lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID:
 <PSAPR06MB49526EEA505F752A2A1E1A97C9222@PSAPR06MB4952.apcprd06.prod.outlook.com>
Date: Tue, 5 Mar 2024 11:24:08 +0800
From: GuanBing Huang <albanhuang@...look.com>
To: Greg KH <gregkh@...uxfoundation.org>
Cc: jirislaby@...nel.org, linux-kernel@...r.kernel.org,
 linux-serial@...r.kernel.org, albanhuang@...cent.com
Subject: Re: [PATCH] serial: 8250_pnp: Support configurable reg shift property

在 2024/3/1 6:00, Greg KH 写道:

> On Thu, Feb 29, 2024 at 07:51:54PM +0800, GuanBing Huang wrote:
>> From: albanhuang <albanhuang@...cent.com>
>>
>> The 16550a serial port based on the ACPI table requires obtaining the
>> reg-shift attribute. In the ACPI scenario, If the reg-shift property
>> is not configured like in DTS, the 16550a serial driver cannot read or
>> write controller registers properly during initialization.
>>
>> Signed-off-by: albanhuang <albanhuang@...cent.com>
>> Signed-off-by: tombinfan <tombinfan@...cent.com>
>> Signed-off-by: dylanlhdu <dylanlhdu@...cent.com>
> "interesting" names, can you not just use your native encoding to make
> this easier?

> ->I'm sorry,this is my first time sending a patch.The names should be changed to the following. Do I need to resend a new patch?

Signed-off-by: Guanbing Huang <albanhuang@...cent.com>
Signed-off-by: Bing Fan <tombinfan@...cent.com>
Signed-off-by: Linheng Du <dylanlhdu@...cent.com>

>> ---
>>   drivers/tty/serial/8250/8250_pnp.c | 1 +
>>   1 file changed, 1 insertion(+)
>>
>> diff --git a/drivers/tty/serial/8250/8250_pnp.c b/drivers/tty/serial/8250/8250_pnp.c
>> index 1974bbadc975..25b4e41e9745 100644
>> --- a/drivers/tty/serial/8250/8250_pnp.c
>> +++ b/drivers/tty/serial/8250/8250_pnp.c
>> @@ -473,6 +473,7 @@ serial_pnp_probe(struct pnp_dev *dev, const struct pnp_device_id *dev_id)
>>   		uart.port.flags |= UPF_SHARE_IRQ;
>>   	uart.port.uartclk = 1843200;
>>   	device_property_read_u32(&dev->dev, "clock-frequency", &uart.port.uartclk);
>> +	device_property_read_u8(&dev->dev, "reg-shift", &uart.port.regshift);
> Is this property documented somewhere?  What happens if the property
> isn't there?
> ->
> 1、In the DTS scenario, certain chips use the reg-shift property, such as:
>
> arch/riscv/boot/dts/microchip/mpfs.dtsi:
>
> 294                 mmuart0: serial@...00000 {
> 295                         compatible = "ns16550a";
> 296                         reg = <0x0 0x20000000 0x0 0x400>;
> 297                         reg-io-width = <4>;
> 298                         reg-shift = <2>;
> 299                         interrupt-parent = <&plic>;
> 300                         interrupts = <90>;
> 301                         current-speed = <115200>;
> 302                         clocks = <&clkcfg CLK_MMUART0>;
> 303                         status = "disabled"; /* Reserved for the HSS */
> 304                 };
>
> drivers/tty/serial/8250/8250_of.c:
>
> of_platform_serial_probe->of_platform_serial_setup:
>
> 150         /* Check for registers offset within the devices address range */
> 151         if (of_property_read_u32(np, "reg-shift", &prop) == 0)
> 152                 port->regshift = prop;
> 2、In the ACPI scenario, 16550a serial port initialization code execution process:
> ->serial_pnp_probe
> ->serial8250_register_8250_port
> ->uart_add_one_port
> ->serial_ctrl_register_port
> ->serial_core_register_port
> ->serial_core_add_one_port
> ->uart_configure_port
> ->serial8250_config_port
> ->autoconfig:
>
> 1194                 scratch = serial_in(up, UART_IER);
> 1195                 serial_out(up, UART_IER, 0);
> 1196 #ifdef __i386__
> 1197                 outb(0xff, 0x080);
> 1198 #endif
> 1199                 /*
> 1200                  * Mask out IER[7:4] bits for test as some UARTs (e.g. TL
> 1201                  * 16C754B) allow only to modify them if an EFR bit is set.
> 1202                  */
> 1203                 scratch2 = serial_in(up, UART_IER) & UART_IER_ALL_INTR;
> 1204                 serial_out(up, UART_IER, UART_IER_ALL_INTR);
> 1205 #ifdef __i386__
> 1206                 outb(0, 0x080);
> 1207 #endif
> 1208                 scratch3 = serial_in(up, UART_IER) & UART_IER_ALL_INTR;
> 1209                 serial_out(up, UART_IER, scratch);
> 1210                 if (scratch2 != 0 || scratch3 != UART_IER_ALL_INTR) {
> 1211                         /*
> 1212                          * We failed; there's nothing here
> 1213                          */
> 1214                         spin_unlock_irqrestore(&port->lock, flags);
> 1215                         DEBUG_AUTOCONF("IER test failed (%02x, %02x) ",
> 1216                                        scratch2, scratch3);
> 1217                         goto out;
> 1218                 }
>
> static unsigned int mem_serial_in(struct uart_port *p, int offset)
> {
>          offset = offset << p->regshift;
>          return readb(p->membase + offset);
> }
>
> static void mem_serial_out(struct uart_port *p, int offset, int value)
> {
>          offset = offset << p->regshift;
>          writeb(value, p->membase + offset);
> }
> The kernel will execute the serial_pnp_probe function to initialize the 16550a serial port during startup.
> When executing the autoconfig function, the serial_in/serial_out function failed to read and write to the UART_IER register,
> causing an abnormal branch entry at line 1210 and an abnormal exit at line 1217, preventing the subsequent 16550a initialization
> process from being executed properly.
> mem_serial_in/mem_serial_out will perform a regshift offset when reading and writing registers,
> because serial_pnp_probe does not have a configured regshift, which is 0 (regshift needs to be configured as 2),
> resulting in invalid register reading and writing.
> Without the reg-shift attribute, print the kernel boot log and identify the serial port as "unknown":
>
> [    1.458722] riscv-plic riscv-plic.0: mapped 64 interrupts with 1 handlers for 1 contexts.
> [    5.342472] Serial: 8250/16550 driver, 4 ports, IRQ sharing disabled
> [    6.038702] 00:02: ttyS0 at MMIO 0x310b0000 (irq = 12, base_baud = 3125000) is a unknown
> Configure the value of the reg-shift attribute to 2, print the kernel boot log, and correctly recognize 16550a:
>
> [    1.459948] riscv-plic riscv-plic.0: mapped 64 interrupts with 1 handlers for 1 contexts.
> [    5.317550] Serial: 8250/16550 driver, 4 ports, IRQ sharing disabled
> [    5.864828] 00:02: ttyS0 at MMIO 0x310b0000 (irq = 12, base_baud = 3125000) is a 16550A
> thanks,
> Guanbing Huang
> thanks,
>
> greg k-h

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ