On Thu, 22 Jul 2010, Ted Ts'o wrote:
> > The kzalloc() in start_this_handle() is failable, so remove __GFP_NOFAIL
> > from its mask.
>
> Unfortunately, while there is error handling in start_this_handle(),
> there isn't in all of the callers of start_this_handle(), which is why
> the __GFP_NOFAIL is there. At the moment, if we get an ENOMEM in the
> delayed writeback code paths, for example, it's a disaster; user data
> can get lost, as a result.
>
I'll change this to
do {
new_transaction = kzalloc(sizeof(*new_transaction),
GFP_NOFS);
} while (!new_transaction);
in the next phase when I introduce __GFP_KILLABLE (that jbd and jbd2 can't
use because they are GFP_NOFS).
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>