bodhi, Only now had the chance to test the latest DTB: root@silverdebian:~# dmesg [ 0.000000] Booting Linux on physical CPU 0x0 [ 0.000000] Linux version 4.12.1-kirkwood-tld-1 (root@tldDebian) (gcc version 4.9.2 (Debian 4.9.2-10) ) #1 PREEMPT Sat Jul 15 21:40:50 PDT 2017 [ 0.000000] CPU: Feroceon 88FR131 [56251311] revision 1 (ARMv5TE), cr=0005397f [ 0.000000] CPU: VIVT databy garagoyun - Debian
@generic, @bodhi, I appreciate you have included me in the DTS, but I have not really contributed to it :-) . You both are the sole contributors.by garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > garagoyun, > > Here is the revised DTS and DTB. Please try. Thanks bodhi, I will try as soon I can find some free time. @generic, You have been using this device with new kernel for years you mentioned. Was just curious, what are you using it mainly for, any specific application?by garagoyun - Debian
generik, Excellent work. I saw a similar code in the original source a few days ago. I thought it was relevant to help overclock the device as the ratio CPU to DDR was 2/5. I thought changing the ratio might increase the CPU frequency, however, it was only affecting the sysclock. CPU clock stayed 1000MHz. ./build/u-boot-3.5.9/board/mv_feroceon/mv_kw/kw_family/ctrlEnv/mvCtrlEnvRegs.h /*by garagoyun - Debian
generik, Thanks for your contribution. Highly appreciated.by garagoyun - Debian
Quotebodhi Wish we knew about before that this is available somewhere :) But it is OK, is was a good excercise for you, I hope :) I wish as well. But it was a good exercise indeed. I have learnt quite alot starting from zero knowledge. Quotebodhi It also mean that we should not ever try to mainline our version of the DTS, as a courtesy to Eric. I believe his DTS is complete. I havent haby garagoyun - Debian
bodhi, The TDS file from Erik has some additional GPIO as well as Silverstore mtdparts. I have attached a diff file, it includes the code from Erik against your DTS and I have slightly modified some other parts. I could not convert dts to dtb so could not try the file from Erik.by garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > @garagoyun, > > Does it say on the chip marking it is 88F6283? I just had to completely disassemble the device to see the marking. It is 88F6283 indeed.by garagoyun - Debian
generik Wrote: ------------------------------------------------------- > Hello there, > > i'm running this device for years with the > mainline kernel myself, maybe i can help out > here. > > as for the cpu frequency the device is based on a > armada 310 (88F6283) this is a low power variant > of the 88F6282 with only up to 1 GHz cpu clock. > http:/by garagoyun - Debian
bodhi, Could you tell me what is the relationship between values in the arrays? It should manipulate the values of cpu freq, l2 and ddr, but i am not sure how exactly: #ifndef MV_ASMLANGUAGE /* CPU clock for 6281,6192,6282 0->Resereved */ #define MV_CPU_CLCK_TBL { 0, 400000000, 0, 0, \ 600000000, 0, 800000000, 1000000000, \ 0, 1200000000, 0, 0, \by garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > So since you don't have serial console on the 2nd > box, this 2nd box u-boot flashing should go like > this: > > 1. Log in the first box. Make sure you are booting > the new u-boot image on flash. Run fw_printenv and > fw_setenv to verify that it works. > 2. Dump the mtd1 where the envs aby garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > Check your /etc/fw_env.config in Debian rootfs on > the 2nd box. Make sure it is correct and the same > as the 1st one. bodhi, Both are the same. Debian: root@silverdebian:~# cat /etc/fw_env.config # MTD device name Device offset Env. size Flash sector size Number of sectors /dev/mby garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > garagoyun, > > A nit pick. Older u-boot used old naming > convention. > > Usually in modern Marvell u-boots we use the > extension .kwb to indicate that the image is > NAND/SPI flash-ready (there is a header in front): > u-boot-3.5.9_HNCN2.kwb. We use .bin to indicate > that the imageby garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > garagoyun, > > Thanks! It looks very good. Would not have been possible without your support.by garagoyun - Debian
bodhi, As promised I have attached the patch and the new u-boot.bin resulted from this patch. I have tested the u-boot with kwboot only, however, I believe it should be fine flashing it as well. Below the boot log: $ ./kwboot -t -B 115200 /dev/ttyUSB0 -b u-boot-3.5.9_HNCN2_bootcmd_exec.bin -p Sending boot message. Please reboot the target.../ Sending boot image... 0 % [.........by garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > the only line you'll need to > change is: > > > + sprintf(buff, "run bootcmd_exec; nand reset; > nand read.e 0x800000 0x%8.8x 0x%8.8x; ", > + nand_layout_ptr->actually_offset + > nand->erasesize, > + (nand_layout_ptr->len + nand->oobblock)& > 0xfffffe0by garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > As I mentioned, it is either: > > 1) They used an old u-boot that they aleady have > built for 1Ghz, and did not build new u-boot so > the CPU was limited to 1Ghz. And they hope normal > users would not notice that. > or > 2) They knew the specific CPUs lot that they > purchase to useby garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > garagoyun, > > Cool :) > > QuotePlease note, I have no knowledge in coding > or u-boot in general, so there might be some > issues with the rebuilt u-boot. But it does boot > Debian, even after I have reset the envvironments > to defaults. I just need to adjust mtdparts env > and maybby garagoyun - Debian
bodhi, I have amended a bit the patches in the source code and rebuild the u-boot. Tests are positive and the new u-boot does boot straight to Debian. $ ./kwboot -t -B 115200 /dev/ttyUSB0 -b u-boot-db88f6282bp_400ddr3db_uart-test1.bin Sending boot message. Please reboot the target...\ Sending boot image... 0 % [......................................................................]by garagoyun - Debian
Martin aus Dortmund, If you would like to install OMV on armv5 the easiest way is via softy script https://github.com/armbian/config. The script is optimised for armv7, however, some of the software installations work on armv5 out of the box, others do after some changes in the script. I have tested OMV. It did install and was running fine, but most of the extra plugins would not work on aby garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > garagoyun , > > Please post ./u-boot-3.5.9/include/config.mk here. That's the config.mk: ARCH = arm CPU = arm926ejs BOARD = config_kw VENDOR = mv_feroceon MV_OUTPUT = db88f6282bp MV_FLAGS += -DMV88F6282 MV_FLAGS += -DDB_88F6282A MV_IMAGE_FLAGS = -DMV_SEC_256K MV_IMAGE_FLAGS += -DMV_BOOTSIby garagoyun - Debian
bodhi, I was looking through configuration files in the source code the other day and inside the generic.mk file I saw the following: /lib/libc.so.6: ifeq ($(strip $(ARCH)),arm) cp -a /scratchbox/compilers/arm-linux-cs2009q3-67/arm-none-linux-gnueabi/libc/lib/* /lib cp -a /scratchbox/compilers/arm-linux-cs2009q3-67/arm-none-linux-gnueabi/libc/usr/lib/libstdc++* /lib endif I thougby garagoyun - Debian
bodhi, I tried building the u-boot natively as well tonight, but got some dependencies as well.by garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > garagoyun, > > Unfortunately, the only Kirkwood box that I have > released u-boot for, which has 6282 and also DDR3 > is the NS325! > > Try the iConnect > (uboot.2017.07-tld-1.iconnect.bodhi.tar). This has > the 6281, 1GHz, but DDR2. I think it will likely > have the same behvior asby garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > garagoyun, > > Stock u-boot make does a whole bunch of > things right. But it did not build u-boot. I don't > think this is possible without doing further > patching in their source tree. Their released GPL > just does not build properly. > > QuoteDid you try kwboot with the NSA325by garagoyun - Debian
bodhi, I have received a reply from the vendor and there is no positive news. Basically, the reply was that I have been forwarded already everything. Would you reckon it is almost impossible to rebuild the u-boot without that extra information? Do I need to start learning how to reverse engineer one now :-)by garagoyun - Debian
bodhi, I have sent a request to the vendor. Fingers crossed.by garagoyun - Debian
Bodhi, I have built from source a few packages in my life :-), but never an u-boot or cross compiled. No experience, so wouldn't even know where to start from. Otherwise, I would have, at least tried, building it personally. If there is anything I could be of help, I would be glad.by garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > garagoyun, > > Looks like something is missing in the GPL > tarball. Is there build instruction in there? > usually there is a README file for some > instruction to build u-boot. That's all I have received as zip file. If there is anything in particular I could request, but the support fromby garagoyun - Debian
bodhi Wrote: ------------------------------------------------------- > FYI, > > I am going to release kernel 4.19 and then get > back to rebuilding this stock u-boot. Thanks, appreciated!by garagoyun - Debian