| 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
| ||
|
Message-ID: <c51e770c-2c97-4050-ae79-7bbc9b64c4e3@yoseli.org>
Date: Mon, 9 Dec 2024 13:57:15 +0100
From: Jean-Michel Hautbois <jeanmichel.hautbois@...eli.org>
To: Greg Ungerer <gerg@...ux-m68k.org>,
Geert Uytterhoeven <geert@...ux-m68k.org>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Jiri Slaby <jirislaby@...nel.org>
Cc: linux-m68k@...ts.linux-m68k.org, linux-kernel@...r.kernel.org,
linux-serial@...r.kernel.org
Subject: Re: [PATCH] m68k: coldfire: Support resources for UART
Hi Greg,
On 12/9/24 13:30, Greg Ungerer wrote:
> Hi JM,
>
> On 5/12/24 23:06, Jean-Michel Hautbois wrote:
>> On 05/12/2024 14:01, Greg Ungerer wrote:
>>> On 4/12/24 21:32, Jean-Michel Hautbois wrote:
>>>> On 04/12/2024 12:15, Greg Ungerer wrote:
>>>>> On 4/12/24 20:58, Jean-Michel Hautbois wrote:
>>>>>> On 04/12/2024 11:54, Greg Ungerer wrote:
>>>>>>> On 2/12/24 20:34, Jean-Michel Hautbois wrote:
>>>>>>>> In order to use the eDMA channels for UART, the
>>>>>>>> mcf_platform_uart needs
>>>>>>>> to be changed. Instead of adding another custom member for the
>>>>>>>> structure, use a resource tree in a platform_device per UART. It
>>>>>>>> then
>>>>>>>> makes it possible to have a device named like "mcfuart.N" with N
>>>>>>>> the
>>>>>>>> UART number.
>>>>>>>>
>>>>>>>> Later, adding the dma channel in the mcf tty driver will also be
>>>>>>>> more
>>>>>>>> straightfoward.
>>>>>>>>
>>>>>>>> Signed-off-by: Jean-Michel Hautbois
>>>>>>>> <jeanmichel.hautbois@...eli.org>
>>>>>>>> ---
>>>>>>>> arch/m68k/coldfire/device.c | 96 +++++++++++++
>>>>>>>> +-------------------------------
>>>>>>>> drivers/tty/serial/mcf.c | 69 +++++++++++++++++++-------------
>>>>>>>> 2 files changed, 70 insertions(+), 95 deletions(-)
>>>>>>>>
>>>>>>>> diff --git a/arch/m68k/coldfire/device.c b/arch/m68k/coldfire/
>>>>>>>> device.c
>>>>>>>> index
>>>>>>>> b6958ec2a220cf91a78a14fc7fa18749451412f7..fd7d0b0ce7eb2970cb8ffe33589fe8d7e88c268d 100644
>>>>>>>> --- a/arch/m68k/coldfire/device.c
>>>>>>>> +++ b/arch/m68k/coldfire/device.c
>>>>>>>> @@ -24,73 +24,35 @@
>>>>>>>> #include <linux/platform_data/dma-mcf-edma.h>
>>>>>>>> #include <linux/platform_data/mmc-esdhc-mcf.h>
>>>>>>>> -/*
>>>>>>>> - * All current ColdFire parts contain from 2, 3, 4 or 10 UARTS.
>>>>>>>> - */
>>>>>>>> -static struct mcf_platform_uart mcf_uart_platform_data[] = {
>>>>>>>> - {
>>>>>>>> - .mapbase = MCFUART_BASE0,
>>>>>>>> - .irq = MCF_IRQ_UART0,
>>>>>>>> - },
>>>>>>>> - {
>>>>>>>> - .mapbase = MCFUART_BASE1,
>>>>>>>> - .irq = MCF_IRQ_UART1,
>>>>>>>> - },
>>>>>>>> -#ifdef MCFUART_BASE2
>>>>>>>> - {
>>>>>>>> - .mapbase = MCFUART_BASE2,
>>>>>>>> - .irq = MCF_IRQ_UART2,
>>>>>>>> - },
>>>>>>>> -#endif
>>>>>>>> -#ifdef MCFUART_BASE3
>>>>>>>> - {
>>>>>>>> - .mapbase = MCFUART_BASE3,
>>>>>>>> - .irq = MCF_IRQ_UART3,
>>>>>>>> - },
>>>>>>>> -#endif
>>>>>>>> -#ifdef MCFUART_BASE4
>>>>>>>> - {
>>>>>>>> - .mapbase = MCFUART_BASE4,
>>>>>>>> - .irq = MCF_IRQ_UART4,
>>>>>>>> - },
>>>>>>>> -#endif
>>>>>>>> -#ifdef MCFUART_BASE5
>>>>>>>> - {
>>>>>>>> - .mapbase = MCFUART_BASE5,
>>>>>>>> - .irq = MCF_IRQ_UART5,
>>>>>>>> - },
>>>>>>>> -#endif
>>>>>>>> -#ifdef MCFUART_BASE6
>>>>>>>> - {
>>>>>>>> - .mapbase = MCFUART_BASE6,
>>>>>>>> - .irq = MCF_IRQ_UART6,
>>>>>>>> - },
>>>>>>>> -#endif
>>>>>>>> -#ifdef MCFUART_BASE7
>>>>>>>> - {
>>>>>>>> - .mapbase = MCFUART_BASE7,
>>>>>>>> - .irq = MCF_IRQ_UART7,
>>>>>>>> +static u64 mcf_uart_mask = DMA_BIT_MASK(32);
>>>>>>>> +
>>>>>>>> +static struct resource mcf_uart0_resource[] = {
>>>>>>>> + [0] = {
>>>>>>>> + .start = MCFUART_BASE0,
>>>>>>>> + .end = MCFUART_BASE0 + 0x3fff,
>>>>>>>> + .flags = IORESOURCE_MEM,
>>>>>>>> },
>>>>>>>> -#endif
>>>>>>>> -#ifdef MCFUART_BASE8
>>>>>>>> - {
>>>>>>>> - .mapbase = MCFUART_BASE8,
>>>>>>>> - .irq = MCF_IRQ_UART8,
>>>>>>>> + [1] = {
>>>>>>>> + .start = 2,
>>>>>>>> + .end = 3,
>>>>>>>> + .flags = IORESOURCE_DMA,
>>>>>>>> },
>>>>>>>> -#endif
>>>>>>>> -#ifdef MCFUART_BASE9
>>>>>>>> - {
>>>>>>>> - .mapbase = MCFUART_BASE9,
>>>>>>>> - .irq = MCF_IRQ_UART9,
>>>>>>>> + [2] = {
>>>>>>>> + .start = MCF_IRQ_UART0,
>>>>>>>> + .end = MCF_IRQ_UART0,
>>>>>>>> + .flags = IORESOURCE_IRQ,
>>>>>>>> },
>>>>>>>> -#endif
>>>>>>>> - { },
>>>>>>>> };
>>>>>>>> -static struct platform_device mcf_uart = {
>>>>>>>> +static struct platform_device mcf_uart0 = {
>>>>>>>> .name = "mcfuart",
>>>>>>>> .id = 0,
>>>>>>>> - .dev.platform_data = mcf_uart_platform_data,
>>>>>>>> + .num_resources = ARRAY_SIZE(mcf_uart0_resource),
>>>>>>>> + .resource = mcf_uart0_resource,
>>>>>>>> + .dev = {
>>>>>>>> + .dma_mask = &mcf_uart_mask,
>>>>>>>> + .coherent_dma_mask = DMA_BIT_MASK(32),
>>>>>>>> + },
>>>>>>>> };
>>>>>>>> #ifdef MCFFEC_BASE0
>>>>>>>> @@ -485,12 +447,12 @@ static struct platform_device mcf_i2c5 = {
>>>>>>>> static const struct dma_slave_map mcf_edma_map[] = {
>>>>>>>> { "dreq0", "rx-tx", MCF_EDMA_FILTER_PARAM(0) },
>>>>>>>> { "dreq1", "rx-tx", MCF_EDMA_FILTER_PARAM(1) },
>>>>>>>> - { "uart.0", "rx", MCF_EDMA_FILTER_PARAM(2) },
>>>>>>>> - { "uart.0", "tx", MCF_EDMA_FILTER_PARAM(3) },
>>>>>>>> - { "uart.1", "rx", MCF_EDMA_FILTER_PARAM(4) },
>>>>>>>> - { "uart.1", "tx", MCF_EDMA_FILTER_PARAM(5) },
>>>>>>>> - { "uart.2", "rx", MCF_EDMA_FILTER_PARAM(6) },
>>>>>>>> - { "uart.2", "tx", MCF_EDMA_FILTER_PARAM(7) },
>>>>>>>> + { "mcfuart.0", "rx", MCF_EDMA_FILTER_PARAM(2) },
>>>>>>>> + { "mcfuart.0", "tx", MCF_EDMA_FILTER_PARAM(3) },
>>>>>>>> + { "mcfuart.1", "rx", MCF_EDMA_FILTER_PARAM(4) },
>>>>>>>> + { "mcfuart.1", "tx", MCF_EDMA_FILTER_PARAM(5) },
>>>>>>>> + { "mcfuart.2", "rx", MCF_EDMA_FILTER_PARAM(6) },
>>>>>>>> + { "mcfuart.2", "tx", MCF_EDMA_FILTER_PARAM(7) },
>>>>>>>> { "timer0", "rx-tx", MCF_EDMA_FILTER_PARAM(8) },
>>>>>>>> { "timer1", "rx-tx", MCF_EDMA_FILTER_PARAM(9) },
>>>>>>>> { "timer2", "rx-tx", MCF_EDMA_FILTER_PARAM(10) },
>>>>>>>> @@ -623,7 +585,7 @@ static struct platform_device mcf_flexcan0 = {
>>>>>>>> #endif /* MCFFLEXCAN_SIZE */
>>>>>>>> static struct platform_device *mcf_devices[] __initdata = {
>>>>>>>> - &mcf_uart,
>>>>>>>> + &mcf_uart0,
>>>>>>>> #ifdef MCFFEC_BASE0
>>>>>>>> &mcf_fec0,
>>>>>>>> #endif
>>>>>>>> diff --git a/drivers/tty/serial/mcf.c b/drivers/tty/serial/mcf.c
>>>>>>>> index
>>>>>>>> 93e7dda4d39acd23daf8c0d4c29ac8d666f263c5..07b8decfdb6005f0265dd130765e45c3fd1715eb 100644
>>>>>>>> --- a/drivers/tty/serial/mcf.c
>>>>>>>> +++ b/drivers/tty/serial/mcf.c
>>>>>>>> @@ -570,31 +570,46 @@ static struct uart_driver mcf_driver = {
>>>>>>>> static int mcf_probe(struct platform_device *pdev)
>>>>>>>> {
>>>>>>>> - struct mcf_platform_uart *platp = dev_get_platdata(&pdev-
>>>>>>>> >dev);
>>>>>>>> struct uart_port *port;
>>>>>>>> - int i;
>>>>>>>> -
>>>>>>>> - for (i = 0; ((i < MCF_MAXPORTS) && (platp[i].mapbase)); i++) {
>>>>>>>> - port = &mcf_ports[i].port;
>>>>>>>> -
>>>>>>>> - port->line = i;
>>>>>>>> - port->type = PORT_MCF;
>>>>>>>> - port->mapbase = platp[i].mapbase;
>>>>>>>> - port->membase = (platp[i].membase) ? platp[i].membase :
>>>>>>>> - (unsigned char __iomem *) platp[i].mapbase;
>>>>>>>> - port->dev = &pdev->dev;
>>>>>>>> - port->iotype = SERIAL_IO_MEM;
>>>>>>>> - port->irq = platp[i].irq;
>>>>>>>> - port->uartclk = MCF_BUSCLK;
>>>>>>>> - port->ops = &mcf_uart_ops;
>>>>>>>> - port->flags = UPF_BOOT_AUTOCONF;
>>>>>>>> - port->rs485_config = mcf_config_rs485;
>>>>>>>> - port->rs485_supported = mcf_rs485_supported;
>>>>>>>> - port->has_sysrq = IS_ENABLED(CONFIG_SERIAL_MCF_CONSOLE);
>>>>>>>> -
>>>>>>>> - uart_add_one_port(&mcf_driver, port);
>>>>>>>> + struct mcf_uart *pp;
>>>>>>>> + struct resource *res;
>>>>>>>> + void __iomem *base;
>>>>>>>> + int id = pdev->id;
>>>>>>>> +
>>>>>>>> + if (id == -1 || id >= MCF_MAXPORTS) {
>>>>>>>> + dev_err(&pdev->dev, "uart%d out of range\n",
>>>>>>>> + id);
>>>>>>>> + return -EINVAL;
>>>>>>>> }
>>>>>>>> + port = &mcf_ports[id].port;
>>>>>>>> + port->line = id;
>>>>>>>> +
>>>>>>>> + base = devm_platform_get_and_ioremap_resource(pdev, 0, &res);
>>>>>>>> + if (IS_ERR(base))
>>>>>>>> + return PTR_ERR(base);
>>>>>>>> +
>>>>>>>> + port->mapbase = res->start;
>>>>>>>> + port->membase = base;
>>>>>>>> +
>>>>>>>> + port->irq = platform_get_irq(pdev, 0);
>>>>>>>> + if (port->irq < 0)
>>>>>>>> + return port->irq;
>>>>>>>> +
>>>>>>>> + port->type = PORT_MCF;
>>>>>>>> + port->dev = &pdev->dev;
>>>>>>>> + port->iotype = SERIAL_IO_MEM;
>>>>>>>> + port->uartclk = MCF_BUSCLK;
>>>>>>>> + port->ops = &mcf_uart_ops;
>>>>>>>> + port->flags = UPF_BOOT_AUTOCONF;
>>>>>>>> + port->rs485_config = mcf_config_rs485;
>>>>>>>> + port->rs485_supported = mcf_rs485_supported;
>>>>>>>> + port->has_sysrq = IS_ENABLED(CONFIG_SERIAL_MCF_CONSOLE);
>>>>>>>> +
>>>>>>>> + pp = container_of(port, struct mcf_uart, port);
>>>>>>>> +
>>>>>>>> + uart_add_one_port(&mcf_driver, port);
>>>>>>>> +
>>>>>>>
>>>>>>> This breaks platforms with more than one UART - which is quite a
>>>>>>> few of
>>>>>>> the ColdFire platforms. Numerous boards bring and use more than
>>>>>>> one UART.
>>>>>>
>>>>>> I don't get why, as I have two uarts here, and each is detected
>>>>>> properly when declaring those in my platform ? I get that it
>>>>>> breaks existing detection (we are parsing all uarts even when only
>>>>>> one or two is used) but it does not prevent it to work ?
>>>>>
>>>>> Building and testing on an M5208EVB platform.
>>>>> With original un-modified code boot console shows:
>>>>>
>>>>> ...
>>>>> [ 0.110000] romfs: ROMFS MTD (C) 2007 Red Hat, Inc.
>>>>> [ 0.110000] ColdFire internal UART serial driver
>>>>> [ 0.110000] mcfuart.0: ttyS0 at MMIO 0xfc060000 (irq = 90,
>>>>> base_baud = 5208333) is a ColdFire UART
>>>>> [ 0.120000] printk: legacy console [ttyS0] enabled
>>>>> [ 0.120000] mcfuart.0: ttyS1 at MMIO 0xfc064000 (irq = 91,
>>>>> base_baud = 5208333) is a ColdFire UART
>>>>> [ 0.120000] mcfuart.0: ttyS2 at MMIO 0xfc068000 (irq = 92,
>>>>> base_baud = 5208333) is a ColdFire UART
>>>>> [ 0.130000] brd: module loaded
>>>>> ...
>>>>>
>>>>>
>>>>> But with this change applied only the first port is probed:
>>>>>
>>>>> ...
>>>>> [ 0.120000] romfs: ROMFS MTD (C) 2007 Red Hat, Inc.
>>>>> [ 0.120000] ColdFire internal UART serial driver
>>>>> [ 0.130000] mcfuart.0: ttyS0 at MMIO 0xfc060000 (irq = 90,
>>>>> base_baud = 5208333) is a ColdFire UART
>>>>> [ 0.130000] printk: legacy console [ttyS0] enabled
>>>>> [ 0.130000] brd: module loaded
>>>>> ...
>>>>
>>>> OK, I see what you mean. Let me try to explain why I did it :-).
>>>>
>>>> The idea is to avoid probing a UART device which may exist as such
>>>> on the core, but not be used as UART at all (on my board, for
>>>> instance, I have uart2 and uart6, I don't need any other UART to be
>>>> probed).
>>>>
>>>> So, based on what I think is the dts philosophy, you declare the
>>>> devices you really need to probe ?
>>>
>>> You can do this too, with the old style platform setups.
>>>
>>> What you want is to have a separate board file just for your board.
>>> There is a few examples already in arch/m68k/coldfire/ like amcore.c,
>>> firebee.c, nettel.c and stmark2.c. None currently specifically extract
>>> out UARTS - no one really seemed to have a need for that in the past.
>>> Most ColdFire parts have 2 or 3 UARTS, the 5441x family is an out-lier
>>> here with 10.
>>>
>>> Anyway, the device.c entries are really just a catch-all for the most
>>> common devices and their most commonly used configurations.
>>
>> Thanks for answering !
>>
>> I know I can have a dedicated file for my board (which i have tbh) but
>> device.c is always built when you select CONFIG_COLDFIRE so, I would
>> end up with 10 UARTs probed anyways ?
>>
>> Is there no way for this patch to find a path ? I mean, I can keep the
>> existing behavior, and have everything probed in device.c if the BASE
>> address is declared. But I don't want my board to have all 10 UARTs
>> and I don't want to locally patch the Makefile to remove the device.c
>> from the built-in ?
>
> Here is one example way to do it.
> Compile tested only - but I am sure you get the idea.
Thank you, it is clear. Yes, I get the idea :-).
I will submit a v2 once I have something (very) inspired ;-).
I just have to find a name for this new configuration, and test it.
JM
Powered by blists - more mailing lists