ssvb has quit [Ping timeout: 245 seconds]
zombah has joined #linux-exynos
ssvb has joined #linux-exynos
afaerber_ is now known as afaerber
dne has quit [*.net *.split]
dne has joined #linux-exynos
mkatic has joined #linux-exynos
<mkatic> hello
<mkatic> so i've seen the chatter here on the past few days
<mkatic> seems like snow is should run quite nicely with linux-next and default exynos_defconfig, correct?
<mkatic> i can't get it to work
<mkatic> i run nv-uboot straight from spi flash
<mkatic> and i can't get it to load the dtb file properly
<mkatic> and this isn't specific to 3.19, i get this with older kernels
<mkatic> i get this:
<mkatic> 3564992 bytes read
<mkatic> ## Booting kernel from Legacy Image at 42000000 ...
<mkatic> Image Name: Linux-3.19.0-next-20150222
<mkatic> Image Type: ARM Linux Kernel Image (uncompressed)
<mkatic> Data Size: 3564928 Bytes = 3.4 MiB
<mkatic> Load Address: 42000000
<mkatic> Entry Point: 42000000
<mkatic> Verifying Checksum ... OK
<mkatic> ## Flattened Device Tree blob at 43000000
<mkatic> Booting using the fdt blob at 0x43000000
<mkatic> XIP Kernel Image ... OK
<mkatic> OK
<mkatic> Loading Device Tree to 5ffee000, end 5ffff47b ... OK
<mkatic> boot_kernel.c: ft_board_setup: warning: fdt_system_setup() fails
<mkatic> Failed to add board information to FDT: FDT_ERR_NOTFOUND
<mkatic> and then uboot just resets
<mkatic> when i try this exact approach with 3.8.11 chromeos kernel, it works
<mkatic> if i try to use the dtb file from 3.8.11 with linux-next it also doesnt work, of course
<mkatic> but at least the dtb file is properly loaded, i don't get the FDT_ERR_NOTFOUND error
<mkatic> Wizzup: You mentioned a few days ago that your snow would randomly turn on
<mkatic> mine does that too
<mkatic> and i suspect that it is a known issue
<mkatic> try to boot something with vboot_two_stop from the uboot cmdline
<mkatic> but keep the lid closed
<mkatic> vboot_two_stop will stop and immediately shut down if it detects a closed lid
<Wizzup> mkatic: did you append the dtb?
<Wizzup> That's an easy way to get this right
afaerber_ has joined #linux-exynos
<mkatic> for chromeos kernel i used Olof's instructions for building a custom kernel
<mkatic> so when i get my zImage, i do the mkimage -f kernel.its kernel.itb thing
<mkatic> and this works with chromeos kernels
<mkatic> so you say i should cat zImage dtb > out
<afaerber_> mkatic, how did you create the uImage and what commands do you use to load things?
<afaerber_> the XIP looks unexpected to me
<afaerber_> I assume it can't modify things in-place for lack of space
afaerber has quit [Ping timeout: 246 seconds]
<mkatic> for uImage i just do:
<mkatic> make ARCH=arm CROSS_COMPILE=../gcc-linaro-arm-linux-gnueabihf-4.9-2014.09_linux/bin/arm-linux-gnueabihf- LOADADDR=0x42000000 uImage dtbs modules -j4
<mkatic> so i just define the LOADADDR, nothing else
<mkatic> then i took the dtb file and the uimage and copied that to my internal mmc boot partition
<Wizzup> of course- for snow you want a different DTB
<mkatic> hm, i think i already tried the approach from peach instructions
<mkatic> i can't remember exactly
<mkatic> well, i'll try again
<Wizzup> do you have nv-uboot
<mkatic> i replaced the default RW firmware with nv -uboot
<mkatic> i can't remember which of the two images i used, the one with simplefb or the other one
<mkatic> should i use the same LOADADDR as defined in the peach guide?
<Wizzup> that should be ok, yes
<mkatic> alright, the peach pit guide helped, i got it to boot
<Wizzup> :)
<mkatic> but the kernel crashes severely
<mkatic> i get no display out
<mkatic> serial output is ok
<Wizzup> display out as in hdmi?
<Wizzup> or internal display
<mkatic> internal
<Wizzup> is this the latest linux-next?
<mkatic> yes
<afaerber_> mkatic, you may need to patches for adding hdmi clock and power-domain
<afaerber_> *two
<Wizzup> afaerber_: it worked a week ago for me though, on snow
<Wizzup> mkatic: just checking, you used the right dtb right?
<mkatic> i'll post the two crashes i get, what's the preferred paste site?
<afaerber_> also I picked up one that avoids a NULL dereference
<afaerber_> Wizzup, a week is ages in linux-next ;)
<Wizzup> heh...
<Wizzup> I guess so
<mkatic> the spinlock lockup thing looks weird
<afaerber_> disable exynos iommu.
<mkatic> i thought i read that it works in linux-next?
<mkatic> alright, i'll try disabling it
<afaerber_> there's a huge series still missing - disabling it fixed things for me, same with exynos thermal
<afaerber_> (but as you seem to have serial output, try one by one)
<mkatic> if i disable iommu, i loose drm aswell, right?
<mkatic> s/loose/lose/g
<afaerber_> no
<Wizzup> I don't think so
<afaerber_> you just need sufficient cma
<mkatic> alright
<mkatic> what's cma?
afaerber_ is now known as afaerber
<afaerber> continuous memory
<afaerber> CONFIG_CMA_SIZE_MBYTES=64
<afaerber> CONFIG_CMA_SIZE_SEL_MBYTES=y
<afaerber> for instance
<mkatic> ah, ok
<afaerber> alternatively, the series I referred to is "[PATCH v5 00/18] Exynos SYSMMU (IOMMU) integration with DT and DMA-mapping subsystem"
<mkatic> thank you, but this isn't really a priority for me right now
<mkatic> i usually run 3.8.11
<mkatic> chromes
<mkatic> sorry, 3.8.11 from the chromeos repo
<mkatic> work fine for me, but one huge problem
<mkatic> it tends to just randomly crash
<mkatic> so i'm hoping that newer kernels might be more stable
<mkatic> success!
<mkatic> disabling iommu got me right into Xorg with linux-next
<Wizzup> :)
<Wizzup> Swabbles: I think we should try to enable anti-spam techniques like sunxi did and then enable reg again
<mkatic> which is surprising since i didn't copy the modules
<Wizzup> I think now it's getting annoying that I have to offer everyone to make a wiki account
<mkatic> suspend is broken
<mkatic> not a big deal for me, just saying
<mkatic> i get this after wakeup:
<mkatic> [ 7.843229] exynos-dp 145b0000.dp-controller: EDID Read success!
<mkatic> [ 7.909122] exynos-dp 145b0000.dp-controller: EDID Read success!
<mkatic> [ 48.347935] exynos-dp 145b0000.dp-controller: EDID Read success!
<mkatic> [ 48.356075] exynos-dp 145b0000.dp-controller: unable to config video
<mkatic> [ 48.356072] exynos-dp 145b0000.dp-controller: Timeout of video streamclk ok
<Wizzup> suspend works for you?
<Wizzup> I didn't get it to work at all
<Wizzup> ah, right
<mkatic> i don't really care about suspend
<afaerber> I'm seeing the same thing on regular bootup on Spring :P
<mkatic> i mean, my snow will boot up ubuntu 14.04 from internal mmc in less than 20 seconds
* afaerber started looking at exynos_dp_core.c but no insights yet
<Wizzup> mkatic: :)
<mkatic> ok, i'll copy the modules now, see if can get sound and wifi going
<Wizzup> wifi may not work, firmware loading broke somewhere, but maybe it's fixed again
<Wizzup> if it's a show stopper, reverting some commit (can look for it in the logs) will restore func.
<Wizzup> audio is still in the works, but starting to work, but not yet in linux-next, iirc
<mkatic> i need wifi now, i can wait for working audio
<afaerber> 68ff2a00dbf590dc9483c20f85fd1f1faa054604.
<mkatic> someone here said that if i compile it as i module, it should work?
<afaerber> but for Snow there was a second issue with mmc timeouts or so
<mkatic> alright then
<afaerber> module> I reported that for Spring, don't have a Snow
<mkatic> ok, i'll try reverting the above mentioned commit
mkatic has quit [Read error: Connection reset by peer]
mkatic has joined #linux-exynos
mkatic has quit [Quit: Leaving]
liquidAcid has joined #linux-exynos
<afaerber> applying the iommu series still breaks boot on Spring
Tenkawa has joined #linux-exynos
Tenkawa has joined #linux-exynos
<javier__> afaerber: the iommu series still breaks on chromebooks due the bootloader leaving the display enabled http://lists.infradead.org/pipermail/linux-arm-kernel/2015-January/319305.html
<javier__> afaerber: marek said that will work on that as a follow-up, that's why I sent a patch to disable exynos iommu on exynos_defconfig
<javier__> afaerber: the mmc timeout problem does not happen on snow but on peach
<afaerber> javier__, yeah I remember reading that. I just assumed that working iommu drivers would help :(
<javier__> afaerber: not yet unfortunately
<afaerber> javier__, regarding my black screen, I rebooted into ChromeOS today and verified that the developer mode etc. is still on - just as with my old branch the stock kernel came up with graphics okay
<afaerber> javier__, do you know how updates work under ChromeOS? I wonder whether there's any firmware/ec updates that I'm missing
<javier__> afaerber: no, I don't know how chromeos updates works
<javier__> regarding your black screen, you can do a git bisect + adding the drm bridge patches on each step
<javier__> it will be tedious but you can know what broke your display
liquidAcid has quit [Quit: Leaving]
<afaerber> javier__, seeing "[PATCH] exynos/drm: fix no hdmi output" I am compiling next-20150222 with less exynos drm options enabled
Tenkawa has quit [Quit: leaving]
<javier__> afaerber: ok
ard has quit [Ping timeout: 265 seconds]
steev has quit [Read error: Network is unreachable]
sjoerd has quit [Ping timeout: 265 seconds]
dne has quit [Ping timeout: 265 seconds]
James_T has quit [Ping timeout: 265 seconds]
sjoerd has joined #linux-exynos
James_T has joined #linux-exynos
dne has joined #linux-exynos
steev has joined #linux-exynos
zombah has quit [Quit: Leaving]
Tenkawa has joined #linux-exynos
Tenkawa has joined #linux-exynos
mkatic has joined #linux-exynos
<mkatic> ...had no luck getting wifi to work on snow running linux-next
<mkatic> i get:
<mkatic> [ 2.706687] mmc_host mmc2: Bus speed (slot 0) = 50000000Hz (slot req 50000000Hz, actual 50000000HZ div = 0)
<mkatic> [ 2.718802] mmc2: new high speed SDIO card at address 0001
<mkatic> [ 2.723230] mwifiex: rx work enabled, cpus 2
<mkatic> [ 2.727235] mwifiex_sdio mmc2:0001:1: Direct firmware load for mrvl/sd8797_uapsta.bin failed with error -2
<mkatic> [ 2.736706] mwifiex_sdio mmc2:0001:1: Failed to get firmware mrvl/sd8797_uapsta.bin
<mkatic> reverting 68ff2a00dbf590dc9483c20f85fd1f1faa054604 had no effect
ard has joined #linux-exynos
<Wizzup> you have to enable the userspace fallback fw loading as well
<Tenkawa> yeah... works great here mkatic
<Tenkawa> 3.19.0-next
<Tenkawa> [ 1.156880] mwifiex: rx work enabled, cpus 2
<Tenkawa> [ 1.774472] mwifiex_sdio mmc2:0001:1: info: FW download over, size 458108 bytes
<Tenkawa> [ 2.106207] mwifiex_sdio mmc2:0001:1: WLAN FW is active
<Tenkawa> [ 2.475925] mwifiex_sdio mmc2:0001:1: driver_version = mwifiex 1.0 (14.66.11.p151)
<mkatic> thank you both, it's working now
<Tenkawa> cool
<Tenkawa> I compiled mine into the kernel
<mkatic> i'll probably do the same once i start trimming the kernel
<Wizzup> :)
<Tenkawa> heheh
<mkatic> hm
<mkatic> can't shut down with this kernel
<Tenkawa> yeah same problem here
<Tenkawa> havent analyzed it yet
<Tenkawa> I've just held power button till it powers off for now
<mkatic> mine goes all the way to
<mkatic> [ 19.620013] reboot: System halted
<Tenkawa> yep
<Tenkawa> same here
<mkatic> display turns off but the power led remains on
<mkatic> alright
<Tenkawa> probable just one address needs to be changed
<Tenkawa> been working on getting my rpi2's online so I havent looked at it at all personally yet
<mkatic> well, it's not a deal breaker, at least not for me
<Tenkawa> here either
<Tenkawa> its been worth it.. 3.8 was way too old for me
<mkatic> dp you get random hangs with 3.8?
<Tenkawa> no
<mkatic> hm
<mkatic> i'm starting to think mine has a hardware issue
<mkatic> even when i was on chrome OS i got random hangs
<Tenkawa> ahh
<mkatic> i _did_ buy it refurbished from an amazon seller
<mkatic> hm
<Wizzup> mkatic: same @ led
<mkatic> can't start chromium, i get a segfault
<Wizzup> there are some errata's, but I am not sure if those are related.
<Wizzup> I only ran into that when recompiling gcc
<Wizzup> errata's missing*
<Wizzup> that you need to enable in mainline u-boot (and also get the mainline u-boot as nv-uboot)
<Wizzup> and I have not tried that yet.
<Wizzup> (no time unfortunately)
<mkatic> i only caught these kernel panics a few times
<mkatic> i mean i caught some dmesg output
<mkatic> hmm, i can't find them now
<Tenkawa> cheers all.. bbl
Tenkawa has quit [Quit: leaving]
<mkatic> bye
mkatic has quit [Quit: Leaving]