Rik van Riel wrote:
Zachary Amsden wrote:
And by NO circumstances, is it required to be a CLOSED source binary
blob. In fact, why can't it be open? In the event of a firmware bug,
in fact, it is very desirable to have this software be open so that
it can be fixed
You're making a very good argument as to why we should probably
require that the code linking against such an interface, if we
decide we want one, should be required to be open source.
Personally, I don't feel a strong requirement that it be open source,
because I don't believe it violates the intent of the GPL license by
crippling free distribution of the kernel, requiring some fee for use,
or doing anything unethical. There have been charges that the VMI layer
is deliberately designed as a GPL circumvention device, which I want to
stamp out now before we try to get any code for integrating to it
upstreamed.
I think you will see why our VMI layer is quite similar to a
traditional ROM, and very dissimilar to an evil GPL-circumvention
device.
(?) There are only two reasonable objections I can see to open
sourcing the binary layer.
Since none of the vendors that might use such a paravirtualized
ROM for Linux actually have one of these reasons for keeping their
paravirtualized ROM blob closed source, I say we might as well
require that it be open source.
I think saying require at this point is a bit preliminary for us -- I'm
trying to prove we're not being evil and subverting the GPL, but I'm
also not guaranteeing yet that we can open-source the code under a
specific license. Sorry about having to doublespeak here - but we have
not yet got a green light to open source the VMI layer under the GPL.
Perhaps there are some other issues I haven't conceived of. We still
have some source separation issues with creating a build environment due
to entangled header files - that is being sorted out, but we're
certainly not ready to distribute an open source buildable VMI layer for
ESX today. I sincerely hope we will be very soon.
As for the evilness of a binary interface - the interface between
kernel and userland is a stable binary interface and is decidedly
non-evil. I could see a similar use for a stable paravirtualization
interface, to make compatibility between Linux and various hypervisor
versions easier.
As long as it's open source so the thing can be debugged :)
Unfortunately, inlining and patching code will break CFI debug
information! I haven't thought of a way to fix this yet other than
using frame pointers. At least the possibility of debugging exists.
Zach
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]
