 2015-09-24 21:20 +0200
          Re: [PATCH v2 2/2] EDAC: Fix sysfs dimm_label store operation Toshi Kani <toshi.kani@hpe.com> - 2015-09-24 22:10 +0200
            Re: [PATCH v2 2/2] EDAC: Fix sysfs dimm_label store operation Borislav Petkov <bp@alien8.de> - 2015-09-25 20:00 +0200

#1230294 — [PATCH v2 2/2] EDAC: Fix sysfs dimm_label store operation

Sysfs "dimm_label" and "chX_dimm_label" have the following issues
in their store operation.

 1) A newline-terminated input string causes redundant newlines

  # echo "test" > /sys/bus/mc0/devices/dimm0/dimm_label
  # cat  /sys/bus/mc0/devices/dimm0/dimm_label
  test

  #  od -bc /sys/bus/mc0/devices/dimm0/dimm_label
  0000000 164 145 163 164 012 012
            t   e   s   t  \n  \n
  0000006

 2) The original label string (31 characters) cannot be stored due to
    an improper size check

  # echo "CPU_SrcID#0_Ha#0_Chan#0_DIMM#0" \
  > /sys/bus/mc0/devices/dimm0/dimm_label
  # cat /sys/bus/mc0/devices/dimm0/dimm_label


  # od -bc /sys/bus/mc0/devices/dimm0/dimm_label
   0000000 012 012
            \n  \n
   0000002

 3) An input string longer than the buffer size results a wrong label
    info as it allows a retry with the remaining string.

  # echo "CPU_SrcID#0_Ha#0_Chan#0_DIMM#0_TEST" \
  > /sys/bus/mc0/devices/dimm0/dimm_label
  # cat  /sys/bus/mc0/devices/dimm0/dimm_label
  _TEST

Fix these issues by making the following changes:
 1) Replace a newline charactor at the end by setting a null. It also
    assures that the string is null-terminated within the size.
 2) Check the label buffer size with 'sizeof(dimm->label)'.
 3) Fail a request if its string exceeds the label buffer size.

Signed-off-by: Toshi Kani <toshi.kani@hpe.com>
Acked-by: Tony Luck <tony.luck@intel.com>
Cc: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Doug Thompson <dougthompson@xmission.com>
Cc: Robert Elliott <elliott@hpe.com>
Cc: Tony Luck <tony.luck@intel.com>
---
 drivers/edac/edac_mc_sysfs.c |   20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

diff --git a/drivers/edac/edac_mc_sysfs.c b/drivers/edac/edac_mc_sysfs.c
index 8983755..01ad279 100644
--- a/drivers/edac/edac_mc_sysfs.c
+++ b/drivers/edac/edac_mc_sysfs.c
@@ -241,13 +241,13 @@ static ssize_t channel_dimm_label_store(struct device *dev,
 	unsigned chan = to_channel(mattr);
 	struct rank_info *rank = csrow->channels[chan];
 
-	ssize_t max_size = 0;
+	if (count == 0 || count > sizeof(rank->dimm->label))
+		return -EINVAL;
 
-	max_size = min((ssize_t) count, (ssize_t) EDAC_MC_LABEL_LEN - 1);
-	strncpy(rank->dimm->label, data, max_size);
-	rank->dimm->label[max_size] = '\0';
+	strncpy(rank->dimm->label, data, count);
+	rank->dimm->label[count - 1] = '\0';
 
-	return max_size;
+	return count;
 }
 
 /* show function for dynamic chX_ce_count attribute */
@@ -495,13 +495,13 @@ static ssize_t dimmdev_label_store(struct device *dev,
 {
 	struct dimm_info *dimm = to_dimm(dev);
 
-	ssize_t max_size = 0;
+	if (count == 0 || count > sizeof(dimm->label))
+		return -EINVAL;
 
-	max_size = min((ssize_t) count, (ssize_t) EDAC_MC_LABEL_LEN - 1);
-	strncpy(dimm->label, data, max_size);
-	dimm->label[max_size] = '\0';
+	strncpy(dimm->label, data, count);
+	dimm->label[count - 1] = '\0';
 
-	return max_size;
+	return count;
 }
 
 static ssize_t dimmdev_size_show(struct device *dev,
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [next] | [standalone]

#1232247

On Tue, Sep 22, 2015 at 08:58:03AM -0600, Toshi Kani wrote:
> Sysfs "dimm_label" and "chX_dimm_label" have the following issues
> in their store operation.
> 
>  1) A newline-terminated input string causes redundant newlines
> 
>   # echo "test" > /sys/bus/mc0/devices/dimm0/dimm_label
>   # cat  /sys/bus/mc0/devices/dimm0/dimm_label
>   test
> 
>   #  od -bc /sys/bus/mc0/devices/dimm0/dimm_label
>   0000000 164 145 163 164 012 012
>             t   e   s   t  \n  \n
>   0000006
> 
>  2) The original label string (31 characters) cannot be stored due to
>     an improper size check
> 
>   # echo "CPU_SrcID#0_Ha#0_Chan#0_DIMM#0" \
>   > /sys/bus/mc0/devices/dimm0/dimm_label
>   # cat /sys/bus/mc0/devices/dimm0/dimm_label
> 
> 
>   # od -bc /sys/bus/mc0/devices/dimm0/dimm_label
>    0000000 012 012
>             \n  \n
>    0000002
> 
>  3) An input string longer than the buffer size results a wrong label
>     info as it allows a retry with the remaining string.
> 
>   # echo "CPU_SrcID#0_Ha#0_Chan#0_DIMM#0_TEST" \
>   > /sys/bus/mc0/devices/dimm0/dimm_label
>   # cat  /sys/bus/mc0/devices/dimm0/dimm_label
>   _TEST
> 
> Fix these issues by making the following changes:
>  1) Replace a newline charactor at the end by setting a null. It also
>     assures that the string is null-terminated within the size.
>  2) Check the label buffer size with 'sizeof(dimm->label)'.
>  3) Fail a request if its string exceeds the label buffer size.
> 
> Signed-off-by: Toshi Kani <toshi.kani@hpe.com>
> Acked-by: Tony Luck <tony.luck@intel.com>
> Cc: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
> Cc: Borislav Petkov <bp@alien8.de>
> Cc: Doug Thompson <dougthompson@xmission.com>
> Cc: Robert Elliott <elliott@hpe.com>
> Cc: Tony Luck <tony.luck@intel.com>
> ---
>  drivers/edac/edac_mc_sysfs.c |   20 ++++++++++----------
>  1 file changed, 10 insertions(+), 10 deletions(-)

...

> @@ -495,13 +495,13 @@ static ssize_t dimmdev_label_store(struct device *dev,
>  {
>  	struct dimm_info *dimm = to_dimm(dev);
>  
> -	ssize_t max_size = 0;
> +	if (count == 0 || count > sizeof(dimm->label))
> +		return -EINVAL;

$ echo "" > /sys/bus/mc0/devices/dimm0/dimm_label
$ od -bc /sys/bus/mc0/devices/dimm0/dimm_label
0000000
$ cat /sys/bus/mc0/devices/dimm0/dimm_label
$

I don't think we want to allow empty labels. I guess something like this
(2 because there's also the additional "\n"):

diff --git a/drivers/edac/edac_mc_sysfs.c b/drivers/edac/edac_mc_sysfs.c
index d4e0bff268d8..e52ba338334b 100644
--- a/drivers/edac/edac_mc_sysfs.c
+++ b/drivers/edac/edac_mc_sysfs.c
@@ -241,7 +241,7 @@ static ssize_t channel_dimm_label_store(struct device *dev,
        unsigned chan = to_channel(mattr);
        struct rank_info *rank = csrow->channels[chan];
 
-       if (count == 0 || count > sizeof(rank->dimm->label))
+       if (count < 2 || count > sizeof(rank->dimm->label))
                return -EINVAL;
 
        strncpy(rank->dimm->label, data, count);
@@ -495,7 +495,7 @@ static ssize_t dimmdev_label_store(struct device *dev,
 {
        struct dimm_info *dimm = to_dimm(dev);
 
-       if (count == 0 || count > sizeof(dimm->label))
+       if (count < 2 || count > sizeof(dimm->label))
                return -EINVAL;
 
        strncpy(dimm->label, data, count);

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

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

#1232351

On Thu, 2015-09-24 at 18:48 +0200, Borislav Petkov wrote:
> On Tue, Sep 22, 2015 at 08:58:03AM -0600, Toshi Kani wrote:
> > Sysfs "dimm_label" and "chX_dimm_label" have the following issues
> > in their store operation.
> > 
> >  1) A newline-terminated input string causes redundant newlines
> > 
> >   # echo "test" > /sys/bus/mc0/devices/dimm0/dimm_label
> >   # cat  /sys/bus/mc0/devices/dimm0/dimm_label
> >   test
> > 
> >   #  od -bc /sys/bus/mc0/devices/dimm0/dimm_label
> >   0000000 164 145 163 164 012 012
> >             t   e   s   t  \n  \n
> >   0000006
> > 
> >  2) The original label string (31 characters) cannot be stored due to
> >     an improper size check
> > 
> >   # echo "CPU_SrcID#0_Ha#0_Chan#0_DIMM#0" \
> >   > /sys/bus/mc0/devices/dimm0/dimm_label
> >   # cat /sys/bus/mc0/devices/dimm0/dimm_label
> > 
> > 
> >   # od -bc /sys/bus/mc0/devices/dimm0/dimm_label
> >    0000000 012 012
> >             \n  \n
> >    0000002
> > 
> >  3) An input string longer than the buffer size results a wrong label
> >     info as it allows a retry with the remaining string.
> > 
> >   # echo "CPU_SrcID#0_Ha#0_Chan#0_DIMM#0_TEST" \
> >   > /sys/bus/mc0/devices/dimm0/dimm_label
> >   # cat  /sys/bus/mc0/devices/dimm0/dimm_label
> >   _TEST
> > 
> > Fix these issues by making the following changes:
> >  1) Replace a newline charactor at the end by setting a null. It also
> >     assures that the string is null-terminated within the size.
> >  2) Check the label buffer size with 'sizeof(dimm->label)'.
> >  3) Fail a request if its string exceeds the label buffer size.
> > 
> > Signed-off-by: Toshi Kani <toshi.kani@hpe.com>
> > Acked-by: Tony Luck <tony.luck@intel.com>
> > Cc: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
> > Cc: Borislav Petkov <bp@alien8.de>
> > Cc: Doug Thompson <dougthompson@xmission.com>
> > Cc: Robert Elliott <elliott@hpe.com>
> > Cc: Tony Luck <tony.luck@intel.com>
> > ---
> >  drivers/edac/edac_mc_sysfs.c |   20 ++++++++++----------
> >  1 file changed, 10 insertions(+), 10 deletions(-)
> 
> ...
> 
> > @@ -495,13 +495,13 @@ static ssize_t dimmdev_label_store(struct device *dev,
> >  {
> >  	struct dimm_info *dimm = to_dimm(dev);
> >  
> > -	ssize_t max_size = 0;
> > +	if (count == 0 || count > sizeof(dimm->label))
> > +		return -EINVAL;
> 
> $ echo "" > /sys/bus/mc0/devices/dimm0/dimm_label
> $ od -bc /sys/bus/mc0/devices/dimm0/dimm_label
> 0000000
> $ cat /sys/bus/mc0/devices/dimm0/dimm_label
> $
> 
> I don't think we want to allow empty labels. I guess something like this
> (2 because there's also the additional "\n"):

edac-utils(1) checks empty labels and shows them as "ch%d" [1].  So,
I think empty labels are supported today, and using 'echo "" >' seems
to be a legitimate way to set them empty if desired.

While looking at the count check, though, I noticed that 'echo -n'
does not send a null-terminated string (hence, the count check needs
to take it into account) since it simply calls putchar() to send '1'
only in the example below. [2]

  $ echo -n "1" | wc
        0       1       1

Attached is updated patch that handles the 'echo -n' case properly.
When an input string is not null-terminated, it appends
'\0' to the
end as long as it fits into the label buffer.

Thanks,
-Toshi

[1] https://github.com/grondo/edac-utils/search?utf8=%E2%9C%93&q=dimm_label_valid&type=Code
[2] http://git.savannah.gnu.org/cgit/coreutils.git/tree/src/echo.c

======
Subject: [PATCH v2 UPDATE 2/2] EDAC: Fix sysfs dimm_label store operation

Sysfs "dimm_label" and "chX_dimm_label" have the following issues
in their store operation.

 1) A newline-terminated input string causes redundant newlines

  # echo "test" > /sys/bus/mc0/devices/dimm0/dimm_label
  # cat  /sys/bus/mc0/devices/dimm0/dimm_label
  test

  #  od -bc /sys/bus/mc0/devices/dimm0/dimm_label
  0000000 164 145 163 164 012 012
            t   e   s   t  \n  \n
  0000006

 2) The original label string (31 characters) cannot be stored due to
    an improper size check

  # echo "CPU_SrcID#0_Ha#0_Chan#0_DIMM#0" \
  > /sys/bus/mc0/devices/dimm0/dimm_label
  # cat /sys/bus/mc0/devices/dimm0/dimm_label


  # od -bc /sys/bus/mc0/devices/dimm0/dimm_label
   0000000 012 012
            \n  \n
   0000002

 3) An input string longer than the buffer size results a wrong label
    info as it allows a retry with the remaining string.

  # echo "CPU_SrcID#0_Ha#0_Chan#0_DIMM#0_TEST" \
  > /sys/bus/mc0/devices/dimm0/dimm_label
  # cat  /sys/bus/mc0/devices/dimm0/dimm_label
  _TEST

Fix these issues by making the following changes:
 1) Replace a newline character at the end by setting a null. It also
    assures that the string is null-terminated in the label buffer.
 2) Check the label buffer size with 'sizeof(dimm->label)'.
 3) Fail a request if its string exceeds the label buffer size.

Signed-off-by: Toshi Kani <toshi.kani@hpe.com>
Acked-by: Tony Luck <tony.luck@intel.com>
Cc: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Doug Thompson <dougthompson@xmission.com>
Cc: Robert Elliott <elliott@hpe.com>
Cc: Tony Luck <tony.luck@intel.com>
---
 drivers/edac/edac_mc_sysfs.c |   34 ++++++++++++++++++++++++----------
 1 file changed, 24 insertions(+), 10 deletions(-)

diff --git a/drivers/edac/edac_mc_sysfs.c b/drivers/edac/edac_mc_sysfs.c
index 8983755..5252fb9 100644
--- a/drivers/edac/edac_mc_sysfs.c
+++ b/drivers/edac/edac_mc_sysfs.c
@@ -240,14 +240,21 @@ static ssize_t channel_dimm_label_store(struct device *dev,
 	struct csrow_info *csrow = to_csrow(dev);
 	unsigned chan = to_channel(mattr);
 	struct rank_info *rank = csrow->channels[chan];
+	size_t copy_count = count;
 
-	ssize_t max_size = 0;
+	if (count == 0)
+		return -EINVAL;
+
+	if (data[count - 1] == '\0' || data[count - 1] == '\n')
+		copy_count -= 1;
 
-	max_size = min((ssize_t) count, (ssize_t) EDAC_MC_LABEL_LEN - 1);
-	strncpy(rank->dimm->label, data, max_size);
-	rank->dimm->label[max_size] = '\0';
+	if (copy_count >= sizeof(rank->dimm->label))
+		return -EINVAL;
 
-	return max_size;
+	strncpy(rank->dimm->label, data, copy_count);
+	rank->dimm->label[copy_count] = '\0';
+
+	return count;
 }
 
 /* show function for dynamic chX_ce_count attribute */
@@ -494,14 +501,21 @@ static ssize_t dimmdev_label_store(struct device *dev,
 				   size_t count)
 {
 	struct dimm_info *dimm = to_dimm(dev);
+	size_t copy_count = count;
 
-	ssize_t max_size = 0;
+	if (count == 0)
+		return -EINVAL;
+
+	if (data[count - 1] == '\0' || data[count - 1] == '\n')
+		copy_count -= 1;
 
-	max_size = min((ssize_t) count, (ssize_t) EDAC_MC_LABEL_LEN - 1);
-	strncpy(dimm->label, data, max_size);
-	dimm->label[max_size] = '\0';
+	if (copy_count >= sizeof(dimm->label))
+		return -EINVAL;
 
-	return max_size;
+	strncpy(dimm->label, data, copy_count);
+	dimm->label[copy_count] = '\0';
+
+	return count;
 }
 
 static ssize_t dimmdev_size_show(struct device *dev,
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

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

#1232366

On Thu, Sep 24, 2015 at 01:06:04PM -0600, Toshi Kani wrote:
> edac-utils(1) checks empty labels and shows them as "ch%d" [1].  So,
> I think empty labels are supported today, and using 'echo "" >' seems
> to be a legitimate way to set them empty if desired.

What would be a sane use case to set a DIMM label to an empty string?

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

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

#1232411

On Thu, 2015-09-24 at 21:15 +0200, Borislav Petkov wrote:
> On Thu, Sep 24, 2015 at 01:06:04PM -0600, Toshi Kani wrote:
> > edac-utils(1) checks empty labels and shows them as "ch%d" [1].  So,
> > I think empty labels are supported today, and using 'echo "" >' seems
> > to be a legitimate way to set them empty if desired.
> 
> What would be a sane use case to set a DIMM label to an empty string?

Well, I guess there isn't a sane use case for it... :-)

Attached is a patch (on top of patch 2/2) to disallow an empty string.
I prefer to make it as a separate patch in case someone comes up with
a valid use-case for it.

Thanks,
-Toshi

====
Subject: [PATCH] EDAC: Fail empty string to sysfs dimm_label store

Updating dimm_label to an empty string does not make much sense.
Change the sysfs dimm_label store option to fail a request when
an input string is empty.

Suggested-by: Borislav Petkov <bp@alien8.de>
Signed-off-by: Toshi Kani <toshi.kani@hpe.com>
---
 drivers/edac/edac_mc_sysfs.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/edac/edac_mc_sysfs.c b/drivers/edac/edac_mc_sysfs.c
index 5252fb9..72ba530 100644
--- a/drivers/edac/edac_mc_sysfs.c
+++ b/drivers/edac/edac_mc_sysfs.c
@@ -248,7 +248,7 @@ static ssize_t channel_dimm_label_store(struct device *dev,
 	if (data[count - 1] == '\0' || data[count - 1] == '\n')
 		copy_count -= 1;
 
-	if (copy_count >= sizeof(rank->dimm->label))
+	if (copy_count == 0 || copy_count >= sizeof(rank->dimm->label))
 		return -EINVAL;
 
 	strncpy(rank->dimm->label, data, copy_count);
@@ -509,7 +509,7 @@ static ssize_t dimmdev_label_store(struct device *dev,
 	if (data[count - 1] == '\0' || data[count - 1] == '\n')
 		copy_count -= 1;
 
-	if (copy_count >= sizeof(dimm->label))
+	if (copy_count == 0 || copy_count >= sizeof(dimm->label))
 		return -EINVAL;
 
 	strncpy(dimm->label, data, copy_count);



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

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

#1232988

On Thu, Sep 24, 2015 at 01:59:27PM -0600, Toshi Kani wrote:
> On Thu, 2015-09-24 at 21:15 +0200, Borislav Petkov wrote:
> > On Thu, Sep 24, 2015 at 01:06:04PM -0600, Toshi Kani wrote:
> > > edac-utils(1) checks empty labels and shows them as "ch%d" [1].  So,
> > > I think empty labels are supported today, and using 'echo "" >' seems
> > > to be a legitimate way to set them empty if desired.
> > 
> > What would be a sane use case to set a DIMM label to an empty string?
> 
> Well, I guess there isn't a sane use case for it... :-)
> 
> Attached is a patch (on top of patch 2/2) to disallow an empty string.
> I prefer to make it as a separate patch in case someone comes up with
> a valid use-case for it.
> 
> Thanks,
> -Toshi
> 
> ====
> Subject: [PATCH] EDAC: Fail empty string to sysfs dimm_label store
> 
> Updating dimm_label to an empty string does not make much sense.
> Change the sysfs dimm_label store option to fail a request when
> an input string is empty.
> 
> Suggested-by: Borislav Petkov <bp@alien8.de>
> Signed-off-by: Toshi Kani <toshi.kani@hpe.com>
> ---
>  drivers/edac/edac_mc_sysfs.c |    4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)

All three applied.

Thanks.

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web