Hi Alex,
On 20 August 2018 at 16:29, Alexander Graf agraf@suse.de wrote:
On 20.08.18 20:54, Simon Glass wrote:
This partially reverts commit 7e21fbca26d18327cf7cabaad08df276a06a07d8.
That change broke sandbox EFI support for unknown reasons. It also changes
Wouldn't it be better to just figure out the reasons? So far all bugs I've found were linker script related and quite obvious once you start to dig into them.
sandbox to use--gc-sections which we don't want.
Why don't we want gc-sections with sandbox?
It is a space optimisation which we don't need for sandbox. It also complicates the object files unnecessarily.
Put another way, why is it desirable?
For now I am just reverting the sandbox portion as presumably this change is safe on other architectures.
Sandbox is your target, so you're free to do whatever you like :). But I'm not sure this is the right path forward. I'd rather like to keep things consistent.
In what sense?
So what do you expect happens with this patch? A resend of a patch 1/18 by itself doesn't really tell me what you're trying to say.
The resend was due to me noticing that people did not get the patch on cc. I only sent this one patch, but I can resend send the whole series if you like.
Regards, Simon