[PATCH] USB: serial/ftdi_sio.c Fix kernel oops

[PATCH] USB: serial/ftdi_sio.c Fix kernel oops

[Lotfi Manseur]

Handle null termios in ftdi_set_termios(), introduced in
commit 552f6bf1bb0eda0011c0525dd587aa9e7ba5b846
This has been corrected in the mainline by
commits c515598e0f5769916c31c00392cc2bfe6af74e55 and
a816e3113b63753c330ca4751ea1d208e93e3015.

This is to be fixed in longterm 2.6.32.60 and 3.4.47.
This bug has been found with coccinelle.

Signed-off-by: Lotfi Manseur <lotfi.manseur@imag.fr>
Signed-off-by: Nicolas Palix <nicolas.palix@imag.fr>
---
 drivers/usb/serial/ftdi_sio.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/usb/serial/ftdi_sio.c b/drivers/usb/serial/ftdi_sio.c
index c374beb..615bd9e 100644
--- a/drivers/usb/serial/ftdi_sio.c
+++ b/drivers/usb/serial/ftdi_sio.c
@@ -2364,7 +2364,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
 
 	cflag = termios->c_cflag;
 
-	if (old_termios->c_cflag == termios->c_cflag
+	if (old_termios
+	    && old_termios->c_cflag == termios->c_cflag
 	    && old_termios->c_ispeed == termios->c_ispeed
 	    && old_termios->c_ospeed == termios->c_ospeed)
 		goto no_c_cflag_changes;
@@ -2373,7 +2374,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
 	   ftdi_sio_read_bulk_callback  - need to examine what this means -
 	   don't see any problems yet */
 
-	if ((old_termios->c_cflag & (CSIZE|PARODD|PARENB|CMSPAR|CSTOPB)) ==
+	if (old_termios &&
+	    (old_termios->c_cflag & (CSIZE|PARODD|PARENB|CMSPAR|CSTOPB)) ==
 	    (termios->c_cflag & (CSIZE|PARODD|PARENB|CMSPAR|CSTOPB)))
 		goto no_data_parity_stop_changes;
 
-- 
1.7.9.5

Re: [PATCH] USB: serial/ftdi_sio.c Fix kernel oops

[Willy Tarreau]

On Fri, Jun 07, 2013 at 03:14:32PM +0200, Lotfi Manseur wrote:
> Handle null termios in ftdi_set_termios(), introduced in
> commit 552f6bf1bb0eda0011c0525dd587aa9e7ba5b846
> This has been corrected in the mainline by
> commits c515598e0f5769916c31c00392cc2bfe6af74e55 and
> a816e3113b63753c330ca4751ea1d208e93e3015.
> 
> This is to be fixed in longterm 2.6.32.60 and 3.4.47.
> This bug has been found with coccinelle.

Thank you Lofti, I have queued it now.

Best regards,
Willy

Re: [PATCH] USB: serial/ftdi_sio.c Fix kernel oops

[Sergei Shtylyov]

Hello.

On 06/07/2013 05:14 PM, Lotfi Manseur wrote:

> Handle null termios in ftdi_set_termios(), introduced in
> commit 552f6bf1bb0eda0011c0525dd587aa9e7ba5b846
> This has been corrected in the mainline by
> commits c515598e0f5769916c31c00392cc2bfe6af74e55 and
> a816e3113b63753c330ca4751ea1d208e93e3015.

    You should also give the summary lines of all mentioned commits in 
parens.

> This is to be fixed in longterm 2.6.32.60 and 3.4.47.
> This bug has been found with coccinelle.
>
> Signed-off-by: Lotfi Manseur <lotfi.manseur@imag.fr>
> Signed-off-by: Nicolas Palix <nicolas.palix@imag.fr>

WBR, Sergei

Re: [PATCH] USB: serial/ftdi_sio.c Fix kernel oops

[Greg Kroah-Hartman]

On Fri, Jun 07, 2013 at 03:14:32PM +0200, Lotfi Manseur wrote:
> Handle null termios in ftdi_set_termios(), introduced in
> commit 552f6bf1bb0eda0011c0525dd587aa9e7ba5b846
> This has been corrected in the mainline by
> commits c515598e0f5769916c31c00392cc2bfe6af74e55 and
> a816e3113b63753c330ca4751ea1d208e93e3015.

Then why shouldn't we just include those commit instead of your patch,
which looks much different from these?

I'd much rather take the originals, not something else.

greg k-h

Re: [PATCH] USB: serial/ftdi_sio.c Fix kernel oops

[Willy Tarreau]

On Fri, Jun 07, 2013 at 10:30:09AM -0700, Greg Kroah-Hartman wrote:
> On Fri, Jun 07, 2013 at 03:14:32PM +0200, Lotfi Manseur wrote:
> > Handle null termios in ftdi_set_termios(), introduced in
> > commit 552f6bf1bb0eda0011c0525dd587aa9e7ba5b846
> > This has been corrected in the mainline by
> > commits c515598e0f5769916c31c00392cc2bfe6af74e55 and
> > a816e3113b63753c330ca4751ea1d208e93e3015.
> 
> Then why shouldn't we just include those commit instead of your patch,
> which looks much different from these?
> 
> I'd much rather take the originals, not something else.

Indeed. I've checked, and the second commit was just cosmetic (silence
a warning reported by sparse for ptr == 0 introduced in the first one).

The context is minimally different (original commit introduces a jump
to a label whose context has changed). I tend to think that the merged
fix has less changes of being incorrectly applied, but given their
triviality we could very well merge the original ones instead, and
keep initial authorship BTW.

Willy

Re: [PATCH] USB: serial/ftdi_sio.c Fix kernel oops

[Greg Kroah-Hartman]

On Fri, Jun 07, 2013 at 03:14:32PM +0200, Lotfi Manseur wrote:
> Handle null termios in ftdi_set_termios(), introduced in
> commit 552f6bf1bb0eda0011c0525dd587aa9e7ba5b846
> This has been corrected in the mainline by
> commits c515598e0f5769916c31c00392cc2bfe6af74e55 and

This commit showed up in 3.3, so it can't go into 3.4 at all.  Please be
more careful when asking for stable patches to be applied.  That is why
I want the _exact_ same patch to apply, don't try going and being smart
by mushing them together into something else, this would have obviously
not been correct for the 3.4 kernel at all.

greg k-h

Re: [PATCH] USB: serial/ftdi_sio.c Fix kernel oops

[Ben Hutchings]

[-- Attachment #1: Type: text/plain, Size: 1958 bytes --]

On Fri, 2013-06-07 at 15:14 +0200, Lotfi Manseur wrote:
> Handle null termios in ftdi_set_termios(), introduced in
> commit 552f6bf1bb0eda0011c0525dd587aa9e7ba5b846
> This has been corrected in the mainline by
> commits c515598e0f5769916c31c00392cc2bfe6af74e55 and
> a816e3113b63753c330ca4751ea1d208e93e3015.
> 
> This is to be fixed in longterm 2.6.32.60 and 3.4.47.
> This bug has been found with coccinelle.
> 
> Signed-off-by: Lotfi Manseur <lotfi.manseur@imag.fr>
> Signed-off-by: Nicolas Palix <nicolas.palix@imag.fr>

I've queued up those changes for 3.2.  This backported version seems
nicer, but we generally prefer to use patches that are as close as
possible to those in mainline.

Ben.

> ---
>  drivers/usb/serial/ftdi_sio.c |    6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/usb/serial/ftdi_sio.c b/drivers/usb/serial/ftdi_sio.c
> index c374beb..615bd9e 100644
> --- a/drivers/usb/serial/ftdi_sio.c
> +++ b/drivers/usb/serial/ftdi_sio.c
> @@ -2364,7 +2364,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
>  
>  	cflag = termios->c_cflag;
>  
> -	if (old_termios->c_cflag == termios->c_cflag
> +	if (old_termios
> +	    && old_termios->c_cflag == termios->c_cflag
>  	    && old_termios->c_ispeed == termios->c_ispeed
>  	    && old_termios->c_ospeed == termios->c_ospeed)
>  		goto no_c_cflag_changes;
> @@ -2373,7 +2374,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
>  	   ftdi_sio_read_bulk_callback  - need to examine what this means -
>  	   don't see any problems yet */
>  
> -	if ((old_termios->c_cflag & (CSIZE|PARODD|PARENB|CMSPAR|CSTOPB)) ==
> +	if (old_termios &&
> +	    (old_termios->c_cflag & (CSIZE|PARODD|PARENB|CMSPAR|CSTOPB)) ==
>  	    (termios->c_cflag & (CSIZE|PARODD|PARENB|CMSPAR|CSTOPB)))
>  		goto no_data_parity_stop_changes;
>  

-- 
Ben Hutchings
friends: People who know you well, but like you anyway.

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

Re: [PATCH] USB: serial/ftdi_sio.c Fix kernel oops

[Nicolas Palix]

Hi,

On Thu, Jun 13, 2013 at 5:35 AM, Ben Hutchings <ben@decadent.org.uk> wrote:
> On Fri, 2013-06-07 at 15:14 +0200, Lotfi Manseur wrote:
>> Handle null termios in ftdi_set_termios(), introduced in
>> commit 552f6bf1bb0eda0011c0525dd587aa9e7ba5b846
>> This has been corrected in the mainline by
>> commits c515598e0f5769916c31c00392cc2bfe6af74e55 and
>> a816e3113b63753c330ca4751ea1d208e93e3015.
>>
>> This is to be fixed in longterm 2.6.32.60 and 3.4.47.
>> This bug has been found with coccinelle.
>>
>> Signed-off-by: Lotfi Manseur <lotfi.manseur@imag.fr>
>> Signed-off-by: Nicolas Palix <nicolas.palix@imag.fr>
>
> I've queued up those changes for 3.2.  This backported version seems
> nicer, but we generally prefer to use patches that are as close as
> possible to those in mainline.

Thank you all for your comments.
We will be more careful for upcoming reports and patches
about the stable branches, and we will privilege mainline
patches next time.

Regards.

>
> Ben.
>
>> ---
>>  drivers/usb/serial/ftdi_sio.c |    6 ++++--
>>  1 file changed, 4 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/usb/serial/ftdi_sio.c b/drivers/usb/serial/ftdi_sio.c
>> index c374beb..615bd9e 100644
>> --- a/drivers/usb/serial/ftdi_sio.c
>> +++ b/drivers/usb/serial/ftdi_sio.c
>> @@ -2364,7 +2364,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
>>
>>       cflag = termios->c_cflag;
>>
>> -     if (old_termios->c_cflag == termios->c_cflag
>> +     if (old_termios
>> +         && old_termios->c_cflag == termios->c_cflag
>>           && old_termios->c_ispeed == termios->c_ispeed
>>           && old_termios->c_ospeed == termios->c_ospeed)
>>               goto no_c_cflag_changes;
>> @@ -2373,7 +2374,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
>>          ftdi_sio_read_bulk_callback  - need to examine what this means -
>>          don't see any problems yet */
>>
>> -     if ((old_termios->c_cflag & (CSIZE|PARODD|PARENB|CMSPAR|CSTOPB)) ==
>> +     if (old_termios &&
>> +         (old_termios->c_cflag & (CSIZE|PARODD|PARENB|CMSPAR|CSTOPB)) ==
>>           (termios->c_cflag & (CSIZE|PARODD|PARENB|CMSPAR|CSTOPB)))
>>               goto no_data_parity_stop_changes;
>>
>
> --
> Ben Hutchings
> friends: People who know you well, but like you anyway.



--
Nicolas Palix
Tel: +33 4 76 51 46 27
http://membres-liglab.imag.fr/palix/