From: Eric Sandeen <sandeen@sandeen.net>
To: Markus Trippelsdorf <markus@trippelsdorf.de>
Cc: "Yongqiang Yang" <xiaoqiangnk@gmail.com>,
"Pádraig Brady" <P@draigbrady.com>, xfs-oss <xfs@oss.sgi.com>,
linux-ext4@vger.kernel.org, coreutils@gnu.org
Subject: Re: Files full of zeros with coreutils-8.11 and xfs (FIEMAP related?)
Date: Thu, 14 Apr 2011 11:49:19 -0500 [thread overview]
Message-ID: <4DA7258F.1080302@sandeen.net> (raw)
In-Reply-To: <20110414164811.GB21658@x4.trippels.de>
On 4/14/11 11:48 AM, Markus Trippelsdorf wrote:
> On 2011.04.14 at 11:31 -0500, Eric Sandeen wrote:
>> On 4/14/11 11:28 AM, Markus Trippelsdorf wrote:
>>
>> <snip>
>>
>>> Yes, but we're still trying to find out what caused the zeros in the
>>> binaries that coreutils installed on my system.
>>>
>>> Now the failure only happens when I use "gold" as my linker. With GNU ld
>>> everything is OK. But I thought this must be a timing issue, because
>>> gold is faster and the binaries in coreutils-8.11/src are all fine.
>>
>> maybe xfs_bmap (or filefrag) of the binaries with both linkers would be instructive; are they laid out significantly differently?
>>
>> does gold preallocate?
>
> Just checked and yes it does. That should explain the issue I was
> seeing.
Well, mystery solved there, at least!
Now for the fixing part :)
Thanks for checking, at least my view of the world is still intact ;)
-Eric
> bool
> Output_file::map_no_anonymous()
> {
> const int o = this->o_;
>
> // If the output file is not a regular file, don't try to mmap it;
> // instead, we'll mmap a block of memory (an anonymous buffer), and
> // then later write the buffer to the file.
> void* base;
> struct stat statbuf;
> if (o == STDOUT_FILENO || o == STDERR_FILENO
> || ::fstat(o, &statbuf) != 0
> || !S_ISREG(statbuf.st_mode)
> || this->is_temporary_)
> return false;
>
> // Ensure that we have disk space available for the file. If we
> // don't do this, it is possible that we will call munmap, close,
> // and exit with dirty buffers still in the cache with no assigned
> // disk blocks. If the disk is out of space at that point, the
> // output file will wind up incomplete, but we will have already
> // exited. The alternative to fallocate would be to use fdatasync,
> // but that would be a more significant performance hit.
> if (::posix_fallocate(o, 0, this->file_size_) < 0)
> gold_fatal(_("%s: %s"), this->name_, strerror(errno));
>
> // Map the file into memory.
> base = ::mmap(NULL, this->file_size_, PROT_READ | PROT_WRITE,
> MAP_SHARED, o, 0);
>
> // The mmap call might fail because of file system issues: the file
> // system might not support mmap at all, or it might not support
> // mmap with PROT_WRITE.
> if (base == MAP_FAILED)
> return false;
>
> this->map_is_anonymous_ = false;
> this->base_ = static_cast<unsigned char*>(base);
> return true;
> }
>
>
WARNING: multiple messages have this Message-ID (diff)
From: Eric Sandeen <sandeen@sandeen.net>
To: Markus Trippelsdorf <markus@trippelsdorf.de>
Cc: "Yongqiang Yang" <xiaoqiangnk@gmail.com>,
xfs-oss <xfs@oss.sgi.com>,
coreutils@gnu.org, linux-ext4@vger.kernel.org,
ady <P@draigbrady.com>,
=?ISO-8859-1?Q?P=E1draig_Br?=@oss.sgi.com
Subject: Re: Files full of zeros with coreutils-8.11 and xfs (FIEMAP related?)
Date: Thu, 14 Apr 2011 11:49:19 -0500 [thread overview]
Message-ID: <4DA7258F.1080302@sandeen.net> (raw)
In-Reply-To: <20110414164811.GB21658@x4.trippels.de>
On 4/14/11 11:48 AM, Markus Trippelsdorf wrote:
> On 2011.04.14 at 11:31 -0500, Eric Sandeen wrote:
>> On 4/14/11 11:28 AM, Markus Trippelsdorf wrote:
>>
>> <snip>
>>
>>> Yes, but we're still trying to find out what caused the zeros in the
>>> binaries that coreutils installed on my system.
>>>
>>> Now the failure only happens when I use "gold" as my linker. With GNU ld
>>> everything is OK. But I thought this must be a timing issue, because
>>> gold is faster and the binaries in coreutils-8.11/src are all fine.
>>
>> maybe xfs_bmap (or filefrag) of the binaries with both linkers would be instructive; are they laid out significantly differently?
>>
>> does gold preallocate?
>
> Just checked and yes it does. That should explain the issue I was
> seeing.
Well, mystery solved there, at least!
Now for the fixing part :)
Thanks for checking, at least my view of the world is still intact ;)
-Eric
> bool
> Output_file::map_no_anonymous()
> {
> const int o = this->o_;
>
> // If the output file is not a regular file, don't try to mmap it;
> // instead, we'll mmap a block of memory (an anonymous buffer), and
> // then later write the buffer to the file.
> void* base;
> struct stat statbuf;
> if (o == STDOUT_FILENO || o == STDERR_FILENO
> || ::fstat(o, &statbuf) != 0
> || !S_ISREG(statbuf.st_mode)
> || this->is_temporary_)
> return false;
>
> // Ensure that we have disk space available for the file. If we
> // don't do this, it is possible that we will call munmap, close,
> // and exit with dirty buffers still in the cache with no assigned
> // disk blocks. If the disk is out of space at that point, the
> // output file will wind up incomplete, but we will have already
> // exited. The alternative to fallocate would be to use fdatasync,
> // but that would be a more significant performance hit.
> if (::posix_fallocate(o, 0, this->file_size_) < 0)
> gold_fatal(_("%s: %s"), this->name_, strerror(errno));
>
> // Map the file into memory.
> base = ::mmap(NULL, this->file_size_, PROT_READ | PROT_WRITE,
> MAP_SHARED, o, 0);
>
> // The mmap call might fail because of file system issues: the file
> // system might not support mmap at all, or it might not support
> // mmap with PROT_WRITE.
> if (base == MAP_FAILED)
> return false;
>
> this->map_is_anonymous_ = false;
> this->base_ = static_cast<unsigned char*>(base);
> return true;
> }
>
>
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
next prev parent reply other threads:[~2011-04-14 16:49 UTC|newest]
Thread overview: 117+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-04-14 10:26 Files full of zeros with coreutils-8.11 and xfs (FIEMAP related?) Markus Trippelsdorf
2011-04-14 12:06 ` Markus Trippelsdorf
2011-04-14 14:02 ` Markus Trippelsdorf
[not found] ` <20110414140222.GB1679-tLCgZGx+iJ+kxVt8IV0GqQ@public.gmane.org>
2011-04-14 14:59 ` Pádraig Brady
2011-04-14 14:59 ` Pádraig Brady
[not found] ` <4DA70BD3.1070409-V8g9lnOeT5ydJdNcDFJN0w@public.gmane.org>
2011-04-14 15:50 ` Eric Sandeen
2011-04-14 15:50 ` Eric Sandeen
[not found] ` <4DA717B2.3020305-+82itfer+wXR7s880joybQ@public.gmane.org>
2011-04-14 15:52 ` Pádraig Brady
2011-04-14 15:52 ` Pádraig Brady
2011-04-14 15:56 ` Eric Sandeen
2011-04-14 15:56 ` Eric Sandeen
2011-04-14 16:03 ` Markus Trippelsdorf
2011-04-14 16:03 ` Markus Trippelsdorf
2011-04-14 16:14 ` Eric Sandeen
2011-04-14 16:14 ` Eric Sandeen
[not found] ` <20110414160343.GA12787-tLCgZGx+iJ+kxVt8IV0GqQ@public.gmane.org>
2011-04-14 16:21 ` Yongqiang Yang
2011-04-14 16:21 ` Yongqiang Yang
[not found] ` <BANLkTimRxvBMp9M7zwiUY_UmmFOY5N58+A-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2011-04-14 16:28 ` Markus Trippelsdorf
2011-04-14 16:28 ` Markus Trippelsdorf
2011-04-14 16:31 ` Eric Sandeen
2011-04-14 16:31 ` Eric Sandeen
2011-04-14 16:48 ` Markus Trippelsdorf
2011-04-14 16:48 ` Markus Trippelsdorf
2011-04-14 16:49 ` Eric Sandeen [this message]
2011-04-14 16:49 ` Eric Sandeen
2011-04-14 16:04 ` Yongqiang Yang
2011-04-14 16:04 ` Yongqiang Yang
2011-04-14 16:10 ` Yongqiang Yang
2011-04-14 16:10 ` Yongqiang Yang
[not found] ` <BANLkTimoLeWMJgNFGW+zdeUeJyZ-_+8fMQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2011-05-05 11:29 ` Pádraig Brady
2011-05-05 11:29 ` Pádraig Brady
2011-05-05 11:47 ` Yongqiang Yang
2011-05-05 11:47 ` Yongqiang Yang
[not found] ` <4DA7182B.8050409-V8g9lnOeT5ydJdNcDFJN0w@public.gmane.org>
2011-04-14 17:27 ` Jim Meyering
2011-04-14 17:27 ` Jim Meyering
2011-04-14 19:13 ` Pádraig Brady
2011-04-14 19:13 ` Pádraig Brady
[not found] ` <878vvcspz0.fsf-CybKA8TIZ99x3y/oJEDuiw@public.gmane.org>
2011-04-14 19:39 ` Jim Meyering
2011-04-14 19:39 ` Jim Meyering
2011-04-14 22:59 ` Dave Chinner
2011-04-14 23:29 ` Pádraig Brady
2011-04-14 23:29 ` Pádraig Brady
2011-04-15 0:09 ` Dave Chinner
2011-04-15 0:09 ` Dave Chinner
2011-04-15 5:01 ` Andreas Dilger
2011-04-15 5:01 ` Andreas Dilger
2011-04-16 0:50 ` Dave Chinner
2011-04-16 0:50 ` Dave Chinner
2011-04-16 5:11 ` Andreas Dilger
2011-04-16 5:11 ` Andreas Dilger
2011-04-16 12:21 ` Theodore Tso
2011-04-16 12:21 ` Theodore Tso
2011-04-18 0:40 ` Dave Chinner
2011-04-18 0:40 ` Dave Chinner
2011-04-18 2:45 ` Andreas Dilger
2011-04-18 2:45 ` Andreas Dilger
2011-04-19 1:58 ` Yongqiang Yang
2011-04-19 1:58 ` Yongqiang Yang
[not found] ` <BANLkTin=WEpSf6ddiOMNMOpCPP-wiEttSw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2011-04-19 2:59 ` Ted Ts'o
2011-04-19 2:59 ` Ted Ts'o
[not found] ` <20110419025949.GA3030-AKGzg7BKzIDYtjvyW6yDsg@public.gmane.org>
2011-04-19 3:05 ` Eric Sandeen
2011-04-19 3:05 ` Eric Sandeen
[not found] ` <4DACFBEB.9040909-+82itfer+wXR7s880joybQ@public.gmane.org>
2011-04-21 20:12 ` Jim Meyering
2011-04-21 20:12 ` Jim Meyering
2011-04-19 3:30 ` Yongqiang Yang
2011-04-19 3:30 ` Yongqiang Yang
2011-04-19 4:14 ` Dave Chinner
2011-04-19 4:14 ` Dave Chinner
2011-04-19 5:27 ` Christoph Hellwig
2011-04-19 5:27 ` Christoph Hellwig
2011-04-19 3:44 ` Dave Chinner
2011-04-19 3:44 ` Dave Chinner
2011-04-19 6:53 ` Yongqiang Yang
2011-04-19 6:53 ` Yongqiang Yang
2011-04-19 7:45 ` Dave Chinner
2011-04-19 7:45 ` Dave Chinner
2011-04-19 8:11 ` Yongqiang Yang
2011-04-19 8:11 ` Yongqiang Yang
2011-04-19 14:05 ` Eric Sandeen
2011-04-19 14:05 ` Eric Sandeen
2011-04-19 14:09 ` Ted Ts'o
2011-04-19 14:09 ` Ted Ts'o
2011-04-19 14:13 ` Eric Sandeen
2011-04-19 14:13 ` Eric Sandeen
2011-04-19 16:01 ` Ted Ts'o
2011-04-19 16:01 ` Ted Ts'o
2011-04-20 1:53 ` Yongqiang Yang
2011-04-20 1:53 ` Yongqiang Yang
2011-04-20 15:21 ` Christoph Hellwig
2011-04-20 15:21 ` Christoph Hellwig
2011-04-20 17:21 ` Ted Ts'o
2011-04-20 17:21 ` Ted Ts'o
[not found] ` <20110419140909.GD3030-AKGzg7BKzIDYtjvyW6yDsg@public.gmane.org>
2011-04-19 21:08 ` Dave Chinner
2011-04-19 21:08 ` Dave Chinner
2011-04-20 15:29 ` Christoph Hellwig
2011-04-20 15:29 ` Christoph Hellwig
2011-04-16 6:05 ` Yongqiang Yang
2011-04-16 6:05 ` Yongqiang Yang
2011-04-18 0:35 ` Dave Chinner
2011-04-18 0:35 ` Dave Chinner
2011-04-15 8:53 ` Jim Meyering
2011-04-15 8:53 ` Jim Meyering
2011-04-15 17:16 ` Christoph Hellwig
2011-04-15 17:16 ` Christoph Hellwig
[not found] ` <20110415171629.GA9088-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2011-04-15 17:24 ` Eric Blake
2011-04-15 17:24 ` Eric Blake
2011-04-15 17:26 ` Christoph Hellwig
2011-04-15 17:26 ` Christoph Hellwig
[not found] ` <20110415172603.GA20086-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2011-04-15 22:28 ` Andreas Dilger
2011-04-15 22:28 ` Andreas Dilger
2011-04-16 0:25 ` Dave Chinner
2011-04-16 0:25 ` Dave Chinner
2011-04-14 14:39 ` Eric Sandeen
[not found] ` <20110414102608.GA1678-tLCgZGx+iJ+kxVt8IV0GqQ@public.gmane.org>
2011-04-20 14:39 ` Jim Meyering
2011-04-20 14:39 ` Jim Meyering
[not found] ` <87d3khugv1.fsf-CybKA8TIZ99x3y/oJEDuiw@public.gmane.org>
2011-04-21 20:01 ` Jim Meyering
2011-04-21 20:01 ` Jim Meyering
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=4DA7258F.1080302@sandeen.net \
--to=sandeen@sandeen.net \
--cc=P@draigbrady.com \
--cc=coreutils@gnu.org \
--cc=linux-ext4@vger.kernel.org \
--cc=markus@trippelsdorf.de \
--cc=xfs@oss.sgi.com \
--cc=xiaoqiangnk@gmail.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.