linux-kernel - Re: [PATCH] max3100 driver

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Sat, 20 Sep 2008 12:35:26 +0200
From:	chri <chripell@...il.com>
To:	"Andrew Morton" <akpm@...ux-foundation.org>
Cc:	linux-kernel@...r.kernel.org, linux-serial@...r.kernel.org
Subject: Re: [PATCH] max3100 driver

Sorry, sent HTML mail by mistake, so resending :-/

On Sat, Sep 20, 2008 at 10:24 AM, Andrew Morton
<akpm@...ux-foundation.org> wrote:

> > +#define MAX3100_MAJOR 204
>
> Allocating a new major is a Big Deal.  It involves getting the major
> registered by contacting device@...ana.org.
>
> It's better to dynamically allocate it - let udev handle it.
>

I looked at other serial driver as an example and checked devices.txt:
if I don't get it wrong major 204 should be already reserved for
serial port. Anyway I choose a minor number already allocated by
mistake (did not see the "...") and will correct that. Is this ok or I
*have to* move to dynamic major (it's a bit a nuisance since max3100
is used in embedded system where udev is not always used)?

> `struct max3100_port' is sufficient, and would be more typical.
>
> > +     struct uart_port port;
> > +     struct spi_device *spi;
> > +
> > +     int cts:1;              /* last CTS received for flow ctrl */
> > +     int tx_empty:1;         /* last TX empty bit */
>
> These two bits will share a word and hence locking is needed to prevent
> modifications to one from trashing modifications to the other on SMP.
>
> That's OK, but it would be best to document that locking right here, and
> to check that it is adhered to.
>

I did not realize this until you explained me. I'm not sure if actual
packing of bit-fields is implementation dependent but I think so. If
this is right I guess it's better to avoid bit-fields in structs that
can be accessed concurrently (or otherwise I have to lock the entire
struct). So, should I avoid bit-fields altogether?

I will correct the patch and resend.

Thanks,

--
Christian Pellegrin, see http://www.evolware.org/chri/
"Real Programmers don't play tennis, or any other sport which requires
you to change clothes. Mountain climbing is OK, and Real Programmers
wear their climbing boots to work in case a mountain should suddenly
spring up in the middle of the computer room."
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/