touch-packages team mailing list archive

[david6 <1333140@xxxxxxxxxxxxxxxxxx>]

I have done repeated installs of 14.04 LTS and 15.04 on an HP Stream 11
notebooks.

This issue generally does NOT occur, other than briefly during initial
load of Ubuntu installer (and sometimes during early stages of install).
This adds no more than 15% to the time it takes to install Ubuntu (with
eMMC storage), and will usually take less than 30 minutes.

However, I realised recently that this was ONLY true for new machines,
or those I had already installed Ubuntu on. Out of the box, I was
interrupting initial startup (at screen backlight on) to go directly to
(BIOS controls for) booting from USB device.

For machines that had were already running Windows 8.1 (or even just
asking for region / user details), it will no longer cleanly install
Ubuntu. I have had it take from 2 hours to 5-7 hours, or just stall
(after about 40 minutes) with continuous errors. This appears to be
spurious interrupts (which cause further errors) due to attempts to
access the RPMB partition. I am NOT even convinced that the Ubuntu
installer is causing these interrupts.

The other possible 'root cause' is that the problem machines (NEW or
used) were older, the same model Notebook but with an earlier generation
/ step release of eMMC chips. I don't have a large enough pool to prove
this, but someone may be able to shed light.

-- 
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.
https://bugs.launchpad.net/bugs/1333140

Title:
  Fix udev rules to consider mmc rpmb partitions

Status in systemd package in Ubuntu:
  Fix Released
Status in udev package in Ubuntu:
  Invalid
Status in udisks package in Ubuntu:
  Triaged
Status in udev source package in Precise:
  Triaged
Status in udisks source package in Precise:
  Confirmed
Status in systemd source package in Trusty:
  Fix Released
Status in udisks source package in Trusty:
  Confirmed

Bug description:
  As per JEDEC 4.5 spec for eMMC devices,
  There is a new partitions as part of eMMC storage devices it self. (Further details please refer eMMC spec)

  *In Linux Kernel@ 3.10.33, mmc driver has created a new partitions
  with "mmcblkXrpmb" if device expresses it support of RPMB.

  Issues observed:

  issue 1:
  RPMB (Replay Protected Memory Block), A signed access to a Replay Protected Memory Block is provided. This function provides means for the system to store data to the specific memory area in an authenticated and replay protected manner.
  In that case, any read/write access to this partition device will report errors.

  issue 2:
  The by-path, line is wrongly mapping to platform-sdhci-tegra.1  -> mmcblk2rpmb were as
  it should be platform-sdhci-tegra.1  -> mmcblk2

  ls -l /dev/disk/by-path/
  total 0
  lrwxrwxrwx 1 root root 17 Jan  3  2000 platform-sdhci-tegra.1 -> ../../mmcblk2rpmb
  lrwxrwxrwx 1 root root 15 Jan  3  2000 platform-sdhci-tegra.1-part1 -> ../../mmcblk2p1
  lrwxrwxrwx 1 root root 13 Jan  3  2000 platform-sdhci-tegra.2 -> ../../mmcblk1
  lrwxrwxrwx 1 root root 15 Jan  3  2000 platform-sdhci-tegra.2-part1 -> ../../mmcblk1p1
  lrwxrwxrwx 1 root root 17 Jan  3  2000 platform-sdhci-tegra.3 -> ../../mmcblk0rpmb
  lrwxrwxrwx 1 root root 15 Jan  3  2000 platform-sdhci-tegra.3-part1 -> ../../mmcblk0p1
  lrwxrwxrwx 1 root root 15 Jan  3  2000 platform-sdhci-tegra.3-part2 -> ../../mmcblk0p2

  We have locally resolved in our platform in this file 60-persistent-
  storage.rules

  For issue 1: (with this rule)
  # skip block read for partitions of type rpmb
  KERNEL=="mmcblk[0-9]rpmb", SUBSYSTEM=="block", GOTO="persistent_storage_end"

  For issue 2:
  ENV{DEVTYPE}=="disk", ENV{ID_PATH}=="?*", KERNEL=="mmcblk[0-9]rpmb", SYMLINK+="disk/by-path/$env{ID_PATH}-rpmb"
  ENV{DEVTYPE}=="disk", ENV{ID_PATH}=="?*", KERNEL!="mmcblk[0-9]rpmb", SYMLINK+="disk/by-path/$env{ID_PATH}"

  Please consider this issues fix in next udev release .

  
  SRU INFO: This is mostly an issue with running backported kernels on 12.04 and 14.04. There is no test case which would reproduce this on arbitrary hardware, but there are several reporters which are in a position to verify a proposed update.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1333140/+subscriptions