lttng-dev.lists.lttng.org archive mirror
 help / color / mirror / Atom feed
From: Mathieu Desnoyers via lttng-dev <lttng-dev@lists.lttng.org>
To: Changqing Li <changqing.li@windriver.com>
Cc: lttng-dev <lttng-dev@lists.lttng.org>
Subject: Re: [lttng-dev] with -Og option, lttng-ust compile failed on 32bit arm
Date: Thu, 30 Sep 2021 09:14:57 -0400 (EDT)	[thread overview]
Message-ID: <613136200.45770.1633007697070.JavaMail.zimbra@efficios.com> (raw)
In-Reply-To: <0c4c0d31-5551-b84a-386c-32990b128f09@windriver.com>


[-- Attachment #1.1: Type: text/plain, Size: 2723 bytes --]

----- On Sep 29, 2021, at 11:47 PM, lttng-dev <lttng-dev@lists.lttng.org> wrote: 

> Hi,

> The problem happened after upgrade to lttng-ust 2.13.0. 2.12.0 don't have this
> issue.

> liburcu version: 0.13.0

> gcc: 11.2.0

> This is my reproduce steps, it is cross compile enviroment based on yocto
> project.

>     1. git clone git://git.yoctoproject.org/poky
>     2. . oe-init-build-env
>     3. echo "MACHINE='qemuarm'" >> conf/local.conf
>     4. echo "DEBUG_BUILD='1'" >> conf/local.conf
>     5. bitbake lttng-ust

> compile failed with error:

>| /work/cortexa15t2hf-neon-poky-linux-gnueabi/lttng-ust/2_2.13.0-r0/recipe-sysroot-native/usr/bin/arm-poky-linux-gnueabi/../../libexec/arm-poky-linux-gnueabi/gcc/arm-poky-linux-gnueabi/11.2.0/ld:
>| ../../../src/lib/lttng-ust/.libs/liblttng-ust.so: undefined reference to
> | `_uatomic_link_error'
> | collect2: error: ld returned 1 exit status
> | Makefile:399: recipe for target 'test_ust_error' failed

> checked with "nm ../../../src/lib/lttng-ust/.libs/liblttng-ust.so" | grep
> atomic, we can see 'U _uatomic_link_error', but since -Og

> is used, liburcu don't define this function.

> [snip]
> #if !defined __OPTIMIZE__ || defined UATOMIC_NO_LINK_ERROR
> static inline __attribute__((always_inline, noreturn))
> void _uatomic_link_error(void)
> {
> #ifdef ILLEGAL_INSTR
> /*
> * generate an illegal instruction. Cannot catch this with
> * linker tricks when optimizations are disabled.
> */
> __asm__ __volatile__(ILLEGAL_INSTR);
> #else
> __builtin_trap();
> #endif
> }
> #else /* #if !defined __OPTIMIZE__ || defined UATOMIC_NO_LINK_ERROR */
> extern void _uatomic_link_error(void);
> #endif /* #else #if !defined __OPTIMIZE__ || defined UATOMIC_NO_LINK_ERROR */

> [snip]

> we cannot see 'U _uatomic_link_error' in following conditions, so compile
> successed:

> 1. without -Og(using -O2), + 32bit arm

> 2. -Og + 64bit arm

> 3. -Og + x86/x86-64

> Do you have any idea about how to fix this? I don't understand why only "-Og +
> 32bit arm" will call function _uatomic_link_error.
I suspect it depends on which optimizations are being enabled at -Og on each architecture. 
The "_uatomic_link_error()" trick indeed depends on the compiler optimizing away 
unreachable calls. 

If you really intend on using "-Og" on arm32, trying building with "-DUATOMIC_NO_LINK_ERROR". 
It should take care of making sure to generate an illegal instruction rather than rely on the linker 
error. 

Thanks, 

Mathieu 

> Thanks

> //Changqing

> _______________________________________________
> lttng-dev mailing list
> lttng-dev@lists.lttng.org
> https://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

-- 
Mathieu Desnoyers 
EfficiOS Inc. 
http://www.efficios.com 

[-- Attachment #1.2: Type: text/html, Size: 4737 bytes --]

[-- Attachment #2: Type: text/plain, Size: 156 bytes --]

_______________________________________________
lttng-dev mailing list
lttng-dev@lists.lttng.org
https://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

  reply	other threads:[~2021-09-30 13:15 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-30  3:47 [lttng-dev] with -Og option, lttng-ust compile failed on 32bit arm Changqing Li via lttng-dev
2021-09-30 13:14 ` Mathieu Desnoyers via lttng-dev [this message]
2021-10-08  8:12   ` Changqing Li via lttng-dev

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=613136200.45770.1633007697070.JavaMail.zimbra@efficios.com \
    --to=lttng-dev@lists.lttng.org \
    --cc=changqing.li@windriver.com \
    --cc=mathieu.desnoyers@efficios.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 NNTP newsgroup(s).