Date: Tue, 01 Jan 2008 13:27:03 -0500 From: Chuck Robey <chuckr@chuckr.org> To: Peter Jeremy <peterjeremy@optushome.com.au> Cc: FreeBSD-Hackers <freebsd-hackers@freebsd.org> Subject: Re: printing boot probe messages Message-ID: <477A85F7.1030906@chuckr.org> In-Reply-To: <20080101035355.GM72458@server.vk2pj.dyndns.org> References: <86abo1g2pu.fsf@ds4.des.no> <476EA521.9020707@chuckr.org> <20071224131425.GF23337@cicely12.cicely.de> <47715E1F.9080502@chuckr.org> <86sl1qikam.fsf@ds4.des.no> <477440A2.4070601@chuckr.org> <86sl1n14vv.fsf@ds4.des.no> <477698D8.2040802@chuckr.org> <20071229213216.GV40785@server.vk2pj.dyndns.org> <47782631.6070305@chuckr.org> <20080101035355.GM72458@server.vk2pj.dyndns.org>
next in thread | previous in thread | raw e-mail | index | archive | help
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Peter Jeremy wrote:
> On Sun, Dec 30, 2007 at 06:13:53PM -0500, Chuck Robey wrote:
>> during booting is to call out a verbose boot. If I do that, then the boot
>> messages DO print during booting, and examination afterwards shows a big
>> file (~ 60K in size). This happens if I use /boot.config==-v, or if I
>> enter option #5 to the beastie menu, or even if I call out a manual loader
>> run, and enter boot -v at the prompt.
>>
>> If I do none of those things, it will not print boot messages.
>
> This doesn't really make sense because the only difference is code like
> if (bootverbose)
> printf("Lots more detail\n");
> and there shouldn't be any side-effects.
>
>> After
>> booting, if I do a "dmesg > dm" then the resulting file is only about 10k
>> in size, and things like the enormous pci probing don't show up.
>
> That makes sense.
>
>> At this
>> moment, I'm in a boot that wasn't printed, and I just got a 0 back from a
>> "sysctl debug.bootverbose"
>
> That is also correct.
>
>> Right now, I moved my /boot.config to /notboot.config. My loader.conf is
>>
>> boot_mute="NO"
>
> I think this is your problem. The kernel doesn't look at the value of
> the environment variable, just whether or not it is set. Try removing
> or commenting out this line and check for other occurrences of "boot_mute"
Fantastic, that was it! There is a strong chance that at some point, I
also had something wrong with my device.hints, because that boot_mute line
wasn't in there when my non-printing problems began, I put it in while
searching for something to make the probes print. During this, I think I
finally got my worst misunderstanding, that of how to formulate a good
device.hints, corrected.
Anyhow, great to finally have this fixed, thanks, now I can get back to my
real problems (which I am not hitting the list with).
>
>> verbose_loading="YES"
>
> verbose_loading is different to boot_verbose. The former makes the forth
> boot loader more verbose, the latter is passed onto the kernel and makes
> it more verbose.
>
>> Sure. "sysctl debug.boothowto" returns -2147418112, any idea what that
>> means?
>
> It's a bitmask of values defined in <sys/reboot.h>
> -2147418112 == 0x80010000 == RB_BOOTINFO | RB_MUTE
>
> The "RB_MUTE" explains the lack of console messages and is caused by
> the boot_mute="NO" in your kenv output. Other than that, the kenv
> all looks reasonable.
>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.4 (FreeBSD)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFHeoX3z62J6PPcoOkRAgsfAJkBgbRWWnT1AOiWhZuRMJKOKGYi+ACbBpTd
ZjjQq71Ll9ZjUNYlGHTByoc=
=AdUW
-----END PGP SIGNATURE-----
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?477A85F7.1030906>