On Wed, Dec 19, 2007 at 07:25:35PM +0100, Wouter Verhelst wrote:
However, since unaligned accesses often cause bus errors, too, my guess is that it's quite a bit more likely for a SIGBUS to be caused by programmer error than a corrupted binary...
Yes, but i386 and amd64 don't require aligned accesses. The bug is manifesting itself on amd64, so it is unlikely to be an unaligned access. In fact, the only case where I've ever seen a SIGBUS on i386 is where a friend accidentally removed the hard disk from the machine whilst it was running.
In general, and especially on certain platforms (sparc comes to mind), you are correct. Unaligned accesses are a very common problem because most programmers don't use architectures that require aligned access and so never notice that anything is amiss.
-- brian m. carlson / brian with sandals: Houston, Texas, US +1 713 440 7475 | http://crustytoothpaste.ath.cx/~bmc | My opinion only a typesetting engine: http://crustytoothpaste.ath.cx/~bmc/code/thwack OpenPGP: RSA v4 4096b 88AC E9B2 9196 305B A994 7552 F1BA 225C 0223 B187
Attachment:
signature.asc
Description: Digital signature