``max_count`` negative values are not used. Since ``port->fifosize``
is an unsigned int, make ``max_count`` the same.
Signed-off-by: Tudor Ambarus <[email protected]>
---
drivers/tty/serial/samsung_tty.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/tty/serial/samsung_tty.c b/drivers/tty/serial/samsung_tty.c
index 90c49197efc7..dbbe6b8e3ceb 100644
--- a/drivers/tty/serial/samsung_tty.c
+++ b/drivers/tty/serial/samsung_tty.c
@@ -760,8 +760,8 @@ static irqreturn_t s3c24xx_serial_rx_chars_dma(void *dev_id)
static void s3c24xx_serial_rx_drain_fifo(struct s3c24xx_uart_port *ourport)
{
struct uart_port *port = &ourport->port;
+ unsigned int max_count = port->fifosize;
unsigned int fifocnt = 0;
- int max_count = port->fifosize;
u32 ufcon, ufstat, uerstat;
u8 ch, flag;
--
2.43.0.472.g3155946c3a-goog
On Wed, Jan 10, 2024 at 4:23 AM Tudor Ambarus <tudor.ambarus@linaroorg> wrote:
>
> ``max_count`` negative values are not used. Since ``port->fifosize``
> is an unsigned int, make ``max_count`` the same.
>
> Signed-off-by: Tudor Ambarus <[email protected]>
> ---
> drivers/tty/serial/samsung_tty.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/tty/serial/samsung_tty.c b/drivers/tty/serial/samsung_tty.c
> index 90c49197efc7..dbbe6b8e3ceb 100644
> --- a/drivers/tty/serial/samsung_tty.c
> +++ b/drivers/tty/serial/samsung_tty.c
> @@ -760,8 +760,8 @@ static irqreturn_t s3c24xx_serial_rx_chars_dma(void *dev_id)
> static void s3c24xx_serial_rx_drain_fifo(struct s3c24xx_uart_port *ourport)
> {
> struct uart_port *port = &ourport->port;
> + unsigned int max_count = port->fifosize;
What if port->fifosize is 0? Then this code below:
while (max_count-- > 0) {
would cause int overflow, if max_count is unsigned?
> unsigned int fifocnt = 0;
> - int max_count = port->fifosize;
> u32 ufcon, ufstat, uerstat;
> u8 ch, flag;
>
> --
> 2.43.0.472.g3155946c3a-goog
>
>
On 1/16/24 18:21, Sam Protsenko wrote:
> On Wed, Jan 10, 2024 at 4:23 AM Tudor Ambarus <[email protected]> wrote:
>>
>> ``max_count`` negative values are not used. Since ``port->fifosize``
>> is an unsigned int, make ``max_count`` the same.
>>
>> Signed-off-by: Tudor Ambarus <[email protected]>
>> ---
>> drivers/tty/serial/samsung_tty.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/tty/serial/samsung_tty.c b/drivers/tty/serial/samsung_tty.c
>> index 90c49197efc7..dbbe6b8e3ceb 100644
>> --- a/drivers/tty/serial/samsung_tty.c
>> +++ b/drivers/tty/serial/samsung_tty.c
>> @@ -760,8 +760,8 @@ static irqreturn_t s3c24xx_serial_rx_chars_dma(void *dev_id)
>> static void s3c24xx_serial_rx_drain_fifo(struct s3c24xx_uart_port *ourport)
>> {
>> struct uart_port *port = &ourport->port;
>> + unsigned int max_count = port->fifosize;
>
> What if port->fifosize is 0? Then this code below:
>
> while (max_count-- > 0) {
>
> would cause int overflow, if max_count is unsigned?
>
good catch, Sam!
I'm thinking of amending this and add at the beginning of the method:
if (!max_count)
return tty_flip_buffer_push(&port->state->port);
Thanks!
ta
Hi,
On Wed, 2024-01-17 at 15:21 +0000, Tudor Ambarus wrote:
>
>
> On 1/16/24 18:21, Sam Protsenko wrote:
> > On Wed, Jan 10, 2024 at 4:23 AM Tudor Ambarus <[email protected]> wrote:
> > >
> > > ``max_count`` negative values are not used. Since ``port->fifosize``
> > > is an unsigned int, make ``max_count`` the same.
> > >
> > > Signed-off-by: Tudor Ambarus <[email protected]>
> > > ---
> > > drivers/tty/serial/samsung_tty.c | 2 +-
> > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/tty/serial/samsung_tty.c b/drivers/tty/serial/samsung_tty.c
> > > index 90c49197efc7..dbbe6b8e3ceb 100644
> > > --- a/drivers/tty/serial/samsung_tty.c
> > > +++ b/drivers/tty/serial/samsung_tty.c
> > > @@ -760,8 +760,8 @@ static irqreturn_t s3c24xx_serial_rx_chars_dma(void *dev_id)
> > > static void s3c24xx_serial_rx_drain_fifo(struct s3c24xx_uart_port *ourport)
> > > {
> > > struct uart_port *port = &ourport->port;
> > > + unsigned int max_count = port->fifosize;
> >
> > What if port->fifosize is 0? Then this code below:
> >
> > while (max_count-- > 0) {
> >
> > would cause int overflow, if max_count is unsigned?
> >
>
> good catch, Sam!
Does it matter, though? As this is a post-decrement, the test is done first, and the
decrement after. Therefore, it'll still bail out as expected.
> I'm thinking of amending this and add at the beginning of the method:
>
> if (!max_count)
> return tty_flip_buffer_push(&port->state->port);
This will not help with overflow. It'll still have wrapped around after completing the
while() (always, no matter what start-value max_count had)
Cheers,
Andre'
On 1/17/24 15:38, André Draszik wrote:
>>>> + unsigned int max_count = port->fifosize;
>>> What if port->fifosize is 0? Then this code below:
>>>
>>> while (max_count-- > 0) {
>>>
>>> would cause int overflow, if max_count is unsigned?
>>>
>> good catch, Sam!
> Does it matter, though? As this is a post-decrement, the test is done first, and the
> decrement after. Therefore, it'll still bail out as expected.
Indeed, it doesn't. This reminds me of stop replying to emails at the
end of the day :)
Cheers Andre'!
ta
On Wed, Jan 17, 2024 at 9:54 AM Tudor Ambarus <tudor.ambarus@linaroorg> wrote:
>
>
>
> On 1/17/24 15:38, André Draszik wrote:
> >>>> + unsigned int max_count = port->fifosize;
> >>> What if port->fifosize is 0? Then this code below:
> >>>
> >>> while (max_count-- > 0) {
> >>>
> >>> would cause int overflow, if max_count is unsigned?
> >>>
> >> good catch, Sam!
> > Does it matter, though? As this is a post-decrement, the test is done first, and the
> > decrement after. Therefore, it'll still bail out as expected.
>
> Indeed, it doesn't. This reminds me of stop replying to emails at the
> end of the day :)
>
And it reminds me to drink some coffee in the morning before doing any
reviews :) With above condition sorted, feel free to add:
Reviewed-by: Sam Protsenko <[email protected]>
> Cheers Andre'!
> ta
On Wed, Jan 17, 2024 at 9:38 AM André Draszik <[email protected]> wrote:
>
> Hi,
>
> On Wed, 2024-01-17 at 15:21 +0000, Tudor Ambarus wrote:
> >
> >
> > On 1/16/24 18:21, Sam Protsenko wrote:
> > > On Wed, Jan 10, 2024 at 4:23 AM Tudor Ambarus <[email protected]> wrote:
> > > >
> > > > ``max_count`` negative values are not used. Since ``port->fifosize``
> > > > is an unsigned int, make ``max_count`` the same.
> > > >
> > > > Signed-off-by: Tudor Ambarus <[email protected]>
> > > > ---
> > > > drivers/tty/serial/samsung_tty.c | 2 +-
> > > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > > >
> > > > diff --git a/drivers/tty/serial/samsung_tty.c b/drivers/tty/serial/samsung_tty.c
> > > > index 90c49197efc7..dbbe6b8e3ceb 100644
> > > > --- a/drivers/tty/serial/samsung_tty.c
> > > > +++ b/drivers/tty/serial/samsung_tty.c
> > > > @@ -760,8 +760,8 @@ static irqreturn_t s3c24xx_serial_rx_chars_dma(void *dev_id)
> > > > static void s3c24xx_serial_rx_drain_fifo(struct s3c24xx_uart_port *ourport)
> > > > {
> > > > struct uart_port *port = &ourport->port;
> > > > + unsigned int max_count = port->fifosize;
> > >
> > > What if port->fifosize is 0? Then this code below:
> > >
> > > while (max_count-- > 0) {
> > >
> > > would cause int overflow, if max_count is unsigned?
> > >
> >
> > good catch, Sam!
>
> Does it matter, though? As this is a post-decrement, the test is done first, and the
> decrement after. Therefore, it'll still bail out as expected.
>
Good catch on my good catch :)
> > I'm thinking of amending this and add at the beginning of the method:
> >
> > if (!max_count)
> > return tty_flip_buffer_push(&port->state->port);
>
> This will not help with overflow. It'll still have wrapped around after completing the
> while() (always, no matter what start-value max_count had)
>
> Cheers,
> Andre'
>