Date: Mon, 3 Oct 2016 02:54:56 -0700 From: Mark Millard <markmi@dsl-only.net> To: Da Rock <freebsd-arm@herveybayaustralia.com.au> Cc: Emmanuel Vadot <manu@bidouilliste.com>, Warner Losh <imp@bsdimp.com>, "freebsd-arm@freebsd.org" <freebsd-arm@freebsd.org> Subject: Re: u-boot efi option Message-ID: <FB252B4B-E1B3-4AF8-962E-0789BB350B3A@dsl-only.net> In-Reply-To: <c6134956-9b3e-421c-d59d-09b08dc7701b@herveybayaustralia.com.au> References: <36160c78-9095-f716-8041-3eb7656642af@herveybayaustralia.com.au> <20161002123700.662f2868688b4b32465395aa@bidouilliste.com> <CANCZdfoaV90H4qSNy64sLgFnnP7hePXbbT6jOX-R1GFiNksR%2Bg@mail.gmail.com> <20161003103247.78ff7596607755ce0f204648@bidouilliste.com> <c6134956-9b3e-421c-d59d-09b08dc7701b@herveybayaustralia.com.au>
index | next in thread | previous in thread | raw e-mail
On 2016-Oct-3, at 2:01 AM, Da Rock <freebsd-arm at herveybayaustralia.com.au> wrote: > A few points that I'd like clarification on: > > 1. I'm not having much success on rpi3 with any images, but I'm working on my own 64bit anyway. I'm also trying to get a handle on what the process of running the system on the rpi3 actually is as well. Fun... :) So the 512kb alignment, why is that an issue? Perhaps I'm a little foggy, but isn't that simply a case of partitioning correctly? > > 2. I should have mentioned this initially (which is why I was checking efi status to see if I was banging on for no reason), but I'm having trouble (still) compiling u-boot version that has the efi capability (2016.09) - can't find <sys/_types.h>. I've tried with clang (x18 error) and gcc to no avail. I tried leveraging ports, but it is not exactly cooperating either (I'll have to try some more diags on that). Anyone have a clue on how to get past this? -I doesn't work, or linking to /usr/local/include. Looks to me like sys/_types.h normally exists: # find /usr/include -name _types.h -print /usr/include/machine/_types.h /usr/include/sys/_types.h These two files are not the same by content and so are not equivalent so /usr/include/sys/_types.h is likely the right one. Using an armv6 stable/11 context the diff shows where the files came from in svn: < * $FreeBSD: stable/11/sys/arm/include/_types.h 301872 2016-06-13 16:48:27Z ian $ --- > * $FreeBSD: stable/11/sys/sys/_types.h 299571 2016-05-12 21:18:17Z cem $ Note that /usr/local is not involved in the earlier /usr/include/sys/_types.h path. [I've never tried to build u-boot 2016.09: the above is just a very generic note about where to find sys/_types.h .] Without more details being presented it is not clear why /usr/include is not being used as a place to look for sys/_types.h in your context. (This wording presumes self-hosted builds instead of cross builds for the path details.) That "not clear" status absent supporting details may well be true for folks familiar with building 2016.09 as well. > 3. My other query was on the EFI boot process on arm. As I understood, the u-boot is equivalent to boot0? So that means boot1 should then be in the fat partition? Is that what you mean by /efi/boot/bootarm.efi? Or doesn't it need to be? > > Apologies if I sound like a complete goose, but I'm still new to all things arm yet - but learning fast :) > > > On 03/10/2016 18:32, Emmanuel Vadot wrote: >> I've never had any problem with it (I know it doesn't really answer to >> your question). >> SuSe have switch to EFI (the main developer for U-Boot >> EFI is from SuSe), OpenBSD too. >> >> The main problem right now for people to test is that the partition >> on the arm images that we provide aren't aligned for it to work. >> >> I either need to fix the bug for non 512kb-aligned partition or >> aligned them in the release scripts. >> >> Also I don't know which ports-tree re@ is using for snapshots. I think >> that they use the latest quarterly for release and stable. >> So we have to be carefull when we will switch the ports to UEFI. >> >> In the meantime don't hesitate to test with my patch. >> >> If you have the correct dtb in /dtb/ on the msdos part U-Boot will load >> it. >> For booting automatically just put boot1.efi as /efi/boot/bootarm.efi >> >> On Sun, 2 Oct 2016 16:06:37 -0600 >> Warner Losh <imp@bsdimp.com> wrote: >> >>> How long do you think until this is mature enough we can cut over to it? >>> There's issues with ubldr on newer u-boot version on some of the platforms >>> we support. If we could cut over to this, that would be great. >>> >>> Warner >>> >>> On Sun, Oct 2, 2016 at 4:37 AM, Emmanuel Vadot <manu@bidouilliste.com> wrote: >>>> Hello, >>>> >>>> I've commited every needed change on our side, for uboot side you will >>>> need this patch : >>>> https://people.freebsd.org/~manu/u-boot_201609_efi.diff >>>> >>>> The only drawback is that you will need to have your partition aligned >>>> on 512kb boundaries on the mmc. >>>> >>>> I've successfully booted my beaglebone black and most of my Allwinner >>>> boards with UEFI. >>>> >>>> Some part of the uboot patch have started to be upstreamed, for >>>> the others I need to rework on some part before I upstream them. >>>> >>>> Cheers, >>>> >>>> On Sun, 2 Oct 2016 19:56:54 +1000 >>>> Da Rock <freebsd-arm@herveybayaustralia.com.au> wrote: >>>> >>>>> Has anyone tried this option yet? I've heard someone got it going for >>>>> slack, but I thought it sounded like it might make it easier for running >>>>> freebsd. >>>>> >>>>> I'm currently trying to build a more current u-boot to test it. . . . === Mark Millard markmi at dsl-only.net _______________________________________________ freebsd-arm@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-arm To unsubscribe, send any mail to "freebsd-arm-unsubscribe@freebsd.org"help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?FB252B4B-E1B3-4AF8-962E-0789BB350B3A>