Xen project Mailing List

Before trying to start from scratch and fixing another issue along the way, we hit, and now back at just One of this previously mentioned bugs.

The first problem that stopped the Xen from working straight from the EFI option is now working, and xen is stuck at the same "second issue":

<> "
<> [global]
<> default=xen>
<>
<> [xen]
<> options=console=vga loglvl=all noreboot
<>
<> kernel=vmlinuz-3.19.0-9-generic ignore_loglevel
<
<No root= etc? I'd expect this to include a bunch more stuff, which is
<mostly the same as your native boot.

I am still using these options, to boot, and they are based mostly on the http://xenbits.xenproject.org/docs/unstable/misc/efi.html page with a few diferences regarding console output to COM ports, what seems like a comment and the kernel version being used.

<> I have also installed the xen entry with:
<> "
<>
<> efibootmgr -c -d /dev/sdf -p 1 -L "Xen" -l "xen-4.5.1-pre.efi"
<> "
<[...]
<> After looking around the internet I've found the rEFInd package with
<> it's PPA, which after installation shows the xen-4.5.1-pre.efi and for
<> this first problem for now, it is an acceptable workaround. Anyway to
<> FIX THIS???
<
<http://wiki.xenproject.org/wiki/Xen_EFI gives a slightly different set
<of options to the efibootmgr command than you do, I don't knwo if they
<are meaningful.

The -c Creates a efi entry, the -d sets the device to my /dev/sdf drive in which the system is being installed, the -p 1, sets the efi to point the EFI partition of the device, -L sets the Label to describe this booting option and the -l sets the path to the file itself within the EFI ESP partition in use,

BUT, since this is actually working and the problem is another, I would probably point that somehow it whas a buggy installation that was fixed during the install from scratch.

<Also it implies that things are in a subdirectory of your EFI ESP, I
<also don't know if that is relevant either, but it might be worth
<reducing the differences.
<
<Where did you place the various files? What filesystem is present there?
<
<How does reFInd describe the Xen boot entry?

Quoting the first e-mail from the "Make World Error Code 2", my partitions are:
"
/dev/sdf1 - fat32 - /boot/efi - 512.00 MB
/dev/sdf2 - ext2 - /boot/ - 244.00 MB
/dev/sdf3 - lvm2 pv - ubuntu-gnome-vg - 465.02 GB

Within that structure in this new install of xen, the xen-4.5.1-pre.efi and .cfg are located inside /boot/efi/EFI/xen/, I have also installed the rEFInd tool just in case it is needed, but since this first issue seems to have been solved I ain't posting it's .cfg file below.

<> Second Problem: After booting with the xen-4.5.1-pre.efi from REFIND
<> Menu entry, the screen goes on loading drivers and everything goes
<> right until these lines from dmesg (without timestamp, and skipping
<> some entries:
<> "
<>
<> ata10.00: failed to IDENTIFY (I/O error, err_mask=0x4)
<> ...
<> ata10: Link is slow to respond, please be patient (ready=0)
<> ata10: COMRESET failed (errno=-16)
<> ...
<> ata10: limiting SATA link speed to 3.0 Gbps
<> ata10: COMRESET failed (errno=-16)
<> ata10: reset failed, giving up
<
<Do these happen on native boot too? Is ata10 the controller which
<connects to your rootfs?

Yes, the ata10 is the controller to my rootfs and, no, they don't happen on my native boot, the ata10 get it's ID correctly, and they don't fail with those COMRESET errors, from dmesg:
"
...
[ÂÂÂ 1.285763] ata10.00: failed to get NCQ Send/Recv Log Emask 0x1
[ÂÂÂ 1.285765] ata10.00: ATA-9: Samsung SSD 840 Series, DXT09B0Q, max UDMA/133
[ÂÂÂ 1.285765] ata10.00: 976773168 sectors, multi 16: LBA48 NCQ (depth 31/32), AA
[ÂÂÂ 1.293764] ata10.00: failed to get NCQ Send/Recv Log Emask 0x1
[ÂÂÂ 1.293772] ata10.00: configured for UDMA/133
...
[ÂÂÂ 1.641670] scsi 9:0:0:0: Direct-AccessÂÂÂÂ ATAÂÂÂÂÂ Samsung SSD 840Â 9B0Q PQ: 0 ANSI: 5
[ÂÂÂ 1.641944] ata10.00: Enabling discard_zeroes_data
[ÂÂÂ 1.641960] sd 9:0:0:0: [sdf] 976773168 512-byte logical blocks: (500 GB/465 GiB)
[ÂÂÂ 1.641978] sd 9:0:0:0: Attached scsi generic sg5 type 0
[ÂÂÂ 1.642051] sd 9:0:0:0: [sdf] Write Protect is off
[ÂÂÂ 1.642053] sd 9:0:0:0: [sdf] Mode Sense: 00 3a 00 00
[ÂÂÂ 1.642129] sd 9:0:0:0: [sdf] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[ÂÂÂ 1.642189] scsi 15:0:0:0: ProcessorÂÂÂÂÂÂÂÂ MarvellÂ 91xx ConfigÂÂÂÂÂ 1.01 PQ: 0 ANSI: 5
[ÂÂÂ 1.642230] ata10.00: Enabling discard_zeroes_data
[ÂÂÂ 1.644823]Â sdf: sdf1 sdf2 sdf3
[ÂÂÂ 1.644969] ata10.00: Enabling discard_zeroes_data
[ÂÂÂ 1.645080] sd 9:0:0:0: [sdf] Attached SCSI disk
...
"

Any idea how to fix this issue?

2015-03-24 10:33 GMT-03:00 Ian Campbell <ian.campbell@xxxxxxxxxx>:

On Sat, 2015-03-21 at 00:06 -0300, Carlos Gustavo Ramirez Rodriguez
wrote:
> Continuing my Installation procedure from last "curses library e-mail"
> that Ian Campbell, so kindly answered, I've got two separate points
> that I got stuck into:
>
> First Problem:
>
>
> After compiling xen-4.5.1-pre.efi and creating the xen-4.5.1-pre.cfg:
> "
>
> [global]
>
> default=xen
>
>
> [xen]
> options=console=vga loglvl=all noreboot
>
> kernel=vmlinuz-3.19.0-9-generic ignore_loglevel

No root= etc? I'd expect this to include a bunch more stuff, which is
mostly the same as your native boot.

> ramdisk=initrd.img-3.19.0-9-generic
> "
>
> I have also installed the xen entry with:
> "
>
> efibootmgr -c -d /dev/sdf -p 1 -L "Xen" -l "xen-4.5.1-pre.efi"
> "
[...]
> After looking around the internet I've found the rEFInd package with
> it's PPA, which after installation shows the xen-4.5.1-pre.efi and for
> this first problem for now, it is an acceptable workaround. Anyway to
> FIX THIS???

http://wiki.xenproject.org/wiki/Xen_EFI gives a slightly different set
of options to the efibootmgr command than you do, I don't knwo if they
are meaningful.

Also it implies that things are in a subdirectory of your EFI ESP, I
also don't know if that is relevant either, but it might be worth
reducing the differences.

Where did you place the various files? What filesystem is present there?

How does reFInd describe the Xen boot entry?

> Second Problem: After booting with the xen-4.5.1-pre.efi from REFIND
> Menu entry, the screen goes on loading drivers and everything goes
> right until these lines from dmesg (without timestamp, and skipping
> some entries:
> "
>
> ata10.00: failed to IDENTIFY (I/O error, err_mask=0x4)
> ...
> ata10: Link is slow to respond, please be patient (ready=0)
> ata10: COMRESET failed (errno=-16)
> ...
> ata10: limiting SATA link speed to 3.0 Gbps
> ata10: COMRESET failed (errno=-16)
> ata10: reset failed, giving up

Do these happen on native boot too? Is ata10 the controller which
connects to your rootfs?

>
> mount: mounting /sys on /root/sys failed: No such file or directory
> mount: mounting /proc on /root/proc failed: No such file or directory
> Target filesystem doesn't have requested /sbin/init.
> No init found. Try passing init= bootarg.

This is the sort of message I would associate with a lack of root= on
your kernel command line (which is why I didn't immediately jump on the
ata errors as the cause).

> "
>
> After this point a am dropped in a initramfs shell.
>
>
> From this perspective, I've started from Google and read that it could
> be possibly a controller-hd related problem, but that doesn't make any
> sense since I can boot from UBUNTU normally even after rEFInd, so the
> only conclusion is that there is a problem while loading the SSD,
> which have already been loaded from using rEFInd, from inside the Xen
> Hypervisor or some other bug that I can't think somehow related to
> loading my SSD on /dev/sdf, and after that nothing else can be mounted
> correctly.

Do you have two disks (HDD and SSD)? Which one contains the ESP?

> IF SOMEONE CAN HELP ME TO FIX THESE OR EVEN DEBUG IT, I WOULD BE VERY
> GRATEFUL,

No need to shout.

>
>
> Thankfull,
>
> Gustavo
>
>
>
>
>
>
> --
> AbraÃos
> Gustavo
> _______________________________________________
> Xen-users mailing list
> Xen-users@xxxxxxxxxxxxx
> http://lists.xen.org/xen-users

Gustavo

Re: [Xen-users] UEFI Booting