verity: do all hashing in dm-bht.c

verity: do all hashing in dm-bht.c

We were doing the hashing of the disk blocks in dm-verity.c and hashing
hash blocks in dm-bht.c. We can simplifiy the code by doing all the
hashing in dm-bht.c. We couldn't do this earlier because we had
to handle unaligned reads (no longer an issue).

I also removed the BUG_ON in dm_bht_get_node because I wanted to use it
in dm_bht_check_block with depth == bht->depth.

The kernel piece has already been reviewed and submitted. What is different
in this CL are the modifications to the tests.

BUG=9752
TEST=Ran dm-verity.git unit tests. Ran platform_DMVerityCorruption on H/W.
kernel.git Review URL: http://codereview.chromium.org/6626037

Change-Id: Ic817a53f383f7727742c77c2064cdb0afe595c43

R=wad@chromium.org,taysom@chromium.org,ups@chromium.org

Committed: http://chrome-svn/viewvc/chromeos?view=rev&revision=77599e8

[Mandeep Singh Baines, 2011-03-16 16:43:32]

Other than the changes to the unittest, all the code in this CL has already been
reviewed.

[Will Drewry, 2011-03-16 16:53:37]

One comment.  Sorry for the delay I clicked on the other link which was the old
review link!

http://codereview.chromium.org/6695038/diff/1/dm-bht_unittest.cc
File dm-bht_unittest.cc (right):

http://codereview.chromium.org/6695038/diff/1/dm-bht_unittest.cc#newcode222
dm-bht_unittest.cc:222: EXPECT_LT(dm_bht_verify_block(bht_.get(), 256,
kBadPage), 0);
This test no longer matches the initial test functionality.  The old one put one
bad page in the blocks, then called verifies around it to trigger failures in
intermediate nodes too.

It's be nice to either add something here or add a second unittest so we don't
lose those tests.  They caught bugs in the past! :)

[Mandeep Singh Baines, 2011-03-16 18:27:17]

Fixed. PTAL.

wad@chromium.org (wad@chromium.org) wrote:
> One comment.  Sorry for the delay I clicked on the other link which
> was the old
> review link!
> 
> 
> http://codereview.chromium.org/6695038/diff/1/dm-bht_unittest.cc
> File dm-bht_unittest.cc (right):
> 
> http://codereview.chromium.org/6695038/diff/1/dm-bht_unittest.cc#newcode222
> dm-bht_unittest.cc:222: EXPECT_LT(dm_bht_verify_block(bht_.get(), 256,
> kBadPage), 0);
> This test no longer matches the initial test functionality.  The old one
> put one bad page in the blocks, then called verifies around it to
> trigger failures in intermediate nodes too.
> 
> It's be nice to either add something here or add a second unittest so we
> don't lose those tests.  They caught bugs in the past! :)
> 

For some reason I thought I needed to change the test because the API changed.
But it was actually quite simple to modify the old test to work with the
new API. Not sure what I was thinking.

Thanks! Fixed.

> http://codereview.chromium.org/6695038/

[Will Drewry, 2011-03-16 19:32:54]

LGTM

THanks! Tests++!

http://codereview.chromium.org/6695038/diff/6001/dm-bht_unittest.cc
File dm-bht_unittest.cc (right):

http://codereview.chromium.org/6695038/diff/6001/dm-bht_unittest.cc#newcode243
dm-bht_unittest.cc:243: 0);
nit: did a tab sneak in? :)