[PATCH v2 3/3] of: request reserved memory regions so other code can't

[Ahmad Fatoum]

On 15.06.22 07:08, Sascha Hauer wrote:
> On Thu, Jun 09, 2022 at 01:18:10PM +0200, Ahmad Fatoum wrote:
>> Add a new of_reserved_mem_walk that can be used to request
>> reserved memory regions. This avoids e.g. bootm trying to
>> place the kernel into a reserved region.
>>
>> Signed-off-by: Rouven Czerwinski <r.czerwinski at pengutronix.de>
>> Signed-off-by: Ahmad Fatoum <a.fatoum at pengutronix.de>
>> ---
>> v1 -> v2:
>>   - don't use of_reserve_entry accounting and instead
>>     directly request memory (Sascha)
>> ---
>>  common/memory.c           | 21 ++++++++++--
>>  drivers/of/Makefile       |  1 +
>>  drivers/of/reserved-mem.c | 71 +++++++++++++++++++++++++++++++++++++++
>>  include/of.h              |  9 +++++
>>  4 files changed, 99 insertions(+), 3 deletions(-)
>>  create mode 100644 drivers/of/reserved-mem.c
>>
>> diff --git a/common/memory.c b/common/memory.c
>> index 95995bb6e310..fd782c7f24f6 100644
>> --- a/common/memory.c
>> +++ b/common/memory.c
>> @@ -3,6 +3,8 @@
>>   * Copyright (c) 2011 Sascha Hauer <s.hauer at pengutronix.de>, Pengutronix
>>   */
>>  
>> +#define pr_fmt(fmt) "memory: " fmt
>> +
>>  #include <common.h>
>>  #include <memory.h>
>>  #include <of.h>
>> @@ -12,6 +14,7 @@
>>  #include <asm-generic/memory_layout.h>
>>  #include <asm/sections.h>
>>  #include <malloc.h>
>> +#include <of.h>
>>  
>>  /*
>>   * Begin and End of memory area for malloc(), and current "brk"
>> @@ -53,9 +56,20 @@ void mem_malloc_init(void *start, void *end)
>>  	mem_malloc_initialized = 1;
>>  }
>>  
>> -#if !defined __SANDBOX__
>> +static int request_reservation(const struct resource *res)
>> +{
>> +	if (!(res->flags & IORESOURCE_EXCLUSIVE))
>> +		return 0;
>> +
>> +	pr_debug("region %s %pa-%pa\n", res->name, &res->start, &res->end);
>> +
>> +	request_sdram_region(res->name, res->start, resource_size(res));
>> +	return 0;
>> +}
>> +
>>  static int mem_malloc_resource(void)
>>  {
>> +#if !defined __SANDBOX__
>>  	/*
>>  	 * Normally it's a bug when one of these fails,
>>  	 * but we have some setups where some of these
>> @@ -77,13 +91,14 @@ static int mem_malloc_resource(void)
>>  			(unsigned long)&__bss_start,
>>  			(unsigned long)&__bss_stop -
>>  			(unsigned long)&__bss_start);
>> +#endif
>>  #ifdef STACK_BASE
>>  	request_sdram_region("stack", STACK_BASE, STACK_SIZE);
>>  #endif
>> -	return 0;
>> +
>> +	return of_reserved_mem_walk(request_reservation);
>>  }
>>  coredevice_initcall(mem_malloc_resource);
>> -#endif
>>  
>>  static void *sbrk_no_zero(ptrdiff_t increment)
>>  {
>> diff --git a/drivers/of/Makefile b/drivers/of/Makefile
>> index ca8da71cb4f0..99b610cba85e 100644
>> --- a/drivers/of/Makefile
>> +++ b/drivers/of/Makefile
>> @@ -5,6 +5,7 @@ obj-$(CONFIG_OF_GPIO) += of_gpio.o
>>  obj-$(CONFIG_OF_PCI) += of_pci.o
>>  obj-y += partition.o
>>  obj-y += of_net.o
>> +obj-$(CONFIG_OFDEVICE) += reserved-mem.o
> 
> Compilation of this file depends on CONFIG_OFDEVICE, but the static
> inline wrapper or not decision depends on CONFIG_OFTREE. This breaks a
> number of defconfigs.
> I took the easiest way out for now and changed to obj-y above to fix
> this.

I am surprised that works. With !CONFIG_OFTREE, <of.h> included in
reserved-mem.c would provide a static inline helper that should
clash with of_reserved_mem_walk defined with external linkage
in that file. Can you change this into obj-$(CONFIG_OFTREE)?

Thanks and sorry for the hassle,
Ahmad

> 
> Sascha
> 


-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |