Attached is an updated version of the patch I sent before that streamined the .mk files that are included in buildrom. This version is updated with the latest and greatest from Myles and everybody.
Tested with all payloads + coreinfo, and hd8mr + kernel (I'm going to test the other platforms with kernel, but I didn't want to wait around for them to finish building).
Jordan
-----Original Message----- From: coreboot-bounces@coreboot.org [mailto:coreboot-bounces@coreboot.org] On Behalf Of Jordan Crouse Sent: Monday, May 05, 2008 3:55 PM To: coreboot@coreboot.org Subject: [coreboot] buildrom: cleanup package includes (updated)
Attached is an updated version of the patch I sent before that streamined the .mk files that are included in buildrom. This version is updated with the latest and greatest from Myles and everybody.
Tested with all payloads + coreinfo, and hd8mr + kernel (I'm going to test the other platforms with kernel, but I didn't want to wait around for them to finish building).
Acked-by: Myles Watson mylesgw@gmail.com
Thanks, Myles
On 05/05/08 17:02 -0600, Myles Watson wrote:
-----Original Message----- From: coreboot-bounces@coreboot.org [mailto:coreboot-bounces@coreboot.org] On Behalf Of Jordan Crouse Sent: Monday, May 05, 2008 3:55 PM To: coreboot@coreboot.org Subject: [coreboot] buildrom: cleanup package includes (updated)
Attached is an updated version of the patch I sent before that streamined the .mk files that are included in buildrom. This version is updated with the latest and greatest from Myles and everybody.
Tested with all payloads + coreinfo, and hd8mr + kernel (I'm going to test the other platforms with kernel, but I didn't want to wait around for them to finish building).
Acked-by: Myles Watson mylesgw@gmail.com
Thanks - I am going to wait on this until tomorrow. Coreboot-v2 + kernel failed on 13 of 17 targets (and only 3 failures are due to missing .mk files). I don't know if this is becuase of my patch, my test harness, or something else. Either way, it is best that I hold off until I can debug it more fully.
Jordan
On 05/05/08 17:29 -0600, Jordan Crouse wrote:
On 05/05/08 17:02 -0600, Myles Watson wrote:
-----Original Message----- From: coreboot-bounces@coreboot.org [mailto:coreboot-bounces@coreboot.org] On Behalf Of Jordan Crouse Sent: Monday, May 05, 2008 3:55 PM To: coreboot@coreboot.org Subject: [coreboot] buildrom: cleanup package includes (updated)
Attached is an updated version of the patch I sent before that streamined the .mk files that are included in buildrom. This version is updated with the latest and greatest from Myles and everybody.
Tested with all payloads + coreinfo, and hd8mr + kernel (I'm going to test the other platforms with kernel, but I didn't want to wait around for them to finish building).
Acked-by: Myles Watson mylesgw@gmail.com
Thanks - I am going to wait on this until tomorrow. Coreboot-v2 + kernel failed on 13 of 17 targets (and only 3 failures are due to missing .mk files). I don't know if this is becuase of my patch, my test harness, or something else. Either way, it is best that I hold off until I can debug it more fully.
I had a second to build one of the failing platforms at random (s2882), and the error ended up being that the kernel was too big for the ROM. I'm going to go ahead and re-run the entire test with verboseness turned on over night to get a clear picture of which platforms need love.
Jordan
--- Jordan Crouse Systems Software Development Engineer Advanced Micro Devices, Inc.
-----Original Message----- From: Jordan Crouse [mailto:jordan.crouse@amd.com] Sent: Monday, May 05, 2008 5:58 PM To: Myles Watson Cc: coreboot@coreboot.org Subject: Re: buildrom: cleanup package includes (updated)
On 05/05/08 17:29 -0600, Jordan Crouse wrote:
On 05/05/08 17:02 -0600, Myles Watson wrote:
-----Original Message----- From: coreboot-bounces@coreboot.org [mailto:coreboot-
bounces@coreboot.org]
On Behalf Of Jordan Crouse Sent: Monday, May 05, 2008 3:55 PM To: coreboot@coreboot.org Subject: [coreboot] buildrom: cleanup package includes (updated)
Attached is an updated version of the patch I sent before that streamined the .mk files that are included in buildrom. This
version
is updated with the latest and greatest from Myles and everybody.
Tested with all payloads + coreinfo, and hd8mr + kernel (I'm going
to
test the other platforms with kernel, but I didn't want to wait
around
for them to finish building).
Acked-by: Myles Watson mylesgw@gmail.com
Thanks - I am going to wait on this until tomorrow. Coreboot-v2 +
kernel
failed on 13 of 17 targets (and only 3 failures are due to missing .mk
files).
I don't know if this is becuase of my patch, my test harness, or
something
else. Either way, it is best that I hold off until I can debug it more fully.
I had a second to build one of the failing platforms at random (s2882), and the error ended up being that the kernel was too big for the ROM. I'm going to go ahead and re-run the entire test with verboseness turned on over night to get a clear picture of which platforms need love.
Several of the platforms have kernels that have never fit for me. I always figured the problem was my broken toolchain. I only test that changes don't break things that were working.
Myles
On Mon, May 05, 2008 at 07:39:01PM -0600, Myles Watson wrote:
Several of the platforms have kernels that have never fit for me.
I've never gotten a pure kernel payload working - but I never really investigated, LAB was more interesting. Time to get that fixed I think.
So; given that we have tight space constraints, is the idea for a kernel payload to have non-essential modules on disk to be loaded later?
In that case we can probably start with the LAB kernel configs. That's actually probably the fastest (only?) way to make the kernel fit.
Thanks, Ward.
On Mon, May 5, 2008 at 8:35 PM, Ward Vandewege ward@gnu.org wrote:
So; given that we have tight space constraints, is the idea for a kernel payload to have non-essential modules on disk to be loaded later?
yes. Just enough to boot from disk or net. I did pure kernel payloads with network drivers only.
I even once experimented with modifying linux main() to put the bootstrap code right into the kernel -- this saves a ton of space, actually.
thanks
ron
On 05/05/08 19:39 -0600, Myles Watson wrote:
-----Original Message----- From: Jordan Crouse [mailto:jordan.crouse@amd.com] Sent: Monday, May 05, 2008 5:58 PM To: Myles Watson Cc: coreboot@coreboot.org Subject: Re: buildrom: cleanup package includes (updated)
On 05/05/08 17:29 -0600, Jordan Crouse wrote:
On 05/05/08 17:02 -0600, Myles Watson wrote:
-----Original Message----- From: coreboot-bounces@coreboot.org [mailto:coreboot-
bounces@coreboot.org]
On Behalf Of Jordan Crouse Sent: Monday, May 05, 2008 3:55 PM To: coreboot@coreboot.org Subject: [coreboot] buildrom: cleanup package includes (updated)
Attached is an updated version of the patch I sent before that streamined the .mk files that are included in buildrom. This
version
is updated with the latest and greatest from Myles and everybody.
Tested with all payloads + coreinfo, and hd8mr + kernel (I'm going
to
test the other platforms with kernel, but I didn't want to wait
around
for them to finish building).
Acked-by: Myles Watson mylesgw@gmail.com
Thanks - I am going to wait on this until tomorrow. Coreboot-v2 +
kernel
failed on 13 of 17 targets (and only 3 failures are due to missing .mk
files).
I don't know if this is becuase of my patch, my test harness, or
something
else. Either way, it is best that I hold off until I can debug it more fully.
I had a second to build one of the failing platforms at random (s2882), and the error ended up being that the kernel was too big for the ROM. I'm going to go ahead and re-run the entire test with verboseness turned on over night to get a clear picture of which platforms need love.
Several of the platforms have kernels that have never fit for me. I always figured the problem was my broken toolchain. I only test that changes don't break things that were working.
Its a good thing I waited - I had a misspelling in coreboot-v2.mk (PLAFORM_GEODE). I fixed that, and the Geode platforms still failed, but at least this time they failed because of size.
So, for the record, the following platforms have kernel configs that are too big (or don't have an accomodating Config.lb):
dbe61 db800 ga-m57sli-s4 serengeti-cheetah norwich msm800sev alix1c s2892 s2881 s2882
The following platforms have no kernel.mk:
a8v-e-se s-c-fam10 ga-2761gxdk
And the following platforms build a kernel successfully:
h8dmr qemu s2891 s2895
All platforms successfully built FILO and coreinfo. I haven't yet added any of the other payloads to the test.
Jordan
Its a good thing I waited - I had a misspelling in coreboot-v2.mk (PLAFORM_GEODE). I fixed that, and the Geode platforms still failed, but at least this time they failed because of size.
So, for the record, the following platforms have kernel configs that are too big (or don't have an accomodating Config.lb):
dbe61 db800 ga-m57sli-s4 serengeti-cheetah norwich msm800sev alix1c s2892 s2881 s2882
The Tyan boards should all work with 64-bit. That kernel is quite a bit smaller.
Thanks, Myles
The following platforms have no kernel.mk:
a8v-e-se s-c-fam10 ga-2761gxdk
And the following platforms build a kernel successfully:
h8dmr qemu s2891 s2895
All platforms successfully built FILO and coreinfo. I haven't yet added any of the other payloads to the test.
Jordan
On 06/05/08 09:50 -0600, Myles Watson wrote:
Its a good thing I waited - I had a misspelling in coreboot-v2.mk (PLAFORM_GEODE). I fixed that, and the Geode platforms still failed, but at least this time they failed because of size.
So, for the record, the following platforms have kernel configs that are too big (or don't have an accomodating Config.lb):
dbe61 db800 ga-m57sli-s4 serengeti-cheetah norwich msm800sev alix1c s2892 s2881 s2882
The Tyan boards should all work with 64-bit. That kernel is quite a bit smaller.
Ahh - good point. I did all these with 32 bit builds. Thats an important data point.
Jordan
On Mon, May 05, 2008 at 05:29:48PM -0600, Jordan Crouse wrote:
On 05/05/08 17:02 -0600, Myles Watson wrote:
-----Original Message----- From: coreboot-bounces@coreboot.org [mailto:coreboot-bounces@coreboot.org] On Behalf Of Jordan Crouse Sent: Monday, May 05, 2008 3:55 PM To: coreboot@coreboot.org Subject: [coreboot] buildrom: cleanup package includes (updated)
Attached is an updated version of the patch I sent before that streamined the .mk files that are included in buildrom. This version is updated with the latest and greatest from Myles and everybody.
Tested with all payloads + coreinfo, and hd8mr + kernel (I'm going to test the other platforms with kernel, but I didn't want to wait around for them to finish building).
Acked-by: Myles Watson mylesgw@gmail.com
Thanks - I am going to wait on this until tomorrow. Coreboot-v2 + kernel failed on 13 of 17 targets (and only 3 failures are due to missing .mk files). I don't know if this is becuase of my patch, my test harness, or something else. Either way, it is best that I hold off until I can debug it more fully.
Do you have a testing script for this? If yes, it would be nice to commit that so everbody can run the tests when changing code or adding new targets/payloads...
Uwe.
On 06/05/08 14:42 +0200, Uwe Hermann wrote:
On Mon, May 05, 2008 at 05:29:48PM -0600, Jordan Crouse wrote:
On 05/05/08 17:02 -0600, Myles Watson wrote:
-----Original Message----- From: coreboot-bounces@coreboot.org [mailto:coreboot-bounces@coreboot.org] On Behalf Of Jordan Crouse Sent: Monday, May 05, 2008 3:55 PM To: coreboot@coreboot.org Subject: [coreboot] buildrom: cleanup package includes (updated)
Attached is an updated version of the patch I sent before that streamined the .mk files that are included in buildrom. This version is updated with the latest and greatest from Myles and everybody.
Tested with all payloads + coreinfo, and hd8mr + kernel (I'm going to test the other platforms with kernel, but I didn't want to wait around for them to finish building).
Acked-by: Myles Watson mylesgw@gmail.com
Thanks - I am going to wait on this until tomorrow. Coreboot-v2 + kernel failed on 13 of 17 targets (and only 3 failures are due to missing .mk files). I don't know if this is becuase of my patch, my test harness, or something else. Either way, it is best that I hold off until I can debug it more fully.
Do you have a testing script for this? If yes, it would be nice to commit that so everbody can run the tests when changing code or adding new targets/payloads...
I'm still working on a testing script. I'm not sure how useful it might end up being, but I will commit it when I am happier with how it works.
Jordan
On 05/05/08 17:02 -0600, Myles Watson wrote:
-----Original Message----- From: coreboot-bounces@coreboot.org [mailto:coreboot-bounces@coreboot.org] On Behalf Of Jordan Crouse Sent: Monday, May 05, 2008 3:55 PM To: coreboot@coreboot.org Subject: [coreboot] buildrom: cleanup package includes (updated)
Attached is an updated version of the patch I sent before that streamined the .mk files that are included in buildrom. This version is updated with the latest and greatest from Myles and everybody.
Tested with all payloads + coreinfo, and hd8mr + kernel (I'm going to test the other platforms with kernel, but I didn't want to wait around for them to finish building).
Acked-by: Myles Watson mylesgw@gmail.com
r182 - patch as posted + corrected mispelling of PLATFORM_GEODE
Thanks, Myles
-
Jordan Crouse -
Myles Watson -
ron minnich -
Uwe Hermann -
Ward Vandewege