All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jens Axboe <axboe@kernel.dk>
To: Thierry BERTAUD <tbertaud@conforama.fr>,
	"fio@vger.kernel.org" <fio@vger.kernel.org>
Subject: Re: fio-f1480f9: AIX 6.1 build KO against gcc function overlap
Date: Wed, 30 Dec 2015 09:38:52 -0700	[thread overview]
Message-ID: <5684089C.7000201@kernel.dk> (raw)
In-Reply-To: <VI1PR05MB13431E57B67F7322189A74E2B1FD0@VI1PR05MB1343.eurprd05.prod.outlook.com>

On 12/30/2015 01:36 AM, Thierry BERTAUD wrote:
> Jens,
>
>> Unless I botched it, it should disable lex on AIX now, unless you
>> explicitly attempt to enable with --enable-lex.
>
> I built against fio-de26b82 and now lex its disable by default. Tha'ts OK  (cf full output below).
>
> Concerning error with  --enable-lex, i think is due that option '-o' is not valid so lex doon't generate the file leyx.yy.c => don't find it after.
> ---
> MTD                           no
> 0: Warning: 1285-300 The o flag is not valid.
> 0: Error: 1285-331 Cannot read input file lex.yy.c
> lex/yacc for arithmetic       yes
> ---
> If i read configure:
>      +1505
>      +1506  # Check if lex fails using -o
>      +1507  if test "$arith" = "yes" ; then
>      +1508  $LEX -o lex.yy.c exp/expression-parser.l 2> /dev/null
>      +1509  if test "$?" = "0" ; then
>      +1510    lex_use_o="yes"
>      +1511  else
>      +1512    lex_use_o="no"
>      +1513  fi
>      +1514  fi
>      +1515
>      +1516  echo "lex/yacc for arithmetic       $arith"
>
> Could you update test condition, if OS is AIX to use "less exp/expression-parser.l" and other case "-o lex.yy.c exp/expression-parser.l"
>     # lex exp/expression-parser.l; echo $?;ls -l lex.yy.c
>     0
>     -rw-r--r--    1 root     staff         48081 Dec 30 09:34 lex.yy.c

But that's what the lex_use_o should control. If lex is enabled, we'll 
output CONFIG_LEX_USE_O=y if lex should use -o and nothing if it should 
not. If CONFIG_LEX_USE_O is in config-host.mak, then the Makefile target 
will use -o. If not, then it should not. But maybe I screwed it up. 
Testing here:

axboe@xps13:/home/axboe/git/fio $ grep USE_O config-host.mak
CONFIG_LEX_USE_O=y
axboe@xps13:/home/axboe/git/fio $ make V=1 lex.yy.c
lex -o lex.yy.c exp/expression-parser.l

which looks correct. I'll manually kill CONFIG_LEX_USE_O=y from 
config-host.mak, and retry:

axboe@xps13:/home/axboe/git/fio $ rm lex.yy.c
axboe@xps13:/home/axboe/git/fio $ grep USE_O config-host.mak
axboe@xps13:/home/axboe/git/fio $ make V=1 lex.yy.c
lex exp/expression-parser.l
axboe@xps13:/home/axboe/git/fio $ ls -al lex.yy.c
-rw-r--r-- 1 axboe axboe 51991 Dec 30 09:37 lex.yy.c

so that logic does look correct. Maybe the configure check for the error 
return of lex -o isn't working. Do you have CONFIG_LEX_USE_O in your 
config-host.mak, if you run configure --enable-lex?

-- 
Jens Axboe



  reply	other threads:[~2015-12-30 16:38 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-12-28 21:45 fio-f1480f9: AIX 6.1 build KO against gcc function overlap Thierry BERTAUD
2015-12-28 22:24 ` Andrey Kuzmin
2015-12-28 23:18   ` Thierry BERTAUD
2015-12-28 23:21 ` Jens Axboe
2015-12-28 23:34   ` Thierry BERTAUD
2015-12-28 23:46     ` Jens Axboe
2015-12-29  0:01       ` Thierry BERTAUD
2015-12-29  0:15         ` Jens Axboe
2015-12-29  0:30           ` Thierry BERTAUD
2015-12-29  0:39             ` Jens Axboe
2015-12-29  0:56               ` Thierry BERTAUD
2015-12-29 15:34                 ` Jens Axboe
2015-12-29 16:16                   ` Thierry BERTAUD
2015-12-29 21:10                     ` Thierry BERTAUD
2015-12-30  3:26                       ` Jens Axboe
2015-12-30  3:59                         ` Jens Axboe
2015-12-30  8:36                           ` Thierry BERTAUD
2015-12-30 16:38                             ` Jens Axboe [this message]
2015-12-30 16:50                               ` Jens Axboe
2015-12-30 16:51                                 ` Jens Axboe
2015-12-30 17:59                                   ` Thierry BERTAUD

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=5684089C.7000201@kernel.dk \
    --to=axboe@kernel.dk \
    --cc=fio@vger.kernel.org \
    --cc=tbertaud@conforama.fr \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.