> >>If you have 4.5GB of free disk space on that ferrari laptop, the following > >>experiment could be interesting: > > There was no crash while copying the files with cp.
OK, when writing to this lofi mounted udfs image the "sd" driver isn't involved. So this can't be bug 6368142. > But I did get two crashes while copying files with > find and cpio to the lofi-mounted image. What exactly did you copy? The same /usr/bin binaries? Something like this? cd /usr/bin; find . -print | cpio -pdum /mnt > The first crash occured after I canceled the cpio > with CTRL-C: > Mar 13 19:59:10 ferrari genunix: [ID 655072 kern.notice] fffffe8000dd9990 > unix:cmntrap+140 () > Mar 13 19:59:10 ferrari genunix: [ID 655072 kern.notice] fffffe8000dd9aa0 > udfs:ud_mark_free+2c () > Mar 13 19:59:10 ferrari genunix: [ID 655072 kern.notice] fffffe8000dd9b30 > udfs:ud_free_space_bmap+145 () > Mar 13 19:59:10 ferrari genunix: [ID 655072 kern.notice] fffffe8000dd9b90 > udfs:ud_free_space+a3 () > Mar 13 19:59:10 ferrari genunix: [ID 655072 kern.notice] fffffe8000dd9c20 > udfs:ud_trunc_ext4+1a1 () Hmm, crashes at "ud_mark_free+2c"... > After that crash I tried to reproduce it: > > - create a new file with dd > - execute the find / cpio command again. > > This time Solaris crashes by itself without pressing > CTRL-C: > Mar 13 20:38:57 ferrari genunix: [ID 655072 kern.notice] fffffe8000a277c0 > unix:cmntrap+140 () > Mar 13 20:38:57 ferrari genunix: [ID 655072 kern.notice] fffffe8000a278d0 > udfs:ud_mark_free+2c () > Mar 13 20:38:57 ferrari genunix: [ID 655072 kern.notice] fffffe8000a27960 > udfs:ud_free_space_bmap+145 () > Mar 13 20:38:57 ferrari genunix: [ID 655072 kern.notice] fffffe8000a279c0 > udfs:ud_free_space+a3 () > Mar 13 20:38:57 ferrari genunix: [ID 655072 kern.notice] fffffe8000a27a00 > udfs:ud_ifree+c7 () Again, crash at "ud_mark_free+2c". Looks like this is: bug 6191228 Synopsis: panic in ud_mark_free http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6191228 Hmm, now the question is: is there a 32<-> 64bit issue with the udfs filesystem? Both your system and the one in bug 6191228 are running a 64-bit kernel. With my 32-bit kernel tests I've been unable to reproduce these crashes... Or is this a problem with the udfs filesystem that is preformatted on Imation / Maxell DVD-RAM media? I believe you already mentioned that you've tried to create a new UDFS filesystem (with "mkfs -F udfs" ?) and the machine still crashed, correct? This message posted from opensolaris.org
