kernel update (relative to 2.6.5)

From: David Mosberger <>
Date: 2004-04-14 11:03:38
At long last, the ia64-patch for 2.6.5...
It's in the usual place:

in file


There is one significant change: I decided to fix signal delivery so
signal handlers are forced to be invoked with little-endian
byte-order.  My position in the past has been that if you turn-on
big-endian byte-order, you're on your own.  Indeed, if you happened to
get a signal in big-endian byte-order, you were pretty much guaranteed
to get a SIGSEGV.  I decided to fix this particular issue because I
found some "legacy" code which was taking advantage of big-endian
byte-order and so that code ran the risk of crashing if a signal
occurred.  Having said that, you're still mostly on your own with
big-endian byte-order.  In particular, things could go horribly wrong
if pointers or frame-related values (such as the return-address) were
stored in big-endian byte-order.  On the other hand, if you turn on
big-endian mode temporarily, say, in a leaf-routine doing some
calculations, you should be OK.

Also, I decided to change the extended config-space addressing support
in SAL such that we can avoid version number checking---that's just
too fragile and has been a repeated source of trouble and lost time.

Oh, there is a big perfmon update with some important fixes.  See the
ChangeLog entry here for details:|ChangeSet|20040409002153|03271


To unsubscribe from this list: send the line "unsubscribe linux-ia64" in
the body of a message to
More majordomo info at
Received on Tue Apr 13 21:03:59 2004

This archive was generated by hypermail 2.1.8 : 2005-08-02 09:20:25 EST