[patch] various bugs and strncpy abuse followup
Bernhard Fischer
rep.nop at aon.at
Thu Jun 8 12:49:13 PDT 2006
On Thu, Jun 08, 2006 at 12:03:57AM +0200, Bernhard Fischer wrote:
>On Tue, Jun 06, 2006 at 03:34:58PM -0700, Erik Hovland wrote:
>>After the very helpful feedback and the continual searching for bugs, I
>>have expanded my patch associated to various bugs all over busybox.
>>
>>This patchset has gotten bigger. So I have taken to splitting them up by
>>file and annotating them in patch. I hope that is useful. Let me know if
>>there is some way I can make this easier to go over or help with
>>acceptance.
>>
>>Feel free to address each patch seperately in responding. I would not be
>>surprised at all if I accidentally introduced regressions. I have not
>>yet tried running the unit testing. But I have attempted to understand
>>why each bug is a bug and addressed it according to its severity.
>>
>>I appreciate any attention you might give this work.
>
>It would be nice if you'd use (GNU)diff's -p if you send a batch
>>E
>>
>> archival/dpkg_deb.c | 6 +++--
>we have xchdir()
>and see the recent discussion about the retval
>
>> archival/rpm.c | 3 +-
>xchdir()
>
>> archival/tar.c | 8 +++----
>Sounds ok except that perhaps the two memcpy add to the size needlessly,
>if compared to just one. Didn't look if that's true
>
>> console-tools/loadkmap.c | 3 +-
>a bold return(1) should do
>
>> coreutils/chown.c | 4 ++-
>Sounds ok
>
>> e2fsprogs/ext2fs/ext_attr.c | 5 +++-
>> e2fsprogs/ext2fs/flushb.c | 3 +-
>> e2fsprogs/ext2fs/mkjournal.c | 3 +-
>> e2fsprogs/ext2fs/read_bb_file.c | 2 -
>> e2fsprogs/mke2fs.c | 8 ++++---
>> e2fsprogs/tune2fs.c | 4 +--
>skipped.
>
>> init/init.c | 10 +++++++--
>msg is far too long (not sure if checking the ioctl is useful at all
>there)
>xchdir() if you really feel like.
>
>I'll look at the rest later.
>
>Let me note that it would be very nice if you could use diff -p if your
>diff supports it and also that a size bar.o.orig bar.o would be nice to
>see how much it adds.
ok?
Let's continue:
>> libbb/concat_subpath_file.c | 2 +
concat_path_file() does already check for path==0
The check for f==0 in concat_subpath_file() you added is redundant.
>> libbb/login.c | 5 +++-
Must have missed that patch.
>> libbb/loop.c | 4 ++-
landley knows this code better.
>> loginutils/getty.c | 9 ++++----
>> loginutils/login.c | 15 +++----------
Both look ok, but i'm not familiar with them.
>> loginutils/vlock.c | 19 +++++++++++------
Not sure how useful these are.
>> miscutils/crontab.c | 2 -
sounds ok.
>> miscutils/devfsd.c | 14 +++++++------
skipped.
>> networking/ifupdown.c | 12 -----------
Sounds ok but i'm not familiar with it (yet, i'll have to look at it
sometimes, i fear)
>> networking/libiproute/iplink.c | 42 ++-------------------------------------
Sounds reasonable.
>> networking/libiproute/iproute.c | 12 ++++++-----
ok.
>> networking/libiproute/iptunnel.c | 2 -
Sounds ok.
>> networking/tftp.c | 25 +++++++++++++----------
use bb_xbind(). The cmd_put is defined to 0 in that case, to it's ok to
leave that part in (drop that part of your patch).
>> networking/udhcp/socket.c | 2 -
Sounds reasonable, but i didn't look.
>> networking/zcip.c | 2 -
ditto.
>> procps/fuser.c | 3 +-
xchdir()
>> shell/lash.c | 2 -
Sounds reasonable.
>> shell/msh.c | 6 ++++-
use bb_dev_null
>> util-linux/fdisk.c | 2 -
I'll leave this to an SGI user..
>> util-linux/mdev.c | 3 +-
Well, the comment says pretty all already. xstat() if you feel like
>> util-linux/nfsmount.c | 2 -
Sounds reasonable, but i didn't look.
>> 33 files changed, 118 insertions(+), 126 deletions(-)
More information about the busybox
mailing list