On 09/26/2013 06:44 PM, Paul Zimmerman wrote: >> From: Suman Anna [mailto:s-anna@xxxxxx] >> Sent: Thursday, September 26, 2013 1:36 PM > >>>>>>>>>> I have an OMAP5432 uEVM which I cannot get to boot with recent mainline >>>>>>>>>> (tried 3.11 and 3.12-rc1). I have the TI GLSDK for this board (v6.0.0.7), >>>>>>>>>> which comes with 3.8.4 which works fine. >>>>>>>>>> >>>>>>>>>> I found this thread: http://marc.info/?l=fedora-arm&m=137717811815777 and >>>>>>>>> >>>>>>>>> Wrong link, should have been http://marc.info/?l=linux-omap&m=137515583214350. >>>>>>>>> >>>>>>>> Its because of commit 03ab349ec{ARM: OMAP5: hwmod data: Add mailbox data} which >>>>>>>> added hwmod data but DT data for mailbox is missing. Reverting that makes >>>>>>>> things work. Looks like mailbox dt patches missed the last merge window. >>>>>>> >>>>>>> I will be respinning the mailbox DT series very soon targeting 3.13, so >>>>>>> should not be an issue when OMAP5 boot is supported directly on mainline. >>>>>> >>>>>> That's good info, but unfortunately it didn't work for me. I must have a >>>>>> different problem, maybe I need a newer version of u-boot. >>>>>> >>>>>> So there is no README or wiki that explains how to get Linux to boot on >>>>>> this board? Is there perhaps a prebuilt SD card image somewhere with a >>>>>> recent kernel that I can grab? >>>>>> >>>>> You don't need anything special as such. Just pull the latest mainline >>>>> denx u-boot build it for 'omap5_uevm' and update your boot-loaders. >>>> >>>> FYI, I was able to boot using the branch posted by Santosh just fine. I >>>> used v2013.07 u-boot. The boot traces show up about 30~45 seconds after >>>> you see the trace "Starting kernel...". >>>> >>>> For using a rootfs from MMC, enable the following in menuconfig: >>>> Device Drivers -> Multifunction device drivers -> TI Palmas series chips >>>> Device Drivers -> Voltage and Current Regulator Support -> TI Palmas >>>> PMIC regulators. >>>> >>>> Depending on your u-boot settings, make sure the mmcroot is also set to >>>> /dev/mmcblk1p2. >>> >>> Hi Suman, >>> >>> Thanks for trying to help. Unfortunately, updating to v2013.07 u-boot, and >>> enabling the kernel config items you mentioned, didn't make a difference. >>> >>> Could you share your boot.scr? I'm still using the one from the TI GLSDK, >>> but maybe something has changed since the 3.8 kernel? Also, how do you >>> build your kernel? I'm doing "make LOADADDR=0x80008000 uImage" because >>> newer kernels won't build without the LOADADDR= bit. >> >> I build the kernel same way as you, and copy the zImage to boot >> partition. I am not using any boot.scr, a simple uEnv.txt to override >> the boot partition variables, and the default environment otherwise. >> >> My boot log is here, including my uEnv.txt at the end. >> http://hastebin.com/yinuxirexu.xml >> >> Hopefully, this should clear out any setup differences.. > > Thanks Suman! > > Using your info I was able to get the board to boot. Glad I could help.. > I added a third line to your uEnv.txt, so I don't have to break into > u-boot and type "setenv mmcroot /dev/mmcblk1p2 rw": > > bootdir= > bootpart=0:1 > mmcroot="/dev/mmcblk1p2 rw" > > Or I suppose you could add that to the kernel command line in the kernel > config. The uEnv.txt approach is good, since you can use the same image on other boards without resorting to other changes in environment for them. > > I'm seeing several stack dumps during the boot, with messages like > "division by 0 in the kernel", but I suppose that has nothing to do with > my boot problems. And it does boot to the login: prompt on the serial > console. Yeah, those I believe are related to the ABE and USB DPLL locking (or lack of) issues. regards Suman -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
