Commit 19e2ad6a09f0c06dbca19c98e5f4584269d913dd ("n_tty: Remove overflow
tests from receive_buf() path") moved the increment of read_head into
the arguments list of read_buf_addr(). Function calls represent a
sequence point in C. Therefore read_head is incremented before the
character c is placed in the buffer. Since the circular read buffer is
a lock-less design since commit 6d76bd2618535c581f1673047b8341fd291abc67
("n_tty: Make N_TTY ldisc receive path lockless"), this creates a race
condition that leads to communication errors.
This patch modifies the code to increment read_head _after_ the data
is placed in the buffer and thus fixes the race for non-SMP machines.
To fix the problem for SMP machines, memory barriers must be added in
a separate patch.
Signed-off-by: Christian Riesch <[email protected]>
Cc: <[email protected]>
---
This is version 2 of the patch in [1].
Changes for v2:
- Rewrote commit message. Since I did not know better, I blamed the compiler
in v1, but actually the code was wrong. See the discussion in [1].
- Removed memory barriers. For non-SMP machines they are not required,
for SMP machines more brainwork and discussions are needed.
Best regards,
Christian
[1] https://lkml.org/lkml/2014/11/6/216
drivers/tty/n_tty.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/tty/n_tty.c b/drivers/tty/n_tty.c
index 2e900a9..b09f326 100644
--- a/drivers/tty/n_tty.c
+++ b/drivers/tty/n_tty.c
@@ -321,7 +321,9 @@ static void n_tty_check_unthrottle(struct tty_struct *tty)
static inline void put_tty_queue(unsigned char c, struct n_tty_data *ldata)
{
- *read_buf_addr(ldata, ldata->read_head++) = c;
+ *read_buf_addr(ldata, ldata->read_head) = c;
+ /* increment read_head _after_ placing the character in the buffer */
+ ldata->read_head++;
}
/**
--
1.7.9.5
Christian Riesch <[email protected]> writes:
> Commit 19e2ad6a09f0c06dbca19c98e5f4584269d913dd ("n_tty: Remove overflow
> tests from receive_buf() path") moved the increment of read_head into
> the arguments list of read_buf_addr(). Function calls represent a
> sequence point in C. Therefore read_head is incremented before the
> character c is placed in the buffer. Since the circular read buffer is
> a lock-less design since commit 6d76bd2618535c581f1673047b8341fd291abc67
> ("n_tty: Make N_TTY ldisc receive path lockless"), this creates a race
> condition that leads to communication errors.
>
> This patch modifies the code to increment read_head _after_ the data
> is placed in the buffer and thus fixes the race for non-SMP machines.
> To fix the problem for SMP machines, memory barriers must be added in
> a separate patch.
>
> Signed-off-by: Christian Riesch <[email protected]>
> Cc: <[email protected]>
> ---
>
> This is version 2 of the patch in [1].
>
> Changes for v2:
> - Rewrote commit message. Since I did not know better, I blamed the compiler
> in v1, but actually the code was wrong. See the discussion in [1].
> - Removed memory barriers. For non-SMP machines they are not required,
> for SMP machines more brainwork and discussions are needed.
>
> Best regards,
> Christian
>
> [1] https://lkml.org/lkml/2014/11/6/216
>
> drivers/tty/n_tty.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/tty/n_tty.c b/drivers/tty/n_tty.c
> index 2e900a9..b09f326 100644
> --- a/drivers/tty/n_tty.c
> +++ b/drivers/tty/n_tty.c
> @@ -321,7 +321,9 @@ static void n_tty_check_unthrottle(struct tty_struct *tty)
>
> static inline void put_tty_queue(unsigned char c, struct n_tty_data *ldata)
> {
> - *read_buf_addr(ldata, ldata->read_head++) = c;
> + *read_buf_addr(ldata, ldata->read_head) = c;
> + /* increment read_head _after_ placing the character in the buffer */
> + ldata->read_head++;
> }
Is that comment really necessary?
--
M?ns Rullg?rd
[email protected]
On Tue, Nov 11, 2014 at 2:04 PM, Måns Rullgård <[email protected]> wrote:
> Christian Riesch <[email protected]> writes:
[...]>> diff --git a/drivers/tty/n_tty.c b/drivers/tty/n_tty.c
>> index 2e900a9..b09f326 100644
>> --- a/drivers/tty/n_tty.c
>> +++ b/drivers/tty/n_tty.c
>> @@ -321,7 +321,9 @@ static void n_tty_check_unthrottle(struct tty_struct *tty)
>>
>> static inline void put_tty_queue(unsigned char c, struct n_tty_data *ldata)
>> {
>> - *read_buf_addr(ldata, ldata->read_head++) = c;
>> + *read_buf_addr(ldata, ldata->read_head) = c;
>> + /* increment read_head _after_ placing the character in the buffer */
>> + ldata->read_head++;
>> }
>
> Is that comment really necessary?
No, I am pretty sure that removing the comment would not break the code ;-)
I just thought it would be good to have some kind of reminder here.
Otherwise someone may think: Hey, it would be a good idea to do the
increment right in the first line. And submit a patch for it.
But I am also ok with removing the comment. So if you like me to post
a v3 without the comment, I'll be happy to do that.
Christian
Christian Riesch <[email protected]> writes:
> On Tue, Nov 11, 2014 at 2:04 PM, M?ns Rullg?rd <[email protected]> wrote:
>> Christian Riesch <[email protected]> writes:
> [...]>> diff --git a/drivers/tty/n_tty.c b/drivers/tty/n_tty.c
>>> index 2e900a9..b09f326 100644
>>> --- a/drivers/tty/n_tty.c
>>> +++ b/drivers/tty/n_tty.c
>>> @@ -321,7 +321,9 @@ static void n_tty_check_unthrottle(struct tty_struct *tty)
>>>
>>> static inline void put_tty_queue(unsigned char c, struct n_tty_data *ldata)
>>> {
>>> - *read_buf_addr(ldata, ldata->read_head++) = c;
>>> + *read_buf_addr(ldata, ldata->read_head) = c;
>>> + /* increment read_head _after_ placing the character in the buffer */
>>> + ldata->read_head++;
>>> }
>>
>> Is that comment really necessary?
>
> No, I am pretty sure that removing the comment would not break the code ;-)
>
> I just thought it would be good to have some kind of reminder here.
> Otherwise someone may think: Hey, it would be a good idea to do the
> increment right in the first line. And submit a patch for it.
The intent all along was to increment after the write. Nobody needs
reminding of that. The problem was a misunderstanding of when the
post-increment takes effect. As much as we'd like for everybody to have
a thorough knowledge of C, a random tty driver doesn't seem the place to
educate them.
--
M?ns Rullg?rd
[email protected]
On Wed, Nov 12, 2014 at 12:53 PM, Måns Rullgård <[email protected]> wrote:
> Christian Riesch <[email protected]> writes:
>
>> On Tue, Nov 11, 2014 at 2:04 PM, Måns Rullgård <[email protected]> wrote:
>>> Christian Riesch <[email protected]> writes:
>> [...]>> diff --git a/drivers/tty/n_tty.c b/drivers/tty/n_tty.c
>>>> index 2e900a9..b09f326 100644
>>>> --- a/drivers/tty/n_tty.c
>>>> +++ b/drivers/tty/n_tty.c
>>>> @@ -321,7 +321,9 @@ static void n_tty_check_unthrottle(struct tty_struct *tty)
>>>>
>>>> static inline void put_tty_queue(unsigned char c, struct n_tty_data *ldata)
>>>> {
>>>> - *read_buf_addr(ldata, ldata->read_head++) = c;
>>>> + *read_buf_addr(ldata, ldata->read_head) = c;
>>>> + /* increment read_head _after_ placing the character in the buffer */
>>>> + ldata->read_head++;
>>>> }
>>>
>>> Is that comment really necessary?
>>
>> No, I am pretty sure that removing the comment would not break the code ;-)
>>
>> I just thought it would be good to have some kind of reminder here.
>> Otherwise someone may think: Hey, it would be a good idea to do the
>> increment right in the first line. And submit a patch for it.
>
> The intent all along was to increment after the write. Nobody needs
> reminding of that. The problem was a misunderstanding of when the
> post-increment takes effect. As much as we'd like for everybody to have
> a thorough knowledge of C, a random tty driver doesn't seem the place to
> educate them.
Ok. I will send a new patch without the comment.
Thanks, Christian