On 11/16/09, Daniel Borkmann <[email protected]> wrote:
> Example:
>
> Within my device driver code I have the following table ...
>
> typedef struct foobar_entry {
> char foo[1024];
> char bar[1024];
> uint32_t x;
> uint32_t y;
> } foobar_entry_t;
>
> foobar_entry_t foobar_table[] =
> {
> { "foo1", "bar1", 232, 0 },
> { "foo2", "bar2", 233, 1 },
> { "foo3", "bar3", 234, 0 },
> { "foo4", "bar4", 235, 1 },
> };
> EXPORT_SYMBOL(foobar_table);
>
> ... and I'd like this to be memory mappped as read-only into user space.
> I tried things with 'remap_pfn_range', but without success. Do you have
> any idea/code example how to manage this?
Hm, what do you mean by "without success"? Do you get segmentation fault?
I guess the basic problem here is fundamental: you're accessing kernel
data from ring 3. And to be honest, I am not sure whether
remap_pfn_range would change the privilege level protection so they
are accessible from user space.
I think, you could do it by providing /proc or /debugfs entry that
read (or possible write) to this array. Robert PJ Day's Linux.com
kernel article series could give you a nice howto on this topic.
NB: Well, maybe they are not {k,v}malloc-ed, but it doesn't matter,
since I bet they are statically allocated (thus resides in .data
section).
--
regards,
Mulyadi Santosa
Freelance Linux trainer and consultant
blog: the-hydra.blogspot.com
training: mulyaditraining.blogspot.com
--
To unsubscribe from this list: send an email with
"unsubscribe kernelnewbies" to [email protected]
Please read the FAQ at http://kernelnewbies.org/FAQ
