[PATCH] ARM: imx: Add iram allocator functions

Uwe Kleine-König u.kleine-koenig at pengutronix.de
Tue Oct 5 03:44:23 EDT 2010 

On Mon, Oct 04, 2010 at 07:03:27PM -0500, Dinh.Nguyen at freescale.com wrote:
> From: Dinh Nguyen <Dinh.Nguyen at freescale.com>
> 
> Add iram allocation functions using GENERIC_ALLOCATOR. The
> allocation size is 4KB multiples to guarantee alignment. The
> idea for these functions is for i.MX platforms to use them
> to dynamically allocate IRAM usage.
> 
> Applies on 2.6.36-rc6
> 
> Signed-off-by: Dinh Nguyen <Dinh.Nguyen at freescale.com>
> Reviewed-by: Amit Kucheria <amit.kucheria at canonical.com>
> ---
>  arch/arm/plat-mxc/Kconfig                   |   10 ++++
>  arch/arm/plat-mxc/Makefile                  |    1 +
>  arch/arm/plat-mxc/include/mach/iram_alloc.h |   35 +++++++++++++++
>  arch/arm/plat-mxc/iram.c                    |   62 +++++++++++++++++++++++++++
>  4 files changed, 108 insertions(+), 0 deletions(-)
>  create mode 100644 arch/arm/plat-mxc/include/mach/iram_alloc.h
>  create mode 100644 arch/arm/plat-mxc/iram.c
> 
> diff --git a/arch/arm/plat-mxc/Kconfig b/arch/arm/plat-mxc/Kconfig
> index 6785db4..5e4ff93 100644
> --- a/arch/arm/plat-mxc/Kconfig
> +++ b/arch/arm/plat-mxc/Kconfig
> @@ -57,6 +57,16 @@ source "arch/arm/mach-mx5/Kconfig"
>  
>  endmenu
>  
> +config IRAM_ALLOC
> +	bool "Enable IRAM allocator"
> +	default y
The iram allocator isn't useful taken alone, no?  So I suggest to make
it 
> +	select GENERIC_ALLOCATOR
> +	help
> +	  Select this to have access to generic IRAM allocation functions that
> +	  use the GENERIC_ALLOCATOR.  The allocation size is 4KB multiples to 
> +	  guarantee alignment. The idea for these functions is for i.MX platforms
> +	  to use them to dynamically allocate IRAM usage.
> +
>  config MXC_IRQ_PRIOR
>  	bool "Use IRQ priority"
>  	help
> diff --git a/arch/arm/plat-mxc/Makefile b/arch/arm/plat-mxc/Makefile
> index 78d405e..35cf07b 100644
> --- a/arch/arm/plat-mxc/Makefile
> +++ b/arch/arm/plat-mxc/Makefile
> @@ -10,6 +10,7 @@ obj-$(CONFIG_MXC_TZIC) += tzic.o
>  
>  obj-$(CONFIG_IMX_HAVE_IOMUX_V1) += iomux-v1.o
>  obj-$(CONFIG_ARCH_MXC_IOMUX_V3) += iomux-v3.o
> +obj-$(CONFIG_IRAM_ALLOC) += iram.o
>  obj-$(CONFIG_MXC_PWM)  += pwm.o
>  obj-$(CONFIG_USB_EHCI_MXC) += ehci.o
>  obj-$(CONFIG_MXC_ULPI) += ulpi.o
> diff --git a/arch/arm/plat-mxc/include/mach/iram_alloc.h b/arch/arm/plat-mxc/include/mach/iram_alloc.h
> new file mode 100644
> index 0000000..12881aa
> --- /dev/null
> +++ b/arch/arm/plat-mxc/include/mach/iram_alloc.h
I'd prefer include/mach/iram.h here

> @@ -0,0 +1,35 @@
> +/*
> + * Copyright (C) 2010 Freescale Semiconductor, Inc. All Rights Reserved.
> + *
> + * This program is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU General Public License
> + * as published by the Free Software Foundation; either version 2
> + * of the License, or (at your option) any later version.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program; if not, write to the Free Software
> + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
> + * MA 02110-1301, USA.
> + */
> +
> +#ifdef CONFIG_IRAM_ALLOC
> +int __init iram_init(unsigned long base, unsigned long size);
> +void *iram_alloc(unsigned int size, unsigned long *dma_addr);
> +void iram_free(unsigned long dma_addr, unsigned int size);
> +#else
> +static inline int __init iram_init(unsigned long base, unsigned long size)
> +{
> +	return -ENOMEM;
I suggest to add an #include to assert ENOMEM is #defined.

> +}
> +static inline void *iram_alloc(unsigned int size, unsigned long *dma_addr)
> +{
> +	return NULL;
> +}
> +static inline void iram_free(unsigned long base, unsigned long size) {}
> +#endif
> +
please remove the trailing newline.  I'd add some new lines around the
cpp lines, though.

> diff --git a/arch/arm/plat-mxc/iram.c b/arch/arm/plat-mxc/iram.c
> new file mode 100644
> index 0000000..3d2a391
> --- /dev/null
> +++ b/arch/arm/plat-mxc/iram.c
> @@ -0,0 +1,62 @@
> +/*
> + * Copyright (C) 2010 Freescale Semiconductor, Inc. All Rights Reserved.
> + *
> + * This program is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU General Public License
> + * as published by the Free Software Foundation; either version 2
> + * of the License, or (at your option) any later version.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program; if not, write to the Free Software
> + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
> + * MA 02110-1301, USA.
> + */
> +
> +#include <linux/kernel.h>
> +#include <linux/io.h>
> +#include <linux/module.h>
Why do you need linux/module.h?  Maybe linux/init.h is enough?

> +#include <linux/spinlock.h>
This can simply go, can't it?

> +#include <linux/genalloc.h>
> +
> +static unsigned long iram_phys_base;
> +static __iomem void *iram_virt_base;
I assume it doesn't make any difference for gcc, but using

	static void __iomem *iram_virt_base

seems to be the more usual form.

> +static struct gen_pool *iram_pool;
> +
> +#define iram_phys_to_virt(p) (iram_virt_base + ((p) - iram_phys_base))
Can you make this a static inline please?

> +
> +void *iram_alloc(unsigned int size, unsigned long *dma_addr)
> +{
> +	if (!iram_pool)
> +		return NULL;
> +
> +	*dma_addr = gen_pool_alloc(iram_pool, size);
> +	pr_debug("iram alloc - %dB at 0x%p\n", size, (void *)*dma_addr);
> +	return iram_phys_to_virt(*dma_addr);
if iram_virt_base is __iomem, the return value of iram_alloc should be
__iomem, too, no?

> +}
> +EXPORT_SYMBOL(iram_alloc);
> +
> +void iram_free(unsigned long addr, unsigned int size)
> +{
> +	if (!iram_pool)
> +		return;
> +
> +	gen_pool_free(iram_pool, addr, size);
> +}
> +EXPORT_SYMBOL(iram_free);
> +
> +int __init iram_init(unsigned long base, unsigned long size)
> +{
> +	iram_phys_base = base;
> +
> +	iram_pool = gen_pool_create(12, -1);
	if (!iram_pool)
		...

> +	gen_pool_add(iram_pool, base, size, -1);
> +	iram_virt_base = ioremap(iram_phys_base, size);
These two can fail, too.

> +	pr_info("i.MX IRAM pool: %ld KB at 0x%p\n", size / 1024, iram_virt_base);
> +	return 0;
> +}
iram_init isn't called yet.  I assume calls to it should be added to the
cpu init routines (later)?

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

More information about the linux-arm-kernel mailing list