Hi bodhi, Quote Ah, that's a bit different, the Mirabox has 88E1510. I have compared Marvells Product Briefs for the two devices. They look very simular. The features highlighted are: 88E1310 (RN102): "The devices support RGMII (Reduced pin count GMII for direct connection) to Copper/Fiber/SGMII with Auto-Media Detect." 88E1510 (Mirabox): "In addition to supporting Enby tme - Debian
Hi Koen, Apparently, 'thumb2' is not set in Netgear's stock firmware: $ grep -i thumb2 netgear/.config # CONFIG_THUMB2_KERNEL is not set $ ls -alF netgear lrwxrwxrwx 1 tme tme 80 nov. 28 09:44 netgear -> /home/tme/ReadyNAS-RN102/src/Netgear/ReadyNASOS_V6.10.3_WW_src/linux-4.4.190-arm/ But maybe it should have been? Regards, Trond Melenby tme - Debian
Hi bodhi, Quote The reasoning: the Mirabox and the RN102 has identical SoC and therefore using the same network chip (is my assumption correct?), and David can get 800's Mbs with iperf test on the Mirabox (running kernel 5.8.5). I don't know about the Mirabox, but the RN102 has a Marvell 88E1318 Gigabit Ethernet PHY (ref). 800 Mb/s is better, but not good. With stock firmware tby tme - Debian
Hi bodhi, This, I believe, are all the Netgear patches relevant to the Ethernet driver: diff -crw mainline/drivers/net/ethernet/marvell/Kconfig netgear/drivers/net/ethernet/marvell/Kconfig *** mainline/drivers/net/ethernet/marvell/Kconfig 2020-11-28 09:39:13.153766986 +0100 --- netgear/drivers/net/ethernet/marvell/Kconfig 2020-02-05 08:32:17.000000000 +0100 *************** *** 72,77 ***by tme - Debian
Hi bodhi, This is off topic now, since the Armade 370 L2 cache issue is solved. The topic has somehow slide into Network performance in stead, so feel free the reorganize as you see fit. I have sucessfully installed your Linux kernel 5.9.3 on my Netgear ReadyNAS RN102: $ ssh debian.local tme@debian.local's password: Linux debian 5.9.3-mvebu-370xp-tld-1 #1.0 SMP PREEMPT Tue Nov 24by tme - Debian
Hi bodhi, I added 'lm75' and 'g762' to '/etc/initramfs-tools/modules' and, as 'root', did update-initramfs -u (cd /boot && mkimage -A arm -O linux -T ramdisk -C gzip -a 0x00000000 -e 0x00000000 -n initramfs-5.8.5-mvebu-tld-1 -d initrd.img-5.8.5-mvebu-370xp-tld-1 uInitrd) to update the initial ram file system. Installed 'lm-sensors'by tme - Debian
Hi bodhi, root@rn102:~# ip link show 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 532 link/ether 12:34:56:78:90:ab brd ff:ff:ff:ff:ff:ff tme@debian:~$ ipby tme - Debian
Hi bodhi, Quote I will build the new kernel 5.9.3 with danitool patch, and see if that will make up for the difference. If you build it for RN102 too, I'll be happy to test it. Quotewhitepawn Driver of g762 fan control chip seems not included in kernel i guess. If you have any idea on how to fix the outstanding fan control issue, please add your best guess for me to test as well.by tme - Debian
Hi bodhi, Here the output of 'ethtool' from both of my Netgear RedyNAS RN102 boxes: root@debian:~# uname -a Linux debian 5.8.5-mvebu-370xp-tld-1 #1.0 SMP PREEMPT Mon Aug 31 00:00:32 PDT 2020 armv7l GNU/Linux root@debian:~# ethtool --version ethtool version 4.19 root@debian:~# ethtool eth0 Settings for eth0: Supported ports: [ TP MII ] Supported link modes: 10baseT/Hby tme - Debian
Hi nickba, If you switched Linux kernel, you may try to run Debian with the old stock kernel. And vice verse. It may not work, but if it one of the two does, at least you know which part of your current firmware to blame. Regards, Trond Melenby tme - Debian
I have repeated the above tests on my second Netgear ReadyNAS RN102 running the current stock firmware version 6.10.3: # uname -a Linux rn102 4.4.190.armada.1 #1 SMP Mon Oct 28 02:07:39 UTC 2019 armv7l GNU/Linux # cat /etc/debian_version 8.11 # dmesg | grep mvebu-pmsu [ 0.002181] mvebu-pmsu: Initializing Power Management Service Unit Register snooping: # busybox devmem 0xd0008100by tme - Debian
Hi AkkJaa, You need a terminal window, for instance "GNOME terminal". You may try holding down both the 'Ctrl' key and the 'Alt' key and then hit the 'T' key. If no terminal window pops up, search the menus on your desktop for a possibility to choose 'terminal' of 'command line'. Regards, Trond Melenby tme - uBoot
Hi AkkJaa, What output did you get from the command bodhi suggested? apt-get install picocom Regards, Trond Melenby tme - uBoot
I repeated the 'iperf' test, but now with the RT102 as the server. This is from the laptop: $ iperf -c 192.168.1.141 ------------------------------------------------------------ Client connecting to 192.168.1.141, TCP port 5001 TCP window size: 612 KByte (default) ------------------------------------------------------------ [ 3] local 192.168.1.200 port 44248 connected with 19by tme - Debian
Quotebodhi One test left to do is the iperf test. In this test, the laptop's IP address was 192.168.1.200. This is from the box: # iperf -c 192.168.1.200 ------------------------------------------------------------ Client connecting to 192.168.1.200, TCP port 5001 TCP window size: 43.8 KByte (default) ------------------------------------------------------------ [ 3] local 192.168by tme - Debian
Hi bodhi, On my Netgear ReadyNAS RN102, I installed and tested Linux kernel '5.8.5-mvebu-370xp-tld-1' with the dtb-file you posted on November 14th. From the boot log: [ 0.000000][ T0] L2C: DT/platform modifies aux control register: 0x12086300 -> 0x1a086302 [ 0.000000][ T0] Aurora cache controller enabled, 4 ways, 256 kB [ 0.000000][ T0] Aurora: CACHE_ID 0x0by tme - Debian
Hi AkkJaa, You are running an outdated, unsupported and obsolete version of Linux, so you can't expect programs depending on current libraries to work. > And if there is a way to upgrade this Fedora to a better version. Or maybe run Ubuntu on it. I don't know what is best. Or Debian. I recommend you do a clean install of Fedora, Ubuntu or Debian from a bootable USB stick witby tme - uBoot
QuoteAkkJaa I have now install a new Fedora C9 on a other computer Hmm. Do you really mean "Fedora Core version 9 (Sulphur)"? According to Wikipedia, this version has not been supported since 2009-07-10. Regards, Trond Melenby tme - uBoot
Hi AkkJaa, The 'picocom' installation RPM files for Fedora 30, 31, 32 and Rawhide are available here, but 'minicom' should work too. Good luck! Regards, Trond Melenby tme - uBoot
Hi whitepawn, I agree with bodhi that your limited Ethernet bandwidth could be caused by the QoS (Quality of Service) settings in the router. If disabling QoS does not help, you may completely remove the router from the test and connect the RN102 directly to your PC. Either both parties should be configured for static IP addresses in the same range (e.g. 192.168.5.xxx), or one of the two maby tme - Debian
Hi bodhi and whitepawn, I was curious how the Netgear Stora released in 2009 with a Marvell 88F6281 Kirkwood Feroceon SoC compares to the 4 years newer ReadyNAS RN102 with a Marvell 88F6710 Armada 370 SoC. Transcript of test session below. My observations: Pulling a 1 GB file over 'sftp' took 2:43 min (10% slower). Pushing a 1 GB file over 'sftp' took 6:37 min (2.5 tiby tme - Debian
Hi whitepawn, To assist in debugging your network issue, here are some more output from 'ethtool' that you may compare to yours: $ sudo ethtool --show-pause eth0 Pause parameters for eth0: Autonegotiate: on RX: off TX: off RX negotiated: on TX negotiated: on $ sudo ethtool --show-coalesce eth0 Coalesce parameters for eth0: Adaptive RX: off TX: off stats-block-usecs:by tme - Debian
Quote Trond, Please run the same iperf tests as whitepawn did. -bodhi Sure bodhi, $ iperf -n2000M -i 3 -c 192.168.1.128 ------------------------------------------------------------ Client connecting to 192.168.1.128, TCP port 5001 TCP window size: 43.8 KByte (default) ------------------------------------------------------------ [ 3] local 192.168.1.109 port 42516 connected with 192.16by tme - Debian
Hi whitepawn, The RN102 is a file server, so I think the only performance measure that really counts is the bandwidth into and out of the box. For the tests below, both the RN102 and my laptop were cabled to the same gigabit router. I have 2 mirroring 2 TB HDDs in the box (RAID-1). I copied the file 'film.mp4' back and forth. Its size is 1 GB and it contains a 3+ hour movie. Sessionby tme - Debian
Hi bodhi, You are right. I was right about the '/dev/sd?1', but wrong about the LEDs. Here, initially, the only disk is attached to the eSATA connector, then a HHD is hot inserted into slot 1, an then another one into slot 2: root@debian:/tmp# lsblk NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT sda 8:0 0 149.1G 0 disk `-sda1 8:1 0 149.1G 0 part / mtdblocby tme - Debian
Hi bodhi and whitepawn, Quotemyself 'backup.mtd0' was created by the command nanddump --noecc --omitoob -l 0x80000 -f backup.mtd0 /dev/mtd0 Why isn't this image suited for 'kwboot -p'? Not much of a mystery, actually. The size of the 'mtd0' partition is 1.5 MiB and more that the first 0x80000 B (512 KiB) are used by the U-Boot image: # cat /proc/mtdby tme - Debian
Hi bodhi and whitepawn, Quote Also in u-boot console, attach an HDD to to eSATA port, and try both of these commands, one of them should initialize all the SATA disks in the system. I have modified my eSATAp to SATA adapter cable so that 5 V power is provided from an external power supply. 3.3 V and 12 V pins are unconnected. The disk is powered up before the box. A HGST 1 TB HDD from Aprilby tme - Debian
Hi whitepawn, Quote `Insert happy smile face emoji here.` Great! Congratulation. I will try to reproduce your result. Regards, Trond Melenby tme - Debian
Hi bodhi, Just to experience the expected behavior, I decided to use 'kwboot' to upload the 'mtd0' backup from my second RN102 which still runs stock Netgear firmware: $ kwboot -b backup.mtd0 -p -t /dev/ttyUSB0 Sending boot message. Please reboot the target...- Sending boot image... 0 % [......................................................................] 1by tme - Debian
Quote Yes, it is no e-SATAp, but neither standard 7 pin e-SATA. If you looked connector inside there are also some pins which I think are power pins. My current understanding is that the power pins are there, but unconnected. Can you please verify? Regards, Trond Melenby tme - Debian