[257] in linux-net channel archive
NE2000 fixes --- 1.2.7 (#3B)
daemon@ATHENA.MIT.EDU (Paul Gortmaker)
Sun Apr 30 02:38:14 1995
From: Paul Gortmaker <paul@rasty.anu.edu.au>
To: linux-net@vger.rutgers.edu
Date: Sun, 30 Apr 1995 15:58:31 +1000 (EST)
This is just a minor nit, and it probably won't make a difference
to anybody except people running 8 bit ne1000 cards in a 386sx-12 :-)
In #3A (now 1.2.7) we are timing the PIO + (wait for ack from card)
with the dma_start/NE_RDC_TIMEOUT combo. This means using a long
(3 jiffies) timeout value to keep the slow boxes happy. However
a 1 jiffy timeout is more than enough for a 16 bit card in a 486.
Here is a one-liner that is a better fix than just increasing the
value of NE_RDC_TIMEOUT to 0x03. Just move the assignment of dma_start
to *after* the PIO, and then it won't make a difference if we are
doing slower outsb on a ne1k vs. faster outsw on a ne2k. We are
then only timing how long before the card "ack"s instead of timing
(PIO + wait for ack) which makes much more sense.
This should allow you to run NE_RDC_TIMEOUT at 0x01 on an old ne1000.
Apply to stock 1.2.7
Paul.
--- ne.c.old Sun Apr 30 14:20:19 1995
+++ ne.c Sun Apr 30 14:21:01 1995
@@ -86,7 +86,7 @@
#define NESM_START_PG 0x40 /* First page of TX buffer */
#define NESM_STOP_PG 0x80 /* Last page +1 of RX ring */
-#define NE_RDC_TIMEOUT 0x03 /* Max wait in jiffies for Tx RDC */
+#define NE_RDC_TIMEOUT 0x01 /* Max wait in jiffies for Tx RDC */
int ne_probe(struct device *dev);
static int ne_probe1(struct device *dev, int ioaddr);
@@ -462,7 +462,6 @@
SLOW_DOWN_IO;
#endif /* rw_bugfix */
- dma_start = jiffies;
outb_p(ENISR_RDC, nic_base + EN0_ISR);
/* Now the normal output. */
@@ -478,6 +477,8 @@
outsb(NE_BASE + NE_DATAPORT, buf, count);
}
+ dma_start = jiffies;
+
#ifdef CONFIG_NE_SANITY
/* This was for the ALPHA version only, but enough people have
been encountering problems so it is still here. */
