| 05:33 | jnettlet[m]> | Ke: yes that should work. It is only the in between speeds like 2.5 and 5ghz that probably need work. |
| 06:25 | Ke> | well at least first attempt failed on linux-4.15 |
| 06:34 | jnettlet[m]> | Ke: 4.15 seems to have introduced some networking bugs. I still need to do the bisect to see what is going on. |
| 06:35 | jnettlet[m]> | currently I am trying to fix a bunch of CF u-boot issues. |
| 06:36 | jnettlet[m]> | I think maybe tomorrow I will have time to do some more mainline work on the 8040 |
| 07:24 | Ke> | thanks |
| 09:57 | Ke> | trying natively compiled EFI next |
| 09:59 | Ke> | (20:51:24) < suihkulok> pft, people just merged a bunch of fixes the compile it with gcc7 |
| 09:59 | Ke> | [GCC7] not defined. No toolchain available for build! |
| 10:01 | Ke> | hmm though I have not defined the prefix |
| 10:05 | Ke> | + build -a AARCH64 -t GCC7 -b RELEASE -p OpenPlatformPkg/Platforms/Marvell/Armada/Armada80x0McBin.dsc |
| 10:05 | Ke> | grepping GCC5 returns loads of matches, while GCC7 or GCC6 return none |
| 10:12 | Ke> | let's try spoofed build |
| 11:43 | Ke> | ok apparently there just is a very recent regression for mcbin, also debian experimental kernel fails right out |
| 11:44 | Ke> | I guess I could bisect |
| 11:44 | Ke> | also my native hack compile fails completely |
| 11:45 | Ke> | so keep building with x86 |
| 11:57 | Ke> | let's try 4.15.7 first (with the ata patch) |
| 13:38 | Ke> | 4.15.7 good with the SCSI patch |
| 14:56 | jnettlet[m]> | there is a 10G regression as well. |
| 14:57 | jnettlet[m]> | vpeter: I would recommend you upgrade to the latest u-boot I posted on your CF. It includes a bunch of fixes that effect the hardware for linux. |
| 14:59 | vpeter> | jnettlet[m]: on my todo list. Atm upgrading my kodi setup to see if something is broken for my special setup. |
| 15:00 | jnettlet[m]> | vpeter: with the new 2018.01 for imx you also get the option for additional boot medium like USB good for testing new images without losing your working setup. |
| 15:01 | jnettlet[m]> | fyi at u-boot break the boot sequence then run `usb start` if usb has been initialized then `env run usb_boot` |
| 15:02 | jnettlet[m]> | I am using that a lot now that it is functional. |
| 15:04 | vpeter> | Currently I'm only interested in u-boot for clearfog you mentioned few days back. All imx6 related stuff was taken over by lucas and he doesn't want help from me. Even if he is totally wrong with some of his thinking. |
| 15:09 | vpeter> | btw: for libreelec on imx6 I'm using nfs boot. Which is handy to change things because I don't have serial console near tv. |
| 15:40 | Ke> | 4.15.9 also boots, anyone see anything in linux-4.15.10 that would explain early boot failure |
| 15:40 | Ke> | board is mcbin |
| 15:45 | Ke> | tmp stuff, objtool stuff, nospec stuff... |
| 15:45 | Ke> | there is no commit "break boot on mcbin" =o( |
| 15:48 | Ke> | objtool I guess can't cause runtime problems |
| 15:49 | jnettlet[m]> | Ke: enable the debugging and earlyprintk and you should get console output that will give some hints |
| 15:50 | Ke> | I could try recompiling 4.15.10 just to see that the previous toolchain was not bad |
| 15:51 | Ke> | debian testing seems to have a new toolchain every day |
| 15:54 | Ke> | https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git/commit/?h=v4.15.10&id=88bf8f7eb708cd3317a52c6edb50d251c15773b7 could it be this one |
| 15:55 | Ke> | I guess pci stuff will be started much later |
| 16:04 | Ke> | I think I'll stress-test the SCSI now, as I have other stuff to do |
| 16:59 | Ke> | total bytes scrubbed: 138.06GiB with 1 errors |
| 16:59 | Ke> | now has gotten a lot further with the patch |