Jump to content


  • Posts

  • Joined

  • Last visited

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. indeed I tried Bionic image and it did mount ROOTFS but Buster not and throws mmc1 (sdcard) TIMEOUTs @balbes150 How can it be since it is supposed that Bionic and Buster images share the same Kernel and DTB ?
  2. Can you tell us which DTB are you using ? Tried /dtb/rockchip/rk3399-rock-pi-4.dtb and it boots but it cannot detect SD card so it cannot mount the rootfs Or you boot with USB pendrive ? TIA
  3. Did that after reverting and didn't work. Reported to Martin and I have to send logs to to linux-amlogic working and not working. Haven't had time to replicate the original incorrect behaviour BTW just been testing X96Max Plus box and it seem it is suffering from the same
  4. See Lastest images from Balbes broke some PHYs. I noted the problem is with a patch that it is used on newer Kernels and how the some PHY parameters are parsed from DTBs https://patchwork.kernel.org/patch/10985167/ So I reverted those changed on DTB and it worked. Please try attached DTB meson-gxm-q200.dtb
  5. I rolled back this patch and worked https://github.com/torvalds/linux/commit/f29cabf240ed6e67993f17594e5e6fffc5bc07e0 Funny think that I tried to increase reset-deassert-us to 1000000 as th old value suggest and didn't worked. Wonder if some manufacturers are using knockoff PHYs
  6. Hi Balbes I'm trying to work around a problem with PHY on a GT1-A Ultimate (gxmm-q200) that used to work on older kernels but not in your last 5.6.0rc image. I'm getting sometimes: [ 15.867041] meson8b-dwmac c9410000.ethernet eth0: no phy at addr -1 [ 15.867079] meson8b-dwmac c9410000.ethernet eth0: stmmac_open: Cannot attach to PHY (error: -19) I'm using the correct DTB since I remember using it on older Kernel before your unified you builds; box uses an external PHY Funny thing is if I boot with on old image of LE it works and after that it works with your latest image after a couple of reboots. Being an external PHY and seeing that DTS talks about a reset signal shared with some led the GT1 does not have a I think it may be the culprit. Anyway, where is your Amlogic branch you're using ? It is https://github.com/150balbes/Amlogic_s905-kernel which branch ? Need to start dissecting the problem. TIA
  7. Balbes Maybe you remind I sent you the gt1-mini-a DTB time ago Now I'm playing with GT1-Mini-2 which is a S905X3, anyway it does boot with the GT1-Mini-A DTB (SEI610 boots but no Ethernet) when I get some time I'll mix both DTBs if some new stuff on the S905X3 merits But as you now this new GT1-Mini-2 has the messed HDMI colors so I installed your u-boot.sd on my card and everything works fine. I noted that basically you load the new uBoot if a u-boot.ext exists but this is on s905_autoscript So I wonder if I burn the eMMC (using your install-aml.sh) will this work also ? Because there is a emmc_autoscript which doesn't check for u-boot.ext existance. Do I need to create a custom emmc_autoscript ? Thanks in advance
  8. Hi, Balbes time ago I sent you a PR for GT1-Mini-A DTS which was accepted, today testing your 5.99 image noted that Kernel is 5.4rc6 and the image does not contain that DTB. Which branches contain which versions on your Github ? LE - > 5.2rc ? Master -> 5.4 ? I'm trying to make a PR for 5.4 and also trying to get support for the GT1 WiFi AP6255: https://gitlab.com/baylibre/amlogic/atv/linux/tree/narmstrong/v5.1/aml/integ-5.1-bcmdhd/drivers/net/wireless/bcmdhd TIA
  9. Ok. So where is the 5.2 branch your're using so I do the PR against that ?
  10. Oleg Why does the DTBs on your 5.2rc images are different than those produces by your master branch on https://github.com/150balbes/Amlogic_s905-kernel ? I was making the PR but my WiFi wasn't working with the DTB generated by the DTS code but worked by decompiled DTBs from your images, after making this changes on DTSI everything fine: juanjo@bionic:~/devel/Amlogic_s905-kernel$ git diff arch/arm64/boot/dts/amlogic/meson-g12a.dtsi diff --git a/arch/arm64/boot/dts/amlogic/meson-g12a.dtsi b/arch/arm64/boot/dts/amlogic/meson-g12a.dtsi index bd85b604c7..6565f5b709 100644 --- a/arch/arm64/boot/dts/amlogic/meson-g12a.dtsi +++ b/arch/arm64/boot/dts/amlogic/meson-g12a.dtsi @@ -200,12 +200,12 @@ apb_efuse: bus@30000 { compatible = "simple-bus"; - reg = <0x0 0x30000 0x0 0x1000>; + reg = <0x0 0x30000 0x0 0x2000>; #address-cells = <2>; #size-cells = <2>; - ranges = <0x0 0x0 0x0 0x30000 0x0 0x1000>; + ranges = <0x0 0x0 0x0 0x30000 0x0 0x2000>; - hwrng: rng { + hwrng: rng@218 { compatible = "amlogic,meson-rng"; reg = <0x0 0x218 0x0 0x4>; }; @@ -2376,7 +2376,7 @@ <&clkc CLKID_FCLK_DIV2>; clock-names = "core", "clkin0", "clkin1"; resets = <&reset RESET_SD_EMMC_A>; - amlogic,ddr-access-quirk; + amlogic,dram-access-quirk; }; Your DTBs do have "dram-access-quirk" but the DTS code on Github not. Upstream they are using "dram-access-quirk" but your master branch uses it on the DTSI but the source of the drivers does check for "ddr-access-quirk" which was older and buggier version Older: https://patchwork.kernel.org/cover/10940585/ Newer: https://patchwork.kernel.org/patch/10962981/ Or simply the 5.2 branch isn't master ??
  11. Seems to be the same case as mine GT1 Mini works and GT1 Mini-A didn't but did with SEI510 which in my case does have a different ETH PHY but I did manage to make a DTS. Now I need to find some time to do a PR for Balbes fork
  12. Ok. Got it working based on SEI510 Device Tree which booted then I changed some bits to enable GbE. By now have GbE and HDMI works; it seems WiFi too since a scan worked. I suppose I need to do a PR on your GitHub repo ?
  13. To start I did pull out the DTB from Android on the GT1 Mini and the GT1 Mini-A converted to DTS and compared them (diff -u). They are identically They are for 4.9 Kernel so they are too many differences to X96 DTS. I attached it anyway gt1_mini.dts
  • Create New...