Messages

16

Zaurus - Void Linux support / Re: who is working on the Zaurus Kernel?

« on: May 14, 2025, 08:42:41 am »

p.s.
wrote a nice tool to automagically get stuff from an internet url, even recursively

Code: [Select]

macmini2-intel irc # echo "https://logs.nslu2-linux.org/livelogs/kexecboot/" > url
macmini2-intel irc # myNET-get-files-from-url url
list_preparing1, getting index.html ... success
preparing /list ... done
downloading [kexecboot_20110601.txt] ... success
downloading [kexecboot_20110602.txt] ... success
downloading [kexecboot_20110603.txt] ... success
downloading [kexecboot_20110604.txt] ... success
downloading [kexecboot_20110605.txt] ... success
downloading [kexecboot_20110606.txt] ... success
downloading [kexecboot_20110607.txt] ... success
downloading [kexecboot_20110608.txt] ... success
downloading [kexecboot_20110609.txt] ... success
downloading [kexecboot_20110610.txt] ... success
downloading [kexecboot_20110611.txt] ... success
downloading [kexecboot_20110612.txt] ... success
downloading [kexecboot_20110613.txt] ... success
downloading [kexecboot_20110614.txt] ... success
downloading [kexecboot_20110615.txt] ... success
downloading [kexecboot_20110616.txt] ... success
downloading [kexecboot_20110617.txt] ... success
downloading [kexecboot_20110618.txt] ... success
downloading [kexecboot_20110619.txt] ... success

I take this opportunity to test it 

17

Zaurus - Void Linux support / Re: who is working on the Zaurus Kernel?

« on: May 14, 2025, 08:39:15 am »

thanks!

I am working on a low level bootloader, plus kernel stuff.
Need info, I can't find around.

I will wget and search into logs

18

Zaurus - Void Linux support / who is working on the Zaurus Kernel?

« on: May 12, 2025, 12:50:37 pm »

hi
is there a mailing list or something to discuss the kernel development? patch? etc
is there still interest?

let me know.
D.

19

For Sale / Wanted / Wanted: TRENDnet TE-CF100, CompactFlash 10/100 Mbps Fast-Ethernet

« on: May 10, 2025, 05:29:54 pm »

HI
I need it for the development of Zaurus C3K
Let me know if you have one for sale.
Thanks

20

Zaurus - Arch Linux ARM support / Re: Update: C1000 / Akita support

« on: May 02, 2025, 02:03:51 pm »

Another thing: I know Xscale can "scale" the CPU frequency.
Can the Linux Kernel reduce the clock frequency at run-time?
It is useful to extend the battery usage time

if I use the PDA to write with { GNU/Nano, VIM, ... }
I do not need 400Mhz, I can very well work at 200Mhz

21

Zaurus - Arch Linux ARM support / Re: Update: C1000 / Akita support

« on: May 02, 2025, 02:03:21 pm »

I would like to comment on the lines of kernel code that activate the suspension when the power/on button is pressed.
to prevent the PDA/Akita from crashing if accidentally pressed

Code: [Select]

disable apm_queue_event() on key_suspend
--- drivers/input/apm-power.c.original  2025-05-02 17:58:24.140000000 -0000
+++ drivers/input/apm-power.c   2025-05-02 17:58:23.788000000 -0000
@@ -20,7 +20,7 @@
 {
        switch (keycode) {
        case KEY_SUSPEND:
-               apm_queue_event(APM_USER_SUSPEND);
+               // apm_queue_event(APM_USER_SUSPEND);
                pr_info("Requesting system suspend...\n");
                break;
        default:


22

Zaurus - Arch Linux ARM support / Re: Update: C1000 / Akita support

« on: May 02, 2025, 02:02:32 pm »

Code: [Select]

macmini2-intel kernel-5.4.229-arm-akita # ./mybuild-2024 2
[step2] compiling
kernel-5.4.229(arm/zaurus-akita)
| vanilla + patches from Void
| status:
|  boots(from SD)
|  boots(from CF-microdrive)
|  usb-lan DM9601 works
|  suspend
|    failure on c1k
|    success on c3k
|  irda/raw serial, added, to be tested
toolchain(armv5tel-softfloat-linux-gnueabi:2.40.0/12)
-----------------------------------------------
checking myhost as qualified_host ... success
checking .config ... success
checking blobs ... success
-----------------------------------------------
gadget_to_do(build) ... overlay/build
cleaning ............... done
gadget_to_do(clean) ... kernel/clean
building kernel ... success
-----------------------------------------------
gadget_to_do(binary) ... overlay/binary
gadget_to_do(sweep) ... kernel/sweep
app_body    =kernel.zImage
app_size    =5003576 byte
app_size_max=6000000 byte
checking size ... passed
gadget_to_do(binary_report) ... overlay/binary_report
macmini2-intel kernel-5.4.229-arm-akita #

that's how I compile kernels for all the platforms I support, mainly MIPS, POWERPC and HPPA2.
I wrote a "builder" in bash-scripting + C/89, to help managing patches, configuration, and checks

Every build comes with a "profile", which tells about

• the toolchain, with constraints on the min and max versions of binutils and gcc
• the ecosystem (it's called "qualified machines", can also be a chroot, or a container), to ensure you are using the right setup, at least "tested"
• the final size of the kernel
• the final file-format { uboot, elf, bin, zImage, ... }
• etc...

Code: [Select]

checking .config ... success

Here, it checks a file in ensures that the .config contains options considered "essential", and that it does not contain options considered "to be avoided"

Code: [Select]

macmini2-intel kernel-5.4.229-arm-akita # cat profile/config_check.list
CONFIG_PANIC_TIMEOUT                 # without this, panic() loops forever

CONFIG_USB_NET_DM9601                # Davicom DM9601 USBv1.1 10/100 Ethernet NIC

not CONFIG_CMDLINE_FROM_BOOTLOADER     # do not allow the bootloader to pass cmdline
CONFIG_CMDLINE_FORCE                 # built-in cmdline
CONFIG_CMDLINE                       #
                                     # console=ttyS0,9600n8 console=tty1
                                     # noinitrd
                                     # root=/dev/hda2    /dev/sda2
                                     # init=/bin/bash
                                     # rootfstype=xfs ro
                                     # video=pxafb,mode:640x480-2
                                     # fbcon=rotate:1
                                     # (debug)
CONFIG_EXT2_FS
CONFIG_EXT3_FS
CONFIG_XFS_FS

CONFIG_IKCONFIG                      # save .config into /proc/config.gz

CONFIG_GPIO_PCA953X                  # set the GPIOs in Akita
....


23

Zaurus - Arch Linux ARM support / Re: kexecboot-size, max-size?

« on: May 02, 2025, 10:49:16 am »

If there was a documented jtag port, I would like to erase the entire flash, and reprogram it from scratch, in a more intelligent way.

strange that no one has ever thought of this ...

... another alternative, physically unsolder the flash chip, reprogram it in the lab, re-solder it.

we'll see.

24

Zaurus - Arch Linux ARM support / Re: kexecboot-size, max-size?

« on: May 02, 2025, 10:31:07 am »

Ok, so it's 1.2Mbyte max.

uboot ? yes, you can build a uboot to fit as a replacement for kexecboot kernel, but you have to port it yourself pxa support in uboot is not that recent I believe for pxa270 or any pxa.
kexecboot was a sensible choice back then...and still is

modern kernel (v5, v6) just don't fit into 1.2Mbyte
uboot is much smaller in size (<800Kbyte) and has much more potential
- tftpboot both kernel and ramrootfs
- ability to script warm boot, or cold boot

it is not a problem for me to support u-boot, even from scratch.
I will look into this.

25

Zaurus - Arch Linux ARM support / Re: Update: C1000 / Akita support

« on: May 02, 2025, 09:59:19 am »

and as per above hint from ant:  CONFIG_GPIO_PCA953X=y is set in your kernel ?

yes

26

C1000/3x00 General discussions / Re: C1000 As A Ir Remote

« on: April 28, 2025, 08:10:34 am »

interesting 

27

Zaurus - Arch Linux ARM support / kexecboot-size, max-size?

« on: April 28, 2025, 07:47:23 am »

I'm a bit confused by what I read online.
I wonder:
1) what is the maximum size of the kexecboot executable? < 2Mbyte? or <1.2 Mbyte? Different things are read online.
2) isn't it possible to repartition the flash and use the entire flash to contain kexecboot or something similar?
3) why is there no interest in u-boot?

28

Zaurus - Arch Linux ARM support / Re: Update: C1000 / Akita support

« on: April 28, 2025, 07:36:26 am »

as you may know, current 3.10.y and previous 3.x kexecboot kernels fail to boot on C1000 / Akita.

why did it fail?
what was/is the problem?

29

Zaurus - Void Linux support / Re: kernel patches? to Vanilla

« on: April 28, 2025, 07:10:20 am »

as far as I observe, the patches are only about GPIO.
Right?

Code: [Select]

        1. 7 Kbyte patch_from_void/01-fix-pxa27x-udc.patch
        3. 8 Kbyte patch_from_void/01_revert_mmc_slot-gpio.patch
        2. 8 Kbyte patch_from_void/02_revert_mmc_spitz_pxamci.patch
        1. 6 Kbyte patch_from_void/03_revert_mmc_pxamci_platform.patch
      295. 0  byte patch_from_void/04_revert_mmc_pxa_platform.patch
        2. 3 Kbyte patch_from_void/05_revert_mmc_pxa_gpio_desc.patch
      759. 0  byte patch_from_void/06_revert_mmc_pxa_gpio_desc.patch
        3. 7 Kbyte patch_from_void/07_revert_mmc_spi_gpio_desc.patch
        3. 8 Kbyte patch_from_void/08_revert_pcmcia_cpufreq.patch
      342. 0  byte patch_from_void/09_usb_power.patch
        1. 1 Kbyte patch_from_void/10_sandisk-cf.patch
      249. 0  byte patch_from_void/11-revert-mmc-slot-gpio.patch
(only these? to Vanilla from kernel dot org)

30

Zaurus - Arch Linux ARM support / kernel from 2.6.23 to kernel 5.4.229: suspend/resume?

« on: April 27, 2025, 06:06:18 pm »

hi
I usually take my C1000 out of the house, as I use it daily, usually on the train(2h/day)/tube(40min/day)/coffee breaks/etc, so I desperately need the suspend/resume feature offered by the power management system, as well as desperately need the battery charging feature to get up to 4/5 hours.

Unfortunately *everything* I've tested after kernel version 2.6.23 doesn't work.

I haven't tested kernels 3.* and 4.*
Kernel 5.* ... ummm, not working.

I don't quite understand: have these two features been fixed, and I simply missing something? Are there any specific patches?
Also kernel 4.* is ok

Please let me know.