NET: Proper return code handling in eth_init() function in file eth.c

PATCH New Upakul Barkakaty 3 participants 9 messages Nov 29, 2007 – Jan 8, 2008 Networking

lore

Share & tools

lore

Upakul Barkakaty

author

2007-11-29 06:04 UTC #0 lore raw

PATCH 1/1 NET: Proper return code handling in eth_init() function in file eth.c

NET: Proper return code handling in eth_init() function in file eth.c

This patch modifies the return code handling in the eth_init() function, to be compatible with the handling of the return codes in the other network stack files. It now returns a 0 on Success and -1 on error.

Signed-off-by Upakul Barkakaty <[email protected]>

net/eth.c +5 -5

diff -Naur a/net/eth.c b/net/eth.c
@@ -404,23 +404,23 @@
 	struct eth_device* old_current;
 
 	if (!eth_current)
-		return 0;
+		return -1;
 
 	old_current = eth_current;
 	do {
 		debug ("Trying %s\n", eth_current->name);
 
-		if (eth_current->init(eth_current, bis)) {
+		if (!eth_current->init(eth_current,bis))
+		{
 			eth_current->state = ETH_STATE_ACTIVE;
-
-			return 1;
+			return 0;
 		}
 		debug  ("FAIL\n");
 
 		eth_try_another(0);
 	} while (old_current != eth_current);
 
-	return 0;
+	return -1;
 }
 
 void eth_halt(void)

Stefan Roese

2007-11-29 06:24 UTC #1 lore raw

PATCH 1/1 NET: Proper return code handling in eth_init() function in file eth.c

Quoted 16 lines · Upakul Barkakaty

NET: Proper return code handling in eth_init() function in file eth.c
This patch modifies the return code handling in the eth_init() function, to be compatible with the handling of the return codes in the other network stack files. It now returns a 0 on Success and -1 on error.
Signed-off-by Upakul Barkakaty <[email protected]>
diff -Naur a/net/eth.c b/net/eth.c
net/eth.c +0 -0
@@ -404,23 +404,23 @@
 	struct eth_device* old_current;
 
 	if (!eth_current)

net/eth.c line 414

Quoted 10 lines

net/eth.c +3 -2

-		return 0;
+		return -1;
 
 	old_current = eth_current;
 	do {
 		debug ("Trying %s\n", eth_current->name);
 
-		if (eth_current->init(eth_current, bis)) {
+		if (!eth_current->init(eth_current,bis))
+		{

Don't move the "{" into the next line.

Please change this and resend the patch. Thanks.

Best regards,
Stefan

=====================================================================
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office at denx.de
=====================================================================

Upakul Barkakaty

author

2007-11-29 06:46 UTC #2 lore raw

PATCH 1/1 NET: Proper return code handling in eth_init() function in file eth.c

NET: Proper return code handling in eth_init() function in file eth.c

Signed-off-by Upakul Barkakaty <[email protected]>

net/eth.c +4 -4

diff -Naur a/net/eth.c b/net/eth.c
@@ -404,23 +404,23 @@
 	struct eth_device* old_current;
 
 	if (!eth_current)
-		return 0;
+		return -1;
 
 	old_current = eth_current;
 	do {
 		debug ("Trying %s\n", eth_current->name);
 
-		if (eth_current->init(eth_current, bis)) {
+		if (!eth_current->init(eth_current,bis)) {
 			eth_current->state = ETH_STATE_ACTIVE;
 
-			return 1;
+			return 0;
 		}
 		debug  ("FAIL\n");
 
 		eth_try_another(0);
 	} while (old_current != eth_current);
 
-	return 0;
+	return -1;
 }
 
 void eth_halt(void)

Stefan Roese

2008-01-08 14:26 UTC #3 lore raw

PATCH 1/1 NET: Proper return code handling in eth_init() function in file eth.c

Quoted 6 lines · Upakul Barkakaty

NET: Proper return code handling in eth_init() function in file eth.c
This patch modifies the return code handling in the eth_init() function, to be compatible with the handling of the return codes in the other network stack files. It now returns a 0 on Success and -1 on error.

I just noticed that this patch breaks TFTP on 4xx platforms:

commit 505be87a65e4f87ad7d8da1d57ea4dcd487d7e32

Author: Upakul Barkakaty <[email protected]>
Date:   Thu Nov 29 12:16:13 2007 +0530

NET: Proper return code handling in eth_init() function in file eth.c

Signed-off-by Upakul Barkakaty <[email protected]>

Signed-off-by Ben Warren <[email protected]>

I just see link auto-negotiation going on, and after that no real transaction is happening! :-(

=> tftp 100000 u-image.bin
Waiting for PHY auto negotiation to complete.... done ENET Speed is 1000 Mbps - FULL duplex connection (EMAC0)

Would be great if someone could look into this ASAP. Unfortunately I have no time left today.

Thanks.

Best regards,
Stefan

Upakul Barkakaty

author

2008-01-08 14:35 UTC #4 lore raw

PATCH 1/1 NET: Proper return code handling in eth_init() function in file eth.c

Quoted 39 lines · Stefan Roese

Quoted 6 lines · Upakul Barkakaty
NET: Proper return code handling in eth_init() function in file eth.c
This patch modifies the return code handling in the eth_init() function, to be compatible with the handling of the return codes in the other network stack files. It now returns a 0 on Success and -1 on error.
I just noticed that this patch breaks TFTP on 4xx platforms:
commit 505be87a65e4f87ad7d8da1d57ea4dcd487d7e32
Author: Upakul Barkakaty <[email protected]>
Date:   Thu Nov 29 12:16:13 2007 +0530
NET: Proper return code handling in eth_init() function in file eth.c
This patch modifies the return code handling in the eth_init() function, to be compatible with the handling of the return codes in the other network stack files. It now returns a 0 on Success and -1 on error.
Signed-off-by Upakul Barkakaty <[email protected]>
Signed-off-by Ben Warren <[email protected]>
I just see link auto-negotiation going on, and after that no real transaction
is happening! :-(
=> tftp 100000 u-image.bin
Waiting for PHY auto negotiation to complete.... done ENET Speed is 1000 Mbps - FULL duplex connection (EMAC0)
Would be great if someone could look into this ASAP. Unfortunately I have no
time left today.
Thanks.

Quoted 8 lines

Best regards,
Stefan
=====================================================================
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office at denx.de
=====================================================================

Hi Stefan,

Could you please tell which ethernet driver is being used? Please mention the filename.

Thanks.

Signature · 5 lines

Regards,
Upakul Barkakaty
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.denx.de/pipermail/u-boot/attachments/20080108/c2a5e301/attachment.htm

Upakul Barkakaty

author

2008-01-08 14:50 UTC #5 lore raw

PATCH 1/1 NET: Proper return code handling in eth_init() function in file eth.c

In reply to Stefan Roese

Quoted 39 lines · Stefan Roese

Quoted 6 lines · Upakul Barkakaty
NET: Proper return code handling in eth_init() function in file eth.c
This patch modifies the return code handling in the eth_init() function, to be compatible with the handling of the return codes in the other network stack files. It now returns a 0 on Success and -1 on error.
I just noticed that this patch breaks TFTP on 4xx platforms:
commit 505be87a65e4f87ad7d8da1d57ea4dcd487d7e32
Author: Upakul Barkakaty <[email protected]>
Date:   Thu Nov 29 12:16:13 2007 +0530
NET: Proper return code handling in eth_init() function in file eth.c
This patch modifies the return code handling in the eth_init() function, to be compatible with the handling of the return codes in the other network stack files. It now returns a 0 on Success and -1 on error.
Signed-off-by Upakul Barkakaty <[email protected]>
Signed-off-by Ben Warren <[email protected]>
I just see link auto-negotiation going on, and after that no real transaction
is happening! :-(
=> tftp 100000 u-image.bin
Waiting for PHY auto negotiation to complete.... done ENET Speed is 1000 Mbps - FULL duplex connection (EMAC0)
Would be great if someone could look into this ASAP. Unfortunately I have no
time left today.
Thanks.

Quoted 9 lines

Best regards,
Stefan
=====================================================================
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office at denx.de
=====================================================================

Hi Stefan,

Probably i got the reason why this is happening. As a convention that is usually followed, the code in eth_init() function in the file eth.c was reworked to handle 0 in case of Success and any non zero value in case of an error.

Most of the Ethernet driver's init routines adhere to the same convention of returning a 0 on Success and a non zero value in case of error. However, the driver you might be using, may be returning a non zero value on Successful initialization, that is why the problem is arising. The solution could be to return 0 in case of Success, from the Ethernet driver's init() routine.

Thanks!!

Signature · 5 lines

Regards,
Upakul Barkakaty
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.denx.de/pipermail/u-boot/attachments/20080108/247a5965/attachment.htm

Ben Warren

2008-01-08 14:56 UTC #6 lore raw

PATCH 1/1 NET: Proper return code handling in eth_init() function in file eth.c

Quoted 66 lines · Upakul Barkakaty

On 1/8/08, *Stefan Roese* <sr at denx.de <mailto:[email protected]>> wrote:
On Thursday 29 November 2007, Upakul Barkakaty wrote:
> NET: Proper return code handling in eth_init() function in file eth.c
>
> This patch modifies the return code handling in the eth_init() > function, to be compatible with the handling of the return codes in > the other network stack files. It now returns a 0 on Success and -1 on
> error.
I just noticed that this patch breaks TFTP on 4xx platforms:
commit 505be87a65e4f87ad7d8da1d57ea4dcd487d7e32
Author: Upakul Barkakaty <upakul at gmail.com <mailto:[email protected]>>
Date: Thu Nov 29 12:16:13 2007 +0530
NET: Proper return code handling in eth_init() function in
file eth.c
This patch modifies the return code handling in the eth_init() function, to be compatible with the handling of the return
codes in
the other network stack files. It now returns a 0 on Success
and -1 on
error.
Signed-off-by: Upakul Barkakaty <
upakul.barkakaty at conexant.com <mailto:[email protected]>>
Signed-off-by Ben Warren <[email protected]
<mailto:[email protected]>>
I just see link auto-negotiation going on, and after that no real transaction
is happening! :-(
=> tftp 100000 u-image.bin
Waiting for PHY auto negotiation to complete.... done ENET Speed is 1000 Mbps - FULL duplex connection (EMAC0)
Would be great if someone could look into this ASAP. Unfortunately I have no
time left today.
Thanks.
Best regards,
Stefan
=====================================================================
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email:
office at denx.de <mailto:[email protected]>
=====================================================================
Hi Stefan,
Probably i got the reason why this is happening. As a convention that is usually followed, the code in eth_init() function in the file eth.c was reworked to handle 0 in case of Success and any non zero value in case of an error.

Quoted 12 lines

Most of the Ethernet driver's init routines adhere to the same convention of returning a 0 on Success and a non zero value in case of error. However, the driver you might be using, may be returning a non zero value on Successful initialization, that is why the problem is arising. The solution could be to return 0 in case of Success, from the Ethernet driver's init() routine.
Thanks!!
Signature · 2 lines
Regards,
Upakul Barkakaty

Yeah, looks like ppc_4xx_eth_init() (cpu/ppc4xx/4xx_enet.c) returns 1 on success and -1 on failure. I don't have hardware to test this. Can you check it out Stefan? Let's look through the other drivers to see if this is an isolated case.

regards,
Ben

Stefan Roese

2008-01-08 17:26 UTC #7 lore raw

PATCH 1/1 NET: Proper return code handling in eth_init() function in file eth.c

Quoted 14 lines · Ben Warren

Quoted 14 lines
Probably i got the reason why this is happening. As a convention that is usually followed, the code in eth_init() function in the file eth.c was reworked to handle 0 in case of Success and any non zero value in case of an error.
Most of the Ethernet driver's init routines adhere to the same convention of returning a 0 on Success and a non zero value in case of error. However, the driver you might be using, may be returning a non zero value on Successful initialization, that is why the problem is arising. The solution could be to return 0 in case of Success, from the Ethernet driver's init() routine.
Thanks!!

Quoted 8 lines

Quoted 3 lines
--
Regards,
Upakul Barkakaty
Yeah, looks like ppc_4xx_eth_init() (cpu/ppc4xx/4xx_enet.c) returns 1 on success and -1 on failure. I don't have hardware to test this. Can you check it out Stefan? Let's look through the other drivers to see if this is an isolated case.

I'm back now and will give it a try this evening. Thanks for pointing this out.

Note: This patch may trigger other platform ethernet driver problems as well. So I strongly encourage all platform maintainers/custodians to check if a change in the return code of the ethernet driver is needed.

Best regards,
Stefan

Ben Warren

2008-01-08 21:10 UTC #8 lore raw

PATCH 1/1 NET: Proper return code handling in eth_init() function in file eth.c

Quoted 29 lines · Stefan Roese

Quoted 14 lines · Ben Warren
Quoted 14 lines
Probably i got the reason why this is happening. As a convention that is usually followed, the code in eth_init() function in the file eth.c was reworked to handle 0 in case of Success and any non zero value in case of an error.
Most of the Ethernet driver's init routines adhere to the same convention of returning a 0 on Success and a non zero value in case of error. However, the driver you might be using, may be returning a non zero value on Successful initialization, that is why the problem is arising. The solution could be to return 0 in case of Success, from the Ethernet driver's init() routine.
Thanks!!
Quoted 9 lines
Quoted 4 lines
--
Regards,
Upakul Barkakaty
Yeah, looks like ppc_4xx_eth_init() (cpu/ppc4xx/4xx_enet.c) returns 1 on success and -1 on failure. I don't have hardware to test this. Can you check it out Stefan? Let's look through the other drivers to see if this is an isolated case.
I'm back now and will give it a try this evening. Thanks for pointing this out.

Quoted 7 lines

Note: This patch may trigger other platform ethernet driver problems as well. So I strongly encourage all platform maintainers/custodians to check if a change in the return code of the ethernet driver is needed.
Best regards,
Stefan

Don't waste too much time on this - every driver returns 1 on success. I'm putting together a patch to clean things up and should have it within an hour or so.

regards,
Ben