internel flashing 4MB only / release management about skulls

[merge]

I wrote a first draft of such a script. It's completely untested! Please, if you can, take a look at it and test it. I'll do so too, sometime "soon". thanks a lot!

[nemanjan00]

I actually do use internal flashing and it proved itself very safe. And yes, you can flash 4mb flash separately...

[merge]

thanks for the feedback. how do you currently do it?

[nemanjan00]

flash.sh

sudo flashrom -p internal --layout x230-layout.txt --image bios --write ./coreboot/build/coreboot.rom

x230-layout.txt

0x00000000:0x007fffff ifdmegbe
0x00800000:0x00bfffff bios

Edit:

Keep in mind coreboot.rom is still 12MB...

Edit 2:

I just figured out this is from your README 🗡️

[merge]

thanks. that's what I meant :) The new "prepare" script takes the (released) 4MB image and creates a 12MB one, along with the layout-file and instructions on how to flash it internally.

Could you have a look at it? That's be of help.

[nemanjan00]

I just woke up so, I am mostly confused on what is happening...

I will take a look at it now...

[nemanjan00]

Looks great to me.

I am thinking that actually it might not be that bad idea to give option to read 8MB image instead of generating it with dd...

So it can later be extended with more features (for example ME removing and using additional space)

[merge]

well. yes. we could do a lot :)

I actually don't want to run flashrom from this "prepare" script. This should be runnable by users, not only by root.

Also I want to be careful here, because it's required to have the 8MB chip flashed externally at least once (for unlocking), before being able to flash internally. Can we detect that when trying to flashrom-read internally? Or does reading already fail if not unlocked?

But you're right...

• we could add an option to provide a full or 8mb (already read) file, that we update instead of prepending zeroes (only do that by default)
• we could create another script that really runs flashrom, and uses the above script. I clearly have to test flashing internally myself. We'd need something that works for everybody. But at least we could come up with something people can test...

In a first step, I'd like to have the current script tested and ready. I updated the README a bit. Future releases won't include a 12MB image.

[nemanjan00]

I am thinking about creating testing platform for this. I have board i do not use. IIt might be interesting to connect it and programmer connected to it online so we can use auto testing... Also, I would connect video capture card to internet...

…

On Fri, Apr 13, 2018, 10:02 Martin Kepplinger ***@***.***> wrote: well. yes. we could do a lot :) I actually don't want to run flashrom from *this* "prepare" script. This should be runnable by users, not only by root. Also I want to be careful here, because it's required to have the 8MB chip flashed externally at least once (for unlocking), before being able to flash internally. Can we detect that when trying to flashrom-read internally? Or does reading already fail if not unlocked? But you're right... - we could add an option to provide a full or 8mb (already read) file, *that* we update instead of prepending zeroes (only do that by default) - we could create *another* script that really runs flashrom, and *uses* the above script. I clearly have to test flashing internally myself. We'd need something that works for everybody. But at least we could come up with something people can test... In a first step, I'd like to have the current script tested and ready. I updated the README a bit. Future releases won't include a 12MB image. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#18 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAbWl4SA9yvPcELDdcB-NUNPuCuoooEEks5toFuYgaJpZM4TS5VN> .

[merge]

I guess we'll always require to write an unlocked image once, using a hardware flasher. But could we use https://github.com/google/pawn to initially get the image so that we could only write the unlocked one once?

That'd save quite some time. Anyone cares to compare original bios images read via hardware with the results from this tool?

[nemanjan00]

Yes, we can not flash internally until we flash unlocked image.

I do not have locked one so I can not compare them...

[merge]

reading is not possible either? I created a new issue #19 for this question.

I tested internal flashing using the script now. seems to be fine... I'll leave this issue open for a while in case we get another test.

[nemanjan00]

I will try reading Bios on T430, when I get home. I do not have locked x230 image...

…

On Fri, Apr 13, 2018, 14:03 Martin Kepplinger ***@***.***> wrote: reading is not possible either? I created a new issue #19 <#19> for this question. I tested internal flashing using the script now. seems to be fine... I'll leave this issue open for a while in case we get another test. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#18 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAbWl76FphTM_t889UMq3lxmXtayCRj9ks5toJQcgaJpZM4TS5VN> .

[merge]

with the 12mb-generating "prepare"-script tested, and a release script added now, this issue is actually resolved. #19 is more interesting anyways :)