http://blog.gmane.org/gmane.linux.busybox hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://comments.gmane.org/gmane.linux.busybox/37778 <pre>I haven't gotten any further feedback from any developers on my proposed UBI feature enhancement patches I resubmitted with Mike's requested changes a few weeks ago. If there's any developer who would like to commit them, that would be great. If there's any developer who would be willing to commit them with any other updates or changes, please let me know what you would like to see. Otherwise, I guess I will just keep them in my local fork and stop trying to contribute them upstream. Thanks… </pre> Paul B. Henson 2013-06-19T18:47:45 http://comments.gmane.org/gmane.linux.busybox/37775 <pre># HG changeset patch # User Thomas De Schampheleire &lt;thomas.de_schampheleire&lt; at &gt;alcatel-lucent.com&gt; # Date 1371558718 -7200 # Node ID 38fb4edf4bb684a9a14ded6c15181e9d490d83ef # Parent 77208178d3ef7ac035f10e5b6e0859be170ff788 lsof: correct check for symbolic link Busybox lsof used the d_type field of a 'struct dirent' to verify whether the entry is a symbolic link. This field, however, is not portable. On at least one board [1] I have seen, that field is 0, and the check fails even though the entry is a link. This patch replaces the check on d_type with a call to lstat and a subsequent check S_ISLNK. This should work in all cases, at the cost of the extra lstat call. [1] A MIPS-based board with glibc 2.9, Linux 2.6.32.27. Signed-off-by: Thomas De Schampheleire &lt;thomas.de.schampheleire&lt; at &gt;gmail.com&gt; --- procps/lsof.c | 9 +++++++-- 1 files changed, 7 insertions(+), 2 deletions(-) diff --git a/procps/lsof.c b/procps/lsof.c --- a/procps/lsof.c +++ b/procps/lsof.c &lt; at &gt;&lt; at &gt; -61,8 +61,13 &lt; at &gt;&lt; at &gt; int lsof_main(int argc UNUSE</pre> Thomas De Schampheleire 2013-06-19T16:52:51 http://comments.gmane.org/gmane.linux.busybox/37774 <pre>Hi all: I'm new to the busybox world, and am using it on an Android device for FTP get/put testing. I noticed there is a patch by Kev Kitchens that will support "EPSV" mode needed for IPv6 FTP transfers. [PATCH] networking/ftpgetput.c: Make ftp client code RFC compliant for IPv4 and IPv6 Will this patch be going into an official distribution soon? I'm hoping to avoid compiling a custom version using the patch, as I have no experience in that area, but I'll make plans to do so if IPv6 FTP will not be "officially" supported anytime soon. thanks for the advice, Mike Harrison </pre> Harrison, Michael J (Mike 2013-06-19T15:43:36 http://comments.gmane.org/gmane.linux.busybox/37765 <pre>Wondering. I don't have contact info for him other than email, and he vanished somewhat abruptly from the list. No git checkins since then either. Never responded to my ping about a #busybox channel on freenode (not that he ever did irc before)... Vacation, do you think? Rob </pre> Rob Landley 2013-06-15T06:04:24 http://comments.gmane.org/gmane.linux.busybox/37757 <pre>Hi, I'm building several different busybox 1.20.2 binaries with various applets, to be run onto a tiny MIPS embedded system. (I'm cross compiling with aboriginal system images and assorted cross tools). Busybox utils I build are setup from makeallnoconfig and then adding applets I want: they are therefore dynamically linked. Applets I built so far are things like gzip, wget, find, cat, echo, du, ... They all work as expected on my target. I decided to do tar applet, with all its options. It compiles, and work on my aboriginal system image, but fails on target with message "unknown file type". I tried with tar options off: same issue. I tried same config but static build: it works! However I find the error message strange would my target really miss a library... How to find which eventual one? Would adding another magic applet/option solve the issue? Any hint, suggestion to fix this, or alternative? As my embedded system is really constrained in storage, I'd rather use dynamically linked version obviou</pre> piquemal michel 2013-06-11T20:45:44 http://comments.gmane.org/gmane.linux.busybox/37754 <pre>Hello! I found that mount will not go to the background and retry mounting if mounting a NFS share with the "bg" options set fails with an error other than ECONNREFUSED. Why does BB mount only go to the background in that case? Is there a specific reason for that? My problem is, that my machines try to mount a NFS share when they're booting up. If the remote host or the network is unreachable during that time, mount will fail (which is why I use the bg option) and never retry (bad). I found this in BusyBox 1.14.4 but it's still present in the current snapshot. The problematic code is at (snapshot) util-linux/mount.c:1580. Out of my naivety: Shouldn't mount retry in case of *any* error? Regards, Andreas </pre> Andreas Wehrmann 2013-06-10T13:46:40 http://comments.gmane.org/gmane.linux.busybox/37753 <pre>In busybox/miscutils/taskset.c the function from_cpuset has following logic for big endian systems: --- #if BB_BIG_ENDIAN /* For big endian, it means LAST bits */ if (CPU_SETSIZE &lt; sizeof(long)) p += CPU_SETSIZE - sizeof(int); else if (CPU_SETSIZE &lt; sizeof(long long)) p += CPU_SETSIZE - sizeof(long); else p += CPU_SETSIZE - sizeof(long long); #endif --- First issue is that CPU_SETSIZE is in bits, while this calculation uses it as bytes, so we overwrite part of the stack from the calling function. Another issue on the platform I'm running it on, is that CPU_SETSIZE is 1024 bits (128 bytes). This logic here takes the last "long long" out of the structure, but the result is actually stored in the first few bytes of the structure. I'm not sure if all big endian systems follow this logic?! The "fancy feature" implementation on the other hand, works fine. </pre> Karl Vogel 2013-06-10T07:21:27 http://comments.gmane.org/gmane.linux.busybox/37720 <pre>Signed-off-by: Paul B. Henson &lt;henson&lt; at &gt;acm.org&gt; --- miscutils/ubi_tools.c | 33 ++++++++++++++++++++++++--------- 1 files changed, 24 insertions(+), 9 deletions(-) diff --git a/miscutils/ubi_tools.c b/miscutils/ubi_tools.c index f09be8c..6b5d384 100644 --- a/miscutils/ubi_tools.c +++ b/miscutils/ubi_tools.c &lt; at &gt;&lt; at &gt; -103,7 +103,8 &lt; at &gt;&lt; at &gt; //usage: "\n-n VOLIDVolume ID, if not specified, it" //usage: "\nwill be assigned automatically" //usage: "\n-N NAMEVolume name" -//usage: "\n-s SIZESize in bytes" +//usage: "\n-s SIZESize in bytes, kilobytes (KiB)," +//usage: "\nmegabytes (MiB), or gigabytes (GiB)" //usage: "\n-t TYPEVolume type (static|dynamic)" //usage: //usage:#define ubirmvol_trivial_usage &lt; at &gt;&lt; at &gt; -117,15 +118,23 &lt; at &gt;&lt; at &gt; //usage:#define ubirsvol_full_usage "\n\n" //usage: "Resize UBI volume\n" //usage: "\n-n VOLIDVolume ID to resize" -//usage: "\n-s SIZESize in bytes" +//usage: "\n-s SIZESize in bytes, kilobytes (KiB)," +//usage: "\n</pre> Paul B. Henson 2013-06-03T01:11:10 http://comments.gmane.org/gmane.linux.busybox/37718 <pre>Signed-off-by: Paul B. Henson &lt;henson&lt; at &gt;acm.org&gt; --- miscutils/ubi_tools.c | 51 +++++++++++++++++++++++++++++++++++++++++++----- 1 files changed, 45 insertions(+), 6 deletions(-) diff --git a/miscutils/ubi_tools.c b/miscutils/ubi_tools.c index f6dfcae..f09be8c 100644 --- a/miscutils/ubi_tools.c +++ b/miscutils/ubi_tools.c &lt; at &gt;&lt; at &gt; -95,10 +95,11 &lt; at &gt;&lt; at &gt; //usage: "\n-d UBI_NUMUBI device number" //usage: //usage:#define ubimkvol_trivial_usage -//usage: "UBI_DEVICE -N NAME -s SIZE" +//usage: "UBI_DEVICE -N NAME [-s SIZE | -m]" //usage:#define ubimkvol_full_usage "\n\n" //usage: "Create UBI volume\n" //usage: "\n-a ALIGNMENTVolume alignment (default 1)" +//usage: "\n-mSet volume size to maximum available" //usage: "\n-n VOLIDVolume ID, if not specified, it" //usage: "\nwill be assigned automatically" //usage: "\n-N NAMEVolume name" &lt; at &gt;&lt; at &gt; -141,11 +142,19 &lt; at &gt;&lt; at &gt; int ubi_tools_main(int argc UNUSED_PARAM, char **argv) int alignment = 1; char *type = NULL; -opt</pre> Paul B. Henson 2013-06-03T00:48:05 http://comments.gmane.org/gmane.linux.busybox/37716 <pre>Signed-off-by: Paul B. Henson &lt;henson&lt; at &gt;acm.org&gt; --- miscutils/ubi_tools.c | 20 ++++++++++++++------ 1 files changed, 14 insertions(+), 6 deletions(-) diff --git a/miscutils/ubi_tools.c b/miscutils/ubi_tools.c index dd99a44..f6dfcae 100644 --- a/miscutils/ubi_tools.c +++ b/miscutils/ubi_tools.c &lt; at &gt;&lt; at &gt; -234,16 +234,24 &lt; at &gt;&lt; at &gt; int ubi_tools_main(int argc UNUSED_PARAM, char **argv) ssize_t len; char *input_data; -// Make assumption that device not is in normal format. -// Removes need for scanning sysfs tree as full libubi does -if (sscanf(ubi_ctrl, "/dev/ubi%u_%u", &amp;ubinum, &amp;volnum) != 2) +// Make assumption that device node is in normal format +// (/dev/ubi%d_%d), removes need for scanning sysfs +// tree as full libubi does +if (strncmp(ubi_ctrl, "/dev/ubi", 8) != 0 || + (input_data = strchr(ubi_ctrl+8, '_')) == NULL) bb_error_msg_and_die("%s volume node not in correct format", "UBI"); +// temporarily tokenize for parsing +*input_data = '\0'; +ubinum = xstrtou(ubi_</pre> Paul B. Henson 2013-06-02T19:49:49 http://comments.gmane.org/gmane.linux.busybox/37714 <pre>Signed-off-by: Paul B. Henson &lt;henson&lt; at &gt;acm.org&gt; --- miscutils/ubi_tools.c | 17 ++++++++++++----- 1 files changed, 12 insertions(+), 5 deletions(-) diff --git a/miscutils/ubi_tools.c b/miscutils/ubi_tools.c index dd99a44..9e3db1e 100644 --- a/miscutils/ubi_tools.c +++ b/miscutils/ubi_tools.c &lt; at &gt;&lt; at &gt; -234,16 +234,23 &lt; at &gt;&lt; at &gt; int ubi_tools_main(int argc UNUSED_PARAM, char **argv) ssize_t len; char *input_data; -// Make assumption that device not is in normal format. -// Removes need for scanning sysfs tree as full libubi does -if (sscanf(ubi_ctrl, "/dev/ubi%u_%u", &amp;ubinum, &amp;volnum) != 2) +// Make assumption that device node is in normal format +// (/dev/ubi%d_%d), removes need for scanning sysfs +// tree as full libubi does +if (strncmp(ubi_ctrl, "/dev/ubi", 8) != 0 || + (input_data = strchr(ubi_ctrl+8, '_')) == NULL) bb_error_msg_and_die("%s volume node not in correct format", "UBI"); +// temporarily tokenize for parsing +*input_data = '\0'; +ubinum = xstrtou(ubi_ctr</pre> Paul B. Henson 2013-06-02T19:49:49 http://comments.gmane.org/gmane.linux.busybox/37705 <pre>man.config can define lines such as DEFINE pager less -s this patch parses them, additionally it always sets pager to less instead of more. Signed-off-by: John Spencer &lt;maillist-busybox&lt; at &gt;barfooze.de&gt; _______________________________________________ busybox mailing list busybox&lt; at &gt;busybox.net http://lists.busybox.net/mailman/listinfo/busybox</pre> John Spencer 2013-06-02T10:51:21 http://comments.gmane.org/gmane.linux.busybox/37704 <pre>the posix manpages ( http://www.kernel.org/pub/linux/docs/man-pages/man-pages-posix/man-pages-posix-2003-a.tar.xz ) are placed in man0p man1p and man3p. Signed-off-by: John Spencer &lt;maillist-busybox&lt; at &gt;barfooze.de&gt; _______________________________________________ busybox mailing list busybox&lt; at &gt;busybox.net http://lists.busybox.net/mailman/listinfo/busybox</pre> John Spencer 2013-06-02T11:10:38 http://comments.gmane.org/gmane.linux.busybox/37700 <pre>Hi! I use busybox 1.19.4 on OpenWRT 12.09 and noticed that search and the -M options do not work. I guess they have been not compiled in, to save space. Can anyone confirm, that in such case the help text (less -h) still shows those options (-M and -I) ? If so, please consider removing them together with the actual function when configured so. Regards, David </pre> David Balažic 2013-06-01T13:08:25 http://comments.gmane.org/gmane.linux.busybox/37692 <pre>Hi, I noticed that thread names are truncated by one character (to 14 characters instead of 15), the below patch fixes the issue. (The attached patch has proper whitespaces !) --- busybox-1.21.0/procps/pstree.c.orig 2013-06-01 02:27:14.008530872 +0200 +++ busybox-1.21.0/procps/pstree.c 2013-06-01 02:31:36.483537110 +0200 &lt; at &gt;&lt; at &gt; -342,7 +342,7 &lt; at &gt;&lt; at &gt; static void handle_thread(const char *comm, pid_t pid, pid_t ppid, uid_t uid) { char threadname[COMM_LEN + 2]; - sprintf(threadname, "{%.*s}", COMM_LEN - 2, comm); + sprintf(threadname, "{%.*s}", COMM_LEN - 1, comm); add_proc(threadname, pid, ppid, uid/*, 1*/); } #endif </pre> Dag Wieers 2013-06-01T00:34:16 http://comments.gmane.org/gmane.linux.busybox/37688 <pre>Signed-off-by: Paul B. Henson &lt;henson&lt; at &gt;acm.org&gt; --- miscutils/ubi_tools.c | 57 +++++++++++++++++++++++++++++++++++++++++------- 1 files changed, 48 insertions(+), 9 deletions(-) diff --git a/miscutils/ubi_tools.c b/miscutils/ubi_tools.c index 6cb7cd5..87dfab0 100644 --- a/miscutils/ubi_tools.c +++ b/miscutils/ubi_tools.c &lt; at &gt;&lt; at &gt; -103,7 +103,8 &lt; at &gt;&lt; at &gt; //usage: "\n-n VOLIDVolume ID, if not specified, it" //usage: "\nwill be assigned automatically" //usage: "\n-N NAMEVolume name" -//usage: "\n-s SIZESize in bytes" +//usage: "\n-s SIZESize in bytes, kilobytes (KiB)," +//usage: "\nmegabytes (MiB), or gigabytes (GiB)" //usage: "\n-t TYPEVolume type (static|dynamic)" //usage: //usage:#define ubirmvol_trivial_usage &lt; at &gt;&lt; at &gt; -117,15 +118,45 &lt; at &gt;&lt; at &gt; //usage:#define ubirsvol_full_usage "\n\n" //usage: "Resize UBI volume\n" //usage: "\n-n VOLIDVolume ID to resize" -//usage: "\n-s SIZESize in bytes" +//usage: "\n-s SIZESize in bytes, kilobytes (KiB)," +//u</pre> Paul B. Henson 2013-05-29T01:17:42 http://comments.gmane.org/gmane.linux.busybox/37687 <pre>Signed-off-by: Paul B. Henson &lt;henson&lt; at &gt;acm.org&gt; --- miscutils/ubi_tools.c | 45 +++++++++++++++++++++++++++++++++++++++------ 1 files changed, 39 insertions(+), 6 deletions(-) diff --git a/miscutils/ubi_tools.c b/miscutils/ubi_tools.c index dd99a44..6cb7cd5 100644 --- a/miscutils/ubi_tools.c +++ b/miscutils/ubi_tools.c &lt; at &gt;&lt; at &gt; -95,10 +95,11 &lt; at &gt;&lt; at &gt; //usage: "\n-d UBI_NUMUBI device number" //usage: //usage:#define ubimkvol_trivial_usage -//usage: "UBI_DEVICE -N NAME -s SIZE" +//usage: "UBI_DEVICE -N NAME [-s SIZE | -m]" //usage:#define ubimkvol_full_usage "\n\n" //usage: "Create UBI volume\n" //usage: "\n-a ALIGNMENTVolume alignment (default 1)" +//usage: "\n-mSet volume size to maximum available" //usage: "\n-n VOLIDVolume ID, if not specified, it" //usage: "\nwill be assigned automatically" //usage: "\n-N NAMEVolume name" &lt; at &gt;&lt; at &gt; -141,11 +142,19 &lt; at &gt;&lt; at &gt; int ubi_tools_main(int argc UNUSED_PARAM, char **argv) int alignment = 1; char *type = NULL; -opt_co</pre> Paul B. Henson 2013-05-28T21:35:29 http://comments.gmane.org/gmane.linux.busybox/37684 <pre>People have been gradually wandering from #uclibc to #busybox on freenode (due ot the incessant git-commit spam from the bot on uclibc about buildroot commits and such, which made everybody stop checking the tab). I finally got around to registering the channel. Denys: glad to op your freenode handle if you'll tell me what it is. Up to you if you want to mention #busybox on irc.freenode.net on the web page... Anyway, thing. (Back to banging on toybox...) Rob </pre> Rob Landley 2013-05-31T17:21:27 http://comments.gmane.org/gmane.linux.busybox/37683 <pre>I haven't seen any feedback on the potential ubi feature additions I sent to the list last weekend, is there any other policy/procedure I should follow for submitting them for consideration other than just posting them to the list? Thanks much… </pre> Paul B. Henson 2013-05-31T19:54:32 http://comments.gmane.org/gmane.linux.busybox/37680 <pre>I've just had a really pleasant experience cross compiling busybox for our embedded PowerPC platform. It reminded me how you can point C compilers at C source and get the right output. It even seemed to do something sensible about libm (temporarily being) missing. Thank you for making this work and please don't let anyone tell you to start using GNU autoconf and friends. They're absolute crap for most packages when cross building. Jon </pre> Jon Schneider 2013-05-31T14:43:17 http://comments.gmane.org/gmane.linux.busybox/37679 <pre>Conceptually, the find utility need not perform lstat on each filename unless it's needed for matching criteria. However, find is implemented based on libbb's recursive_action, which always performs stat or lstat. This makes busybox's find excruciatingly slow compared to GNU find. I think the easiest way to fix this is to add an extra argument to recursive_action so that the dirent d_type field can be passed, with fallback code to call [l]stat when d_type is blank or the flags are such that d_type is insufficient to determine the preferred action (i.e. when the type is a symlink and symlinks are to be followed). There may also be solutions that involve just trying an operation (e.g. opendir), but I don't think these approaches handle the full range of flags supported by recursive_action. Comments? Anyone interested in looking into this? Rich </pre> Rich Felker 2013-05-29T04:12:40 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.linux.busybox