Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 08 Oct 2014 21:46:29 +0200
From:      marekrud@gmail.com
To:        John Baldwin <jhb@freebsd.org>
Cc:        Kevin Oberman <rkoberman@gmail.com>, freebsd-stable@freebsd.org
Subject:   Re: vt does not resume properly after zzz
Message-ID:  <86d2a29wfe.fsf@gmail.com>
In-Reply-To: <1526283.Onb3mfWQFN@ralph.baldwin.cx>
References:  <864mvqfabt.fsf@gmail.com> <1506050.d54hhbRZ6V@ralph.baldwin.cx> <86h9ze9zi9.fsf@gmail.com> <1526283.Onb3mfWQFN@ralph.baldwin.cx>

next in thread | previous in thread | raw e-mail | index | archive | help
John Baldwin <jhb@freebsd.org> writes:

> On Wednesday, October 08, 2014 08:39:58 PM marekrud@gmail.com wrote:
>> John Baldwin <jhb@freebsd.org> writes:
>> > On Wednesday, October 08, 2014 05:47:14 PM marekrud@gmail.com wrote:
>> >> John Baldwin <jhb@freebsd.org> writes:
>> >> > On Saturday, October 04, 2014 04:24:11 PM Kevin Oberman wrote:
>> >> >> On Thu, Oct 2, 2014 at 1:52 PM, <marekrud@gmail.com> wrote:
>> >> >> >  - DELL M1330 with Intel graphics card (Xorg used to work with intel
>> >> >> >
>> >> >> >    driver
>> >> >>
>> >> >> All of the information I can find says that this unit has nVidia
>> >> >> graphics,
>> >> >> but Intel may be a low-priced option, as well. Again,
>> >> >> /var/log/Xorg.0.log
>> >> >> should have this information as should "pciconf -lv | grep -A3 vga".
>> >> >> This
>> >> >> laptop goes back to 2007, so it should be using the old UMS Intel
>> >> >> graphics.
>> >> >> It should not be using VESA, but if it is, that might point out a
>> >> >> common
>> >> >> thread.
>> >> >>
>> >> >> I don't know the details and the actual problem was never identified,
>> >> >> but
>> >> >> I
>> >> >> know that some systems needed to have a kernel built with "NOOPTION
>> >> >> VESA"
>> >> >> to get it to resume. I had this problem on my Lenovo T520 (which I am
>> >> >> using
>> >> >> to send this reply).
>> >> >
>> >> > That doesn't apply to his system.  I have an older HP netbook (i386)
>> >> > that
>> >> > resumes fine in text mode with syscons, but does not resume in text
>> >> > mode
>> >> > in
>> >> > vt(4).  (This is a case where the VESA bits actually help rather than
>> >> > hurt.) However, if I kldload the kms driver ('kldload i915kms') when
>> >> > using vt(4), then resume works fine (and it also works fine in X).
>> >> >
>> >> > Marek,
>> >> >
>> >> > Can you try 'kldload i915kms' before you suspend and see if that fixes
>> >> > your
>> >> > issue?
>> >>
>> >> 1. On the laptop with Intel graphics card, when i915kms is loaded, vt
>> >> resumes without problems.  On the laptop with the ATI graphics card
>> >> (Radeon HD 8240), I use vesa driver at the moment, and it does not
>> >> resume the screen.
>> >
>> > Let's just be clear:
>> >
>> > For laptop 1 with Intel graphics card:
>> > - syscons + text mode: resumes ok
>> > - syscons + X: resumes ok?
>> > - vt: blank screen
>> > - vt + i915kms + text mode: resumes ok
>> > - vt + X: resumes ok?
>>
>> For laptop 1 with Intel graphics card:
>> - syscons + text mode: resumes ok
>> - syscons + X: resumes ok
>> - vt: blank screen
>> - vt + i915kms + text mode: resumes ok
>> - vt + X: blank screen
>> - vt + i915kms + X: resumes ok
>
> Ok, this is similar to my little HP netbook where-in suspend/resume on the
> console without kms loaded is a regression relative to syscons.  This is not
> trivial to fix I'm afraid, though the idea would be to take the existing VESA
> code for suspend and resume from syscons and port it to vt, but in a way that
> i915kms can disable it when it is loaded.
>
>> > For laptop 2 with ATI graphics card:
>> > - syscons + text mode: resumes ok
>> > - syscons + X: resumes ok?
>> > - vt + text mode: blank screen
>> > - vt + X: ???
>>
>> For laptop 2 with ATI graphics card:
>> - syscons + text mode: resumes ok
>> - syscons + X: resumes ok
>> - vt + text mode: blank screen
>> - vt + X: blank screen
>
> I assume with X here that radeonkms is not being auto-loaded?  (i.e. you are
> using the vesa driver or some such)?

Yes, using vesa.


>> > For ATI, there is a radeonkms driver I believe.  Can you try loading that
>> > when you are using vt(4) to see if it makes a difference?
>>
>> - vt + text mode + radeonkms: blank screen
>
> Ok.  You might try e-mailing dumbbell@ about the laptop with an ATI card as he
> is maintaining the radeonkms bits.

Thank you, I'll try that.


Marek



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?86d2a29wfe.fsf>