On Tuesday 29 September 2015 07:06:46, Jörg Frings-Fürst wrote: > Hello Stefan, > > I have forward your bug and the first answer was: > > Mon Sep 28 15:54:38 2015, comment #1: > > > > > > What does the "extracted smbios file" look like? How was it > > generated? The DMI table is beyond the first MB of memory on this > > laptop. Unless you dumped all 4 GB of memory to the file, the > > table is not in it. The SIGBUS happens because mmap() is called > > beyond the end of the file (as documented in mmap(2).) > > > > > > The problem here is that you are abusing option -d. If you want to > > keep a copy of the DMI table and/or decode the DMI table on a > > different machine, please use options --dump-bin and --from-dump > > which were introduced in dmidecode version 2.10 (November 2008.) > > Option -d only works reliably when operating on an actual device > > file. > Please can you check it?
I don't doubt that the analysis is correct, even though I don't have the hardware available right now. We had created the file by dumping the 0xe0000-0xfffff area just below 1MB and putting it in a sparse file at the same offset. But exiting with an error message is much better than a SIGBUS. Thanks for fixing that. Cheers, Stefan
