linux-kernel - Re: serial: earlycon: Extend earlycon command line option to support 64-bit addresses

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]

Date:   Sat, 10 Sep 2016 17:45:54 -0700
From:   Guenter Roeck <linux@...ck-us.net>
To:     Alexander Sverdlin <alexander.sverdlin@...ia.com>
Cc:     linux-kernel@...r.kernel.org, linux-serial@...r.kernel.org,
        Jiri Slaby <jslaby@...e.com>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Subject: Re: serial: earlycon: Extend earlycon command line option to support
 64-bit addresses

Hi,

On Fri, Sep 02, 2016 at 01:20:21PM +0200, Alexander Sverdlin wrote:
> earlycon implementation used "unsigned long" internally, but there are systems
> (ARM with LPAE) where sizeof(unsigned long) == 4 and uart is mapped beyond 4GiB
> address range.
> 
> Switch to resource_size_t internally and replace obsoleted simple_strtoul() with
> kstrtoull().
> 

This patch breaks a qemu emulation using
	console=uart,mmio,0x90000000,115200
as command line argument. Problem appears to be the replacement of
simple_strtoul() with kstrtoull(). It looks like simple_strtoul() parsed
the address, but kstrtoull() doesn't. Reverting the kstrtoull() back to
simple_strtoul() fixes the problem for me.

The command line in question is found in arch/openrisc/boot/dts/or1ksim.dts.
There are similar command lines in other dts files, which makes me suspect
that this construct is also used by out-of-kernel command lines. Besides, the
code looks for ',' after the address, which suggests that the above should be
a valid argument.

Is this change to kstrtoull really necessary ? If yes, can you make it work
with existing command lines ?

Thanks,
Guenter

> Signed-off-by: Alexander Sverdlin <alexander.sverdlin@...ia.com>
> ---
>  drivers/tty/serial/8250/8250_core.c |  2 +-
>  drivers/tty/serial/earlycon.c       |  7 +++----
>  drivers/tty/serial/serial_core.c    | 12 +++++++++---
>  include/linux/serial_core.h         |  2 +-
>  4 files changed, 14 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/tty/serial/8250/8250_core.c b/drivers/tty/serial/8250/8250_core.c
> index 13ad5c3..f64d6cd 100644
> --- a/drivers/tty/serial/8250/8250_core.c
> +++ b/drivers/tty/serial/8250/8250_core.c
> @@ -639,7 +639,7 @@ static int univ8250_console_match(struct console *co, char *name, int idx,
>  {
>  	char match[] = "uart";	/* 8250-specific earlycon name */
>  	unsigned char iotype;
> -	unsigned long addr;
> +	resource_size_t addr;
>  	int i;
>  
>  	if (strncmp(name, match, 4) != 0)
> diff --git a/drivers/tty/serial/earlycon.c b/drivers/tty/serial/earlycon.c
> index 067783f..3940280 100644
> --- a/drivers/tty/serial/earlycon.c
> +++ b/drivers/tty/serial/earlycon.c
> @@ -38,7 +38,7 @@ static struct earlycon_device early_console_dev = {
>  	.con = &early_con,
>  };
>  
> -static void __iomem * __init earlycon_map(unsigned long paddr, size_t size)
> +static void __iomem * __init earlycon_map(resource_size_t paddr, size_t size)
>  {
>  	void __iomem *base;
>  #ifdef CONFIG_FIX_EARLYCON_MEM
> @@ -49,8 +49,7 @@ static void __iomem * __init earlycon_map(unsigned long paddr, size_t size)
>  	base = ioremap(paddr, size);
>  #endif
>  	if (!base)
> -		pr_err("%s: Couldn't map 0x%llx\n", __func__,
> -		       (unsigned long long)paddr);
> +		pr_err("%s: Couldn't map %pa\n", __func__, &paddr);
>  
>  	return base;
>  }
> @@ -92,7 +91,7 @@ static int __init parse_options(struct earlycon_device *device, char *options)
>  {
>  	struct uart_port *port = &device->port;
>  	int length;
> -	unsigned long addr;
> +	resource_size_t addr;
>  
>  	if (uart_parse_earlycon(options, &port->iotype, &addr, &options))
>  		return -EINVAL;
> diff --git a/drivers/tty/serial/serial_core.c b/drivers/tty/serial/serial_core.c
> index 9fc1533..89f5d6a 100644
> --- a/drivers/tty/serial/serial_core.c
> +++ b/drivers/tty/serial/serial_core.c
> @@ -1938,11 +1938,14 @@ uart_get_console(struct uart_port *ports, int nr, struct console *co)
>   *	   console=<name>,0x<addr>,<options>
>   *	is also accepted; the returned @iotype will be UPIO_MEM.
>   *
> - *	Returns 0 on success or -EINVAL on failure
> + *	Returns 0 on success, -EINVAL or -ERANGE on failure
>   */
> -int uart_parse_earlycon(char *p, unsigned char *iotype, unsigned long *addr,
> +int uart_parse_earlycon(char *p, unsigned char *iotype, resource_size_t *addr,
>  			char **options)
>  {
> +	int ret;
> +	unsigned long long tmp;
> +
>  	if (strncmp(p, "mmio,", 5) == 0) {
>  		*iotype = UPIO_MEM;
>  		p += 5;
> @@ -1968,7 +1971,10 @@ int uart_parse_earlycon(char *p, unsigned char *iotype, unsigned long *addr,
>  		return -EINVAL;
>  	}
>  
> -	*addr = simple_strtoul(p, NULL, 0);
> +	ret = kstrtoull(p, 0, &tmp);
> +	if (ret)
> +		return ret;
> +	*addr = tmp;
>  	p = strchr(p, ',');
>  	if (p)
>  		p++;
> diff --git a/include/linux/serial_core.h b/include/linux/serial_core.h
> index 2f44e20..cdba6f1 100644
> --- a/include/linux/serial_core.h
> +++ b/include/linux/serial_core.h
> @@ -374,7 +374,7 @@ extern int of_setup_earlycon(const struct earlycon_id *match,
>  
>  struct uart_port *uart_get_console(struct uart_port *ports, int nr,
>  				   struct console *c);
> -int uart_parse_earlycon(char *p, unsigned char *iotype, unsigned long *addr,
> +int uart_parse_earlycon(char *p, unsigned char *iotype, resource_size_t *addr,
>  			char **options);
>  void uart_parse_options(char *options, int *baud, int *parity, int *bits,
>  			int *flow);