<html> <head> <meta content="text/html; charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#FFFFFF" text="#000000"> <div class="moz-cite-prefix"> Thank you. Is that patch going to be added to the main source code base? I think I found an error (segfault), see below. ==> I have a few more improvement suggestions. * IMHO vgs/lvs should include a column called seg_size_pe, similar to seg_start and seg_start_pe * IMHO vgs/lvs/pvs should have an command line parameter, or option modifier allowing to not display empty columns, i.e. the following command might skip displaying origin, snap%, move, log, copy%, convert, because they have no values. That would allow me to show copy_percent column only when copy is in progress. Would be nice. <tt>lvs --all -o +vg_extent_size</tt><tt> </tt><tt> LV VG Attr LSize Origin Snap% Move Log Copy% Convert Ext </tt><tt> </tt><tt> boot volume -wi-ao 1.00g 4.00m</tt><tt> </tt><tt> copy volume -wi-ao 2.00t 4.00m</tt><tt> </tt><tt> emulate volume -wi-ao 50.00g 4.00m</tt><tt> </tt><tt> home volume -wi-ao 150.00g 4.00m</tt><tt> </tt><tt> p2p volume -wi-a- 300.00g 4.00m</tt><tt> </tt><tt> root volume -wi-ao 4.00g 4.00m</tt><tt> </tt><tt> swap volume -wi-ao 8.00g 4.00m</tt><tt> </tt><tt> tmp volume -wi-ao 2.00g 4.00m</tt><tt> </tt><tt> usr volume -wi-ao 20.00g 4.00m</tt><tt> </tt><tt> var volume -wi-ao 20.00g 4.00m</tt> Possible error: <tt>root@localhost# lvs --rows --align</tt><tt> </tt><tt>Does not align, should it?</tt><tt> </tt><tt> </tt>==> SEGFAULT INFO <tt>root@localhost# pvs --all -o +vg_fmt</tt><tt> </tt><tt>Segmentation fault (core dumped)</tt><tt> </tt><tt> However the following seems to work..... root@localhost# pvs --all -o +vg_name</tt><tt> </tt><tt> PV VG Fmt Attr PSize PFree VG </tt><tt> </tt><tt> /dev/dm-0 -- 0 0 </tt><tt> </tt><tt> --- some entries removed for brevity ----</tt><tt> </tt><tt> /dev/md127 volume lvm2 a- 5.46t 2.77t volume</tt><tt> </tt><tt> --- some entries removed for brevity ----</tt><tt> </tt><tt> /dev/sde1 -- 0 0 </tt><tt> </tt><tt> </tt><tt>Additional information that might help. </tt><tt></tt><tt> </tt><tt>root@localhost# </tt><tt>pvs --version</tt><tt> </tt><tt> LVM version: 2.02.66(2) (2010-05-20)</tt><tt> </tt><tt> Library version: 1.02.48 (2010-05-20)</tt><tt> </tt><tt> Driver version: 4.22.0</tt><tt> </tt><tt>root@localhost# cat /etc/lsb-release </tt><tt> </tt><tt>DISTRIB_ID=Ubuntu</tt><tt> </tt><tt>DISTRIB_RELEASE=12.04</tt><tt> </tt><tt>DISTRIB_CODENAME=precise</tt><tt> </tt><tt>DISTRIB_DESCRIPTION="Ubuntu 12.04.2 LTS"</tt><tt> </tt><tt> </tt><tt>root@localhost#</tt><tt> valgrind pvs --all -o +vg_fmt</tt><tt> </tt><tt>==14584== Memcheck, a memory error detector</tt><tt> </tt><tt>==14584== Copyright (C) 2002-2011, and GNU GPL'd, by Julian Seward et al.</tt><tt> </tt><tt>==14584== Using Valgrind-3.7.0 and LibVEX; rerun with -h for copyright info</tt><tt> </tt><tt>==14584== Command: pvs --all -o +vg_fmt</tt><tt> </tt><tt>==14584== </tt><tt> </tt><tt>==14584== Invalid read of size 8</tt><tt> </tt><tt>==14584== at 0x548ED77: dm_report_field_string (in /lib/libdevmapper.so.1.02.1)</tt><tt> </tt><tt>==14584== by 0x549006F: dm_report_object (in /lib/libdevmapper.so.1.02.1)</tt><tt> </tt><tt>==14584== by 0x46B138: report_object (in /sbin/lvm)</tt><tt> </tt><tt>==14584== by 0x423618: ??? (in /sbin/lvm)</tt><tt> </tt><tt>==14584== by 0x4254B2: process_each_pv (in /sbin/lvm)</tt><tt> </tt><tt>==14584== by 0x4230D3: ??? (in /sbin/lvm)</tt><tt> </tt><tt>==14584== by 0x41AE60: lvm_run_command (in /sbin/lvm)</tt><tt> </tt><tt>==14584== by 0x41B307: lvm2_main (in /sbin/lvm)</tt><tt> </tt><tt>==14584== by 0x58CE76C: (below main) (libc-start.c:226)</tt><tt> </tt><tt>==14584== Address 0x28 is not stack'd, malloc'd or (recently) free'd</tt><tt> </tt><tt>==14584== </tt><tt> </tt><tt>==14584== </tt><tt> </tt><tt>==14584== Process terminating with default action of signal 11 (SIGSEGV)</tt><tt> </tt><tt>==14584== Access not within mapped region at address 0x28</tt><tt> </tt><tt>==14584== at 0x548ED77: dm_report_field_string (in /lib/libdevmapper.so.1.02.1)</tt><tt> </tt><tt>==14584== by 0x549006F: dm_report_object (in /lib/libdevmapper.so.1.02.1)</tt><tt> </tt><tt>==14584== by 0x46B138: report_object (in /sbin/lvm)</tt><tt> </tt><tt>==14584== by 0x423618: ??? (in /sbin/lvm)</tt><tt> </tt><tt>==14584== by 0x4254B2: process_each_pv (in /sbin/lvm)</tt><tt> </tt><tt>==14584== by 0x4230D3: ??? (in /sbin/lvm)</tt><tt> </tt><tt>==14584== by 0x41AE60: lvm_run_command (in /sbin/lvm)</tt><tt> </tt><tt>==14584== by 0x41B307: lvm2_main (in /sbin/lvm)</tt><tt> </tt><tt>==14584== by 0x58CE76C: (below main) (libc-start.c:226)</tt><tt> </tt><tt>==14584== If you believe this happened as a result of a stack</tt><tt> </tt><tt>==14584== overflow in your program's main thread (unlikely but</tt><tt> </tt><tt>==14584== possible), you can try to increase the size of the</tt><tt> </tt><tt>==14584== main thread stack using the --main-stacksize= flag.</tt><tt> </tt><tt>==14584== The main thread stack size used in this run was 8388608.</tt><tt> </tt><tt>==14584== </tt><tt> </tt><tt> </tt><tt>root@localhost# </tt><tt>ltrace pvs --all -o +vg_fmt 2>&1 | tail</tt><tt> </tt><tt>__ctype_toupper_loc() = 0x7f2f0e83f768</tt><tt> </tt><tt>__ctype_toupper_loc() = 0x7f2f0e83f768</tt><tt> </tt><tt>__ctype_toupper_loc() = 0x7f2f0e83f768</tt><tt> </tt><tt>__sprintf_chk(0x1e86e88, 1, -1, 0x48daa5, 0x48dab8) = 2</tt><tt> </tt><tt>dm_pool_strdup(0x1e88120, 0x1e86e88, 0x1e86e8a, 0, 0x48daa8) = 0x1e8e538</tt><tt> </tt><tt>dm_pool_alloc(0x1e88120, 8, 2, 0x7f2f0d892030, 0x48daa8) = 0x1e8e540</tt><tt> </tt><tt>dm_report_field_set_value(0x1e8e508, 0x1e8e538, 0x1e8e540, 0x7f2f0d892030, 0x48daa8) = 0x1e8e540</tt><tt> </tt><tt>dm_report_field_string(0x1e88080, 0x1e8e550, 40, 0x6b1fe0, 0x1e6cfe0 <unfinished ...></tt><tt> </tt><tt>--- SIGSEGV (Segmentation fault) ---</tt><tt> </tt><tt>+++ killed by SIGSEGV +++</tt> <tt>root@localhost# </tt><tt><tt>root@localhost#</tt> gdb --args pvs --all -o +vg_fmt</tt><tt> </tt><tt>GNU gdb (Ubuntu/Linaro 7.4-2012.04-0ubuntu2.1) 7.4-2012.04</tt><tt> </tt><tt>Copyright (C) 2012 Free Software Foundation, Inc.</tt><tt> </tt><tt>License GPLv3+: GNU GPL version 3 or later <a class="moz-txt-link-rfc2396E" href="http://gnu.org/licenses/gpl.html"><http://gnu.org/licenses/gpl.html></a></tt><tt> </tt><tt>This is free software: you are free to change and redistribute it.</tt><tt> </tt><tt>There is NO WARRANTY, to the extent permitted by law. Type "show copying"</tt><tt> </tt><tt>and "show warranty" for details.</tt><tt> </tt><tt>This GDB was configured as "x86_64-linux-gnu".</tt><tt> </tt><tt>For bug reporting instructions, please see:</tt><tt> </tt><tt><a class="moz-txt-link-rfc2396E" href="http://bugs.launchpad.net/gdb-linaro/"><http://bugs.launchpad.net/gdb-linaro/></a>...</tt><tt> </tt><tt>Reading symbols from /sbin/pvs...(no debugging symbols found)...done.</tt><tt> </tt><tt>(gdb) run</tt><tt> </tt><tt>Starting program: /sbin/pvs --all -o +vg_fmt</tt><tt> </tt><tt>[Thread debugging using libthread_db enabled]</tt><tt> </tt><tt>Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".</tt><tt> </tt><tt>File descriptor 3 (socket:[46992174]) leaked on pvs invocation. Parent PID 14623: gdb</tt><tt> </tt><tt>File descriptor 4 (socket:[46992175]) leaked on pvs invocation. Parent PID 14623: gdb</tt><tt> </tt><tt>File descriptor 5 (pipe:[46992176]) leaked on pvs invocation. Parent PID 14623: gdb</tt><tt> </tt><tt>File descriptor 6 (pipe:[46992176]) leaked on pvs invocation. Parent PID 14623: gdb</tt><tt> </tt><tt> </tt><tt>Program received signal SIGSEGV, Segmentation fault.</tt><tt> </tt><tt>0x00007ffff757dd77 in dm_report_field_string () from /lib/libdevmapper.so.1.02.1</tt><tt> </tt><tt>(gdb) disassemble dm_report_field_string</tt><tt> </tt><tt>Dump of assembler code for function dm_report_field_string:</tt><tt> </tt><tt> 0x00007ffff757dd70 <+0>: push %rbx</tt><tt> </tt><tt> 0x00007ffff757dd71 <+1>: mov %rsi,%rbx</tt><tt> </tt><tt> 0x00007ffff757dd74 <+4>: mov (%rdi),%rdi</tt><tt> </tt><tt>=> 0x00007ffff757dd77 <+7>: mov (%rdx),%rsi</tt><tt> </tt><tt> 0x00007ffff757dd7a <+10>: callq 0x7ffff7570a30 <dm_pool_strdup@plt></tt><tt> </tt><tt> 0x00007ffff757dd7f <+15>: test %rax,%rax</tt><tt> </tt><tt>(gdb) info registers </tt><tt> </tt><tt>rax 0x6dd550 7198032</tt><tt> </tt><tt>rbx 0x6dd550 7198032</tt><tt> </tt><tt>rcx 0x6b1fe0 7020512</tt><tt> </tt><tt>rdx 0x28 40</tt><tt> </tt><tt>rsi 0x6dd550 7198032</tt><tt> </tt><tt>rdi 0x6d7120 7172384</tt><tt> </tt><tt>rbp 0x6d7080 0x6d7080</tt><tt> </tt> On 07/20/2013 02:20 AM, Alasdair G Kergon wrote: </div> <blockquote cite="mid:20130719232052.GB23574@agk-dp.fab.redhat.com" type="cite"> <pre wrap="">It uses lists of ranges separated by colons. :x-y If x is missing, 0 is inserted; if y is missing, the last extent on the device is inserted. Man pages can always be improved with more examples! For +length, try the patch below. Alasdair --- a/tools/toollib.c +++ b/tools/toollib.c @@ -1001,7 +1001,7 @@ static int _parse_pes(struct dm_pool *mem, char *c, struct dm_list *pe_ranges, const char *pvname, uint32_t size) { char *endptr; - uint32_t start, end; + uint32_t start, end, len; /* Default to whole PV */ if (!c) { @@ -1041,7 +1041,16 @@ static int _parse_pes(struct dm_pool *mem, char *c, struct dm_list *pe_ranges, goto error; c = endptr; } + } else if (*c == '+') { /* Length? */ + c++; + if (isdigit(*c)) { + if (!xstrtouint32(c, &endptr, 10, &len)) + goto error; + c = endptr; + end = start + (len ? (len - 1) : 0); + } } + if (*c && *c != ':') goto error; </pre> </blockquote> </body> </html>