[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZWAGhy4ZuYvKQeh2@ntb.petris.klfree.czf>
Date: Fri, 24 Nov 2023 03:12:23 +0100
From: Petr Malat <oss@...at.biz>
To: Masami Hiramatsu <mhiramat@...nel.org>
Cc: paulmck@...nel.org, Randy Dunlap <rdunlap@...radead.org>,
linux-kernel@...r.kernel.org, rostedt@...dmis.org
Subject: Re: [PATCH 2/2] bootconfig: Apply early options from embedded config
Hi!
On Thu, Nov 23, 2023 at 11:18:05PM +0900, Masami Hiramatsu wrote:
> On Thu, 23 Nov 2023 11:04:00 +0100
> > On Thu, Nov 23, 2023 at 11:22:07AM +0900, Masami Hiramatsu wrote:
> > > BTW, we also need to make a block-list for some early params. some of those
> > > MUST be passed from the bootloader. E.g. initrd address and size will be
> > > passed from the bootloader via commandline. Thus such params in the embedded
> > > bootconfig should be filtered at the build time.
> >
> > It's ok to configure these in the embedded bootconfig - in a case they are
> > provided by the bootloader, the bootconfig value is overridden, if not, the
> > value from bootconfig is used, so it works as expected.
>
> I meant some params only bootloader knows, like where the initrd is loaded.
> Anyway, if user sets such value, it will break the kernel boot as expected :P.
If somebody sets these by hand, he probably knows what is he doing. I
remember myself hardcoding initrd size in Linux to workaround broken
bootloader.
I find it more important to warn the user when he tries to set early
options in initrd (assuming my change gets merged), as that's what
burned me.
Petr
Powered by blists - more mailing lists