IJAH400@INDYVAX.IUPUI.EDU ("James A. Harvey") (04/07/90)
Here's a patch for the 3C503 lockup problem. This is a different lockup problem than the one Carl Beame reported. The symptom is a lockup upon the first receipt of a broadcast packet when the driver is installed. Most systems appear to "recover" but the time taken to do this may be quite a while on a system with not much broadcast traffic. Assuming you have either the 6.x beta test packet driver or the 5.x with the sti/cli swap fix already in place the only code affected is the same and is in etopen where the shared memory is setup: OLD CODE: call set_8390_eaddr ; Now set the address in the 8390 chip call set_8390_multi ; Put the right stuff into 8390's multicast masks loadport ; Base of I/O regs setport EN1_CURPAG ; Set current shared page for RX to work on mov al, SM_RSTART_PG+1 out dx, al NEW CODE: call set_8390_eaddr ; Now set the address in the 8390 chip call set_8390_multi ; Put the right stuff into 8390's multicast masks loadport ; Base of I/O regs setport EN_CCMD ;[jah] Switch to page 1 registers mov al,ENC_NODMA+ENC_PAGE1 ;[jah] out dx, al ;[jah] setport EN1_CURPAG ; Set current shared page for RX to work on mov al, SM_RSTART_PG+1 out dx, al setport EN1_CURPAG ; Set current shared page for RX to work on I sent this into Russ Nelson, it should get into the "official" 6.x release of the Clarkson packet drivers... - Jim Harvey IUPUI Computing Services, DEC Systems Group Internet: ijah400@indyvax.iupui.edu BITNET: ijah400@indyvax Disclaimer: Opinions expressed here are mine, not necessarily those of Indiana University "If you don't know where you're going, you'll probably end up somewhere else"