[9front] [patch] nusb/ether followup to cs 8387

[9front] [patch] nusb/ether followup to cs 8387

[Romano]

[-- Attachment #1: Type: text/plain, Size: 215 bytes --]

Attached is a patch to address minor issues introduced by changset
8397, or when no ethernet exists at boot.  If this should go to some
other mailing list (9front-bugs?), let me know, or if there are issues
with it.

[-- Attachment #2: Type: text/plain, Size: 1683 bytes --]

changeset:   8397:c13e74f20026
branch:      mbp-2011
user:        Romano <unobe@cpan.org>
date:        Thu Apr 08 00:58:00 2021 -0700
summary:     Don't attempt to load an ether interface if none exists. nusb/serial does not handle RNDIS CSP devices, so do not load those. follow-up doc fix for changeset 6370:8d9b9553c65c.

diff -r 5c114f6f0d21 -r c13e74f20026 rc/bin/cpurc
--- a/rc/bin/cpurc	Thu Apr 08 00:49:59 2021 -0700
+++ b/rc/bin/cpurc	Thu Apr 08 00:58:00 2021 -0700
@@ -59,7 +59,8 @@
 		addrs=`{ndb/query -a sys $sysname ether}
 		if(! ~ $#addrs 0){
 			for(ether in /net/ether*){
-				addr=`{cat $ether/addr}
+				if(test -d $ether)
+					addr=`{cat $ether/addr}
 				switch($addr){
 				case $addrs
 					# try /lib/ndb first, then do dhcp/slaac
diff -r 5c114f6f0d21 -r c13e74f20026 rc/bin/nusbrc
--- a/rc/bin/nusbrc	Thu Apr 08 00:49:59 2021 -0700
+++ b/rc/bin/nusbrc	Thu Apr 08 00:58:00 2021 -0700
@@ -14,8 +14,9 @@
 		case *01
 			nusb/audio $id
 		case *02
-			# serial and ethernet
-			nusb/serial $id
+			# serial and ethernet: 0xff0202 is RNDIS, handled by nusbrc
+			if(! ~ $4 ff0202)
+				nusb/serial $id
 			# handled /sys/src/9/boot/nusbrc
 			# nusb/ether $id
 		case *03
diff -r 5c114f6f0d21 -r c13e74f20026 sys/src/cmd/nusb/ether/ether.c
--- a/sys/src/cmd/nusb/ether/ether.c	Thu Apr 08 00:49:59 2021 -0700
+++ b/sys/src/cmd/nusb/ether/ether.c	Thu Apr 08 00:58:00 2021 -0700
@@ -849,7 +849,7 @@
 
 	nerr = 0;
 	for(;;){
-		/* receive allocates buffer and calls etheriq(b, 1); */
+		/* receive allocates buffer and calls etheriq(b); */
 		if((*epreceive)(epin) < 0){
 			rerrstr(err, sizeof(err));
 			if(strstr(err, "interrupted") || strstr(err, "timed out"))

Re: [9front] [patch] nusb/ether followup to cs 8387

[cinap_lenrek]

ok, but it nusbrc can be solved much simpler:

diff -r 4dfbef4fa4ac rc/bin/nusbrc
--- a/rc/bin/nusbrc	Sat Apr 03 19:32:47 2021 +0200
+++ b/rc/bin/nusbrc	Thu Apr 08 21:49:21 2021 +0200
@@ -13,6 +13,8 @@
 		switch($4){
 		case *01
 			nusb/audio $id
+		case ff0202
+			# RNDIS; handled by /sys/src/9/boot/nusbrc
 		case *02
 			# serial and ethernet
 			nusb/serial $id

it is not clear to me why the test -d is needed in cpurc. 
you ust want to silence the error message from cat?

--
cinap

Re: [9front] [patch] nusb/ether followup to cs 8387

[Romano]

On April 8, 2021 7:51:00 PM UTC, cinap_lenrek@felloff.net wrote:
>ok, but it nusbrc can be solved much simpler:
>
>diff -r 4dfbef4fa4ac rc/bin/nusbrc
>--- a/rc/bin/nusbrc	Sat Apr 03 19:32:47 2021 +0200
>+++ b/rc/bin/nusbrc	Thu Apr 08 21:49:21 2021 +0200
>@@ -13,6 +13,8 @@
> 		switch($4){
> 		case *01
> 			nusb/audio $id
>+		case ff0202
>+			# RNDIS; handled by /sys/src/9/boot/nusbrc
> 		case *02
> 			# serial and ethernet
> 			nusb/serial $id

Yes, that's better.

>it is not clear to me why the test -d is needed in cpurc. 
>you ust want to silence the error message from cat?

Yes. I don't consider it an error to not have an ethernet interface at boot. If there's no ethernet interface, $ether will be '/net/ether*' because * won't expand. $ether/addr will not exist, because $ether is not a directory.

Re: [9front] [patch] nusb/ether followup to cs 8387

[cinap_lenrek]

> Yes. I don't consider it an error to not have an ethernet interface at boot.
> If there's no ethernet interface, $ether will be '/net/ether*' because * won't expand.
> $ether/addr will not exist, because $ether is not a directory.

ah, i see. i think in that case, it is faster to just redirect stderr to /dev/null.

--
cinap