startup()/shutdown() callbacks access SIFIVE_SERIAL_IE_OFFS.
The register is also accessed from write() callback.

If console were printing and startup()/shutdown() callback
gets called, its access to the register could be overwritten.

Add port->lock to startup()/shutdown() callbacks to make sure
their access to SIFIVE_SERIAL_IE_OFFS is synchronized against
write() callback.

Fixes: 45c054d0815b ("tty: serial: add driver for the SiFive UART")
Signed-off-by: Ryo Takakura <ryotkkr98@gmail.com>
Cc: stable@vger.kernel.org
---

This patch used be part of a series for converting sifive driver to
nbcon[0]. It's now sent seperatly as the rest of the series does not
need be applied to the stable branch.

Sincerely,
Ryo Takakura

[0] https://lore.kernel.org/all/20250405043833.397020-1-ryotkkr98@gmail.com/

---
 drivers/tty/serial/sifive.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/tty/serial/sifive.c b/drivers/tty/serial/sifive.c
index 5904a2d4c..054a8e630 100644
--- a/drivers/tty/serial/sifive.c
+++ b/drivers/tty/serial/sifive.c
@@ -563,8 +563,11 @@ static void sifive_serial_break_ctl(struct uart_port *port, int break_state)
 static int sifive_serial_startup(struct uart_port *port)
 {
 	struct sifive_serial_port *ssp = port_to_sifive_serial_port(port);
+	unsigned long flags;
 
+	uart_port_lock_irqsave(&ssp->port, &flags);
 	__ssp_enable_rxwm(ssp);
+	uart_port_unlock_irqrestore(&ssp->port, flags);
 
 	return 0;
 }
@@ -572,9 +575,12 @@ static int sifive_serial_startup(struct uart_port *port)
 static void sifive_serial_shutdown(struct uart_port *port)
 {
 	struct sifive_serial_port *ssp = port_to_sifive_serial_port(port);
+	unsigned long flags;
 
+	uart_port_lock_irqsave(&ssp->port, &flags);
 	__ssp_disable_rxwm(ssp);
 	__ssp_disable_txwm(ssp);
+	uart_port_unlock_irqrestore(&ssp->port, flags);
 }
 
 /**
-- 
2.34.1

On Sat, Apr 05, 2025 at 10:24:58PM +0900, Ryo Takakura wrote:
> startup()/shutdown() callbacks access SIFIVE_SERIAL_IE_OFFS.
> The register is also accessed from write() callback.
> 
> If console were printing and startup()/shutdown() callback
> gets called, its access to the register could be overwritten.
> 
> Add port->lock to startup()/shutdown() callbacks to make sure
> their access to SIFIVE_SERIAL_IE_OFFS is synchronized against
> write() callback.
> 
> Fixes: 45c054d0815b ("tty: serial: add driver for the SiFive UART")
> Signed-off-by: Ryo Takakura <ryotkkr98@gmail.com>
> Cc: stable@vger.kernel.org
> ---
> 
> This patch used be part of a series for converting sifive driver to
> nbcon[0]. It's now sent seperatly as the rest of the series does not
> need be applied to the stable branch.

That means this is a v2 patch, and you should also send the other patch
as a v2 as well, right?

thanks,

greg k-h

Hi Greg,

On Sat, 5 Apr 2025 14:40:33 +0100, Greg KH wrote:
>On Sat, Apr 05, 2025 at 10:24:58PM +0900, Ryo Takakura wrote:
>> startup()/shutdown() callbacks access SIFIVE_SERIAL_IE_OFFS.
>> The register is also accessed from write() callback.
>> 
>> If console were printing and startup()/shutdown() callback
>> gets called, its access to the register could be overwritten.
>> 
>> Add port->lock to startup()/shutdown() callbacks to make sure
>> their access to SIFIVE_SERIAL_IE_OFFS is synchronized against
>> write() callback.
>> 
>> Fixes: 45c054d0815b ("tty: serial: add driver for the SiFive UART")
>> Signed-off-by: Ryo Takakura <ryotkkr98@gmail.com>
>> Cc: stable@vger.kernel.org
>> ---
>> 
>> This patch used be part of a series for converting sifive driver to
>> nbcon[0]. It's now sent seperatly as the rest of the series does not
>> need be applied to the stable branch.
>
>That means this is a v2 patch, and you should also send the other patch
>as a v2 as well, right?

Oh yes. I wasn't sure about the versioning for this patch. Let me resend
this patch as v2.
Also for the other patch, as now its sent as a single standalone patch,
I'll resend it as v2.

>thanks,
>
>greg k-h