Re: "do not modify" blobs
On Tue, Oct 03, 2006 at 03:19:51PM -0700, firstname.lastname@example.org wrote:
> >Has anyone done a survey to see how many "do not modify" blobs
> >we are talking about here?
> Not counting files already removed in 2.6.17,
> drivers/net/appletalk/cops_ffdrv.h use-only (2)
> drivers/net/appletalk/cops_ltdrv.h use-only (2)
> drivers/net/tg3.c redistr-only
> drivers/net/tokenring/3c359_microcode.h use-only (2)
> drivers/net/typhoon-firmware.h redistr-only
> drivers/scsi/qla2xxx/ql2100_fw.c redistr-only (1)
> drivers/scsi/qla2xxx/ql2200_fw.c redistr-only (1)
> drivers/scsi/qla2xxx/ql2300_fw.c redistr-only (1)
> drivers/scsi/qla2xxx/ql2322_fw.c redistr-only (1)
> drivers/scsi/qla2xxx/ql2400_fw.c redistr-only (1)
> drivers/usb/misc/emi26_fw.h redistr-only (2,3)
> (1) deprecated upstream, removed upstream as of 2.6.18?
> (2) marked for deletion by recent "Kernel team position statement"
> (3) redistributable as part of a Linux or other Open Source operating
> system kernel
Which means that, at most, we have two installer-relevant firmware blobs
distributed in the upstream kernel (typhoon and tg3) that are not covered by
the exception in Frederik's revised proposal.
If typhoon-firmware.h isn't needed to run any hardware, and is used only for
performance reasons (I don't know if this is the case), that count is down
I'm inclined to think that the amount of work being spent on drafting a GR
with principles allowing us to include tg3 and typhoon in main is
disproportionate. Would it not be more straightforward to add these two
firmware blobs to the exception by name? That way, there's no doubt left to
voters that this GR will be used as a justification for adding a bunch of
new firmware to main on the grounds that it's "needed" by the installer for
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.