Skip to contentSkip to conversation

SGI Seeq cleanup

PATCH In review Ladislav Michl 3 participants 4 messages May 7, 2003 – May 8, 2003

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 SGI Seeq cleanup

read eaddr using NVRAM access fuctions and make various cleanups so driver can be build as module

drivers/net/Config.in +1 -1
diff -Naur linux_2_4/drivers/net/Config.in linux_2_4-fucked/drivers/net/Config.in
@@ -217,7 +217,7 @@
       dep_tristate '    D-Link DE620 pocket adapter support' CONFIG_DE620 $CONFIG_ISA
    fi
    if [ "$CONFIG_SGI_IP22" = "y" ]; then
-      bool '  SGI Seeq ethernet controller support' CONFIG_SGISEEQ
+      tristate '  SGI Seeq ethernet controller support' CONFIG_SGISEEQ
    fi
    if [ "$CONFIG_DECSTATION" = "y" ]; then
       tristate '  DEC LANCE ethernet controller support' CONFIG_DECLANCE
drivers/net/Space.c +0 -11
diff -Naur linux_2_4/drivers/net/Space.c linux_2_4-fucked/drivers/net/Space.c
@@ -83,7 +83,6 @@
 extern int SK_init(struct net_device *);
 extern int seeq8005_probe(struct net_device *);
 extern int smc_init( struct net_device * );
-extern int sgiseeq_probe(struct net_device *);
 extern int atarilance_probe(struct net_device *);
 extern int sun3lance_probe(struct net_device *);
 extern int sun3_82586_probe(struct net_device *);
@@ -366,14 +365,6 @@
 	{NULL, 0},
 };
 
-
-static struct devprobe sgi_probes[] __initdata = {
-#ifdef CONFIG_SGISEEQ
-	{sgiseeq_probe, 0},
-#endif
-	{NULL, 0},
-};
-
 static struct devprobe mips_probes[] __initdata = {
 #ifdef CONFIG_MIPS_JAZZ_SONIC
 	{sonic_probe, 0},
@@ -407,8 +398,6 @@
 	if (probe_list(dev, m68k_probes) == 0)
 		return 0;
 	if (probe_list(dev, mips_probes) == 0)
-		return 0;
-	if (probe_list(dev, sgi_probes) == 0)
 		return 0;
 	if (probe_list(dev, eisa_probes) == 0)
 		return 0;
drivers/net/sgiseeq.c +99 -114

409 lines not on this page · raw message

#1 lore raw
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"ladis" == Ladislav Michl <[email protected]> writes:

ladis> read eaddr using NVRAM access fuctions and make various cleanups so driver ladis> can be build as module

You are my hero!

[ Removal of Space.c entry ]

Hero++

ladis> @@ -96,8 +97,8 @@
ladis> struct sgiseeq_private {
ladis> volatile struct sgiseeq_init_block srings;
ladis> char *name;
ladis> - volatile struct hpc3_ethregs *hregs;
ladis> - volatile struct sgiseeq_regs *sregs;
ladis> + struct hpc3_ethregs *hregs;
ladis> + struct sgiseeq_regs *sregs;

I read through all the patch, and I didn't understand why volatile is not needed anymore :(

Althought not that I did understand why it was needed in the first place :)

ladis> @@ -435,7 +439,7 @@
ladis> /* Always check for received packets. */
ladis> sgiseeq_rx(dev, sp, hregs, sregs);

ladis> - /* Only check for tx acks iff we have something queued. */
ladis> + /* Only check for tx acks if we have something queued. */
ladis> if (sp->tx_old != sp->tx_new)
ladis> sgiseeq_tx(dev, sp, hregs, sregs);

iff == Math speak for if and only if. Not sure if iff is needed in that context at all.

Later, Juan.

Signature · 2 lines
In theory, practice and theory are the same, but in practice they 
are different -- Larry McVoy
#2 lore raw
Quoted 8 lines · Juan Quintela
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"ladis" == Ladislav Michl <[email protected]> writes:

ladis> read eaddr using NVRAM access fuctions and make various cleanups so driver ladis> can be build as module

You are my hero!

[ Removal of Space.c entry ]

Question comes in mind: It's possible to have ISA eth card installed in Indigo2. How to ensure that onboard Seeq will be always eth0 without using Space.c? (and don't advice me to build driver for ISA card as module :))

Quoted 11 lines

Hero++

ladis> @@ -96,8 +97,8 @@
ladis> struct sgiseeq_private {
ladis> volatile struct sgiseeq_init_block srings;
ladis> char *name;
ladis> - volatile struct hpc3_ethregs *hregs;
ladis> - volatile struct sgiseeq_regs *sregs;
ladis> + struct hpc3_ethregs *hregs;
ladis> + struct sgiseeq_regs *sregs;

Quoted 5 lines

I read through all the patch, and I didn't understand why volatile is not needed anymore :(

Althought not that I did understand why it was needed in the first place :)

I think struct members have to be volatile, not struct itself.

Quoted 11 lines

ladis> @@ -435,7 +439,7 @@
ladis> /* Always check for received packets. */
ladis> sgiseeq_rx(dev, sp, hregs, sregs);

ladis> - /* Only check for tx acks iff we have something queued. */
ladis> + /* Only check for tx acks if we have something queued. */
ladis> if (sp->tx_old != sp->tx_new)
ladis> sgiseeq_tx(dev, sp, hregs, sregs);

iff == Math speak for if and only if. Not sure if iff is needed in that context at all.

Ah, didn't know it. But i'd say simple if is enough in this case :)

ladis

#3 lore raw
Quoted 3 lines · Ladislav Michl
Quoted 3 lines

You are my hero!

[ Removal of Space.c entry ]

Absolutely, Space.c, RIP - Rot In Pieces.

Quoted 4 lines

Question comes in mind: It's possible to have ISA eth card installed in Indigo2. How to ensure that onboard Seeq will be always eth0 without using Space.c? (and don't advice me to build driver for ISA card as module :))

You just need to make sure sgiseeq.o is linked in before any of the ISA drivers.

ISA link ordering is sort of emprical rocket science but fortunately the IP22 Seeq isn't ISA.

Ralf