[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <832fe56f-1688-d128-50dd-3f668815a248@loongson.cn>
Date: Wed, 11 Dec 2024 11:11:28 +0800
From: Tiezhu Yang <yangtiezhu@...ngson.cn>
To: Josh Poimboeuf <jpoimboe@...nel.org>
Cc: Huacai Chen <chenhuacai@...nel.org>, Peter Zijlstra
<peterz@...radead.org>, loongarch@...ts.linux.dev,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v5 02/10] objtool: Handle different entry size of rodata
On 12/10/2024 03:54 AM, Josh Poimboeuf wrote:
> On Sat, Dec 07, 2024 at 09:59:07AM +0800, Tiezhu Yang wrote:
>> +__weak unsigned int arch_reloc_size(struct reloc *reloc)
>> +{
>> + return 8;
>> +}
>
> Instead of making a weak function, each arch should have an explicit
> definition of this function, as it's not always 8:
>
> - x86 has R_X86_64_PC32 and R_X86_64_PLT32 which are 4 bytes.
>
> - 32-bit powerpc
>
> - 64-bit powerpc has R_PPC64_REL32
For x86, like this, I am not sure "case R_X86_64_32" is necessary,
leave it as is if no objections, please let me know if it should be
removed.
unsigned int arch_reloc_size(struct reloc *reloc)
{
switch (reloc_type(reloc)) {
case R_X86_64_32:
case R_X86_64_PC32:
case R_X86_64_PLT32:
return 4;
default:
return 8;
}
}
For ppc, like this:
unsigned int arch_reloc_size(struct reloc *reloc)
{
switch (reloc_type(reloc)) {
case R_PPC_REL32:
case R_PPC64_REL32:
return 4;
default:
return 8;
}
}
>
>> @@ -1967,8 +1973,10 @@ static int add_jump_table(struct objtool_file *file, struct instruction *insn,
>> if (reloc != table && reloc == next_table)
>> break;
>>
>> + entry_size = arch_reloc_size(reloc);
>> +
>> /* Make sure the table entries are consecutive: */
>> - if (prev_offset && reloc_offset(reloc) != prev_offset + 8)
>> + if (prev_offset && reloc_offset(reloc) != prev_offset + entry_size)
>> break;
>
> No need to use a variable here, just call arch_reloc_size() directly.
OK, will do it.
Thanks,
Tiezhu
Powered by blists - more mailing lists