Hello, Stephen,
It seem that in the presence of several jornalling file systems in the
kernel some mechanism is needed for proper sharing of the ->journal_info
pointer in the task_struct.
One journalled file system can indirectly call another during page fault
in copy_from_user, quota file modification, or
__alloc_pages()->balance_classzone()->shrink_cache()->vm_writeback()
sequence. (Actually, I don't understand when prevents ext3 from
re-entering itself in such a case.)
Can we replace
struct task_struct {
...
void *journal_info;
};
with
struct fs_activation;
struct fs_activation {
struct super_block *super;
struct fs_activation *parent;
};
struct task_struct {
...
struct fs_activation *journal_info;
};
Such that each file system can store original ->journal_info in the
->parent field and restore it on exit.
All fs_activation for nested file system calls are then chained through
->parent's and file system can easily check whether it is already in the
call chain, etc.
Of course, fs_activation will be actually embedded into file system
specific data structure (handle_t for ext3, reiser4_context for reiser4).
Does this make sense?
Nikita.
_______________________________________________
Ext3-users@redhat.com
https://listman.redhat.com/mailman/listinfo/ext3-users
