Re: [PATCH v2] bootp: add native DHCPv4 support

[Daniel Kiper]

On Fri, Nov 18, 2016 at 08:51:43PM +0300, Andrei Borzenkov wrote:
> 17.11.2016 22:06, Daniel Kiper ??????????:
> > On Wed, Apr 27, 2016 at 04:34:19PM +0100, Mark Rutland wrote:
> >> Hi,
> >>
> >> On Tue, Mar 22, 2016 at 09:26:20PM +0300, Andrei Borzenkov wrote:
> >>> This patch adds support for native DHCPv4 and removes requirement for
> >>> BOOTP compatibility support in DHCP server.
> >>
> >> Sorry for the delay in testing this; I'd meant to take a look at this
> >> but it managed to fall off my radar.
> >>
> >> I've just given this a spin atop of 7a5b301e3adb8e05 ("build: Use
> >> AC_HEADER_MAJOR to find device macros"), with a small conflict manually
> >> fixed up (I assume that the NULL check for inter still stands).
> >>
> >> Otherwise, this seems to work perfectly for me in my BOOT-free work
> >> environment on an AArch64 UEFI platform:
> >>
> >> grub> insmod efinet
> >> grub> net_ls_cards
> >> efinet0 00:02:f7:00:61:79
> >> grub> net_dhcp efinet0
> >> grub> net_default_server=10.1.205.151
> >> grub> linux (tftp)/Image.gz
> >> grub> boot
> >> EFI stub: Booting Linux Kernel...
> >> EFI stub: Using DTB from configuration table
> >> EFI stub: Exiting boot services and installing virtual address map...
> >> [    0.000000] Booting Linux on physical CPU 0x100
> >> [    0.000000] Linux version 4.6.0-rc5-next-20160427+ (address@hidden) 
> >> (gcc version 5.1.1 20150608 (Linaro GCC 5.1-2015.08) ) #8 SMP PREEMPT Wed 
> >> Apr 27 14:53:05 BST 2016
> >> ...
> >>
> >> As previously, I don't have a BOOTP server to hand to test with; sorry
> >> for not being able to test that.
> >>
> >> It would be great to see this committed!
> >
> > Vladimir, Andrei, should we take it into 2.02 train?
> >
>
> I do not think so, not at this time, sorry. It is far too intrusive
> change, and patch is obviously incomplete, in particular, it does not

OK, added to list of tasks after 2.02 release.

> touch lease renewal at all. I also think we should keep net_bootp doing
> what it does now, and provide new net_dhcp, not replacing exiting
> functionality.

That it good idea.

> What we do need is to fix manual to not claim we are doing DHCP right now.

Make sense for me.

Daniel