All of lore.kernel.org
 help / color / mirror / Atom feed
From: Dave Chinner <david@fromorbit.com>
To: Angelo Dureghello <angelo.dureghello@nomovok.com>
Cc: xfs@oss.sgi.com
Subject: Re: xfstests, xfs test 080 fails on arm 32bit
Date: Wed, 30 Sep 2015 15:04:26 +1000	[thread overview]
Message-ID: <20150930050425.GC3902@dastard> (raw)
In-Reply-To: <560A9307.6080600@nomovok.com>

On Tue, Sep 29, 2015 at 03:32:55PM +0200, Angelo Dureghello wrote:
> Hi,
> 
> executing this test on a 32bit arm. i get (i traced mmapping sizes):
> 
> # ./start_xfs_test.sh
> QA output created by 080
> [  161.827446] XFS (mmcblk0p5): Mounting V4 Filesystem
> [  162.357952] XFS (mmcblk0p5): Ending clean mount
> 
> mmapping 512000000
> mmapping 512000000
> mmapping 512000000
> mmapping 512000000
> mmapping 512000000
> mmapping 512000000
> 
> oio ( 3241) 15:18:17
> ---------------------
> mmap() failed - 0xffffffff 12
> 
> doio ( 3241) 15:18:17
> ---------------------
> mmap-read() request failed:  Cannot allocate memory (12)
> Request number 15
>           fd 4 is file /media/p5/rwtest.file - open flags are 0 O_RDONLY,
>           read done at file offset 978498
>           number of requests is 1, strides per request is 1
>           i/o byte count = 119505
>           memory alignment is unaligned
> 
> syscall:  mmap-read(NULL, 512000000, PROT_READ, MAP_SHARED, 4, 0)
>     file is mmaped to: 0x0
>     file-mem=0xeee42, length=119505, buffer=0x484a6

Does it close/unmap the other files it mmaps? i.e. 32 bit only has a
3GB address space for user applications (IIRC), so the 6th 512MB
mapping will run the process out of address space and hence fail....

> doio ( 3241) 15:18:17
> ---------------------
> doio(): operation 120 returned != 0
> rwtest.sh : iogen reported errors (r=141)
> 
> 
> Is it possible the test was thought to pass on 64bit systems ?

Works just fine on 64 bit systems:

 $ sudo ./check xfs/080
 FSTYP         -- xfs (debug)
 PLATFORM      -- Linux/x86_64 test2 4.3.0-rc1-dgc+
 MKFS_OPTIONS  -- -f -bsize=4096 /dev/sdg
 MOUNT_OPTIONS -- -o context=system_u:object_r:nfs_t:s0 /dev/sdg
 /mnt/scratch

 xfs/080 2s ... 7s
 Ran: xfs/080
 Passed all 1 tests

And looking at the memory footprint:

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
 3300 root      20   0 3010672   1612   1344 D   4.7  0.0   0:00.14 doio

Yup, 3GB of virtual address space used, 1.6MB of actual memory used.

_require_64bit_userspace is probably needed here...

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

      reply	other threads:[~2015-09-30  5:04 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-09-29 13:32 xfstests, xfs test 080 fails on arm 32bit Angelo Dureghello
2015-09-30  5:04 ` Dave Chinner [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=20150930050425.GC3902@dastard \
    --to=david@fromorbit.com \
    --cc=angelo.dureghello@nomovok.com \
    --cc=xfs@oss.sgi.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 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.