RE: [iPAQ] Decreasing frame rate on an h3600

Regarding the flicker:

Please update to 2.4.18-rmk3-hh6. The problem is due to a bug in the SA1110 that gives pcmcia priority over the DMA to the LCD. I added a hook so that we could control the bus timing from usermode to adjust for little or no flicker and working PCMCIA/CF cards. If you also update to the latest version of hotplug, the scripts will adjust the timing to values that seem to work pretty well. The actual hook is in

        /proc/sys/bus/pcmcia/timing_increment_ns

As it says, it adjusts the bus timing by the specified number of nanoseconds. The hotplug script that adjusts the timing is /etc/hotplug/sleeve.agent, iirc. Other values can be specified in /etc/sysconfig/hotplug.

Regarding framerate: This is not controlled by the h3600_hal, it is controlled by the framebuffer driver in drivers/video/sa1100fb.

The asset information in the eeprom on H3600/h3700 and in flash on H3800 includes the model number, serial number, language, etc.

Jamey

> -----Original Message-----
> From: Justin (Gus) Hurwitz [mailto:ghurwitz@dyndns.com]
> Sent: Monday, May 06, 2002 3:13 PM
> To: ipaq@handhelds.org
> Subject: [iPAQ] Decreasing frame rate on an h3600
>
>
>
> I've put together a wonderfully well-working system on my iPAQ, for a
> special-purpose monitoring application. It's running
> 2.4.18-rmk3-hh2, with
> Xvnc/fbvnc and dillo. Network connectivity is provided via a PCMCIA
> network card (and USB networking for development).
>
> As a few others have noted, under PCMCIA activity the screen flickers.
> Looking at previous research into this problem, it appears that a high
> frame rate for the LCD controller is a culprit (the real problem being
> caused by bus saturation). Indeed, /proc/hal/assets shows the
> frame rate
> to be a huge 70 fps. I've been trying to reduce this to see
> if that will
> solve the problem.
>
> The first thing I tried was adding a /proc/sys/hal/frame_rate
> interface. I
> added to arch/arm/mach-sa1100/h3600_hal.c: (sorry- I don't have the
> sources available to generate a diff)
>
> /*************************************************************
> **********************/
> /* Proc filesystem interface
> */
> /*************************************************************
> **********************/
>
> +static int frame_rate = 70;
>
> +static int
> +frame_rate_procctl(ctl_table *ctl, int write, struct file *filp,
> + void *buffer, size_t *lenp)
> +{
> + char buf[16];
> + int len;
> +
> + if (write) {
> + h3600_micro_eeprom_write( 82, buffer, lenp);
> + } else {
> + len = sprintf(buf, "%d\n",
> h3600_micro_eeprom_read( 82, buf, 2 ));
> + if (copy_to_user(buffer, buf, len))
> + return -EFAULT;
> + }
> + return 0;
> +}
> +
> static struct ctl_table h3600_hal_table[3] =
> {
> {3, "max_flite_brightness", &max_flite_brightness,
> sizeof(max_flite_brightness),
> 0666, NULL, &proc_dointvec},
> + {4, "frame_rate", NULL, NULL,
> + 0666, NULL, &frame_rate_procctl},
> {0}
> };
>
>
> Now, this didn't work. It didn't work by a long shot. I've
> played around
> with a few different ways of passing data to
> h3600_micro_eeprom_write(),
> but have had no success. The Good News(TM) is that most of the data
> reported by /proc/hal/assets is now garbage. Yay. So, my
> first question
> is: does h3600_micro_eeprom_write() work? If so, how? It's not used
> anywhere in the kernel, as far as I can tell, so I've no
> refrence for it-
> as far as I know, it might have never been tested.
>
> Next question is- is this a Right Way(TM) to do this, or can
> anyone think
> of a bettwe way to update the frame rate. I assume that I
> need to reboot
> and/or reset the LCD controller (which being integrated will probably
> require a reboot, anyhow, eh?) for this change to take effect.
>
> And, my final question is: can someone send me a copy of their
> /proc/hal/assets file :) Once I get writes to the eeprom
> working again,
> I'd really like to make the data in the eeprom right again. I
> just hope
> that I didn't overwrite anything I didn't want to.
>
> Thanks,
> --Gus
>
> PS- I should probably say what happens when I try to write to
> the eeprom
> with the above patch:
> sh-2.05# echo "30" > /proc/sys/hal/frame_rate
> [sits here endlessly, or at least for 5+minutes. During this time, I
> cannot access the eeprom at all- cat /proc/hal/assets on
> another terminal
> just sits there until I kill the above echo command.]
>
> --
> Justin (Gus) Hurwitz
> ghurwitz@dyndns.com
> ghurwitz@lanl.gov
>
>
> _______________________________________________
> Open Source Software for Compaq iPAQ Handhelds.
> iPAQ mailing list
> iPAQ@handhelds.org
> http://handhelds.org/mailman/listinfo/ipaq
> irc://irc.openprojects.net #ipaq
>
Received on Tue May 07 12:23:11 2002