If buffer was allocated linearly, hard to detect invalid buffer access.
So, don't use preallocation if BUFFER_PARANOIA_DEBUG is defined.
---
user/test/buffer.c | 31 +--
1 file changed, 29 insertions(+), 2 deletions(-)
diff -puN user/test/buffer.c~buffer-debug
can handle unaligned accesses on all
architectures.
I see. BTW, Linux has macros for it - get_unaligned_be*/_le*.
--
OGAWA Hirofumi [EMAIL PROTECTED]
___
Tux3 mailing list
Tux3@tux3.org
http://tux3.org/cgi-bin/mailman/listinfo/tux3
Daniel Phillips [EMAIL PROTECTED] writes:
On Friday 17 October 2008 02:59, OGAWA Hirofumi wrote:
Daniel Phillips [EMAIL PROTECTED] writes:
-struct ileaf { u16 magic, count; inum_t ibase; char table[]; };
+struct ileaf { u16 magic, count; u32 pad; inum_t ibase; char table
---
user/test/inode.c |3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff -puN user/test/inode.c~inode-purge_inum-fix user/test/inode.c
--- tux3/user/test/inode.c~inode-purge_inum-fix 2008-10-16 01:20:44.0
+0900
+++ tux3-hirofumi/user/test/inode.c 2008-10-16
. Looks good.
--
OGAWA Hirofumi [EMAIL PROTECTED]
___
Tux3 mailing list
Tux3@tux3.org
http://tux3.org/cgi-bin/mailman/listinfo/tux3
is supposed to generate code to access
the fields bytewise on architectures that generate alignment faults,
when the compiler can't prove statically that the fields are aligned.
I could be completely wrong about that...
Um.. We will need to check recent gcc...
--
OGAWA Hirofumi [EMAIL PROTECTED
---
user/test/ileaf.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff -puN user/test/ileaf.c~ileaf-split-use-entries_per_leaf user/test/ileaf.c
--- tux3/user/test/ileaf.c~ileaf-split-use-entries_per_leaf 2008-10-16
01:20:43.0 +0900
+++
---
user/test/btree.c |1 +
user/test/filemap.c |2 ++
2 files changed, 3 insertions(+)
diff -puN user/test/filemap.c~filemap-dtree-dirty user/test/filemap.c
--- tux3/user/test/filemap.c~filemap-dtree-dirty2008-10-16
01:20:45.0 +0900
+++
Let's use extent for local value.
---
user/test/dleaf.c |2 +-
user/test/filemap.c |6 +++---
2 files changed, 4 insertions(+), 4 deletions(-)
diff -puN user/test/dleaf.c~extent-rename user/test/dleaf.c
--- tux3/user/test/dleaf.c~extent-rename2008-10-16 01:20:45.0
]));
index += extent_count(seg[i]);
}
printf( (%i)\n, segs);
Whoops, sorry about that. gcc-4.3.2 didn't warn those. If warning flags
is bothering us, we can just reject that patch.
--
OGAWA Hirofumi [EMAIL PROTECTED]
___
Tux3 mailing list
Tux3
be wrong, or
there may be more good fix.
--
OGAWA Hirofumi [EMAIL PROTECTED]
Yes, the change look correct to me. I merged all your patches,
Thanks.
inserting this one between 2 and 3 and all the tests passed after each
one. I also now have a semi-automated system for testing patches and
patch
# HG changeset patch
# User OGAWA Hirofumi [EMAIL PROTECTED]
# Date 1226380667 -32400
# Node ID 24f3b0f20899b2c5ad5554b7512f647f75c5d3be
# Parent dab895e2e896189f1764a7ec3330b6473757138e
Use tuxtime() to update timestamp
And this moves time functions to tux3.h to use those without including
# HG changeset patch
# User OGAWA Hirofumi [EMAIL PROTECTED]
# Date 1226380749 -32400
# Node ID daac4717e4cc70df5efde5bf82df1841dcc99537
# Parent 55013c67b722e7b218c348f7cda68c766c62fd31
Add billionths() and use it, instead of millionths()
diff -r 55013c67b722 -r daac4717e4cc user/tux3.h
# HG changeset patch
# User OGAWA Hirofumi [EMAIL PROTECTED]
# Date 1226387681 -32400
# Node ID 6b615de4d62bea052bb57474e754233f911edf1e
# Parent dab895e2e896189f1764a7ec3330b6473757138e
Use tuxtime() to update timestamp
And this moves time functions to tux3.h to use those without including
# HG changeset patch
# User OGAWA Hirofumi [EMAIL PROTECTED]
# Date 1226387746 -32400
# Node ID d680e7bc61353f6c4de5ef24e30b4f5d8245150d
# Parent 6b615de4d62bea052bb57474e754233f911edf1e
Convert more timestamp to high resolution in tux3fuse
diff -r 6b615de4d62b -r d680e7bc6135 user/tux3fuse.c
# HG changeset patch
# User OGAWA Hirofumi [EMAIL PROTECTED]
# Date 1226387771 -32400
# Node ID e50b06dfcea7f1d56578e12e91d50e5de47983cc
# Parent d680e7bc61353f6c4de5ef24e30b4f5d8245150d
Add billionths() and use it, instead of millionths()
diff -r d680e7bc6135 -r e50b06dfcea7 user/tux3.h
:1b9f30bfd9043fc008e6ad32141001144306089f
tag: tip
user:OGAWA Hirofumi [EMAIL PROTECTED]
date:Thu Nov 13 20:56:30 2008 +0900
files: user/tux3graph.c
description:
tux3graph update to current format
The order of fields in struct extent was changed, use same order.
changeset: 420
Daniel Phillips [EMAIL PROTECTED] writes:
On Wednesday 19 November 2008 04:17, OGAWA Hirofumi wrote:
Unfortunately, quick search was right. SLUB didn't warn by
test. (probably, I should send a patch to lkml) So, I tweaked malloc()
like the following.
might_sleep() is only for CONFIG_SLUB
Daniel Phillips [EMAIL PROTECTED] writes:
Hi Hirofumi,
Hi
This patch has some code from user/inode.c that also belongs in kernel.
It has a lot of minor problems still, please take a look at it if you
wish.
Thanks. I'll tackle based on this.
--
OGAWA Hirofumi [EMAIL PROTECTED
)
This functions may rewrite in kernel completely, I'm not sure. Well,
this difference wouldn't matter at least for now.
Thanks.
--
OGAWA Hirofumi [EMAIL PROTECTED]
___
Tux3 mailing list
Tux3@tux3.org
http://tux3.org/cgi-bin/mailman/listinfo/tux3
, and maybe forever as they evolve and develop better
interfaces.
Ok, I'll move it to inside __KERNEL__.
Is there are reason to change tux_path[] to *tux_path?
We are allocate it by alloc_path(), so it's not array actually now.
--
OGAWA Hirofumi [EMAIL PROTECTED
memory. Well, much work is remaining, enjoy.
--
OGAWA Hirofumi [EMAIL PROTECTED]
___
Tux3 mailing list
Tux3@tux3.org
http://tux3.org/cgi-bin/mailman/listinfo/tux3
, e.g. kmap_atomic()
or something? Or we use GFP_USERS or something (at least wouldn't use
for file data)? At least I was not thinking about it.
--
OGAWA Hirofumi [EMAIL PROTECTED]
___
Tux3 mailing list
Tux3@tux3.org
http://tux3.org/cgi-bin/mailman/listinfo
, otherwise the user can open it via cached_lookup() without
inode-i_op-lookup().
So, I think it's necessary.
Maybe, test case is:
$ echo 111 test.txt
$ sleep 30 test.txt
$ rm test.txt
$ cat test.txt
Thanks.
--
OGAWA Hirofumi [EMAIL PROTECTED
://userweb.kernel.org/~hirofumi/tux3-new.png
Please check it.
--
OGAWA Hirofumi [EMAIL PROTECTED]
___
Tux3 mailing list
Tux3@tux3.org
http://tux3.org/cgi-bin/mailman/listinfo/tux3
OGAWA Hirofumi [EMAIL PROTECTED] writes:
Those patches are trivial cleanup, and tux3graph improvement. Well, the
output of tux3graph is not changed much, however, perhaps the
preparation to dump the data which is pointed by extents is done almost.
static-http://userweb.kernel.org
png. Can we make the svg support depend on the graphviz version?
Yes. However, graphviz seems to be already supporting svg at 2004.
$ tux3graph [-v] volname
$ dot -Tsvg volname.dot foo.svg
$ viewer foo.svg
We can use -Tsvg, instead of -Tpng.
--
OGAWA Hirofumi [EMAIL
)
return -(long)(ptr - error_base);
IS_ERR(ptr)
return (error_base ptr ptr = error_base + MAX_ERRNO);
--
OGAWA Hirofumi [EMAIL PROTECTED]
___
Tux3 mailing list
Tux3@tux3.org
http://tux3.org/cgi-bin/mailman/listinfo/tux3
/scripts/checkpatch.pl /path/to/your/foo.patch
or
linux/scripts/checkpatch.pl --file /path/to/your/foo.c
--
OGAWA Hirofumi [EMAIL PROTECTED]
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
) {
+ if (d_negative(nd-path.dentry)) {
path_put(old_path);
return 1;
}
Although I'm not seeing whole path, before I forget, this should be
!d_negative(nd-path.dentry).
--
OGAWA Hirofumi [EMAIL PROTECTED
to be fixed soon or later.
static-http://userweb.kernel.org/~hirofumi/tux3/
Please review it.
--
OGAWA Hirofumi [EMAIL PROTECTED]
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
of inode. We have to
provide own -get_name() handler.
Thanks.
--
OGAWA Hirofumi [EMAIL PROTECTED]
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
Daniel Phillips phill...@phunq.net writes:
On Thursday 11 December 2008 11:20, OGAWA Hirofumi wrote:
static-http://userweb.kernel.org/~hirofumi/tux3/
I upgraded from mercurial 0.9.1 to 1.0.1 to get some of the fancy new
extensions, and was not able to pull/clone your repository any more
and last sector to zero?
Ah, I was forgetting about it. I think, first 1024bytes (FAT, NTFS, and EFI),
and last 64kbytes (EFI, and MD RAID), should be enough. Right?
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http
viewer datafile.svg
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
, mkfs clears first 1024b (or 4096b) and last 64kb, at least
for now, it is enough?
Thanks.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
become the whole volume or
at least more blocks.
E.g. why do we care only reiserfs? In the same reason, we would want to
clear blocks on any offset which is using storage? E.g. ntfs would use
the middle of volume. Ugh.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
OGAWA Hirofumi hirof...@mail.parknet.co.jp writes:
I can see what are you saying, and of course, it is reasonable. But, if
we use that rule perfectly, the cleaner would become the whole volume or
at least more blocks.
E.g. why do we care only reiserfs? In the same reason, we would want
reiserfs and NTFS. Because tux3
put superblock at 4096 offset for now, and other blocks can be used for
any purpose.
So I think, mkfs clears first 1024b (or 4096b) and last 64kb, at least
for now, it is enough?
Thanks.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
OGAWA Hirofumi hirof...@mail.parknet.co.jp writes:
Anyway, please apply this patch instead.
That is incremental patch of attached patch.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi
the following instead.
Does the following patch work for your gcc version too?
Thanks.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
diff -puN user/kernel/dleaf.c~dleaf-cleanup user/kernel/dleaf.c
--- tux3/user/kernel/dleaf.c~dleaf-cleanup 2008-12-31 03:17:33.0
+0900
+++ tux3-hirofumi
guess it shouldn't be hard.
Thanks.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
after 2.6.28. And we shouldn't add needless #ifdef.
[BTW, if we really want to do this, we should add compat layer like
compat.h of external network drivers.]
Thanks.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
Cleanup/Fix COW credential
COW credential was added after 2.6.28. And if we really
didn't break anything. Your improvement is also fine. I
would like to keep the #if on VERSION.
Oh. I'll just queue the patches for lastest kernel until tux3.git.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http
OGAWA Hirofumi hirof...@mail.parknet.co.jp writes:
Well, this make simple sb/inode user, on stack sb/inode is setted up by
macros, so we can avoid strace thing by unsetup sb/inode.
And, with this, we can share tux_new_volmap() in both of userland and
kernel.
static-http
OGAWA Hirofumi hirof...@mail.parknet.co.jp writes:
Just demonstration though for now. With this hack, volmap/tux_vol_apos
seem to work.
diff -puN user/kernel/inode.c~vol_bread user/kernel/inode.c
--- tux3/user/kernel/inode.c~vol_bread 2009-01-10 00:24:46.0 +0900
+++ tux3-hirofumi/user
: not very much. It is looking like a good decision.
Yes, I'm running fsx-linux from 2:42 am in JST (Now, 4:32 am).
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
OGAWA Hirofumi hirof...@mail.parknet.co.jp writes:
Daniel Phillips phill...@phunq.net writes:
One thing I was trying to do is keep buffer.c from knowing details of
Tux3, such as inode format. Mainly so that it could be used for some
other project, the same way I borrowed it from ddsnap. I
Hi,
This is mainly for debugging buffer state, and cleanup for it. I think
we will play with buffer soon, to get bugs more early, I hope this is
useful for it.
static-http://userweb.kernel.org/~hirofumi/tux3/
Please review.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
see. Sorry, I didn't notice it.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
Hi,
This start to add redirect support to map_region(). It is not big
change though.
And others is unrelated changes almost. It cleans userland build, and
improves debugging.
static-http://userweb.kernel.org/~hirofumi/tux3/
Please review.
Thanks.
--
OGAWA Hirofumi hirof
OGAWA Hirofumi hirof...@mail.parknet.co.jp writes:
This start to add redirect support to map_region(). It is not big
change though.
And others is unrelated changes almost. It cleans userland build, and
improves debugging.
static-http://userweb.kernel.org/~hirofumi/tux3/
This fixes
soon. Well, anyway, this is
for someone want to build (uml) kernel for tux3.
Signed-off-by: OGAWA Hirofumi hirof...@mail.parknet.co.jp
---
fs/Kconfig |1 +
fs/Makefile |1 +
2 files changed, 2 insertions(+)
diff -puN fs/Kconfig~tux3 fs/Kconfig
--- tux3fs-2.6/fs/Kconfig~tux3 2009-01-27 12
, and please pull if it's ok.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
/~hirofumi/tux3/
I'm not reviewing those by myself, it is for the review. So, please
don't pull it yet. However, with this patchset, it passed the fsstress
test for several hours.
Thanks.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing
OGAWA Hirofumi hirof...@mail.parknet.co.jp writes:
Hi,
I've found several bugs in kernel/ileaf.c. Even if all inodes on ileaf
was removed, we don't remove ileaf itself, at least for now. So, we
have to handle empty ileaf, but in that case, some places is not
handling it correctly
changing interface for it. So, crypto stuff may be good one.
BTW, this issue has git too. So git is using the openssl one and MPL
replacement (from mozilla).
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http
tux3_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
_
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
), \
+ .pinned = LIST_HEAD_INIT((sb).pinned)
#define rapid_open_inode(sb, io, mode, init_defs...) ({\
struct inode *__inode = (struct inode){}; \
_
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
,
LOG_REDIRECT };
struct commit_entry { be_u64 previous; };
_
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
*argv[])
goto eek;
}
+ /* sb-bitmap is always dirty */
+ invalidate_buffers(sb-bitmap-map);
//printf( show state \n);
//show_buffers(sb-rootdir-map);
//show_buffers(sb-volmap-map);
_
--
OGAWA Hirofumi hirof
Signed-off-by: Daniel Phillips tux3@tux3.org
I guess it is not good. Especially Signed-off-by, tux3@tux3.org is
mailing-list address, not yourself.
Thanks.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3
Andrew Morton a...@linux-foundation.org writes:
- When 'bh' is known to be non-NULL, use put_bh() rather than brelse().
It sounds strange. Almost all bh is non-NULL. This means we are going to
replace almost all brelse() by put_bh()?
Thanks.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
, people join
to develop those, and see the progress? Or people just want to see the
result of various work later? ...
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
it, I guess (long long) will bother devlopers.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
transactions to that job]
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
Samuel Harrington robot...@gmail.com writes:
On Tue, May 5, 2009 at 12:49 PM, Martin Steigerwald mar...@lichtvoll.de
wrote:
Am Dienstag 05 Mai 2009 schrieb OGAWA Hirofumi:
http://userweb.kernel.org/~hirofumi/note.defree-logs
Hmmm, the URL can't be found on the server. Thats at least
OGAWA Hirofumi hirof...@mail.parknet.co.jp writes:
But, it is not true. So, this patch fixes the cursor position only if
splited new block has target child.
This means the following,
But, it is not true. So, this patch fixes the cursor position by
changing the position only if splited new
Hi,
I've picked the bug-fix patches up from my patchset. And this changes is
including the patch for packaging and reported bugs from Mario
Fetka. Thanks Mario.
static-http://userweb.kernel.org/~hirofumi/tux3/
Please review, and pull if ok.
Thanks.
--
OGAWA Hirofumi hirof
not a great idea. In tux3.c, it is a (minor) bug that should
be cleaned up, even though we have to do #define assert(expr) nothing to
trigger it.
Yes, I agree. And I think it would also be good to use #include
assert.h with the patch of fixing assert() usage.
Thanks.
--
OGAWA Hirofumi hirof
or less.
static-http://userweb.kernel.org/~hirofumi/tux3/
Please review, and pull if ok.
Thanks.
--
OGAWA Hirofumi hirof...@mail.parknet.co.jp
___
Tux3 mailing list
Tux3@tux3.org
http://mailman.tux3.org/cgi-bin/mailman/listinfo/tux3
71 matches
Mail list logo