[U-Boot] [PATCH 3/7] malloc_simple: Add Kconfig option for using only malloc_simple in the SPL
Hans de Goede
hdegoede at redhat.com
Tue Sep 22 11:38:46 CEST 2015
Hi,
On 22-09-15 06:00, Simon Glass wrote:
> Hi Hans,
>
> On 13 September 2015 at 09:42, Hans de Goede <hdegoede at redhat.com> wrote:
>> common/dlmalloc.c is quite big, both in .text and .data usage, therefor
>> on some boards the SPL is build to use only malloc_simple.c and not the
>> dlmalloc.c code. This is done in various include/configs/foo.h with the
>> following construct:
>>
>
> Was there a '#' at the start of this missing line?
Yep, I keep falling over git commit eating up any lines starting with
a # when quoting code snippets in commit messages...
I will fix this and send a v2.
>> This commit introduces a SPL_MALLOC_SIMPLE Kconfig bool which allows
>> selecting this functionality through Kconfig instead.
>>
>> Signed-off-by: Hans de Goede <hdegoede at redhat.com>
>> ---
>> Kconfig | 10 ++++++++++
>> common/malloc_simple.c | 3 ++-
>> include/_exports.h | 3 ++-
>> include/exports.h | 3 ++-
>> include/malloc.h | 3 ++-
>> 5 files changed, 18 insertions(+), 4 deletions(-)
>>
>> diff --git a/Kconfig b/Kconfig
>> index 05a34f7..0ae4fab 100644
>> --- a/Kconfig
>> +++ b/Kconfig
>> @@ -114,6 +114,16 @@ config SPL
>> help
>> If you want to build SPL as well as the normal image, say Y.
>>
>> +config SPL_MALLOC_SIMPLE
>
> If you made it SPL_SYS_MALLOC_SIMPLE...
>
>> + bool
>> + depends on SPL
>> + prompt "Only use malloc_simple functions in the spl"
>> + help
>> + Say Y here to only use the *_simple malloc functions from
>> + malloc_simple.c, rather then using the versions from dlmalloc.c
>> + this will make the SPL binary smaller at the cost of more heap
>> + usage as the *_simple malloc functions do not re-use free-ed mem.
>> +
>> config SPL_STACK_R
>> depends on SPL
>> bool "Enable SDRAM location for SPL stack"
>> diff --git a/common/malloc_simple.c b/common/malloc_simple.c
>> index c745863..e9c1eaa 100644
>> --- a/common/malloc_simple.c
>> +++ b/common/malloc_simple.c
>> @@ -40,7 +40,8 @@ void *memalign_simple(size_t align, size_t bytes)
>> return ptr;
>> }
>>
>> -#ifdef CONFIG_SYS_MALLOC_SIMPLE
>> +#if (defined CONFIG_SPL_BUILD && defined CONFIG_SPL_MALLOC_SIMPLE) || \
>> + (defined CONFIG_SYS_MALLOC_SIMPLE)
>
> then I think this could become:
>
> #if CONFIG_IS_ENABLED(SYS_MALLOC_SIMPLE)
I think you're right, good one. I'll give this a spin and
if it works include the change in V2.
>> void *calloc(size_t nmemb, size_t elem_size)
>> {
>> size_t size = nmemb * elem_size;
>> diff --git a/include/_exports.h b/include/_exports.h
>> index 74a882a..f811c5d 100644
>> --- a/include/_exports.h
>> +++ b/include/_exports.h
>> @@ -23,7 +23,8 @@
>> EXPORT_FUNC(dummy, void, free_hdlr, void)
>> #endif
>> EXPORT_FUNC(malloc, void *, malloc, size_t)
>> -#ifndef CONFIG_SYS_MALLOC_SIMPLE
>> +#if !(defined CONFIG_SPL_BUILD && defined CONFIG_SPL_MALLOC_SIMPLE) && \
>> + !(defined CONFIG_SYS_MALLOC_SIMPLE)
>> EXPORT_FUNC(free, void, free, void *)
>> #endif
>> EXPORT_FUNC(udelay, void, udelay, unsigned long)
>> diff --git a/include/exports.h b/include/exports.h
>> index a3e0469..8171b31 100644
>> --- a/include/exports.h
>> +++ b/include/exports.h
>> @@ -19,7 +19,8 @@ int printf(const char* fmt, ...);
>> void install_hdlr(int, interrupt_handler_t, void*);
>> void free_hdlr(int);
>> void *malloc(size_t);
>> -#ifndef CONFIG_SYS_MALLOC_SIMPLE
>> +#if !(defined CONFIG_SPL_BUILD && defined CONFIG_SPL_MALLOC_SIMPLE) && \
>> + !(defined CONFIG_SYS_MALLOC_SIMPLE)
>> void free(void*);
>> #endif
>> void __udelay(unsigned long);
>> diff --git a/include/malloc.h b/include/malloc.h
>> index f4da9e6..e5592fc 100644
>> --- a/include/malloc.h
>> +++ b/include/malloc.h
>> @@ -872,7 +872,8 @@ extern Void_t* sbrk();
>>
>> #else
>>
>> -#ifdef CONFIG_SYS_MALLOC_SIMPLE
>> +#if (defined CONFIG_SPL_BUILD && defined CONFIG_SPL_MALLOC_SIMPLE) || \
>> + (defined CONFIG_SYS_MALLOC_SIMPLE)
>> #define malloc malloc_simple
>> #define realloc realloc_simple
>> #define memalign memalign_simple
>> --
>> 2.4.3
>>
>
> Regards,
> Simon
>
Regards,
Hans
More information about the U-Boot
mailing list