Thread (3 messages) 3 messages, 2 authors, 2011-08-23

Re: attempt to fix race in bigalloc code path

From: Aditya Kali <hidden>
Date: 2011-08-22 17:30:23

On Mon, Aug 22, 2011 at 6:54 AM, Yongqiang Yang [off-list ref] wrote:
Hi,

I can not find the patch - attempt to fix race in bigalloc code path -
in the mailing list.

I don't think this patch can work.  set_buffers_da_mapped() is called
under i_data_sem, so it can not lock page.  If the page is not locked,
calling page_has_buffers() is not secure as Jan pointed out
previously.  If a page is partially mapped(block size < page size),
setting BH_Da_Mapped without locking page is not safe too.
The BH_Da_Mapped is set only during page writeback when the page is
locked (in write_cache_pages_da). So, the setting of BH_Da_Mapped
should be safe, no ?
BTW:  find_delalloc_range() does not lock page as ext4_fiemap_cb().
As Jan said, it can crash system.  BUT we can not lock page in both
scenarios, because i_data_sem is hold. So delayed extent list is
necessary.  I have finished the code implementing delayed extent list,
which needs some further testing.


What are your opinions?
I agree that having a delayed extent list will greatly simply this
solution in this case.
--
Best Wishes
Yongqiang Yang
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help