Re: [coreboot] From wildcat linking to exploring new languages
On 29.08.2015 14:09, Peter Stuge wrote: ron minnich wrote: and I still think it belongs in the tree. No way. This is a library of device drivers. It has no place whatsoever as a subdirectory lost somewhere in the already too big coreboot repository. libsparkhw needs to live in its own repository, and be pulled into coreboot as well as other consumers (e.g. FILO as Nico mentioned) as a submodule. But putting the submodule under 3rdparty isn't a great fit, because the library is very much from the 1stparty, namely from the coreboot community. So it should be somewhere else, I suggest in lib/[lib]sparkhw. That's exactly what I had in mind. Thanks for your support. In the end, I don't care as much about a dedicated repo than about the path (i.e. not under 3rdparty). Nico -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
On 29.08.2015 21:58, ron minnich wrote: If people feel strongly enough about this then we can do an external repo for now. Either way around, we would have to learn how to best integrate SPARK code in coreboot. There would still be some steps to go from linking with an adalib, to SPARK beeing a first class citizen of our build system. Nico, do you want to set up a github repo and we can work out procedures people use to try this out? As you know I'm more interested in Muen as a replacement for the ramstage but this is a good first step. I guess, I'll just push things to gerrit, when the code is ready. Nico -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
Ron, ron minnich wrote: I think this could prove to be a very important new direction for coreboot, I agree. and I still think it belongs in the tree. No way. This is a library of device drivers. It has no place whatsoever as a subdirectory lost somewhere in the already too big coreboot repository. libsparkhw needs to live in its own repository, and be pulled into coreboot as well as other consumers (e.g. FILO as Nico mentioned) as a submodule. But putting the submodule under 3rdparty isn't a great fit, because the library is very much from the 1stparty, namely from the coreboot community. So it should be somewhere else, I suggest in lib/[lib]sparkhw. License-wise, in order to be maximally compatible and reusable together with libpayload I would welcome and support a simplified BSD license. I haven't ever used Ada, but used lots of Pascal long long ago. Fun! //Peter -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
If people feel strongly enough about this then we can do an external repo for now. Nico, do you want to set up a github repo and we can work out procedures people use to try this out? As you know I'm more interested in Muen as a replacement for the ramstage but this is a good first step. ron -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
ron minnich wrote: If people feel strongly enough about this then we can do an external repo for now. Nico, do you want to set up a github repo There are more options than the coreboot repository and github. If Nico is willing to push the code to coreboot.org then I think a separate repository on coreboot infrastructure is the perfect place. //Peter -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
Am 21.08.2015 22:47, schrieb ron minnich: This sounds wonderful. I'm all for it. Did you just kill any discussion with that line? :) Things live currently in a small library that I called `libsparkhw` (came spontaneously, I don't care much about the name). It contains some framework (Timer, I/O, Debugging) to do hardware stuff in SPARK, plus the mentioned graphics initialization. I falter to push it directly for inclusion into the coreboot repo. Maybe we want a separate reposi- tory for it (like libpayload has)? It would also make it easier to use it in other components (e.g. bootloaders and such, and of course our own products where the code stems from). We have discussed the licensing and will push this under GPLv2 + later if there are no reasonable concerns. Nico -- M. Sc. Nico Huber Senior Berater SINA-Softwareentwicklung Netzwerk- Client-Sicherheit / Network Client Security Division Öffentliche Auftraggeber / Public Authorities secunet Security Networks AG Tel.: +49-201-5454-3635, Fax: +49-201-5454-1325 E-Mail: nico.hu...@secunet.com Mergenthalerallee 77, 65760 Eschborn, Deutschland www.secunet.com __ Sitz: Kronprinzenstraße 30, 45128 Essen, Deutschland Amtsgericht Essen HRB 13615 Vorstand: Dr. Rainer Baumgart (Vors.), Thomas Pleines Aufsichtsratsvorsitzender: Dr. Peter Zattler __ -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
Am 21.08.2015 22:47, schrieb ron minnich: This sounds wonderful. I'm all for it. Did you just kill any discussion with that line? :) Things live currently in a small library that I called `libsparkhw` (came spontaneously, I don't care much about the name). It contains some framework (Timer, I/O, Debugging) to do hardware stuff in SPARK, plus the mentioned graphics initialization. I falter to push it directly for inclusion into the coreboot repo. Maybe we want a separate reposi- tory for it (like libpayload has)? It would also make it easier to use it in other components (e.g. bootloaders and such, and of course our own products where the code stems from). We have discussed the licensing and will push this under GPLv2 + later if there are no reasonable concerns. Nico -- M. Sc. Nico Huber Senior Berater SINA-Softwareentwicklung Netzwerk- Client-Sicherheit / Network Client Security Division Öffentliche Auftraggeber / Public Authorities secunet Security Networks AG Tel.: +49-201-5454-3635, Fax: +49-201-5454-1325 E-Mail: nico.hu...@secunet.com Mergenthalerallee 77, 65760 Eschborn, Deutschland www.secunet.com __ Sitz: Kronprinzenstraße 30, 45128 Essen, Deutschland Amtsgericht Essen HRB 13615 Vorstand: Dr. Rainer Baumgart (Vors.), Thomas Pleines Aufsichtsratsvorsitzender: Dr. Peter Zattler __ -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
On Fri, Aug 28, 2015 at 9:12 AM, ron minnich rminn...@gmail.com wrote: I would say bring it in to the repo. We have assembly, we have our own C compiler even, so a higher level language hardly seems a bad idea. If this can make our code base better in some way, why not use it? Does this mean we can start writing our userland tools in golang? ron On Fri, Aug 28, 2015 at 4:55 AM Nico Huber nico.hu...@secunet.com wrote: Am 21.08.2015 22:47, schrieb ron minnich: This sounds wonderful. I'm all for it. Did you just kill any discussion with that line? :) Things live currently in a small library that I called `libsparkhw` (came spontaneously, I don't care much about the name). It contains some framework (Timer, I/O, Debugging) to do hardware stuff in SPARK, plus the mentioned graphics initialization. I falter to push it directly for inclusion into the coreboot repo. Maybe we want a separate reposi- tory for it (like libpayload has)? It would also make it easier to use it in other components (e.g. bootloaders and such, and of course our own products where the code stems from). We have discussed the licensing and will push this under GPLv2 + later if there are no reasonable concerns. Nico -- M. Sc. Nico Huber Senior Berater SINA-Softwareentwicklung Netzwerk- Client-Sicherheit / Network Client Security Division Öffentliche Auftraggeber / Public Authorities secunet Security Networks AG Tel.: +49-201-5454-3635, Fax: +49-201-5454-1325 E-Mail: nico.hu...@secunet.com Mergenthalerallee 77, 65760 Eschborn, Deutschland www.secunet.com __ Sitz: Kronprinzenstraße 30, 45128 Essen, Deutschland Amtsgericht Essen HRB 13615 Vorstand: Dr. Rainer Baumgart (Vors.), Thomas Pleines Aufsichtsratsvorsitzender: Dr. Peter Zattler __ -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
I would say bring it in to the repo. We have assembly, we have our own C compiler even, so a higher level language hardly seems a bad idea. If this can make our code base better in some way, why not use it? ron On Fri, Aug 28, 2015 at 4:55 AM Nico Huber nico.hu...@secunet.com wrote: Am 21.08.2015 22:47, schrieb ron minnich: This sounds wonderful. I'm all for it. Did you just kill any discussion with that line? :) Things live currently in a small library that I called `libsparkhw` (came spontaneously, I don't care much about the name). It contains some framework (Timer, I/O, Debugging) to do hardware stuff in SPARK, plus the mentioned graphics initialization. I falter to push it directly for inclusion into the coreboot repo. Maybe we want a separate reposi- tory for it (like libpayload has)? It would also make it easier to use it in other components (e.g. bootloaders and such, and of course our own products where the code stems from). We have discussed the licensing and will push this under GPLv2 + later if there are no reasonable concerns. Nico -- M. Sc. Nico Huber Senior Berater SINA-Softwareentwicklung Netzwerk- Client-Sicherheit / Network Client Security Division Öffentliche Auftraggeber / Public Authorities secunet Security Networks AG Tel.: +49-201-5454-3635, Fax: +49-201-5454-1325 E-Mail: nico.hu...@secunet.com Mergenthalerallee 77, 65760 Eschborn, Deutschland www.secunet.com __ Sitz: Kronprinzenstraße 30, 45128 Essen, Deutschland Amtsgericht Essen HRB 13615 Vorstand: Dr. Rainer Baumgart (Vors.), Thomas Pleines Aufsichtsratsvorsitzender: Dr. Peter Zattler __ -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
Am 28.08.2015 16:12, schrieb ron minnich: I would say bring it in to the repo. We have assembly, we have our own C compiler even, so a higher level language hardly seems a bad idea. If this can make our code base better in some way, why not use it? I'm not concerned about the language. More about the content `device driver`, which could be used anywhere (actually, I first planed to use it in FILO, why is coreboot doing gfx init anyway?). Also, I was bothered some- times when code got copied from coreboot to libpayload or the other way around. I don't know if the real problem was that libpayload resides in it's own repository. But having simple device drivers in it's own place seems to be a good idea, for me. Anyway, I'll push it next week (I guess) when the code got cleared. Nico ron On Fri, Aug 28, 2015 at 4:55 AM Nico Huber nico.hu...@secunet.com wrote: Am 21.08.2015 22:47, schrieb ron minnich: This sounds wonderful. I'm all for it. Did you just kill any discussion with that line? :) Things live currently in a small library that I called `libsparkhw` (came spontaneously, I don't care much about the name). It contains some framework (Timer, I/O, Debugging) to do hardware stuff in SPARK, plus the mentioned graphics initialization. I falter to push it directly for inclusion into the coreboot repo. Maybe we want a separate reposi- tory for it (like libpayload has)? It would also make it easier to use it in other components (e.g. bootloaders and such, and of course our own products where the code stems from). We have discussed the licensing and will push this under GPLv2 + later if there are no reasonable concerns. Nico -- M. Sc. Nico Huber Senior Berater SINA-Softwareentwicklung Netzwerk- Client-Sicherheit / Network Client Security Division Öffentliche Auftraggeber / Public Authorities secunet Security Networks AG Tel.: +49-201-5454-3635, Fax: +49-201-5454-1325 E-Mail: nico.hu...@secunet.com Mergenthalerallee 77, 65760 Eschborn, Deutschland www.secunet.com __ Sitz: Kronprinzenstraße 30, 45128 Essen, Deutschland Amtsgericht Essen HRB 13615 Vorstand: Dr. Rainer Baumgart (Vors.), Thomas Pleines Aufsichtsratsvorsitzender: Dr. Peter Zattler __ -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
2015-08-28 16:13 GMT+02:00 Aaron Durbin adur...@google.com: Does this mean we can start writing our userland tools in golang? If you can provide portability guarantees for userland compilers like we do for target compilers... (which is one of my main worries right now: gnat requires an ada compiler to build itself, but even that is pretty portable compared to what's going on with golang) Patrick -- Google Germany GmbH, ABC-Str. 19, 20354 Hamburg Registergericht und -nummer: Hamburg, HRB 86891, Sitz der Gesellschaft: Hamburg Geschäftsführer: Graham Law, Christine Elizabeth Flores -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
2015-08-28 17:35 GMT+02:00 Nico Huber nico.hu...@secunet.com: I don't know if the real problem was that libpayload resides in it's own repository. libpayload still shares a repo with coreboot. There were ideas to separate them, but that never materialized. But having simple device drivers in it's own place seems to be a good idea, for me. If you want, I can setup a repo at coreboot.org, mirrored to github.com/coreboot, and we hook it up as submodule. That would place it somewhere below 3rdparty/ for consistency with the other submodules. Your code, your decision. Patrick -- Google Germany GmbH, ABC-Str. 19, 20354 Hamburg Registergericht und -nummer: Hamburg, HRB 86891, Sitz der Gesellschaft: Hamburg Geschäftsführer: Graham Law, Christine Elizabeth Flores -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
Am 28.08.2015 17:51, schrieb Patrick Georgi: 2015-08-28 17:35 GMT+02:00 Nico Huber nico.hu...@secunet.com: I don't know if the real problem was that libpayload resides in it's own repository. libpayload still shares a repo with coreboot. There were ideas to separate them, but that never materialized. Ah, sorry. Got confused about that. What I meant is that they are not sharing a (code) tree. But having simple device drivers in it's own place seems to be a good idea, for me. If you want, I can setup a repo at coreboot.org, mirrored to github.com/coreboot, and we hook it up as submodule. That would be convenient for external development... That would place it somewhere below 3rdparty/ for consistency with the other submodules. That might hinder people from taking part... Your code, your decision. I can't decide! why did I dare to ask :) Nico -- M. Sc. Nico Huber Senior Berater SINA-Softwareentwicklung Netzwerk- Client-Sicherheit / Network Client Security Division Öffentliche Auftraggeber / Public Authorities secunet Security Networks AG Tel.: +49-201-5454-3635, Fax: +49-201-5454-1325 E-Mail: nico.hu...@secunet.com Mergenthalerallee 77, 65760 Eschborn, Deutschland www.secunet.com __ Sitz: Kronprinzenstraße 30, 45128 Essen, Deutschland Amtsgericht Essen HRB 13615 Vorstand: Dr. Rainer Baumgart (Vors.), Thomas Pleines Aufsichtsratsvorsitzender: Dr. Peter Zattler __ -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
I would really like this to be in the tree. It gives us a chance to do things in coreboot that go beyond C and assembly. So that's my $.02. What harm would it do? ron On Fri, Aug 28, 2015 at 9:00 AM Nico Huber nico.hu...@secunet.com wrote: Am 28.08.2015 17:51, schrieb Patrick Georgi: 2015-08-28 17:35 GMT+02:00 Nico Huber nico.hu...@secunet.com: I don't know if the real problem was that libpayload resides in it's own repository. libpayload still shares a repo with coreboot. There were ideas to separate them, but that never materialized. Ah, sorry. Got confused about that. What I meant is that they are not sharing a (code) tree. But having simple device drivers in it's own place seems to be a good idea, for me. If you want, I can setup a repo at coreboot.org, mirrored to github.com/coreboot, and we hook it up as submodule. That would be convenient for external development... That would place it somewhere below 3rdparty/ for consistency with the other submodules. That might hinder people from taking part... Your code, your decision. I can't decide! why did I dare to ask :) Nico -- M. Sc. Nico Huber Senior Berater SINA-Softwareentwicklung Netzwerk- Client-Sicherheit / Network Client Security Division Öffentliche Auftraggeber / Public Authorities secunet Security Networks AG Tel.: +49-201-5454-3635, Fax: +49-201-5454-1325 E-Mail: nico.hu...@secunet.com Mergenthalerallee 77, 65760 Eschborn, Deutschland www.secunet.com __ Sitz: Kronprinzenstraße 30, 45128 Essen, Deutschland Amtsgericht Essen HRB 13615 Vorstand: Dr. Rainer Baumgart (Vors.), Thomas Pleines Aufsichtsratsvorsitzender: Dr. Peter Zattler __ -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
we can work those problems. We've worked lots harder ones. . Heck, we build, what, 5 different cross compilers now, including the riscv one that's not even official? And our *own* C compiler? I think we're OK. Oh, wait, I forgot: we have the ACPI language too. If we can put something that ugly in the tree, I think we have room for SPARK :-) I think this could prove to be a very important new direction for coreboot, and I still think it belongs in the tree. ron On Fri, Aug 28, 2015 at 2:07 PM Carl-Daniel Hailfinger c-d.hailfinger.devel.2...@gmx.net wrote: On 28.08.2015 18:29, ron minnich wrote: I would really like this to be in the tree. It gives us a chance to do things in coreboot that go beyond C and assembly. So that's my $.02. Same here. Besides that, SPARK gives us easier provability of code. That is something governments and safety engineers care about, and it makes for great marketing. What harm would it do? Having it in the tree would be beneficial because it will be easier to bisect than two separate trees. The only thing I'm worried about is whether we can guarantee the complete tree can be built on all of the platforms where it can be built now, but that worry applies in the git submodule case as well. Regards, Carl-Daniel -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
On 28.08.2015 18:29, ron minnich wrote: I would really like this to be in the tree. It gives us a chance to do things in coreboot that go beyond C and assembly. So that's my $.02. Same here. Besides that, SPARK gives us easier provability of code. That is something governments and safety engineers care about, and it makes for great marketing. What harm would it do? Having it in the tree would be beneficial because it will be easier to bisect than two separate trees. The only thing I'm worried about is whether we can guarantee the complete tree can be built on all of the platforms where it can be built now, but that worry applies in the git submodule case as well. Regards, Carl-Daniel -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
[coreboot] From wildcat linking to exploring new languages
Hi coreboot folks! I dared to link a coreboot ramstage with, well, not only C but also some fancy Ada code. Things worked out well enough and I'd like to release the code and upstream it. So, first question: Would it be accepted? being written in a foreign language? This code is essentially another native graphics init for Intel hardware, currently supporting eDP on Ivybridge (tested as native graphics init), and everything but VGA on Haswell/Broadwell (only tested after VBIOS ran). It's actually written in SPARK (Ada subset, with some provability extras). Some internal discussion concluded that we may release it if there is either real interest in the community to take part in further development or at least the will to review it and give feedback. As this is one of the areas where I work on nearly the same code in C and Ada, I hope for the former :) Thoughts? And here comes the disclaimer: Supporting the upstreaming of this code may result in further development in Ada (currently I only have simple device drivers in mind). I hereby wash my hands of any damage that may arise from that :P Nico -- M. Sc. Nico Huber Senior Berater SINA-Softwareentwicklung Netzwerk- Client-Sicherheit / Network Client Security Division Öffentliche Auftraggeber / Public Authorities secunet Security Networks AG Tel.: +49-201-5454-3635, Fax: +49-201-5454-1325 E-Mail: nico.hu...@secunet.com Mergenthalerallee 77, 65760 Eschborn, Deutschland www.secunet.com __ Sitz: Kronprinzenstraße 30, 45128 Essen, Deutschland Amtsgericht Essen HRB 13615 Vorstand: Dr. Rainer Baumgart (Vors.), Thomas Pleines Aufsichtsratsvorsitzender: Dr. Peter Zattler __ -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] From wildcat linking to exploring new languages
This sounds wonderful. I'm all for it. ron On Fri, Aug 21, 2015 at 1:46 PM Nico Huber nico.hu...@secunet.com wrote: Hi coreboot folks! I dared to link a coreboot ramstage with, well, not only C but also some fancy Ada code. Things worked out well enough and I'd like to release the code and upstream it. So, first question: Would it be accepted? being written in a foreign language? This code is essentially another native graphics init for Intel hardware, currently supporting eDP on Ivybridge (tested as native graphics init), and everything but VGA on Haswell/Broadwell (only tested after VBIOS ran). It's actually written in SPARK (Ada subset, with some provability extras). Some internal discussion concluded that we may release it if there is either real interest in the community to take part in further development or at least the will to review it and give feedback. As this is one of the areas where I work on nearly the same code in C and Ada, I hope for the former :) Thoughts? And here comes the disclaimer: Supporting the upstreaming of this code may result in further development in Ada (currently I only have simple device drivers in mind). I hereby wash my hands of any damage that may arise from that :P Nico -- M. Sc. Nico Huber Senior Berater SINA-Softwareentwicklung Netzwerk- Client-Sicherheit / Network Client Security Division Öffentliche Auftraggeber / Public Authorities secunet Security Networks AG Tel.: +49-201-5454-3635, Fax: +49-201-5454-1325 E-Mail: nico.hu...@secunet.com Mergenthalerallee 77, 65760 Eschborn, Deutschland www.secunet.com __ Sitz: Kronprinzenstraße 30, 45128 Essen, Deutschland Amtsgericht Essen HRB 13615 Vorstand: Dr. Rainer Baumgart (Vors.), Thomas Pleines Aufsichtsratsvorsitzender: Dr. Peter Zattler __ -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot