Konstantin Belousov <> changed:

           What    |Removed                     |Added
                 CC|                            |

--- Comment #8 from Konstantin Belousov <> ---
(In reply to Fabian Keil from comment #7)
I think the only needed part of the change is the move of n calculation after
the error check.

In head, cluster_read() and bread() relibably reset bp to NULL on error, _and_
brelse ignores NULL argument.  So it may be argued that the fix also could
remobe brelse() call from the error path.

You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________ mailing list
To unsubscribe, send any mail to ""

Reply via email to