Posted: Wed Nov 30, 2011 15:36 Post subject: Re: Mega bins with samba no nmbd
ascott wrote:
Taomyn, I get it. Thanks. That makes sense. It sounds like that when you apply changes to the network, it brings down and up the interface again, which loses the setting. This makes me think that setting interference mode should be done from the networking services routines: when an interface is brought up, check a nvram variable for interference mode and turn it on if it's set.
We were thinking we were going to put this setting into a init script, but now I see that's not going to work. I'll take a look at adding it in networking services and see if I can formulate a patch to submit. I'm thinking this might be something the dd-wrt devs would want to add to their UI: setting interference mode nvram variables that are picked up from networking services.
-ascott
Cool.
Is this setting related to the "Bluetooth Coexistence" value under WL0/WL1-Advanced in the GUI?
Device Boot Start End Blocks Id System
/dev/sdb1 1 71 527308+ 83 Linux
/dev/sdb2 71 80 68913 82 Linux swap / Solaris
/dev/sdb3 80 97 130169 b W95 FAT32
/dev/sdb4 97 132 260338 83 Linux
root@DD-WRT 45:d3:de:~# mount
rootfs on / type rootfs (rw)
/dev/root on / type squashfs (ro)
none on /dev type devfs (rw)
proc on /proc type proc (rw)
sysfs on /sys type sysfs (rw)
ramfs on /tmp type ramfs (rw)
devpts on /dev/pts type devpts (rw)
devpts on /proc/bus/usb type usbfs (rw)
/dev/discs/disc0/part1 on /opt type ext3 (rw,data=ordered)
/dev/discs/disc0/part4 on /opt type ext2 (rw)
root@DD-WRT 45:d3:de:~# cat /tmp/disktype.dump
--- /dev/discs/disc0/disc
Block device, size 7.521 GiB (8075120640 bytes)
DOS/MBR partition map
Partition 1: 514.9 MiB (539963904 bytes, 1054617 sectors from 2048)
Type 0x83 (Linux)
Ext3 file system
UUID 92DA4140-DEC1-4A64-AC15-7EE3EBA9477B (DCE, v4)
Volume size 514.9 MiB (539963392 bytes, 131827 blocks of 4 KiB)
Partition 2: 67.30 MiB (70566912 bytes, 137826 sectors from 1056665)
Type 0x82 (Linux swap / Solaris)
Linux swap, version 2, subversion 1, 4 KiB pages, little-endian
Swap size 67.29 MiB (70557696 bytes, 17226 pages of 4 KiB)
Partition 3: 127.1 MiB (133293056 bytes, 260338 sectors from 1194491)
Type 0x0B (Win95 FAT32)
FAT16 file system (hints score 5 of 5)
Volume size 126.9 MiB (133015552 bytes, 64949 clusters of 2 KiB)
Volume name ""
Partition 4: 254.2 MiB (266586112 bytes, 520676 sectors from 1454829)
Type 0x83 (Linux)
Ext2 file system
UUID 0572FBBD-6AD9-4BBD-A8CE-7F08CAC1FE22 (DCE, v4)
Volume size 254.2 MiB (266584064 bytes, 260336 blocks of 1 KiB)
Status: <b>Mounted on /opt</b>
Status: <b>Mounted on /opt</b>
Seems like the code is mounting all ext partitions to /opt/.
Posted: Thu Dec 01, 2011 0:23 Post subject: Re: Mega bins with samba no nmbd
Taomyn wrote:
Is this setting related to the "Bluetooth Coexistence" value under WL0/WL1-Advanced in the GUI?
I think bluetooth coexistence mode is different than interference mode. The btc mode is specifically for bluetooth devices. The interference mode however, is undocumented and is driver level interference detection that we've been playing with, settable from wl.
Is wl working for you on that latest build? It was not for me, but than I did a make clean somewhere and rolled again, and now it's working.
Posted: Thu Dec 01, 2011 7:54 Post subject: Re: Mega bins with samba no nmbd
ascott wrote:
Is wl working for you on that latest build? It was not for me, but than I did a make clean somewhere and rolled again, and now it's working.
Yes it works as in since it was discovered to fix the signal stability in various recent builds, but no in that the setting never stays set unless I use my sleep loop.
I perform 2 30/30/30 resets, one before the upgrade and one afterwards so I hope that's enough to ensure a clean start each time.
It's the one without Samba if it helps narrow it down further - I was not told if the rebuilt 17949 added anything more than Samba so I haven't bother wasting my time installing it as I can live without Samba until another version is released with hopefully the USB fixes.
Seems like the code is mounting all ext partitions to /opt/.
-ascott
This is my output from fsdisk and blkid
Haven't had time to change extensions will have more time in the weekend.
Code:
# fdisk -l /dev/sda
Disk /dev/sda: 1992 MB, 1992294400 bytes
62 heads, 62 sectors/track, 1012 cylinders
Units = cylinders of 3844 * 512 = 1968128 bytes
Disk identifier: 0x00028a37
Device Boot Start End Blocks Id System
/dev/sda1 1 274 526597 83 Linux
/dev/sda2 275 309 67270 82 Linux swap / Solaris
/dev/sda4 310 1012 1351166 83 Linux
thanks Fractal. any chance of a openvpn_small compile as well? seems like brainslayer has abandoned his regular builds, and we have many E1000 customers who may benefit.
daveM: attached is an openvpn_small 17952 build using 4.1.2 toolchain. I tested on a WNR3500L so I can guarantee it will not brick, but I cannot guarantee that all the openvpn stuff works. Let us know.
Posted: Thu Dec 01, 2011 9:53 Post subject: Patch for USB fix - Successful testing
Hello,
Here's the preprocessor-based solution to the USB fix that I had mentioned.
The reasoning behind it is that 17841 added support for ext4 to disktype which checks the kernel for features which are not fully supported until kernel 2.6.26 or later. For Broadcom-based users who are on 2.6.24.111 kernel, this means automounting suddenly stopped working.
We have rolled and tested, here on the forums, current revisions with 17841 manually reverted. The testing has been successful, so I'm proposing the following patch.
17841 modified the ext23 function and changed it to ext234. I added preprocessor conditionals to build the old body of the function if < 2.6.26 kernel, otherwise build new version of the function.
In order to use it you have to include linux/version.h. In order to do that you have to add -I$(LINUXDIR)/include to CFLAGS in the disktype/Makefile. I have rolled firmware with this patch and tested it on e3k and WNR3500L.
The dual-mounting is not so much a bug as the way that usb_hotplug.c appears to be coded for multiple partitions. I'm going to look at it some more and Kong's patches again. Maybe the best way to handle multiple partitions is to name them with a label. I was thinking it might be cool to just use nvram variables. usb_mntpoint=opt, usb_mntpoint2=jffs, usb_mntpoint3=store. (see code snippet at bottom for relevant lines of usb_hotplug.c -- everything falls into the else)
I view dual-mounting opt as another problem to look into, so if people are cool with this proposed patch, I'll submit it to Trac.
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,26)
+
/*
- * ext2/ext3/ext4 file system
+ * ext2/ext3/ext4 file system - two different functions depending on kernel version
*/
this is changing bin size a little, better compression I am guessing.
I am running the nv60k on my E3000 right now... seems to work well... I am using wl interference 2 to lower noise and its terrific.... I have 100mb/s upload/download and over wifi I am getting 78mb/s down and 65mb/s up... that is great 2.4ghz -N
-Fractal.
Fractal,
Thank you for the good work. This build is running stable with my E3000 router (I was running build 15962). Good wifi range. Keep the good work.
Posted: Thu Dec 01, 2011 19:21 Post subject: Build test 17949
this and my HD mounted
--- / Dev/discs/disc0/disc
Block device, size 232.9 MiB (250,059,350,016 bytes)
DOS / MBR partition map
Partition 1: 232.9 MiB (250,056,705,024 bytes, 488392002 sectors from 63)
Type 0x83 (Linux)
Ext3 file system
UUID 763D7CDE-2F5B-419F-B751-513F0BFB54A5 (DCE, v4)
Volume size 232.9 MiB (250,056,704,000 bytes, 61049000 blocks of 4 KiB)
Status: Mounted on / mnt
Status: Not mounted
Okay?
Congratulations Fractal continues ..
It would be nice if you do a custom build for E4200 as a dimension where it can get even at 12MB ....
_________________ Buffalo WZR-HP-AG300H DD-WRT v24-sp2 (03/19/12) std - build 18777 + modem Digicom combo cx + Nas Synology DS-209 da 2 T + LG DP1W+ APC Back-UPS RS-Power Saving Back-UPS Pro 550 + switch Zyxel Gs105B + http://forum.emulesecurity.net/topic.asp?TOPIC_ID=3441
thanks Fractal. any chance of a openvpn_small compile as well? seems like brainslayer has abandoned his regular builds, and we have many E1000 customers who may benefit.
daveM: attached is an openvpn_small 17952 build using 4.1.2 toolchain. I tested on a WNR3500L so I can guarantee it will not brick, but I cannot guarantee that all the openvpn stuff works. Let us know.
-ascott
this appears to be working fine with my E1000 v2.1 test unit. thanks!