To clarify, can you give an exact procedure to reproduce?
(E.g. an ftp transfer of a 100MB file from the internet to
another box, routed through onboard ethernet on the Soekris)
I've been getting seemingly random occasional hangs with ral
in HostAP mode that I haven't been able to correlate with
I've had no problems with an rt2860 in this configuration;
hostname.ral0:
mode 11g mediaopt hostap chan 1 nwid gen wpa \
wpapsk `wpa-psk gen chilledbrains` \
group internal up
(This interface happens to be bridged to an re(4) which is why it
has no address assigned here.)
But
Tested on an F71882 (at 0x2e; GENERIC only configs it at 4e).
The voltage dividers register is full of lies, so the part of this
patch that reads it can be ripped out (along with the temperature
diode offset bits, I guess). These chips also have a watchdog
timer. I can set it and watch it
Well, building a kernel with DRM and removing option MigrationHeuristic
gives some impressive results:
GtkEntry - time: 0.26
GtkComboBox - time: 6.69
GtkComboBoxEntry - time: 5.76
GtkSpinButton - time: 0.91
GtkProgressBar - time: 0.38
GtkToggleButton - time: 2.17
GtkCheckButton - time:
vga1 at pci1 dev 5 function 0 ATI Radeon X1250 rev 0x00
The default AccelMethod is apparently still XAA (even though it says
it's unsupported and to please use EXA instead). Since this is
an IGP, the optimal MigrationHeuristic is probably never, but
greedy seems to be the best approximation :-)
but the point can be made to the authors. -E exports *all* non-static
symbols, which will allow loadable modules to gain easy access to symbols
they are not supposed to.
perhaps a more elegant solution: http://gcc.gnu.org/wiki/Visibility
6 matches
Mail list logo
