First I must say I appreciate your effort, at LOT!
I have booted the binary 'linwizard0.1.0' and i looks fine.
I can see clearly on the LCD, i can use the keyboard, I can even login to it using ssh via the usb-gadget!
I have used gentoo's crossdev to compile for the armv5te-linwizard-linux-gnueabi target
then I compiled the kernel with:
make distclean && make htcwizard_defconfig && make
I copied the arch/arm/boot/Image to my /mnt/usb (being my wizard mounted with WM5torage running)
I edited default.txt to look like this:
set MTYPE 1229
set KERNEL "Image"
set INITRD "initrd"
set CMDLINE "root=/dev/ram0 ramdisk_size=20504 init=/linuxrc"
bootlinux
And here is the problem:
->> When I boot 'my own' kernel, the freshly compiled one,
the LCD looks weird and after a while becomes completely unreadable...
Is there something wrong with the LCD driver, or am I doing something wrong?
Thanks for the support!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
First of all, thanks for the post and interest in Linwizard.
The git repos are in active development, and if you want something stable, your best bet is the release. That being said, the stable tag on the repo was tested against a WIZA200. We are working with the WIZA100, WIZA200 and Prophet hardware. The "stable" tag uses a completely different framebuffer from that which was used in the 0.1.0 release. We do have those issues that you're talking about fixed in a later commit, but USB is now broken. The stable tag was the last known kernel with the new framebuffer driver and working USB.
You may also want to add "The Walrus" to your repo, as a lot of developemnt work gets pushed there.
If you don't mind the USB being broken you can try the "upstream" branch which is our current effort to sync up to the kernel development mainline. Last reported working commit was 24520b15537bc3fb26775d008b40530bfe9289a6. Keep in mind that this is changing all the time.
You can also try cherry-picking the commit e42b32dfda84c8fb64e846b7a2e20d0aa80bf35d, it has the updated LCD timings that are believed to work across all the different devices.
---------------------- drivers/video/omap/lcd_htcwizard.c ----------------------
similarity index 90%
rename from drivers/video/omap/lcd_wizard.c
rename to drivers/video/omap/lcd_htcwizard.c
index cafc277..d63f1a8 100644
@@ -64,19 +64,22 @@ static unsigned long wizard_panel_get_caps(void)
}
Hi there!
First I must say I appreciate your effort, at LOT!
I have booted the binary 'linwizard0.1.0' and i looks fine.
I can see clearly on the LCD, i can use the keyboard, I can even login to it using ssh via the usb-gadget!
But now I have fetched the omap-linux repo from mvista using this HOWTO:
http://www.muru.com/linux/omap/README_OMAP_GIT
then i updated to this repo:
http://miknix.homelinux.com/linwizard/linwizard-kernel.git/
I have used gentoo's crossdev to compile for the armv5te-linwizard-linux-gnueabi target
then I compiled the kernel with:
make distclean && make htcwizard_defconfig && make
I copied the arch/arm/boot/Image to my /mnt/usb (being my wizard mounted with WM5torage running)
I edited default.txt to look like this:
set MTYPE 1229
set KERNEL "Image"
set INITRD "initrd"
set CMDLINE "root=/dev/ram0 ramdisk_size=20504 init=/linuxrc"
bootlinux
And here is the problem:
->> When I boot 'my own' kernel, the freshly compiled one,
the LCD looks weird and after a while becomes completely unreadable...
Is there something wrong with the LCD driver, or am I doing something wrong?
Thanks for the support!
First of all, thanks for the post and interest in Linwizard.
The git repos are in active development, and if you want something stable, your best bet is the release. That being said, the stable tag on the repo was tested against a WIZA200. We are working with the WIZA100, WIZA200 and Prophet hardware. The "stable" tag uses a completely different framebuffer from that which was used in the 0.1.0 release. We do have those issues that you're talking about fixed in a later commit, but USB is now broken. The stable tag was the last known kernel with the new framebuffer driver and working USB.
You may also want to add "The Walrus" to your repo, as a lot of developemnt work gets pushed there.
"git remote add walrus git://walrus.centrixrp.com/linwizard/linux-omap-2.6.git/"
If you don't mind the USB being broken you can try the "upstream" branch which is our current effort to sync up to the kernel development mainline. Last reported working commit was 24520b15537bc3fb26775d008b40530bfe9289a6. Keep in mind that this is changing all the time.
You can also try cherry-picking the commit e42b32dfda84c8fb64e846b7a2e20d0aa80bf35d, it has the updated LCD timings that are believed to work across all the different devices.
---------------------- drivers/video/omap/lcd_htcwizard.c ----------------------
similarity index 90%
rename from drivers/video/omap/lcd_wizard.c
rename to drivers/video/omap/lcd_htcwizard.c
index cafc277..d63f1a8 100644
@@ -64,19 +64,22 @@ static unsigned long wizard_panel_get_caps(void)
}
struct lcd_panel wizard_panel = {
- .name = "wizard",
- .config = OMAP_LCDC_PANEL_TFT,
+ .name = "htcwizard",
+ .config = OMAP_LCDC_PANEL_TFT |
+ OMAP_LCDC_INV_VSYNC |
+ OMAP_LCDC_INV_HSYNC |
+ OMAP_LCDC_HSVS_OPPOSITE,
.bpp = 16,
.data_lines = 16,
.x_res = 240,
.y_res = 320,
- .pixel_clock = 6250,
- .hsw = 40,
- .hfp = 40,
- .hbp = 29,
- .vsw = 1,
- .vfp = 1,
+ .pixel_clock = 5500,
+ .hsw = 10,
+ .hfp = 10,
+ .hbp = 20,
+ .vsw = 2,
+ .vfp = 2,
.vbp = 3,
.init = wizard_panel_init,
Thanks again for your interest,
Any questions you may have, let us know!
Zebediah C. McClure