tree ad03c769feff22925d3ddcf2b12ee2f5e37aef73
parent b5ba797ba9ee97130d56f1457ba239721d5b8177
author Brian Norris <briannorris@chromium.org> 1561568615 -0700
committer Commit Bot <commit-bot@chromium.org> 1563596898 +0000

update_engine: drop MTD logic

It has no users.

Note that I'm also dropping unit tests of the form:

  EXPECT_TRUE(utils::SplitPartitionName("/dev/loop10_0", &disk, &part_num));
  EXPECT_EQ("/dev/loop", disk);
  EXPECT_EQ(10, part_num);

  EXPECT_TRUE(utils::SplitPartitionName("/dev/loop28p11_0", &disk, &part_num));
  EXPECT_EQ("/dev/loop28", disk);
  EXPECT_EQ(11, part_num);

AFAICT, the part of the change that introduced these
(https://crrev.com/c/191785) was not based on any real issue; it was
(correctly) handling partition suffixes for loop devices (e.g., the 'p1'
in '/dev/loop0p1'), but the underscore syntax is specific to
ubi/ubiblock and should not apply to loop devices.

See the ubiblock naming code [1], analogous loop device naming code [2],
and the partition-name generation code [3].

[1] https://elixir.bootlin.com/linux/v5.1.15/source/drivers/mtd/ubi/block.c#L402
[2] https://elixir.bootlin.com/linux/v5.1.15/source/drivers/block/loop.c#L2012
[3] https://elixir.bootlin.com/linux/v5.1.15/source/block/partition-generic.c#L35

BUG=chromium:978563
TEST=unit tests

Change-Id: I38754a5060ed3c9e6b11fb53d82ff6fb79149c72