Am 02/09/2014 19:36, schrieb Alexander G. M. Smith:
On 02/09/14 12:25 PM, Earl Pottinger (Redacted sender earl_colby_pottinger@xxxxxxxxx for DMARC) wrote:But why would Haiku write to the SuperBlock *AFTER* sending its final cache-flush during shut-down operations?BeOS compatibility? :-) Actually, BeOS was similarly weird during shutdown, as I found out when debugging my old block based RAM disk device driver.
No, the reason is much simpler: the file system is a module of its own. The kernel doesn't have absolute control over it.
Anyway, after the super block gets written, there should definitely be another cache flush of the underlying device. If that doesn't happen, there is likely to be something wrong.
Bye, Axel.