[PATCH v2] drivers/base: Fix length checks in create_syslog_header()/dev_vprintk_emit()

[PATCH v2] drivers/base: Fix length checks in create_syslog_header()/dev_vprintk_emit()

[Ben Hutchings]

[-- Attachment #1: Type: text/plain, Size: 1616 bytes --]

snprintf() returns the number of bytes that could have been written
(excluding the null), not the actual number of bytes written.  Given a
long enough subsystem or device name, these functions will advance
beyond the end of the on-stack buffer in dev_vprintk_exit(), resulting
in an information leak or stack corruption.  I don't know whether such
a long name is currently possible.

In case snprintf() returns a value >= the buffer size, do not add
structured logging information.  Also WARN if this happens, so we can
fix the driver or increase the buffer size.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
---
v2: use dev_WARN() not dev_WARN_ON()

 drivers/base/core.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/drivers/base/core.c b/drivers/base/core.c
index 20da3ad..062b7bf 100644
--- a/drivers/base/core.c
+++ b/drivers/base/core.c
@@ -2007,6 +2007,8 @@ create_syslog_header(const struct device *dev, char *hdr, size_t hdrlen)
 		return 0;
 
 	pos += snprintf(hdr + pos, hdrlen - pos, "SUBSYSTEM=%s", subsys);
+	if (pos >= hdrlen)
+		goto overflow;
 
 	/*
 	 * Add device identifier DEVICE=:
@@ -2038,7 +2040,14 @@ create_syslog_header(const struct device *dev, char *hdr, size_t hdrlen)
 				"DEVICE=+%s:%s", subsys, dev_name(dev));
 	}
 
+	if (pos >= hdrlen)
+		goto overflow;
+
 	return pos;
+
+overflow:
+	dev_WARN(dev, 1, "device/subsystem name too long");
+	return 0;
 }
 
 int dev_vprintk_emit(int level, const struct device *dev,

-- 
Ben Hutchings
Never attribute to conspiracy what can adequately be explained by stupidity.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 828 bytes --]

Re: [PATCH v2] drivers/base: Fix length checks in create_syslog_header()/dev_vprintk_emit()

[Joe Perches]

On Sun, 2014-06-08 at 23:51 +0100, Ben Hutchings wrote:
> snprintf() returns the number of bytes that could have been written
> (excluding the null), not the actual number of bytes written.  Given a
> long enough subsystem or device name, these functions will advance
> beyond the end of the on-stack buffer in dev_vprintk_exit(), resulting
> in an information leak or stack corruption.  I don't know whether such
> a long name is currently possible.
> In case snprintf() returns a value >= the buffer size, do not add
> structured logging information.

I believe this is guaranteed to be < 128, but
I also suppose it'd be better to emit whatever
buffer is < 128 and add a WARN_ON_ONCE instead.

Re: [PATCH v2] drivers/base: Fix length checks in create_syslog_header()/dev_vprintk_emit()

[Greg Kroah-Hartman]

On Sun, Jun 08, 2014 at 11:51:43PM +0100, Ben Hutchings wrote:
> snprintf() returns the number of bytes that could have been written
> (excluding the null), not the actual number of bytes written.  Given a
> long enough subsystem or device name, these functions will advance
> beyond the end of the on-stack buffer in dev_vprintk_exit(), resulting
> in an information leak or stack corruption.  I don't know whether such
> a long name is currently possible.
> 
> In case snprintf() returns a value >= the buffer size, do not add
> structured logging information.  Also WARN if this happens, so we can
> fix the driver or increase the buffer size.
> 
> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
> ---
> v2: use dev_WARN() not dev_WARN_ON()

This patch breaks the build in a huge way:

drivers/base/core.c: In function ‘create_syslog_header’:
drivers/base/core.c:2049:16: error: expected ‘)’ before numeric constant
  dev_WARN(dev, 1, "device/subsystem name too long");
                  ^

is the start of it, it goes on for a page or so after that :(

greg k-h

Re: [PATCH v2] drivers/base: Fix length checks in create_syslog_header()/dev_vprintk_emit()

[Ben Hutchings]

[-- Attachment #1: Type: text/plain, Size: 1504 bytes --]

On Wed, 2014-07-09 at 16:57 -0700, Greg Kroah-Hartman wrote:
> On Sun, Jun 08, 2014 at 11:51:43PM +0100, Ben Hutchings wrote:
> > snprintf() returns the number of bytes that could have been written
> > (excluding the null), not the actual number of bytes written.  Given a
> > long enough subsystem or device name, these functions will advance
> > beyond the end of the on-stack buffer in dev_vprintk_exit(), resulting
> > in an information leak or stack corruption.  I don't know whether such
> > a long name is currently possible.
> > 
> > In case snprintf() returns a value >= the buffer size, do not add
> > structured logging information.  Also WARN if this happens, so we can
> > fix the driver or increase the buffer size.
> > 
> > Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
> > ---
> > v2: use dev_WARN() not dev_WARN_ON()
> 
> This patch breaks the build in a huge way:
> 
> drivers/base/core.c: In function ‘create_syslog_header’:
> drivers/base/core.c:2049:16: error: expected ‘)’ before numeric constant
>   dev_WARN(dev, 1, "device/subsystem name too long");
>                   ^
> 
> is the start of it, it goes on for a page or so after that :(

Sorry about that, I must not have committed the working version before
mailing it.  If you delete the '1, ' it should work.  If you've already
deleted the patch, I'll check and send the working version.

Ben.

-- 
Ben Hutchings
To err is human; to really foul things up requires a computer.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 811 bytes --]

Re: [PATCH v2] drivers/base: Fix length checks in create_syslog_header()/dev_vprintk_emit()

[Greg Kroah-Hartman]

On Fri, Jul 11, 2014 at 06:20:58PM +0100, Ben Hutchings wrote:
> On Wed, 2014-07-09 at 16:57 -0700, Greg Kroah-Hartman wrote:
> > On Sun, Jun 08, 2014 at 11:51:43PM +0100, Ben Hutchings wrote:
> > > snprintf() returns the number of bytes that could have been written
> > > (excluding the null), not the actual number of bytes written.  Given a
> > > long enough subsystem or device name, these functions will advance
> > > beyond the end of the on-stack buffer in dev_vprintk_exit(), resulting
> > > in an information leak or stack corruption.  I don't know whether such
> > > a long name is currently possible.
> > > 
> > > In case snprintf() returns a value >= the buffer size, do not add
> > > structured logging information.  Also WARN if this happens, so we can
> > > fix the driver or increase the buffer size.
> > > 
> > > Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
> > > ---
> > > v2: use dev_WARN() not dev_WARN_ON()
> > 
> > This patch breaks the build in a huge way:
> > 
> > drivers/base/core.c: In function ‘create_syslog_header’:
> > drivers/base/core.c:2049:16: error: expected ‘)’ before numeric constant
> >   dev_WARN(dev, 1, "device/subsystem name too long");
> >                   ^
> > 
> > is the start of it, it goes on for a page or so after that :(
> 
> Sorry about that, I must not have committed the working version before
> mailing it.  If you delete the '1, ' it should work.  If you've already
> deleted the patch, I'll check and send the working version.

It's long gone, can you please fix and resend?

thanks,

greg k-h