• Home
  • Help
  • Search
  • Login
  • Register
  Show Posts
Pages: 1 2 [3] 4 5 ... 12
31  Hardware and U-Boot firmware / U-Boot stuff / How do the various mtd devices function? on: September 05, 2009, 11:48:20 AM
I see the following here.  I assume the second type is serial interface and the 3rs serial read only?  How are the serial interfaces used?

Code:
/dev/mtd*
brw-rw---- 1 root disk 31, 0 Sep  3 17:00 /dev/mtdblock0
brw-rw---- 1 root disk 31, 1 Sep  3 17:00 /dev/mtdblock1
crw-rw---- 1 root root 90, 0 Sep  3 17:01 /dev/mtd0
crw-rw---- 1 root root 90, 2 Sep  3 17:01 /dev/mtd1
crw-rw---- 1 root root 90, 1 Sep  3 17:01 /dev/mtd0ro
crw-rw---- 1 root root 90, 3 Sep  3 17:01 /dev/mtd1ro
32  Hardware and U-Boot firmware / Hardware / Re: Spinning down usb attached hard drive on: September 05, 2009, 11:10:52 AM
Yes, Maxtor, is one brand (the only one I know of) which has a built in spin down feaure; even that needs to be enabled thru the interface.  A method is to use an ide to get that done, and the configuration can be set to stick over power cycle.  Since I have been studying this area for a while, my view is that there has been no definite answers to any of the questions I asked; I believe the answers to most of them could be yes or no with equal chance.  At this point I am looking for first hand experience; hearsay is less useful --- internet is full of it Smiley
33  Hardware and U-Boot firmware / Hardware / Re: Spinning down usb attached hard drive on: September 05, 2009, 10:06:13 AM
restamp,
Yes, it was hdparm that I had always used.  So if you have first hand data that you succeeded in spinning down a usb disk, that is very encouraging, and first time I heard that in 8+ years looking at this on and off.  So that means our ShPg's usb interface IS good for that.
  Now the critical question:  what is the exact brand model of the usb case and drive?  I have about 10 usb drives and cases, but since my previous linux (the Slug) had a bad usb interface I do not know if any of the drive is good for this particular function.  They are all in various other current use, I only tried one with ShPg, which returned error on many hdparm commands; but many other hdparm commands (non related to control and settings of this) do work fine.  BTW, I was thinking of making another thread on use of hdparm for other very useful purposes, which I shall do soon.
34  Hardware and U-Boot firmware / Hardware / Spinning down usb attached hard drive on: September 05, 2009, 08:26:26 AM
I know many (all?) usb attached drives do not honor (receive?) the spin down command.  There has been a lot of discussions about this on the net.  Some say it is the usb interface, in our case, the one in the Plug, that has the deficiency, although the interface in most notebook drives that come in usb cases seem to have a similar problem with the spin down.  So both the drive, and plug interfaces have to work. [However, we know almost all notebooks can spin down their drives under idle conditions, but they use ide interface.]

Does anybody know if our ShPg's usb is capable of spin down?  I have tried one drive I have, and gets the error msg with the command.  But most likely my usb drive does not respond to the command.

BTW, there has been a work around in some [rare] drives, where a timer can be configured so that it will shut down after the elapsed time of activity.  I do know that even for those drives, configuring that timer cannot be done through many usb interfaces.

Has any body here succeeded with spin down?  Is there any expert here who knows the real scoop about our ShPg usb interface.
35  General Category / General Discussion / Re: Post your working SD cards here! on: September 05, 2009, 08:02:37 AM
reksa, IMHO, unplugging an sdcard while running, is at best a gamble; it is a bit like: most often you would get to the other side when walking across a not so busy road even if you did not look left or right. Depending on the activity level, and the type fs, most times we may not experience a problem even if a set of files got corrupted, since that file may not be needed by us for sometime to come.

I think, if you first issue the command 'poweroff' and wait ~30secs, your card safety is immensely improved.
36  Hardware and U-Boot firmware / U-Boot stuff / Re: Env setup for: uboot from mmccard, if fails then boot from usb sequence on: September 03, 2009, 09:14:22 PM
restamp,

Is uImage  directory or a file?  Do I just copy a file over, of directory copy as when I did cp the rootfs to the usb?

Do you have handy the uImage mount line parameter, after you did the Alpha-6 install?  Also the boot line env variable you currently use?  Yes I know they are all at various places here already.  I shall write a short Wiki page after this, if such a Wiki is not there already.
37  Linux Stuff / General Linux questions / Re: Huge rootfs copy to sd-card? on: September 02, 2009, 05:57:13 PM
restamp, That was a very bad eye I had here (sheepishly grining)  Embarrassed  When I was quickly looking at my saved log, which was not in fixed space font, and the two numbers aligned differently and I had thought they were similar in the 30k area for me Smiley  This is good to know.

Now that I learned this, I cannot resist asking:  can we not use the similar/same compressed fs on the external sd card?  Would be interested in knowing about this, and side effects if any.  I would not really be sharing my rootfs with anything else like NFS, Samba; all interface is directly to the ShPg.
38  Linux Stuff / General Linux questions / Re: Huge rootfs copy to sd-card? on: September 02, 2009, 10:55:43 AM
riel,  When I did copy many weeks ago, I kept a copy of my directory numbers.  I copied from the internal nand, just the rootfs while running from the nand, to an external sdcard in the sdcard slot.  Here is the before after numbers, which are very close, the small size difference in my case is because of the larger directory space of my 2GB sdcard.  I was USING EXT2 FILE SYSTEM.

Before the copy
Code:
>mkfs -t ext2 /dev/mmcblk0p1
>mkdir /mnt/sdcard
>mount /dev/mmcblk0p1 /mnt/sdcard
>df
Filesystem           1K-blocks      Used Available Use% Mounted on
rootfs                  519168    151100    368068  30% /
tmpfs                   257816         0    257816   0% /lib/init/rw
varrun                  257816        36    257780   1% /var/run
varlock                 257816         0    257816   0% /var/lock
udev                    257816        12    257804   1% /dev
tmpfs                   257816         0    257816   0% /dev/shm
tmpfs                   257816         0    257816   0% /var/cache/apt
/dev/mmcblk0p1         1965176      2928   1862420   1% /mnt/sdcard

After copy:
Code:
>cp -ax / /mnt/sdcard # takes 13 minutes
>cp -a /dev /mnt/sdcard
>df
Filesystem           1K-blocks      Used Available Use% Mounted on
rootfs                  519168    151100    368068  30% /
tmpfs                   257816         0    257816   0% /lib/init/rw
varrun                  257816        36    257780   1% /var/run
varlock                 257816         0    257816   0% /var/lock
udev                    257816        12    257804   1% /dev
tmpfs                   257816         0    257816   0% /dev/shm
tmpfs                   257816         0    257816   0% /var/cache/apt
/dev/mmcblk0p1         1965176    300072   1565276  17% /mnt/sdcard

Does this not look different from what you got?  Were you using ext3 ??
39  Linux Stuff / General Linux questions / Re: Huge rootfs copy to sd-card? on: September 02, 2009, 08:54:17 AM
Hmm...  I do not think I see this type of expansion when copying just the nand rootfs to an sdcard.  Are you copying just the rootfs only, or the uImage too?  Once when I saw this type of double size, it was caused by a pilot error here, because in a second time copy the whole rootfs got copied under in subdirectory.  Perhaps I am misundertanding this?  Am eager to follow this up...
40  Hardware and U-Boot firmware / Hardware / Re: Power Plug rev 1.3 j6 uart and j7 on: September 02, 2009, 08:44:43 AM
Is there a quick how to open the case?  I may be able to force open it, but open the little plastic locks break if you don't try the proper way.

As for 2.5v, the general trend is towards lower voltage for all electronics; one of the impediments has been the large number/variety of external devices which uses the higher voltages.  As mobile devices become wide spread, I predict all interfaces will move towards sub 2v -- lower power, less heat, long battery life.  In one of the boards I was involved in there were 6 internal tiny power converters (switching type to be efficient) to supply all the interfaces, and that board was not much bigger than the Plug.
41  General Category / General Discussion / Re: Post your working SD cards here! on: September 02, 2009, 07:52:26 AM
I usually go for the lesser brand ones --- I am in the semiconductor area.  For compatibility when used in non windows system (especially if it has to be reformatted as for our use in the Plug, and when booting is needed) generic ones seems to be the least troublesome.  Many of the brands try to distinguish themselves with features, especially the security oriented ones, making their interface less standard.  Many brands use the same 3 or 4 brand chips internally, sometimes with tiny customization -- ip cores.  However some vendors, especially totally generic ones from Asia, may be using some reject chips with serious flaws.  I am in the Silicon Valley, now a days I buy them only if we agree to take them back if returned within 2 days.

All of the following work here, he higher speed last two need bootdelay if used as the bootable rootfs.

SSDCards  (some are micro ones -- small formfactor, with an adapter)
MicroCener 2GB (have 2)
MicroCenter 1GB
Kensington 4GB
Kensingon 2GB
Polaroid 2GB
Patrist 4GB

I have a number of usb sticks too, though have not tried them in bootable mode.

I would later come back with a quick speed testing script(s)


42  Hardware and U-Boot firmware / U-Boot stuff / Re: Env setup for: uboot from mmccard, if fails then boot from usb sequence on: August 31, 2009, 09:15:14 PM
I wish to check with you before doing.  So now I can simply rsync copy the uImage into a file /uImage on my USB drive or sdcard which is the rootfs (ext2).  Then adjust the bootcmd to load this image with ext2load instead of my current nand read. Will that work?  I have no risk here because I would have a complete duplicate of the target sdcard or usb drive volume before I copy.  Kindly check if there is any hole in what I am going to do.  Thanks.

BTW, added a note at the Wiki multi boot about the kernel to be in at the same place.
43  Hardware and U-Boot firmware / U-Boot stuff / Re: Env setup for: uboot from mmccard, if fails then boot from usb sequence on: August 31, 2009, 04:41:29 PM
This won't work.  Multi-boot means trying to boot the Uimage (kernel) from the USB/SDcard/NAND in some order, not booting a kernel and then picking where you want to mount the root FS from after a kernel is booted.  Once the kernel takes control, it won't give it up.
Wow!  I had not realized this Sad  Yes I had my kernel in nand.  Without this msg, I would spent many more hours, since I thought I simply was not typing things in right Smiley

Well, no worries, mate (as my friends on the other side of the big pond says).  I would do the moving of the uboot image (uImage?) to my disk first.  I was thinking of doing that, but did not have enough of an incentive to do that up until now Smiley  Let me go looking for your tips on that.

BTW, I am now guessing that it was because I had to use that $(mtdpartitions) in my boot line (since I had to read nand) that my line was too long to fit in one env variable?  Anyway this is moot now.
 
44  Hardware and U-Boot firmware / U-Boot stuff / Re: Env setup for: uboot from mmccard, if fails then boot from usb sequence on: August 31, 2009, 03:05:24 PM
restamp,

Tried your suggestion to run from the uboot cmdline manually.  Well, when the device is not there, it fails as before, but the problem is I do not get the command line back  -- equivalent to not returning to the subsequent part of the multi boot command line.  At this point reset is the only think that would work.

The last few lines of this manually typed in here:
Code:
... Setting system clock to ...
Waiting 10sec before mounting root device [Yes I have a rootdelay]
Root-NFS: No NFS server available, giving up.
VFS: Unable to mount root fs via NFS, trying floppy.
VFS: Cannot open root device "sda1" or unknown-block(2,0)
Please append a correct "root=" boot opton; here are the avilable partitions:
1f00      4096 mtdblock0 (driver?)
1f01   519168 mtdblock1 (driver?)
b300 1974272 mmcblk0 driver: mmcblk
b301  1974090 mmcblk0p1
Kernel panic -- not syncing: VFS: Iman;e to mount roof fs on unknown-block(2,0)
[<c00809c0>] (unwind_backtrace+0x0/0xe0) from [<c008db20] (panic+0x50/0x120)
[...]
[...]
[...]
[<c0008734>] (kernel_init+0xc0/0xec) from [<c0040b18>] (do_exit+0x0/0x6ac)
It is hung here, no command line...
45  Hardware and U-Boot firmware / U-Boot stuff / Re: Env setup for: uboot from mmccard, if fails then boot from usb sequence on: August 31, 2009, 01:56:15 PM
restamp,

I think I follow everything you said.  The contents of o_bootcmd was that of the last bootcmd when it was wokring.  The saveenv was added after the command was failing (failing exactly the same way before and after adding it), just to see if it returned to there; because the contents of bootargs_root on the next boot, even after power a cycle will retain the last value.

You introduced one new good idea, I would try the manual run of each part and the whole; that might give a better insight into if returning or possibly what error.

BTW, I would like to keep my current 2-part boot of each device to a simpler one part, but could not get that into one variable without the >16 error, though I tried intermediate variables.

I also had an intermediate goal, if possible, to change very little of the exiting and introduce just one or two variables, and use the original bootcmd with minimal retyping, for simplifying it for others.  Of course first I have to make it functional.  BTW, I first editing the command line in uboot very inconvenient:  is there a way to get the last command or history to re-edit?  Even cut and paste is quite error prone, because of the limited size of the terminal window I have to use.

I am doing this from another place; when I get to the Plug I shall report back.
Pages: 1 2 [3] 4 5 ... 12