One of the main goals of the IPC Logging is to allow extracting logs
from memory dumps. The current implementation has the following
limitations:
1) if logs are being dumped through debugfs at the time of the crash,
it is not possible to extract logs that were already dumped
2) it is not always possible to tell the difference between empty,
partially full, and full log pages resulting in extraction
complications
3) if the system memory is not cleared between reboots, then the same
log ID and log page may be extracted leading to duplicate log
pages
Add additional debugfs read index, timestamps for log pages, and change
the read/write indices to easily differentiate between empty, partially
full, and full log pages.
CRs-Fixed: 670345
Change-Id: I695f6e0605f3755b868ede59851f5ca919f8fb0d
Signed-off-by: Eric Holmberg <eholmber@codeaurora.org>
If clients use custom serialization functions, then they may need to
define a version for deserialization support for log extraction.
Add client version support.
Change-Id: Id135f06d4142de39275b5d0caab88708d5496b5e
Signed-off-by: Eric Holmberg <eholmber@codeaurora.org>
Add IPC Logging documentation that was missed when the initial
driver was added.
Change-Id: Ib3caffb4da98aa453f72d65643ab9c61d7774434
Signed-off-by: Eric Holmberg <eholmber@codeaurora.org>