Re: [PATCH 0/3] pstore: pass allocated memory region back to caller

[Kees Cook]

On Mon, Nov 14, 2011 at 3:55 PM, Tony Luck <tony.luck@xxxxxxxxx> wrote:
> On Mon, Nov 14, 2011 at 3:12 PM, Kees Cook <keescook@xxxxxxxxxxxx> wrote:
>> Since pstore's buf_lock cannot be held while doing the VFS population,
>> these patches make it the backend's responsibility to pass up an allocated
>> buffer instead during reads.
>
> The split into three parts is good for review purposes - but I'll
> squish them all together
> into one commit to avoid breaking bisectability.

Okay, sounds good.

> I'm trying to work out whether I should be worried about the removal
> of the locks
> around the:
>
>   (*open)(...)
>   while (size = (*read)(...)) {
>   }
>   (*close)()
>
> section in pstore_get_records().  If someone tries to simultaneously mount
> pstore in multiple places - will the upper level mount code serialize? I don't
> think the backends will be happy if multiple callers execute that at the same
> time.

How should it be handled? We can't hold any lock while doing the VFS work.

-Kees

-- 
Kees Cook
ChromeOS Security
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/