String literals in __init functions

[Mason]

On 25/03/2015 19:01, Joe Perches wrote:

> On Wed, 2015-03-25 at 18:56 +0100, Mason wrote:
>
>> AFAIU, functions only used at system init are tagged __init to have
>> the linker store them in a separate .init.text section, so memory can
>> be reclaimed once initialization is complete. Is that correct?
>>
>> The corresponding tag for data is __initdata (section .init.data)
>>
>> I started wondering if the string literals used in an __init functions
>> were automatically marked __initdata.
>>
>> Looking at the objdump output, I see that the string literals are,
>> in fact, stored in the .rodata section. I suppose that .rodata is NOT
>> reclaimed after init?
>>
>> This way seems to work:
>>
>> static       char XyZa[] __initdata  = KERN_ALERT "foo";
>> static const char XyZb[] __initconst = KERN_ALERT "bar";
>> void __init XyZc(void) { printk(XyZa); printk(XyZb); }
>>
>> $ arm-linux-gnueabihf-objdump -xd arch/arm/mach-tangox/time.o | grep XyZ
>> 00000000 l     O .init.data	00000006 XyZa
>> 00000000 l     O .init.rodata	00000006 XyZb
>> 00000000 g     F .init.text	00000028 XyZc
>> 00000000 <XyZc>:
>>
>> $ arm-linux-gnueabihf-objdump -xd vmlinux | grep XyZ
>> c021e360 l     O .init.data	00000006 XyZa
>> c0220090 l     O .init.data	00000006 XyZb
>> c020d928 g     F .init.text	00000028 XyZc
>> c020d928 <XyZc>:
>>
>> c020d928 <XyZc>:
>> c020d928:       e1a0c00d        mov     ip, sp
>> c020d92c:       e92dd800        push    {fp, ip, lr, pc}
>> c020d930:       e24cb004        sub     fp, ip, #4
>> c020d934:       e30e0360        movw    r0, #58208      ; 0xe360
>> c020d938:       e34c0021        movt    r0, #49185      ; 0xc021
>> c020d93c:       ebfe00c9        bl      c018dc68 <printk>
>> c020d940:       e3000090        movw    r0, #144        ; 0x90
>> c020d944:       e34c0022        movt    r0, #49186      ; 0xc022
>> c020d948:       ebfe00c6        bl      c018dc68 <printk>
>> c020d94c:       e89da800        ldm     sp, {fp, sp, pc}
>>
>> Did I miss something in init.h?
>> Or should it be done like above to reclaim string literals?
>
> No, you didn't miss anything.
>
> One proposal:
>
> https://lkml.org/lkml/2014/8/21/255

Thanks for the link!

Here's the equivalent gmane link for my own reference:
http://thread.gmane.org/gmane.linux.kernel/1771969

Basically, if I understand correctly, Ingo NAKed the patch, saying
this should be done automatically by the toolchain. That would make
for an interesting side-project...

For the record, I wrote a trivial wrapper for my limited use-case.

#define printk_init(format, ...) do { \
   static char fmt[] __initdata = format; printk(fmt, ## __VA_ARGS__); \
} while(0)

(I dislike the "statement-in-expression" extension, because vim thinks
there's a syntax error, and flashes a bright red block.)

https://gcc.gnu.org/onlinedocs/cpp/Variadic-Macros.html
https://gcc.gnu.org/onlinedocs/gcc/Statement-Exprs.html

Regards.