Re: [Qemu-devel] [PATCH v4.1] Add optionrom compatible with fw_cfg DMA version

[Laszlo Ersek <lersek@redhat.com>]

On 04/04/16 17:21, Richard W.M. Jones wrote:
> On Mon, Apr 04, 2016 at 04:02:04PM +0100, Stefan Hajnoczi wrote:
>>   (1) initrd loading is broken, kernel complains it finds only gibberish:
>>
>>   [    0.934582] Unpacking initramfs...
>>   [    1.166983] Initramfs unpacking failed: junk in compressed archive
>>   [    1.168458] Freeing initrd memory: 32812k freed
> 
> That's strange.  I certainly never saw anything like this.  I wonder
> if it's because your initrd is particularly large?

If you search your mailbox for Message-Id <54C2D2C4.3010806@redhat.com>
(subject: "Test libguestfs on RHELSA"), you'll find an earlier occurence
of the same.

(You had also replied to that email, so you simply didn't remember it
this time.)

In that case, the problem was an oversized initrd, relative to the
amount of RAM that the guest had. You eliminated (worked around?) the
issue in libguestfs commit c24f242521e88.

Of course, without reviewing the patch under discussion (which I won't
volunteer for now, apologies), I cannot exclude that the patch makes the
kernel *think* that the initrd is too large.

>>   (2) going back to non-dma boot via -M pc-$old doesn't work, appearently
>>       fw_cfg dma is enabled even for old machine types.
> 
> IIRC there was a thread about how we accidentally added DMA to old
> machine types.  Does this matter?

What Gerd said -- see e6915b5f3a87.

(I have so few commits in QEMU that if I look for something I wrote, I
can find it simply with "--author=lersek". I don't know if I should
smile or cry...)

Thanks
Laszlo