OpenWrt Forum Archive

Topic: davidc502 1900ac 3200acm builds

The content of this topic has been archived between 26 Feb 2018 and 7 May 2018. Unfortunately there are posts – most likely complete pages – missing.

yes, i am talking about patching dts for 4.9 kernel by adding :
  flash@d0000 {
                               ................................................
                               nand-ecc-strength = <4>;
                               nand-ecc-step-size = <512>;
                               ................................................

maybe i am wrong, but seems that last r6302 does not include this dts patch.

hnyman wrote:
davidc502 wrote:

Patch 005-linksys_hardcode_nand_ecc_settings.patch will be introduced in the next kernel 4.14.x as I don't see any patches for 4.4 or 4.9.

To my understanding, for 4.4 and 4.9 only the DTS file needs to be patched (as the DTS is part of Openwrt sources).

For 4.14 there is a separate patch, as the DTS itself is upstreamed to Linux sources.

https://git.openwrt.org/?p=openwrt/stag … 08d3403a83

I can confirm this, I spoke by email with Kaloz.  He said patching the DTS files will fix the WRT3200ACM Winbond flash issue on 4.4 and 4.9 Kernel builds. The difference is that on builds before 4.11 kernel, the flash chip will not be recogized and show as 'unknown' in the devices.  This is purely cosmetic however.

I have no issues adding those lines to the rango.dts.  Thanks for telling me.

Hello friends,

happened something strange after the installation and configuration of the last build, I did the installation last night and today in the course of the day the WRT1900ACS started re-starting alone, did it for about three times before I shut it down and go back to the firmware of manufactures.

I do not know if the problem is directly related to the LEDE, but what I noticed abnormal was that the temperature was very high in 83C and with peaks of 87C, I had heard that these processors work in these temperature media, although if this is I really do not understand why Linksys did not do a better preparation for thermal protection.

well after returning to the factory has been an hour and WRT is stable, but I can not know to what temp, the same is working, at night I will re-install the LEDE again and try to isolate the problem, that this information can help the project.

If you have any suggestions or suggestion of how I can do to speed up the process, I am grateful.

gu3d3s wrote:

Hello friends,

happened something strange after the installation and configuration of the last build, I did the installation last night and today in the course of the day the WRT1900ACS started re-starting alone, did it for about three times before I shut it down and go back to the firmware of manufactures.

I do not know if the problem is directly related to the LEDE, but what I noticed abnormal was that the temperature was very high in 83C and with peaks of 87C, I had heard that these processors work in these temperature media, although if this is I really do not understand why Linksys did not do a better preparation for thermal protection.

well after returning to the factory has been an hour and WRT is stable, but I can not know to what temp, the same is working, at night I will re-install the LEDE again and try to isolate the problem, that this information can help the project.

If you have any suggestions or suggestion of how I can do to speed up the process, I am grateful.

After the install, take a look at htop and see if there's a process or processes running 100% of the processor.

(Last edited by davidc502 on 27 Feb 2018, 21:39)

Hi David,
I have wrt3200acm LEDE 4.9.82.
I could not install minidlna on r6302-0f54d96d24.

Lede SNAPSHOT, r6302-0f54d96d24
~# opkg install minidlna
Unknown package 'minidlna'.
Collected errors:
* opkg_install_cmd: Cannot install package minidlna.
~# opkg install luci-app-minidlna
Unknown package 'luci-app-minidlna'.
Collected errors:
* opkg_install_cmd: Cannot install package luci-app-minidlna.

(Last edited by o9uz on 27 Feb 2018, 22:03)

davidc502 wrote:

After the install, take a look at htop and see if there's a process or processes running 100% of the processor.

http://i68.tinypic.com/11abj93.jpg


living and learning, there is the screen, re-installed the previous vs by precuation because could not leave the router stopped, at the end of the week back to new and testo from home with calm.

Thanks for help and patience  wink

seb101 wrote:

I can confirm this, I spoke by email with Kaloz.  He said patching the DTS files will fix the WRT3200ACM Winbond flash issue on 4.4 and 4.9 Kernel builds. The difference is that on builds before 4.11 kernel, the flash chip will not be recogized and show as 'unknown' in the devices.  This is purely cosmetic however.

  it is not only cosmetic. units with new winbond nand cant boot ( it will be stuck after kernel load by uboot. kernel will not recognize nand flash and cant mount rootfs. this was reported for a while and nobody could make it working on new units until this patch).

davidc502 wrote:

I have no issues adding those lines to the rango.dts.  Thanks for telling me.

Thanks

I think @seb101 reference to cosmetic was the report in the log output, but that the patches from github for 4.4, 4.9, 4.14 all work. I think all have bee proven at this point.

Hi David,
May I ask a question about compile mwlwifi?

I'm using your image.
Model    Linksys WRT3200ACM
Firmware Version    Lede SNAPSHOT r5917-36f1978a70 / LuCI Master (git-18.020.50283-194e429)
Kernel Version    4.9.77

steps:
1. git clone lede source and git checkout -b for_r5917 36f1978a70

2. download your (config.seed) r5917, rename to .config

3. update package/kernel/mwlwifi/Makefile

PKG_VERSION:=10.3.4.0-20180226
PKG_SOURCE_VERSION:=51350a3e085a811179a270baddcae03ec9384f07

4. make -j10

5. scp kmod-mwlwifi_4.9.77+10.3.4.0-20180226-1_arm_cortex-a9_vfpv3.ipk to my router

6. opkg install kmod-mwlwifi_4.9.77+10.3.4.0-20180226-1_arm_cortex-a9_vfpv3.ipk

Here is the output

root@LEDE:~# opkg install kmod-mwlwifi_4.9.77+10.3.4.0-20180226-1_arm_cortex-a9_vfpv3.ipk
Upgrading kmod-mwlwifi on root from 4.9.77+10.3.4.0-20180118-1 to 4.9.77+10.3.4.0-20180226-1...
Collected errors:
 * satisfy_dependencies_for: Cannot satisfy the following dependencies for kmod-mwlwifi:
 *     kernel (= 4.9.77-1-de92b0c8319b0c317e238fde62507e64) *
 * opkg_install_cmd: Cannot install package kmod-mwlwifi.


I don't know why the kernel version magic is not the same.

Here some infos for you.

root@LEDE:~# opkg info kmod-mwlwifi
Package: kmod-mwlwifi
Version: 4.9.77+10.3.4.0-20180118-1
Depends: kernel (= 4.9.77-1-b629ab20ff6ea798f323eff7cd5beb94), kmod-mac80211
Status: install user installed
Section: kernel
Architecture: arm_cortex-a9_vfpv3
Size: 717300
Filename: kmod-mwlwifi_4.9.77+10.3.4.0-20180118-1_arm_cortex-a9_vfpv3.ipk
Description: Marvell 88W8864 wireless driver
Installed-Time: 1516487765

Please help me and point out where  I did sth wrong.

Thanks.

(Last edited by hungrybirder on 28 Feb 2018, 09:10)

To get ipkg that satisfies dependencies you should follow below steps:

1. git clone lede source

2. git checkout -b r5917 36f1978a70

3. ./scripts/feeds update -a

4. ./scripts/feeds install -a

5. download config.seed from r5917, mv config.seed .config

6. make defconfig

7. update package/kernel/mwlwifi/Makefile
PKG_VERSION:=10.3.4.0-20180226
PKG_SOURCE_VERSION:=51350a3e085a811179a270baddcae03ec9384f07

8. make -j10 (to compile everything)

9. scp kmod-mwlwifi_4.9.77+10.3.4.0-20180226-1_arm_cortex-a9_vfpv3.ipk to my router

10. opkg install kmod-mwlwifi_4.9.77+10.3.4.0-20180226-1_arm_cortex-a9_vfpv3.ipk


You were missing 3,4 and 6 (I think latter is just simple omission) and . The important thing is that you have to do ./scripts/feeds update/install -a before you do make defconfig. Without it you won't get matching kernel configuration with official build, despite of using proper config.seed file.

Please let me know if it helped.

(Last edited by gpiszczek on 28 Feb 2018, 10:12)

@gpiszczek

It works with your steps.

Thanks a lot.

^_^

Villeneuve wrote:

I think @seb101 reference to cosmetic was the report in the log output, but that the patches from github for 4.4, 4.9, 4.14 all work. I think all have bee proven at this point.

Yes exactly that, thanks Villeneuve. 

The original issue was present in all kernels when using Winbond flash.  After patching DTS files the issue is fixed in 4.4 and 4.9 but will report as 'Unknown' flash chip in the Kernel log.  After 4.11 it will correctly report as 'Winbond' flash in the Kernel log.

[    0.951515] nand: device found, Manufacturer ID: 0xef, Chip ID: 0xda
[    0.957896] nand: Winbond W29N02GV
[    0.961321] nand: 256 MiB, SLC, erase size: 128 KiB, page size: 2048, OOB size: 64
[    0.968926] pxa3xx-nand f10d0000.flash: ECC strength 16, ECC step size 2048

@seb101 : got it. thanks

Can't wait for next release with the new wifi drivers.

Keltere wrote:

Can't wait for next release with the new wifi drivers.


*EDIT*

Never mind... The pull request after the latest wifi version change deals with "Cleaned up styling. Moved text around to and added more words…"

I swear the list changes some times depending on the time you check.

The latest driver has already been available for download.

(Last edited by davidc502 on 28 Feb 2018, 20:06)

David, i have not followed every detail
On the driver issues, and figured it is with he 3200 series, not the 1900acs. So anyway, just wondering are the 1900 series pretty much wrapped up with wifi drivers? I tested a previous firmware late last month with great results, just minor issues and waited for a more final release.

davidc502 wrote:
Keltere wrote:

Can't wait for next release with the new wifi drivers.


*EDIT*

Never mind... The pull request after the latest wifi version change deals with "Cleaned up styling. Moved text around to and added more words…"

I swear the list changes some times depending on the time you check.

The latest driver has already been available for download.

Yup i was talking about 10.3.4.0-20180226, i can't replace them manually so i need to wait for an update.
Btw good job with this firmware

Notsofast wrote:

David, i have not followed every detail
On the driver issues, and figured it is with he 3200 series, not the 1900acs. So anyway, just wondering are the 1900 series pretty much wrapped up with wifi drivers? I tested a previous firmware late last month with great results, just minor issues and waited for a more final release.

The firmware changes certainly has been 3200acm driven, but the wifi driver from what I recall may affect more than just the 3200acm.

Keltere wrote:
davidc502 wrote:
Keltere wrote:

Can't wait for next release with the new wifi drivers.


*EDIT*

Never mind... The pull request after the latest wifi version change deals with "Cleaned up styling. Moved text around to and added more words…"

I swear the list changes some times depending on the time you check.

The latest driver has already been available for download.

Yup i was talking about 10.3.4.0-20180226, i can't replace them manually so i need to wait for an update.
Btw good job with this firmware

If you can download "putty", and log into the router with the same user/pass as LuCi.. I think I can get you on the new wifi driver with only 4 commands. It would be a copy paste for you. That's assuming you have the latest build installed.

You can look at FAQ here >> https://davidc502sis.dynamic-dns.net/faq/

FAQ #5

(Last edited by davidc502 on 28 Feb 2018, 20:35)

Ah... So no need to disable the wifi before?

EDIT: never mind... wifi and internet is not more working...Nice.

(Last edited by Keltere on 28 Feb 2018, 21:59)

Keltere wrote:

Ah... So no need to disable the wifi before?

EDIT: never mind... wifi and internet is not more working...Nice.

No need to disable wifi as the driver is already loaded.

Sounds like other issues?

davidc502 wrote:
Keltere wrote:

Ah... So no need to disable the wifi before?

EDIT: never mind... wifi and internet is not more working...Nice.

No need to disable wifi as the driver is already loaded.

Sounds like other issues?

Maybe a permission problem?

Here is my steps:

cd /tmp
wget davidc502sis.dynamic-dns.net/releases/kmod-mwlwifi_4.9.82+10.3.4.0-20180226-1_arm_cortex-a9_vfpv3.ipk
opkg install kmod-mwlwifi_4.9.82+10.3.4.0-20180226-1_arm_cortex-a9_vfpv3.ipk
cd /lib/firmware/mwlwifi/
mv 88W8964.bin 88W8964.bin.old
wget github.com/kaloz/mwlwifi/blob/51350a3e085a811179a270baddcae03ec9384f07/bin/firmware/88W8964.bin
reboot

EDIT: now it seems it's working, i did nothing except waiting... Strange.

(Last edited by Keltere on 1 Mar 2018, 00:04)

davidc502 wrote:

r6302 has been uploaded to the server and is available for download for the 3200acm, 1200ac, 1900acs, and 1900acv2.

Very nice.  I note that 1900ac was not listed, nor is it linked to
from your firmware page.  Yet I do see a build for the 1900ac
was none the less posted under snapshots/r6302.

Since I own a 1900acv1, would I be foolish to try it?

OK,  i found all the latest files... checked the dates and they are all 2-24-18. I am assuming the img file has all the updates, and does not need manual wifi driver insertion package.  If not I am wondering what happened to keep the package from the build when compiling.