Bug#883089: udev: on alpha architecture, /dev/disk/by-XXX nodes not created

2017-11-29 Thread Bob Tracy
Package: udev
Version: 235-3
Severity: important

Dear Maintainer,

With kernel version 4.13, udev worked correctly at least two releases ago.  The
current release (and maybe the prior one) fails to detect/enumerate SCSI disk
partitions (both real filesystems and swap partitions) except for "/" and "/usr"
which are mounted immediately upon boot by the initial ramdisk.

"journalctl -xb" has multiple messages of the form

dev-sdaX.device: Job dev-sdaX.device/start timed out.
Timed out waiting for device dev-sdaX.device.

When this problem appeared nearly three years ago, the issue was CONFIG_FHANDLE
not being set in the kernel configuration: it is currently set, and has been for
every kernel build since that time.

Possibly related issue: the USB subsystem failed to initialize as well.

In systemd's emergency mode, all the various filesystems and swap partitions
can be mounted or enabled as appropriate.  After doing this and exiting the
emergency shell with a cntl-d, the system comes up in a mostly usable state.
Networking can be manually started, as can the USB subsystem.

The situation is *much* worse with the 4.14 kernel.  "cat /proc/scsi/scsi"
in the resulting emergency shell shows only the libata-mapped IDE cdrom
device, i.e., the real SCSI host adapter and attached disk aren't seen at
all.


-- Package-specific info:

-- System Information:
Debian Release: buster/sid
  APT prefers unreleased
  APT policy: (500, 'unreleased'), (500, 'unstable')
Architecture: alpha

Kernel: Linux 4.13.0
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), 
LANGUAGE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash
Init: systemd (via /run/systemd/system)

Versions of packages udev depends on:
ii  adduser  3.116
ii  libacl1  2.2.52-3
ii  libblkid12.30.2-0.1
ii  libc6.1  2.25-1
ii  libkmod2 24-1
ii  libselinux1  2.7-2
ii  libudev1 235-3
ii  lsb-base 9.20170808
ii  procps   2:3.3.12-3
ii  util-linux   2.30.2-0.1

udev recommends no packages.

udev suggests no packages.

Versions of packages udev is related to:
ii  systemd  235-3

-- debconf information:
  udev/title/upgrade:
  udev/sysfs_deprecated_incompatibility:
  udev/reboot_needed:
  udev/new_kernel_needed: false
P: /devices/parport0
E: DEVPATH=/devices/parport0
E: DEVTYPE=parport
E: SUBSYSTEM=parport

P: /devices/pci:00/:00:03.0
E: DEVPATH=/devices/pci:00/:00:03.0
E: DRIVER=tulip
E: MODALIAS=pci:v1011d0019svsdbc02sc00i00
E: PCI_CLASS=2
E: PCI_ID=1011:0019
E: PCI_SLOT_NAME=:00:03.0
E: PCI_SUBSYS_ID=:
E: SUBSYSTEM=pci

P: /devices/pci:00/:00:03.0/net/eth0
E: DEVPATH=/devices/pci:00/:00:03.0/net/eth0
E: IFINDEX=2
E: INTERFACE=eth0
E: SUBSYSTEM=net

P: /devices/pci:00/:00:04.0
E: DEVPATH=/devices/pci:00/:00:04.0
E: DRIVER=pata_cmd64x
E: MODALIAS=pci:v1095d0646svsdbc01sc01i80
E: PCI_CLASS=10180
E: PCI_ID=1095:0646
E: PCI_SLOT_NAME=:00:04.0
E: PCI_SUBSYS_ID=:
E: SUBSYSTEM=pci

P: /devices/pci:00/:00:04.0/ata1/ata_port/ata1
E: DEVPATH=/devices/pci:00/:00:04.0/ata1/ata_port/ata1
E: SUBSYSTEM=ata_port

P: /devices/pci:00/:00:04.0/ata1/host1
E: DEVPATH=/devices/pci:00/:00:04.0/ata1/host1
E: DEVTYPE=scsi_host
E: SUBSYSTEM=scsi

P: /devices/pci:00/:00:04.0/ata1/host1/scsi_host/host1
E: DEVPATH=/devices/pci:00/:00:04.0/ata1/host1/scsi_host/host1
E: SUBSYSTEM=scsi_host

P: /devices/pci:00/:00:04.0/ata1/host1/target1:0:0
E: DEVPATH=/devices/pci:00/:00:04.0/ata1/host1/target1:0:0
E: DEVTYPE=scsi_target
E: SUBSYSTEM=scsi

P: /devices/pci:00/:00:04.0/ata1/host1/target1:0:0/1:0:0:0
E: DEVPATH=/devices/pci:00/:00:04.0/ata1/host1/target1:0:0/1:0:0:0
E: DEVTYPE=scsi_device
E: MODALIAS=scsi:t-0x05
E: SUBSYSTEM=scsi

P: /devices/pci:00/:00:04.0/ata1/host1/target1:0:0/1:0:0:0/bsg/1:0:0:0
N: bsg/1:0:0:0
E: DEVNAME=/dev/bsg/1:0:0:0
E: 
DEVPATH=/devices/pci:00/:00:04.0/ata1/host1/target1:0:0/1:0:0:0/bsg/1:0:0:0
E: MAJOR=254
E: MINOR=1
E: SUBSYSTEM=bsg

P: 
/devices/pci:00/:00:04.0/ata1/host1/target1:0:0/1:0:0:0/scsi_device/1:0:0:0
E: 
DEVPATH=/devices/pci:00/:00:04.0/ata1/host1/target1:0:0/1:0:0:0/scsi_device/1:0:0:0
E: SUBSYSTEM=scsi_device

P: /devices/pci:00/:00:04.0/ata1/link1/ata_link/link1
E: DEVPATH=/devices/pci:00/:00:04.0/ata1/link1/ata_link/link1
E: SUBSYSTEM=ata_link

P: /devices/pci:00/:00:04.0/ata1/link1/dev1.0/ata_device/dev1.0
E: DEVPATH=/devices/pci:00/:00:04.0/ata1/link1/dev1.0/ata_device/dev1.0
E: SUBSYSTEM=ata_device

P: /devices/pci:00/:00:04.0/ata1/link1/dev1.1/ata_device/dev1.1
E: DEVPATH=/devices/pci:00/:00:04.0/ata1/link1/dev1.1/ata_device/dev1.1
E: SUBSYSTEM=ata_device

P: /devices/pci:00/:00:04.0/ata2/ata_port/ata2
E: DEVPATH=/devices/pci:00/:00:04.0/ata2/ata_port/ata2
E: SUBSYSTEM=ata_port

P: 

Bug#883089: udev: on alpha architecture, /dev/disk/by-XXX nodes not created

2017-11-29 Thread Bob Tracy
On Wed, Nov 29, 2017 at 05:12:58PM +0100, Michael Biebl wrote:
> Please check all other kernel config options as well that are listed in
>  /usr/share/doc/systemd/README.gz

A few of the listed config options do not exist, at least in 4.14:
CONFIG_DMIID
CONFIG_DEVPTS_MULTIPLE_INSTANCES

There were a few namespace and crypto config settings that were
configured as modules rather than built-in, so I fixed those and am
in the process of rebuilding the 4.14 kernel.

> Please test the kernel that is provided by Debian and report back if it
> works with that one.

The Debian-supplied kernels for alpha have been problematic in the past,
to the point I've been building my own from the mainline kernel.org
source tree for several years now.

On a possibly related note, three years ago bug #652459 was opened
because systemd and udev failed to work properly with "/usr" being
a separate partition from "/".  Has anything happened with systemd
and/or initramfs development that broke the resolution of that bug?

Thanks for the assist!  I *do* appreciate your time and trouble.

--Bob

___
Pkg-systemd-maintainers mailing list
Pkg-systemd-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-systemd-maintainers


Bug#883089: udev: on alpha architecture, /dev/disk/by-XXX nodes not created

2017-12-01 Thread Bob Tracy
On Fri, Dec 01, 2017 at 06:13:01PM -0600, Bob Tracy wrote:
> Not ignoring you...  PWS 433au takes overnight to build a kernel these
> days.
> 
> Rebuilt 4.14 with the few changes noted in my previous message.  Will
> try it tomorrow when I've got a couple of free hours to collect data and
> deal with the consequences if it doesn't work.  Report to follow...

Request closure of this bug.  Effective with the 4.14 kernel, firmware
(if applicable) for built-in drivers must be included as part of the
kernel build, and my qlogic SCSI card had been relying on the firmware
being loaded from disk at boot time.  This kernel configuration change,
and possibly a few other minor configuration changes as suggested by the
"systemd" README file, got me back on track.  Particular thanks to
Michael Cree for letting me know about the probable firmware issue.

--Bob

___
Pkg-systemd-maintainers mailing list
Pkg-systemd-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-systemd-maintainers


Bug#883089: udev: on alpha architecture, /dev/disk/by-XXX nodes not created

2017-12-01 Thread Bob Tracy
On Wed, Nov 29, 2017 at 11:10:41PM -0600, Bob Tracy wrote:
> On Wed, Nov 29, 2017 at 05:12:58PM +0100, Michael Biebl wrote:
> > Please test the kernel that is provided by Debian and report back if it
> > works with that one.

Not ignoring you...  PWS 433au takes overnight to build a kernel these
days.

Rebuilt 4.14 with the few changes noted in my previous message.  Will
try it tomorrow when I've got a couple of free hours to collect data and
deal with the consequences if it doesn't work.  Report to follow...

--Bob

___
Pkg-systemd-maintainers mailing list
Pkg-systemd-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-systemd-maintainers