How to define a symbol with absolute address for AArch64?
Sebastian Huber
sebastian.huber@embedded-brains.de
Thu Sep 14 07:30:01 GMT 2023
On 13.09.23 12:02, Richard Earnshaw (lists) wrote:
> On 13/09/2023 07:04, Sebastian Huber wrote:
>> On 12.09.23 18:18, Richard Earnshaw wrote:
>>>
>>> On 12/09/2023 15:23, Sebastian Huber wrote:
>>>> On 12.09.23 15:46, Xi Ruoyao wrote:
>>>>> On Tue, 2023-09-12 at 14:58 +0200, Sebastian Huber wrote:
>>>>>> On 12.09.23 13:21, Richard Earnshaw wrote:
>>>>>>> On 12/09/2023 11:02, Sebastian Huber wrote:
>>>>>>>> Hello,
>>>>>>>>
>>>>>>>> I would like to define a global symbol with an absolute address in an
>>>>>>>> assembly/C source file for the AArch64 target. This works for all
>>>>>>>> other architectures I tried so far, but not for AArch64:
>>>>>>>>
>>>>>>>> extern char abs_symbol[];
>>>>>>>> extern char abs_symbol_2[];
>>>>>>>>
>>>>>>>> __asm__(
>>>>>>>> "\t.globl abs_symbol\n"
>>>>>>>> "\t.set abs_symbol, 0x123\n"
>>>>>>>> );
>>>>>>>>
>>>>>>>> unsigned long f_abs_symbol(void)
>>>>>>>> {
>>>>>>>> return (unsigned long)abs_symbol;
>>>>>>>> }
>>>>>>>>
>>>>>>>> unsigned long f_abs_symbol_2(void)
>>>>>>>> {
>>>>>>>> return (unsigned long)abs_symbol_2;
>>>>>>>> }
>>>>>>>>
>>>>>>>> unsigned long _start(void)
>>>>>>>> {
>>>>>>>> return f_abs_symbol() + f_abs_symbol_2();
>>>>>>>> }
>>>>>>>>
>>>>>>>> aarch64-rtems6-gcc abs.c -Wl,--gc-sections -Wl,--defsym=abs_symbol_2=291
>>>>>>> Have you tried -mcmodel=large? With that I get:
>>>>>> I get the same result with -mcmodel=large. With -mcmodel=tiny I get:
>>>>>>
>>>>>> aarch64-rtems6-gcc abs.c -Wl,--gc-sections -Wl,--defsym=abs_symbol_2=291
>>>>>> -mcmodel=tiny
>>>>>> /tmp/ccKUnvyq.o: in function `f_abs_symbol_2':
>>>>>> abs.c:(.text+0x8): relocation truncated to fit: R_AARCH64_ADR_PREL_LO21
>>>>>> against symbol `abs_symbol_2' defined in*ABS* section in a.out
>>>>>> collect2: error: ld returned 1 exit status
>>>>>>
>>>>>> Is this a tool bug?
>>>>> No, it's how code models are defined. GCC documentation says clearly:
>>>>>
>>>>> -mcmodel=tiny
>>>>> Generate code for the tiny code model. The program and its
>>>>> statically defined symbols must be within 1MB of each other.
>>>>> Programs can be statically or dynamically linked.
>>>>>
>>>>> Here the text is located at 0x400000 but abs_symbol_2 is at 0x123, thus
>>>>> violating the definition of -mcmodel=tiny.
>>>> Yes, this makes sense.
>>>>
>>>>>> Is there some way to make this working with -mcmodel=small?
>>>>> No because -mcmodel=small only assumes the program and the symbols are
>>>>> within*a* 4GB range - for example it's allowed to be [47GB, 51GB).
>>>>> This is different from the default code model of RISC-V (-
>>>>> mcmodel=medlow) where the symbols must be in [-2GB, 2GB).
>>>>>
>>>>> If you really think GCC should support this you can ask GCC for adding a
>>>>> new code model. Anyway this is not a linker issue because the adrp-add
>>>>> pairs are generated by GCC.
>>>> Thanks for the explanation.
>>>>
>>>> I guess what reliably works across code models is using the address of an existing symbol and then add a constant which fits into the code model. To retrieve the constant, we just have to subtract the address of the other symbol.
>>>>
>>> There are various ways you can code this:
>>>
>>> 1) An alternative when using the large code model is to just write
>>>
>>> extern char abs_symbol[];
>>>
>>> char * const abs_sym_ptr = abs_symbol;
>>>
>>> unsigned long f_abs_symbol(void)
>>> {
>>> return (unsigned long)abs_sym_ptr;
>>> }
>>>
>>> Which assembles to:
>>> 0000000000400000 <f_abs_symbol>:
>>> 400000: 90000000 adrp x0, 400000 <f_abs_symbol>
>>> 400004: 91004000 add x0, x0, #0x10
>>> 400008: f9400000 ldr x0, [x0]
>>> 40000c: d65f03c0 ret
>>> 400010: 00000123 .word 0x00000123
>>> 400014: 00000000 .word 0x00000000
>>>
>>> 2) If you really want to use other code models, you can try things like:
>>>
>>> extern char abs_symbol[];
>>>
>>> char * const volatile abs_sym_ptr = abs_symbol;
>>>
>>> unsigned long f_abs_symbol(void)
>>> {
>>> return (unsigned long)abs_sym_ptr;
>>> }
>>>
>>> The "const volatile" forces the compiler not to try to inline the pointer value dereference, so you end up with:
>>>
>>> 0000000000400000 <f_abs_symbol>:
>>> 400000: 90000000 adrp x0, 400000 <f_abs_symbol>
>>> 400004: f9400c00 ldr x0, [x0, #24]
>>> 400008: d65f03c0 ret
>>>
>>> ...
>>>
>>> Disassembly of section .rodata:
>>>
>>> 0000000000400018 <abs_sym_ptr>:
>>> 400018: 00000123 00000000 #.......
>>>
>>>
>>> In other respects, though, the pointer is const, so attempts to modify it via normal C code will be inhibited by the compiler.
>>>
>>> This should work with both the small and tiny models as well as the large model.
>> For my use case, I can't use things in the data or read-only data section. The RTEMS real-time operating system is deployed as a start.o start file and a set of static libraries. Applications are statically linked with the operating system. The operating system is configured by the application through a set of objects and some symbols those addresses are configuration values. One of these symbols is the size of the interrupt and system initialization stack. Setting up the system initialization stack is usually one of the first things done by the startup code in start.o. This is done before the data sections are copied from the load memory area to the virtual (runtime) memory area. For example on AArch64 it uses this code currently:
>>
>> /* Calculate interrupt stack area end for current processor */
>> #ifdef AARCH64_MULTILIB_ARCH_V8_ILP32
>> ldr w1, =_ISR_Stack_size
>> #else
>> ldr x1, =_ISR_Stack_size
>> #endif
>> #ifdef RTEMS_SMP
>> add x3, x7, #1
>> mul x1, x1, x3
>> #endif
>> #ifdef AARCH64_MULTILIB_ARCH_V8_ILP32
>> ldr w2, =_ISR_Stack_area_begin
>> #else
>> ldr x2, =_ISR_Stack_area_begin
>> #endif
>> add x3, x1, x2
>>
>> I adjusted our test cases so that they work now also for AArch64:
>>
>> https://git.rtems.org/rtems/commit/?id=d7a6e803984e5508c30aa9e3625c76460beb807b
>>
> So force the symbol into the text section. It's a bit naughty, but it should work, assuming the text section has read permission:
>
> extern char abs_symbol[];
>
> char * const volatile __attribute__((section (".text.constdata")))
> abs_sym_ptr = abs_symbol;
>
> unsigned long f_abs_symbol(void)
> {
> return (unsigned long)abs_sym_ptr;
> }
>
> unsigned long _start(void)
> {
> return f_abs_symbol();
> }
>
> $ aarch64-none-elf-gcc -O -nostartfiles -nostdlib start.c -Wl,--gc-sections -Wl,--defsym=abs_symbol=291
> $ aarch64-linux-gnu-objdump -d a.out
>
> a.out: file format elf64-littleaarch64
>
>
> Disassembly of section .text:
>
> 0000000000400000 <f_abs_symbol>:
> 400000: 90000000 adrp x0, 400000 <f_abs_symbol>
> 400004: f9400c00 ldr x0, [x0, #24]
> 400008: d65f03c0 ret
>
> 000000000040000c <_start>:
> 40000c: 90000000 adrp x0, 400000 <f_abs_symbol>
> 400010: f9400c00 ldr x0, [x0, #24]
> 400014: d65f03c0 ret
>
> 0000000000400018 <abs_sym_ptr>:
> 400018: 00000123 00000000 #.......
Thanks for the trick with the const volatile object. We had some other
places which used symbols with an arbitrary address directly. Using the
const volatile objects worked fine.
Independent of this, I think in the -mcmodel=small case some tool should
issue an error instead of getting an invalid value at runtime. Just like
in the -mcmodel=tiny case, where you get a relocation truncated error.
--
embedded brains GmbH
Herr Sebastian HUBER
Dornierstr. 4
82178 Puchheim
Germany
email: sebastian.huber@embedded-brains.de
phone: +49-89-18 94 741 - 16
fax: +49-89-18 94 741 - 08
Registergericht: Amtsgericht München
Registernummer: HRB 157899
Vertretungsberechtigte Geschäftsführer: Peter Rasmussen, Thomas Dörfler
Unsere Datenschutzerklärung finden Sie hier:
https://embedded-brains.de/datenschutzerklaerung/
More information about the Binutils
mailing list