Christian, thank you for the investigation. It proves that we are not to blame. It remains to see what should be done about it. Disabling the offending function by default for all future doesn't seem to be a good idea as the information retreived thru this function may be helpful when people send in their logs with their bug reports. I'd therefore suggest to enable it and provide information on the web and in the man page how the hang can be avoided if somebody encounters it.
Egbert. Christian Zietz writes: > > I made some research with a little tool of mine which allows me to load > and use another BIOS than the pre-installed one. (This tool is NOT going > to be released.) > The BIOS dump I got from an Inspiron 510m crashes under DOS, too. This > crash only occurs for CX (device number) 0x01. It's not an endless loop > but a jump to "nowhere". Let me explain: Depending on the device number > the address of the function to call is taken from a table. This fails > for device 1 and isn't properly caught so a jump to the middle of some > other (totally unrelated) function is executed. That obviously leads to > the crash. > BIOS build 3197 (which can be obtained from the Intel web site) doesn't > have that bug. Since the table with the addresses is in the (partially) > user configurable BIOS data block, the missing table entry could also be > a configuration error made by Dell. But as Dell's policy towards Linux > is "we don't support it", chances are low to get a new BIOS asking there. > _______________________________________________ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
