#
45274420 |
|
16-May-2018 |
Sean Klein <smklein@google.com> |
[blobfs] Use size_t instead of uint64_t for bitmap API Change-Id: I70cb9a1d8941abe56faaec72e2bfeb34769a3240
|
#
7c243942 |
|
16-May-2018 |
Sean Klein <smklein@google.com> |
[blobfs][fsck] Validate that all blob blocks are allocated Additionally, identify the blocks that are not allocated if any are missing. Change-Id: I2e2a3689e4f46b9b9b4afd2f1d81e4889300a558
|
#
50908ceb |
|
16-May-2018 |
Tricia Landers <planders@google.com> |
[blobfs] Remove usage of %lu Change-Id: Ibfd9bd63f0f70eba639bbf573ed3c2f4fc412b92
|
#
dd1523b0 |
|
16-May-2018 |
Tricia Landers <planders@google.com> |
[fvm] Add host-side tool option to decompress sparse file. Change-Id: I76bac360f304b8bbde1f1054983f360d3b1a8829
|
#
3b972eed |
|
18-Apr-2018 |
Sean Klein <smklein@google.com> |
[minfs][blobfs] Re-implement filesystem teardown Rather than simply "sync + exit process", implement a fully featured filesystem teardown procedure, capable of: - Tearing down all active connections. - Syncing underlying media using a non-blocking dispatch loop. - Use closures to signal both mounters and unmounters on filesystem destruction. As a consequence of the controlled filesystem teardown procedure, make the "filesystem" objects in minfs and blobfs no longer refcounted, to clarify that "Shutdown" is the single point of destruction after mounting. ZX-1577 #done ZX-1923 #done ZX-1959 #done ZX-1576 #comment In Progress Change-Id: I3f8d71af76e391c41c665434baf86a245a7cd223
|
#
0e5dbc4e |
|
17-Apr-2018 |
Tricia Landers <planders@google.com> |
[blobfs] [host] Copy inode on VerifyBlob. Reading the blocks belonging to an inode overwrites the blobfs' block cache, so the previous inode data is erased. This patch modifies VerifyBlob to keep a local copy of the inode data. ZX-2017 #done Change-Id: I65875f1ac7e515213567737baf293a35af25ee2d
|
#
60d96f82 |
|
30-Jan-2018 |
Tricia Landers <planders@google.com> |
[blobstore] Track reserved blocks and nodes When a blob is first created, we allocate a node and blocks in memory to write the blob contents to. The allocated resources should not be persisted to disk until the client has finished writing all data to the blob. Previously there existed the possibility that upon completion of another blob, allocated nodes/blocks of blobs still being written would also be persisted to disk. Upon successful cleanup of the blobs, these blocks would be released; however, in the event of a disk failure, the disk would be left in an inconsistent state. This change allows reserved nodes/blocks for blobs that are not yet completed to be stored safely in memory, without potential compromise of the integrity of on-disk structures. ZX-1898 #done Change-Id: Ic424951154e0b145b498aa777aa8f2ccc0a55d54
|
#
c99c8c07 |
|
04-Mar-2018 |
Adam Barth <abarth@chromium.org> |
[blobstore] Rename blobstore to blobfs Also, mount blobfs at /blob rather than /blobstore. Change-Id: Ifdc473752ede37b909b30e35dbaac1397565a659
|