[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <140d4869-3578-43e4-975f-f52b46711aaa@daynix.com>
Date: Fri, 3 Jan 2025 14:25:00 +0900
From: Akihiko Odaki <akihiko.odaki@...nix.com>
To: Dave Martin <Dave.Martin@....com>
Cc: Eric Biederman <ebiederm@...ssion.com>, Kees Cook <kees@...nel.org>,
Catalin Marinas <catalin.marinas@....com>, Mark Brown <broonie@...nel.org>,
linux-mm@...ck.org, linux-kernel@...r.kernel.org, devel@...nix.com
Subject: Re: [PATCH] elf: Correct note name comment
On 2025/01/02 23:40, Dave Martin wrote:
> Hi,
>
> On Wed, Dec 25, 2024 at 03:46:44PM +0900, Akihiko Odaki wrote:
>> NT_PRSTATUS note is also named "CORE". Correct the comment accordingly.
>>
>> Fixes: 00e19ceec80b ("ELF: Add ELF program property parsing support")
>> Signed-off-by: Akihiko Odaki <akihiko.odaki@...nix.com>
>> ---
>> include/uapi/linux/elf.h | 4 ++--
>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/include/uapi/linux/elf.h b/include/uapi/linux/elf.h
>> index b54b313bcf07..4f00cdca38b2 100644
>> --- a/include/uapi/linux/elf.h
>> +++ b/include/uapi/linux/elf.h
>> @@ -372,8 +372,8 @@ typedef struct elf64_shdr {
>> * Notes used in ET_CORE. Architectures export some of the arch register sets
>> * using the corresponding note types via the PTRACE_GETREGSET and
>> * PTRACE_SETREGSET requests.
>> - * The note name for these types is "LINUX", except NT_PRFPREG that is named
>> - * "CORE".
>> + * The note name for these types is "LINUX", except NT_PRSTATUS and NT_PRFPREG
>> + * that are named "CORE".
>> */
>> #define NT_PRSTATUS 1
>> #define NT_PRFPREG 2
>
> [...]
>
> This still seems rather confusing. It's not clear which note types are
> being referred to in "for these types". I think this statement was
> supposed to refer only to the architectural regset notes.
I'm not sure about the original intention, but this comment starts with
"notes used in ET_CORE" so I think it should ideally describe all types
used in ET_CORE.
>
> I guess "CORE" was for generic coredump notes goverened by common
> specs, and LINUX was for Linux-specific stuff, but I suspect that this
> distinction may have bitrotted. It looks like the ELF specs never
> defined the core dump format, so the concept of non-OS-specific
> coredump notes may not make much sense.
>
> The ELF specs _do_ explicitly say [1] that the note name must be taken
> into account when identifying the type of a note, so the note name for
> each kind if note should really be documented explicitly.
>
> Is it worth adding explicit #defines for the note name of each kind
> of note, to make the ABI contract explicit?
Maybe so. I don't have a particular idea how such #defines should be
written though as I don't have actual code that may utilize them.
Regards,
Akihiko Odaki
Powered by blists - more mailing lists