RFC: Add ___tls_get_addr

[H.J. Lu]

On Thu, Jul 6, 2017 at 1:06 AM, Szabolcs Nagy <szabolcs.nagy@arm.com> wrote:
> On 05/07/17 17:18, H.J. Lu wrote:
>> On Wed, Jul 5, 2017 at 8:53 AM, Szabolcs Nagy <szabolcs.nagy@arm.com> wrote:
>>> On 05/07/17 16:38, H.J. Lu wrote:
>>>> On x86-64, __tls_get_addr has to realigns stack so that binaries compiled by
>>>> GCCs older than GCC 4.9.4:
>>>>
>>>> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58066
>>>>
>>>> continue to work even if vector instructions are used by functions called
>>>> from __tls_get_addr, which assumes 16-byte stack alignment as specified
>>>> by x86-64 psABI.
>>>>
>>>> We are considering to add an alternative interface, ___tls_get_addr, to
>>>> glibc, which doesn't realign stack.  Compilers, which properly align stack
>>>> for TLS, call generate call to ___tls_get_addr, instead of __tls_get_addr,
>>>> if ___tls_get_addr is available.
>>>>
>>>> Any comments?
>>>>
>>>>
>>>
>>> what happens when new compiler generating the new symbol
>>> is used with old glibc?
>>>
>>
>> Compiler shouldn't do that.
>>
>
> i don't see how can the compiler not do that
>
> e.g. somebody building an old glibc from
> source with new compiler, then runs the tests,
> all tls tests would fail because the compiler
> generated the new symbol.

Using  ___tls_get_addr should be controlled by an option
selected by compiler build time as well as run-time.

> or users interposing __tls_get_addr (asan) need
> to update their code.

Yes.  That is true.

> or there are cases when libraries built against
> one libc is used with another (e.g. musl can
> mostly use a libstdc++ compiled against glibc
> on x86_64)

This happens every time when a new version of a function
is added to glibc.   musl has to deal with it.

> i think introducing new libc<->compiler abi

This is no different from adding a new version of a function
to glibc.

> should be done conservatively when it is really
> necessary and from Rich's mail it seems there
> is no need for new abi here.
>

See:

https://sourceware.org/ml/libc-alpha/2017-07/msg00086.html

-- 
H.J.