From: Adhemerval Zanella <adhemerval.zanella@linaro.org>
To: Szabolcs Nagy <szabolcs.nagy@arm.com>,
Florian Weimer <fweimer@redhat.com>
Cc: John Mellor-Crummey <johnmc@rice.edu>, libc-alpha@sourceware.org
Subject: Re: [PATCH v6 20/20] elf: Add SVE support for aarch64 rtld-audit
Date: Tue, 21 Dec 2021 14:08:05 -0300 [thread overview]
Message-ID: <71a3b9ed-85bd-0d9b-a3b3-ef787fb36cb7@linaro.org> (raw)
In-Reply-To: <20211221164500.GC3294453@arm.com>
On 21/12/2021 13:45, Szabolcs Nagy wrote:
> The 12/21/2021 15:27, Florian Weimer via Libc-alpha wrote:
>> * Adhemerval Zanella:
>>
>>> To implement lazy binding is enabled when profiling or auditing used,
>>> even when STO_AARCH64_VARIANT_PCS is set. Also, to not incur in
>>> performance penalties on architecture without SVE, the PLT entrypoint
>>> is set to a newer one, _dl_runtime_profile_sve, which is used iff
>>> 'hwcap' has HWCAP_SVE bit set.
>>>
>>> This should be a fair assumption since SVE has a defined set of
>>> registers for argument passing and return values. A new ABI with either
>>> different argument passing or different registers would require a
>>> different PLT entry, but I assume this would require another symbol flag
>>> anyway (or at least a different ELF mark to indicate so).
>>
>> Someone familiar with SVE needs to review this.
>>
>> The restriction to the currently-specified vector calling convention
>> seems to be a bit arbitrary to me. To me, STO_AARCH64_VARIANT_PCS
>> strongly suggests that the toolchain makes every conceivable attempt to
>> support arbitrary calling conventions, but that's probably just me.
>
>
> thanks for working on this, but yeah this approach does not work:
>
> A PLT hook for STO_AARCH64_VARIANT_PCS symbol has to preserve *all*
> registers (except x16,x17,cc), this is the same strict requirement
> as linker inserted veneers follow. (historically the dynamic linker
> lazy binding code made assumptions about the call convention, but if
> there are extern calls with unusual call conventions then this does
> not work and it has to behave like the static linker. this is the
> meaning of variant_pcs, it is not just for sve calls.)
>
> there is no way to tell which call is an sve call, so i'd suggest
> having 2 profile entries: normal and variant_pcs and the latter is
> used if a module is marked with DT_AARCH64_VARIANT_PCS and it
> preserves all registers.
>
> when "all registers" increase via arch extensions then the PLT hook
> code has to be updated, but ideally only the variant_pcs hooks
> will require a glibc update to work on a new arch.
I was afraid we would require a change like that and based on
STO_AARCH64_VARIANT_PCS definition is does seem the only option.
>
> e.g. the user callback can take
>
> struct regs {
> // base arg regs inline;
> ...
> void *vpcs_regs;
> };
> > where vpcs_regs!=0 for variant_pcs calls and contains all the regs
> (in a way that can represent various combinations of arch extensions,
> i don't think we have to make it userfriendly to access).
Yeah, that was one idea I had when coding it. It will make the
_dl_runtime_profile somewhat more complex for SVE, but it the cost
of having a non-specific ABIs for lazy binding.
I think we will need to have this new field on BZ #26643 fix at least,
and chage current code to zero it. It will make it easier to provide
SVE support on 2.36.
>
> (i think we don't have to get the aarch64 fixes done in 2.35 as this
> is a big design work.)
Agreed, I will drop this patch for v7.
prev parent reply other threads:[~2021-12-21 17:08 UTC|newest]
Thread overview: 89+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-15 18:37 [PATCH v6 00/20] Multiple rtld-audit fixes Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 01/20] elf: Suppress audit calls when a (new) namespace is empty (BZ #28062) Adhemerval Zanella
2021-11-15 19:01 ` Florian Weimer
2021-11-16 13:14 ` Adhemerval Zanella
2021-11-16 13:15 ` Florian Weimer
2021-11-16 13:45 ` Andreas Schwab
2021-11-16 13:48 ` Florian Weimer
2021-11-16 14:16 ` Andreas Schwab
2021-11-18 19:58 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 02/20] elf: Add _dl_audit_objopen Adhemerval Zanella
2021-12-10 12:40 ` Florian Weimer
2021-12-10 12:48 ` Florian Weimer
2021-12-10 13:45 ` Adhemerval Zanella
2021-12-10 14:11 ` Adhemerval Zanella
2021-12-10 14:15 ` Florian Weimer
2021-12-10 14:41 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 03/20] elf: Add _dl_audit_activity_map and _dl_audit_activity_nsid Adhemerval Zanella
2021-12-10 13:54 ` Florian Weimer
2021-12-10 14:58 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 04/20] elf: Add _dl_audit_objsearch Adhemerval Zanella
2021-12-17 12:21 ` Florian Weimer
2021-12-17 16:09 ` Adhemerval Zanella
2021-12-17 16:12 ` Florian Weimer
2021-12-17 16:33 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 05/20] elf: Add _dl_audit_objclose Adhemerval Zanella
2021-12-17 16:29 ` Florian Weimer
2021-12-17 16:39 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 06/20] elf: Add _dl_audit_symbind_alt and _dl_audit_symbind Adhemerval Zanella
2021-12-17 16:50 ` Florian Weimer
2021-12-17 18:46 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 07/20] elf: Add _dl_audit_preinit Adhemerval Zanella
2021-12-17 17:13 ` Florian Weimer
2021-12-17 18:54 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 08/20] elf: Add _dl_audit_pltenter Adhemerval Zanella
2021-12-17 17:21 ` Florian Weimer
2021-11-15 18:37 ` [PATCH v6 09/20] elf: Add _dl_audit_pltexit Adhemerval Zanella
2021-12-17 17:43 ` Florian Weimer
2021-12-17 19:02 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 10/20] elf: Avoid unnecessary slowdown from profiling with audit (BZ#15533) Adhemerval Zanella
2021-12-18 18:45 ` Florian Weimer
2021-12-20 12:10 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 11/20] elf: Add audit tests for modules with TLSDESC Adhemerval Zanella
2021-12-18 18:53 ` Florian Weimer
2021-12-20 12:23 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 12/20] elf: Do not fail for failed dlmopen on audit modules (BZ #28061) Adhemerval Zanella
2021-12-18 18:59 ` Florian Weimer
2021-12-20 12:24 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 13/20] elf: Fix initial-exec TLS access on audit modules (BZ #28096) Adhemerval Zanella
2021-12-18 18:01 ` Florian Weimer
2021-12-20 13:25 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 14/20] elf: Issue audit la_objopen() for vDSO Adhemerval Zanella
2021-12-18 20:00 ` Florian Weimer
2021-12-20 12:50 ` Adhemerval Zanella
2021-12-20 13:18 ` Florian Weimer
2021-11-15 18:37 ` [PATCH v6 15/20] elf: Run constructors if executable has a soname of a dependency Adhemerval Zanella
2021-12-18 20:08 ` Florian Weimer
2021-12-20 16:49 ` Adhemerval Zanella
2021-12-20 16:52 ` Florian Weimer
2021-12-20 16:55 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 16/20] elf: Add main application on main_map l_name Adhemerval Zanella
2021-12-20 13:32 ` Florian Weimer
2021-12-20 18:04 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 17/20] elf: Add la_activity during application exit Adhemerval Zanella
2021-12-20 13:34 ` Florian Weimer
2021-12-20 19:46 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 18/20] elf: Issue la_symbind() for bind-now (BZ #23734) Adhemerval Zanella
2021-12-20 19:18 ` Florian Weimer
2021-12-20 20:43 ` Adhemerval Zanella
2021-12-20 21:04 ` Florian Weimer
2021-12-20 21:09 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 19/20] elf: Fix runtime linker auditing on aarch64 (BZ #26643) Adhemerval Zanella
2021-12-20 21:32 ` Florian Weimer
2021-12-21 14:33 ` Adhemerval Zanella
2021-12-21 14:44 ` Florian Weimer
2021-12-21 14:48 ` Adhemerval Zanella
2021-12-21 14:54 ` Florian Weimer
2021-12-21 17:03 ` Adhemerval Zanella
2021-12-21 17:22 ` Florian Weimer
2021-12-21 17:38 ` Adhemerval Zanella
2021-12-21 18:11 ` Florian Weimer
2021-12-21 18:19 ` Adhemerval Zanella
2021-12-20 21:34 ` Florian Weimer
2021-12-21 17:47 ` Szabolcs Nagy
2021-12-21 17:49 ` Adhemerval Zanella
2021-11-15 18:37 ` [PATCH v6 20/20] elf: Add SVE support for aarch64 rtld-audit Adhemerval Zanella
2021-12-21 14:27 ` Florian Weimer
2021-12-21 14:37 ` Adhemerval Zanella
2021-12-21 16:45 ` Szabolcs Nagy
2021-12-21 17:08 ` Adhemerval Zanella [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=71a3b9ed-85bd-0d9b-a3b3-ef787fb36cb7@linaro.org \
--to=adhemerval.zanella@linaro.org \
--cc=fweimer@redhat.com \
--cc=johnmc@rice.edu \
--cc=libc-alpha@sourceware.org \
--cc=szabolcs.nagy@arm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).