linux-kernel - Re: Display class

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:	Thu, 7 Dec 2006 17:21:13 +0100
From:	"Miguel Ojeda" <maxextreme@...il.com>
To:	"James Simmons" <jsimmons@...radead.org>
Cc:	"Linux Kernel Mailing List" <linux-kernel@...r.kernel.org>,
	Luming.yu@...el.com, zap@...elink.ru, randy.dunlap@...cle.com,
	kernel-discuss@...dhelds.org
Subject: Re: Display class

On 12/7/06, James Simmons <jsimmons@...radead.org> wrote:
>
> >   - I would remove "struct device *dev, void *devdata" of display_device_register()
> >     Are they neccesary for other display drivers? I have to pass NULL right now.
>
> Yes. Passing in a struct device allows you a link between the device and
> the class. If you pass in the device for the parport a link to the parport
> device would exist in you displayX directory. The devdata is used by the
> probe function to get data about the monitor if it is not NULL.
>
> In the case of most desktop monitors they have EDID blocks. You can
> retrieve them (devdata) and it gets parsed thus you have detail data
> about the monitor. In your case it can be null.
>
> >   - I would add a paramtere ("char *name") to display_device_register() so we
> >     set the name when registering. Right now I have to set my name after inited,
> >     and this is a Linux module and not a person borning, right? ;)
>
> The probe function gets this for you. In your case you would have a probe
> method that would just fill in the name of the LCD. For me using the ACPI
> video driver I get the name for my monitor
>
> LEN  15"XGA 200nit
>
> Which is the manufacturer - monitor id - ascii block.
>
> >   - I would add a read/writeable attr called "rate" for set/unset the refresh rate
> >     of a display.
>
> I suggest creating a group for your driver. See device.h for
> group_attributes.
>
> >   - I was going to maintain the drivers/auxdisplay/* tree.
> >     Are you going to maintain the driver? I think so, just for being sure.
>
> Yes. I need it for the ACPI video and fbdev layer. Remember its in the
> early stages yet.
>
> > P.S.
> >
> >   When I was working at 2.6.19-rc6-mm2 it worked all fine, but now
> >   I have copied it to git7 I'm getting some weird segmentation faults
> >   (oops) when at cfag12864bfb_init, at mutex_lock() in
> >   display_device_unregister module... I think unrelated (?), but I will
> >   look for some mistake I made.
>
> Did you solve the problem?
>

I didn't look further, but I will be able to try again in some hours.
Anyway, the problem is probably at cfag12864bfb.c (as it was almost
the only file I modified from -mm2); but I can't tell where the
problem is as I tried just a few times.

Please review cfag12864bfb_init/_exit to check if your code/class is
intended to be used that way, I will focus on the segfaults.

Greets!

-- 
Miguel Ojeda
http://maxextreme.googlepages.com/index.htm
-
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/