Re: [coreboot] Building 4.5 from release sources
On Thu, Jan 5, 2017 at 3:28 PM, Trammell Hudsonwrote: > On Thu, Jan 05, 2017 at 06:34:42AM -0700, Trammell Hudson wrote: >> When I build coreboot 4.5 from the release sources it is necessary >> to download the coreboot-blobs-4.5.tar.xz file and it looks like there >> might be a dependency now on the 3rdparty/vboot tree as well since >> cbfs.h includes vb2_api.h [...] > > Ah, I see what happened. My script fetched the tar file from the > github release tag: > > https://github.com/coreboot/coreboot/archive/4.5.tar.gz > > Instead of fetching from the official release server: > > https://www.coreboot.org/releases/coreboot-4.5.tar.xz > > That's why the submodules are not included in the source tar. > > What's the distinction between coreboot/src/vboot and the > coreboot/3rdparty/vboot directories? As Peter pointed out, > having the necessary headers self contained in the coreboot > tree would be nice to have. > src/vboot calls into 3rdparty/vboot. src/vboot is the runtime binding for the vboot implementation. That's orthogonal to the other utility functions vboot supplies that the coreboot uses (hashing constructs). I hope that helps. > -- > Trammell > > > -- > coreboot mailing list: coreboot@coreboot.org > https://www.coreboot.org/mailman/listinfo/coreboot -- coreboot mailing list: coreboot@coreboot.org https://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] "FCH SPI: Too much to write" on KGPE-D16
Hello Timothy, On 2017年01月06日 01:14, Timothy Pearson wrote: > On 01/05/2017 04:33 AM, Iru Cai wrote: > > Hello, > > > I've set up a computer with Asus KGPE-D16 recently. I use an Winbond > > W25Q64.V 8M flash chip and tested with this config with only ROM size > > changed: > > > > https://review.coreboot.org/gitweb/cgit/board-status.git/tree/asus/kgpe-d16/4.5-759-gab8f7d3/2017-01-04T21_18_55Z/config.short.txt > > > And it can boot to my OS on disk. However, after I enabled the serial > > console, I can only see: > > "FCH SPI: Too much to write. Does your SPI chip driver use > > spi_crop_chunk()?" > > Do you see any other messages at all from a cold start (i.e. AC power > removed from the system, then reapplied after 15 seconds while the > serial console is being logged)? > I have other console logs, but not about this SPI issue, I came into other problems when I try other configurations. I don't know if it's related to serial console. The config on board status and the ROMs from libreboot all boots, but they all have no serial console output. > I have not personally tested the 8M chips; it could be that there is an > issue with coreboot's generic driver for those devices. I know it's not > a general issue with large Flash chips since the 16M chips work fine. > Thanks for the info. > Thanks! > Iru -- coreboot mailing list: coreboot@coreboot.org https://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Building 4.5 from release sources
On Thu, Jan 05, 2017 at 06:34:42AM -0700, Trammell Hudson wrote: > When I build coreboot 4.5 from the release sources it is necessary > to download the coreboot-blobs-4.5.tar.xz file and it looks like there > might be a dependency now on the 3rdparty/vboot tree as well since > cbfs.h includes vb2_api.h [...] Ah, I see what happened. My script fetched the tar file from the github release tag: https://github.com/coreboot/coreboot/archive/4.5.tar.gz Instead of fetching from the official release server: https://www.coreboot.org/releases/coreboot-4.5.tar.xz That's why the submodules are not included in the source tar. What's the distinction between coreboot/src/vboot and the coreboot/3rdparty/vboot directories? As Peter pointed out, having the necessary headers self contained in the coreboot tree would be nice to have. -- Trammell -- coreboot mailing list: coreboot@coreboot.org https://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Building 4.5 from release sources
On Thu, Jan 5, 2017 at 3:26 PM, Peter Stugewrote: > Aaron Durbin via coreboot wrote: >> This was discussed around a month ago on #coreboot, iirc. > > That's not really good enough, IMHO. I was providing background on a recent discussion related to the topic at hand. > > So far, coreboot was always self-contained. It even goes so far as to > build its own toolchain. > > Creating mandatory dependencies which may actually not always be > neccessary is simply poor software engineering, and not something > I think that we should be doing. Let's try to fix it! We actually control the submodules. They live on coreboot.org. It's the release process that doesn't include the dependencies (at least that's my understanding). There are a hoard a mandatory dependencies depending on what board you are trying to build that likely aren't packaged from all the submodules. Where is the line drawn for board not working/building because of an unmet dependency not included in the release tarball? > Is there any technical reason *not* to decouple coreboot source from > vboot source completely? > No. It's just that we use functionality of vboot library currently with no replacement to plug that functionality. One could sprinkle all kinds of #ifdefs around, but that's just papering over a deficiency in the current release process. As I noted there's supposed to be work done to provide the crypto/hash interface that would decouple things. I assumed people were working on it. Maybe they aren't? > > //Peter > > -- > coreboot mailing list: coreboot@coreboot.org > https://www.coreboot.org/mailman/listinfo/coreboot -- coreboot mailing list: coreboot@coreboot.org https://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Building 4.5 from release sources
Aaron Durbin via coreboot wrote: > This was discussed around a month ago on #coreboot, iirc. That's not really good enough, IMHO. So far, coreboot was always self-contained. It even goes so far as to build its own toolchain. Creating mandatory dependencies which may actually not always be neccessary is simply poor software engineering, and not something I think that we should be doing. Let's try to fix it! Is there any technical reason *not* to decouple coreboot source from vboot source completely? //Peter -- coreboot mailing list: coreboot@coreboot.org https://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Building 4.5 from release sources
On Thu, Jan 5, 2017 at 1:45 PM, Peter Stugewrote: > Trammell Hudson wrote: >> When I build coreboot 4.5 from the release sources it is necessary >> to download the coreboot-blobs-4.5.tar.xz file and it looks like there >> might be a dependency now on the 3rdparty/vboot tree as well since >> cbfs.h includes vb2_api.h: >> >> https://github.com/coreboot/coreboot/blob/master/src/commonlib/include/commonlib/cbfs.h#L21 > > This seems like a bug to me. I think the vboot header should either > be conditionally included according to configuration, or if it's an > ABI then I would suggest to keep a local copy. I assume the ABI is > versioned, so back+forward compatibility should be no problem. Whether or not that include is removed from that file doesn't matter. cbfstool uses vb2 for the hashing routines internally -- as well as coreboot proper. This was discussed around a month ago on #coreboot, iirc. I don't recall with whom aside from Damien, but people were going to look into providing a crypto interface that can be implemented for hashing, etc. This CL spawned that discussion: https://review.coreboot.org/#/c/17469/ However, I haven't seen that work at all yet. I tried removing that API a while back, but I can't find that CL. I'm not sure why. I think Zaolin wanted that API. Here's a CL to remove it that will fail to build because of what I described: https://review.coreboot.org/18045 > > > //Peter > > -- > coreboot mailing list: coreboot@coreboot.org > https://www.coreboot.org/mailman/listinfo/coreboot -- coreboot mailing list: coreboot@coreboot.org https://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Building 4.5 from release sources
Trammell Hudson wrote: > When I build coreboot 4.5 from the release sources it is necessary > to download the coreboot-blobs-4.5.tar.xz file and it looks like there > might be a dependency now on the 3rdparty/vboot tree as well since > cbfs.h includes vb2_api.h: > > https://github.com/coreboot/coreboot/blob/master/src/commonlib/include/commonlib/cbfs.h#L21 This seems like a bug to me. I think the vboot header should either be conditionally included according to configuration, or if it's an ABI then I would suggest to keep a local copy. I assume the ABI is versioned, so back+forward compatibility should be no problem. //Peter -- coreboot mailing list: coreboot@coreboot.org https://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] "FCH SPI: Too much to write" on KGPE-D16
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 01/05/2017 04:33 AM, Iru Cai wrote: > Hello, > > I've set up a computer with Asus KGPE-D16 recently. I use an Winbond > W25Q64.V 8M flash chip and tested with this config with only ROM size > changed: > > https://review.coreboot.org/gitweb/cgit/board-status.git/tree/asus/kgpe-d16/4.5-759-gab8f7d3/2017-01-04T21_18_55Z/config.short.txt > > And it can boot to my OS on disk. However, after I enabled the serial > console, I can only see: > "FCH SPI: Too much to write. Does your SPI chip driver use > spi_crop_chunk()?" Do you see any other messages at all from a cold start (i.e. AC power removed from the system, then reapplied after 15 seconds while the serial console is being logged)? I have not personally tested the 8M chips; it could be that there is an issue with coreboot's generic driver for those devices. I know it's not a general issue with large Flash chips since the 16M chips work fine. Thanks! - -- Timothy Pearson Raptor Engineering +1 (415) 727-8645 (direct line) +1 (512) 690-0200 (switchboard) https://www.raptorengineering.com -BEGIN PGP SIGNATURE- Version: GnuPG v1 Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQEcBAEBAgAGBQJYbn7XAAoJEK+E3vEXDOFbIasIAIb1myve1e7FuuGcsOsPztM7 OUiIl2ru/jcowbIpAjpTmvWeuGg+FMyf/qmVAZC0dDDoj0prGxmax/SeiLp3N2E7 m/Dms9MKa6p90XuZe+KBATwQi7uBoMrrYQzJ5ZDPlaPsaiPsD7jPQA7sl+522b+m Y58pD1b4ZjLFydjHp2VPBqpcrLaVzv8Trubka2jqfDFsYmrwQeWswFG+LS4BdwgM dcop2VpHybY6tX/PEoGyOYYprOiNbmH4kY//HSWfK7aRLUazShW0+svFHGYjzcuK MdfBIJoAnuNc9Dk4uauJC2uFVHzYkXWeG23YyjZg2guGrhTmOLwEDetXeV1V3yg= =rd9i -END PGP SIGNATURE- -- coreboot mailing list: coreboot@coreboot.org https://www.coreboot.org/mailman/listinfo/coreboot
[coreboot] Building 4.5 from release sources
When I build coreboot 4.5 from the release sources it is necessary to download the coreboot-blobs-4.5.tar.xz file and it looks like there might be a dependency now on the 3rdparty/vboot tree as well since cbfs.h includes vb2_api.h: https://github.com/coreboot/coreboot/blob/master/src/commonlib/include/commonlib/cbfs.h#L21 When building from git it checks out the correct submodule, so this is only a problem for me with releases. Is there an official release tar for vboot sources? I don't see one in https://www.coreboot.org/releases/. -- Trammel -- coreboot mailing list: coreboot@coreboot.org https://www.coreboot.org/mailman/listinfo/coreboot
[coreboot] "FCH SPI: Too much to write" on KGPE-D16
Hello, I've set up a computer with Asus KGPE-D16 recently. I use an Winbond W25Q64.V 8M flash chip and tested with this config with only ROM size changed: https://review.coreboot.org/gitweb/cgit/board-status.git/tree/asus/kgpe-d16/4.5-759-gab8f7d3/2017-01-04T21_18_55Z/config.short.txt And it can boot to my OS on disk. However, after I enabled the serial console, I can only see: "FCH SPI: Too much to write. Does your SPI chip driver use spi_crop_chunk()?" And the machine cannot boot. Has anyone seen this issue before? Thanks, Iru -- Please do not send me Microsoft Office/Apple iWork documents. Send OpenDocument instead! http://fsf.org/campaigns/opendocument/ -- coreboot mailing list: coreboot@coreboot.org https://www.coreboot.org/mailman/listinfo/coreboot