[PATCH] serial: uartlite: Use dynamic allocation for major number

[PATCH] serial: uartlite: Use dynamic allocation for major number

[Manikanta Guntupalli]

Device number 204 has a range of minors on major number.
uart_register_driver is failing due to lack of minor numbers
when more number of uart ports used. So, use dynamic allocation
for major number to avoid minor number limitation on 204 major
number.

https://docs.kernel.org/arch/arm/sa1100/serial_uart.html

Signed-off-by: Manikanta Guntupalli <manikanta.guntupalli@amd.com>
---
 drivers/tty/serial/uartlite.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/tty/serial/uartlite.c b/drivers/tty/serial/uartlite.c
index 404c14acafa5..c80ed0373b44 100644
--- a/drivers/tty/serial/uartlite.c
+++ b/drivers/tty/serial/uartlite.c
@@ -24,8 +24,8 @@
 #include <linux/pm_runtime.h>
 
 #define ULITE_NAME		"ttyUL"
-#define ULITE_MAJOR		204
-#define ULITE_MINOR		187
+#define ULITE_MAJOR		0	/* use dynamic node allocation */
+#define ULITE_MINOR		0
 #define ULITE_NR_UARTS		CONFIG_SERIAL_UARTLITE_NR_UARTS
 
 /* ---------------------------------------------------------------------
-- 
2.25.1

Re: [PATCH] serial: uartlite: Use dynamic allocation for major number

[Greg KH]

On Thu, Nov 09, 2023 at 06:06:40PM +0530, Manikanta Guntupalli wrote:
> Device number 204 has a range of minors on major number.
> uart_register_driver is failing due to lack of minor numbers
> when more number of uart ports used.

So you need more than the 4 allocated to you?

> So, use dynamic allocation
> for major number to avoid minor number limitation on 204 major
> number.
> 
> https://docs.kernel.org/arch/arm/sa1100/serial_uart.html

What does this break by doing this?

Also, you forgot to update the documentation :(

And how was this tested?  What about older systems with static device
nodes, are you sure none are out there for this old hardware anymore?

thanks,

greg k-h

RE: [PATCH] serial: uartlite: Use dynamic allocation for major number

[Guntupalli, Manikanta]

Hi Greg,

> -----Original Message-----
> From: Greg KH <gregkh@linuxfoundation.org>
> Sent: Thursday, November 9, 2023 6:50 PM
> To: Guntupalli, Manikanta <manikanta.guntupalli@amd.com>
> Cc: git (AMD-Xilinx) <git@amd.com>; Simek, Michal
> <michal.simek@amd.com>; jacmet@sunsite.dk; jirislaby@kernel.org; linux-
> serial@vger.kernel.org; linux-kernel@vger.kernel.org; Pandey, Radhey Shyam
> <radhey.shyam.pandey@amd.com>; Goud, Srinivas
> <srinivas.goud@amd.com>; Datta, Shubhrajyoti
> <shubhrajyoti.datta@amd.com>; manion05gk@gmail.com
> Subject: Re: [PATCH] serial: uartlite: Use dynamic allocation for major number
> 
> On Thu, Nov 09, 2023 at 06:06:40PM +0530, Manikanta Guntupalli wrote:
> > Device number 204 has a range of minors on major number.
> > uart_register_driver is failing due to lack of minor numbers when more
> > number of uart ports used.
> 
> So you need more than the 4 allocated to you?
Yes, we have a customer who has 32 uartlite instances in his board.
> 
> > So, use dynamic allocation
> > for major number to avoid minor number limitation on 204 major number.
> >
> > https://docs.kernel.org/arch/arm/sa1100/serial_uart.html
> 
> What does this break by doing this?
uart_register_driver() is failing due to lack of minor numbers when the customer
has 32 uartlite instances in his board.
> 
> Also, you forgot to update the documentation :(
We will update the documentation.
> 
> And how was this tested?  
We tested on both ZCU106 AMD/Xilinx evaluation board with 32 uartlite instances with customer design.

>What about older systems with static device nodes,
> are you sure none are out there for this old hardware anymore?
Shall we use below approach to support both legacy hardware and hardware with more number of uartlite instances use case. Please suggest.
diff --git a/drivers/tty/serial/uartlite.c b/drivers/tty/serial/uartlite.c
index 404c14acafa5..517f1f34143d 100644
--- a/drivers/tty/serial/uartlite.c
+++ b/drivers/tty/serial/uartlite.c
@@ -24,8 +24,13 @@
 #include <linux/pm_runtime.h>
 
 #define ULITE_NAME             "ttyUL"
+#if (CONFIG_SERIAL_UARTLITE_NR_UARTS > 4)
+#define ULITE_MAJOR             0       /* use dynamic node allocation */
+#define ULITE_MINOR             0
+#else
 #define ULITE_MAJOR            204
 #define ULITE_MINOR            187
+#endif
 #define ULITE_NR_UARTS         CONFIG_SERIAL_UARTLITE_NR_UARTS


Thanks,
Manikanta.

Re: [PATCH] serial: uartlite: Use dynamic allocation for major number

[Greg KH]

On Fri, Nov 10, 2023 at 09:28:40AM +0000, Guntupalli, Manikanta wrote:
> Hi Greg,
> 
> > -----Original Message-----
> > From: Greg KH <gregkh@linuxfoundation.org>
> > Sent: Thursday, November 9, 2023 6:50 PM
> > To: Guntupalli, Manikanta <manikanta.guntupalli@amd.com>
> > Cc: git (AMD-Xilinx) <git@amd.com>; Simek, Michal
> > <michal.simek@amd.com>; jacmet@sunsite.dk; jirislaby@kernel.org; linux-
> > serial@vger.kernel.org; linux-kernel@vger.kernel.org; Pandey, Radhey Shyam
> > <radhey.shyam.pandey@amd.com>; Goud, Srinivas
> > <srinivas.goud@amd.com>; Datta, Shubhrajyoti
> > <shubhrajyoti.datta@amd.com>; manion05gk@gmail.com
> > Subject: Re: [PATCH] serial: uartlite: Use dynamic allocation for major number
> > 
> > On Thu, Nov 09, 2023 at 06:06:40PM +0530, Manikanta Guntupalli wrote:
> > > Device number 204 has a range of minors on major number.
> > > uart_register_driver is failing due to lack of minor numbers when more
> > > number of uart ports used.
> > 
> > So you need more than the 4 allocated to you?
> Yes, we have a customer who has 32 uartlite instances in his board.
> > 
> > > So, use dynamic allocation
> > > for major number to avoid minor number limitation on 204 major number.
> > >
> > > https://docs.kernel.org/arch/arm/sa1100/serial_uart.html
> > 
> > What does this break by doing this?
> uart_register_driver() is failing due to lack of minor numbers when the customer
> has 32 uartlite instances in his board.
> > 
> > Also, you forgot to update the documentation :(
> We will update the documentation.
> > 
> > And how was this tested?  
> We tested on both ZCU106 AMD/Xilinx evaluation board with 32 uartlite instances with customer design.
> 
> >What about older systems with static device nodes,
> > are you sure none are out there for this old hardware anymore?
> Shall we use below approach to support both legacy hardware and hardware with more number of uartlite instances use case. Please suggest.

Yes, that looks much better, also update the Kconfig entry for
CONFIG_SERIAL_UARTLITE_NR_UARTS as well so that people know the
major/minor will be dynamic and will not be the other entry if they ask
for over 4.

thanks,

greg k-h