Reply
Thread Tools
Posts: 9 | Thanked: 0 times | Joined on Mar 2008
#1
Hello,

I'm a N800 newbie and have been having some trouble with MMC
cards in my N800. I've been trying to get two 8G A-data SDHC
to work without any luck. I haven't been able to write anything to
them, nor have I been able to format them.

In an attempt to start simpler, I removed those cards out and put in
a 2G card (from Transcend) in the internal slot.

I also re-flashed the latest 2008 release.

Things seemed to go pretty well - I've installed maemo-mapper
and begun loading maps to a directory under /media/mmc1.

I wanted to see how much space the maps were tacking so I
ssh-ed into the N800 and did a 'df' and a 'mount'

Both of those say that the card is mounted on /media/mmc2 and
that nothing is mounted on /media/mmc1. Weird.

When I look at the mmc2 directory, it is empty - but /media/mmc1 is
still showing the mapper directory. (And the m-mapper file under
mmc1 is still growing as mapper writes into its cache.)

If I look at /etc/fstab, it lists the /dev/mmcblk0p1 as mounting at
/media/mmc1.

So, I'm confused as to why the 'df' and 'mount' commands are
showing what I think is the wrong directory.

Could this be behind my problems with trying to start with the 2 new
cards? (Perhaps the system was getting mixed up concerning
which card was mounted where.)

Can someone clear things up for me? (Or post a pointer to a thread
that covers this. My searching uncovered lots on mounting, but not
anything on this.)

Regards,

Ed
 
Posts: 2,152 | Thanked: 1,490 times | Joined on Jan 2006 @ Czech Republic
#2
/media/mmc2 is card in internal slot
/media/mmc1 is card in external slot
the reason is that Nokia 770 had only external slot mounted to /media/mmc1 so they kept the naming for better compatibility

if you are writing to /media/mmc1 with no card inserted (not sure if it is possible, most probably yes) you are writing to internal flash, in that case you may move data to card via 'mv /media/mmc1/* /media/mmc2/'

the mmcblkX confusion may be caused by this http://bugs.maemo.org/show_bug.cgi?id=2747
__________________
Newbies click here before posting. Thanks.

If you really need to PM me with troubleshooting question please consider posting it to the forum instead. It is OK to PM me a link to such post then. Thank you.
 
Posts: 9 | Thanked: 0 times | Joined on Mar 2008
#3
Thanks for the insight Fanoush. I understand the need to keep
things consistent with previous versions. (Although I would have
thought that /media/mmc0 might have been a better file name
for the internal slot).

I understand the mapping to be:

external slot -> /dev/mmcblk1 -> /media/mmc1
internal slot -> /dev/mmcblk0 -> /media/mmc2

Now my question is, why does /etc/fstab have the line:

/dev/mmcblk0p1 /media/mmc1 vfat rw,noauto,nodev,noexec,nosuid,utf8,uid-29999 0 0

Should I remove it?

And just where is the mounting taking place, given whatever
is doing it is not paying attention to fstab?

Ed
 
Posts: 47 | Thanked: 5 times | Joined on Mar 2008
#4
How do you SSH?
 
Posts: 2,152 | Thanked: 1,490 times | Joined on Jan 2006 @ Czech Republic
#5
Originally Posted by Ed C View Post
Now my question is, why does /etc/fstab have the line:

/dev/mmcblk0p1 /media/mmc1 vfat rw,noauto,nodev,noexec,nosuid,utf8,uid-29999 0 0

Should I remove it?
Not sure. I have it too on my N810. Looks like bug. This line is valid for 770 but makes no sense on N8x0. Maybe they forgot to remove it because it does not hurt. In theory removing it should not break anything.

Card mounting and unmounting is done automagically by /usr/sbin/osso-mmc-[u]mount.sh scripts. These are possibly called from ke-recv or hal stuff in /usr/lib/hal/ when kernel detects opening/closing card doors or inserting/removing usb cable. Not sure how this stuff (ke-recv, hald, /usr/lib/hal/* ) works exactly.
__________________
Newbies click here before posting. Thanks.

If you really need to PM me with troubleshooting question please consider posting it to the forum instead. It is OK to PM me a link to such post then. Thank you.
 
Posts: 9 | Thanked: 0 times | Joined on Mar 2008
#6
Originally Posted by Inate00 View Post
How do you SSH?
I'm not sure if you are asking to learn or because this may impact my
troubleshooting? If it is to learn, then I am probably not the best teacher
as I'm pretty much a novice here. But I would suggest that you look at
the wiki page
HTML Code:
[URL="http://maemo.org/community/wiki/installssh2006/"]http://maemo.org/community/wiki/installssh2006/[/URL]
 
Posts: 9 | Thanked: 0 times | Joined on Mar 2008
#7
oops, and I forgot to add. If this is to help with the troubleshooting, then I
just did a 'apt-get install ssh' to get the SSH server up and running and
then come in using a client on a Linux box or Windows (depending on which
one I happen to be sitting at).

Ed
 
Posts: 9 | Thanked: 0 times | Joined on Mar 2008
#8
Originally Posted by fanoush View Post
Not sure. I have it too on my N810. Looks like bug. This line is valid for 770 but makes no sense on N8x0. Maybe they forgot to remove it because it does not hurt. In theory removing it should not break anything.

Card mounting and unmounting is done automagically by /usr/sbin/osso-mmc-[u]mount.sh scripts. These are possibly called from ke-recv or hal stuff in /usr/lib/hal/ when kernel detects opening/closing card doors or inserting/removing usb cable. Not sure how this stuff (ke-recv, hald, /usr/lib/hal/* ) works exactly.
I am not so sure that having that having the line there doesn't break things.

With the line there then when I power up without any cards in, mount will
still show a mmcblk device mounted on mmc1.

If I start up maemo-mapper in this mode, it seems to die. (Seems to die
because although its windows never comes up, the process is still there if
you do a 'ps'.) I think that this is because I have configured mapper to
store its map cache under the /media/mmc1 directory.

If I start mapper without any cards installed *and* with the mmc line in fstab
removed, then mapper complains about not being able to create the cache
file (becuase it can't get the the "/media/mmc1/Maps" directory), but still
succeeds in bring up its window - and can be terminated.

Once I get into this mode where mapper is hung, I cannot shut the box
down. I have to pop the battery out to reset it. (Fortunately, this doesn't
seem to cause any problems. Perhaps I've just been luck in this part.)

Although the screen eventually shuts off, the box is still running and I can
ssh into it.

A 'ps' shows that the mini-shutdown routine is (supposedly) running, but
nothing much is happening. Trying to reboot/shutdown from the ssh
session also does nothing.

(I also cannot 'kill -9' the mapper process.)

To me It has the feel that the system is trying to unmount the filesystems as
part of the shutdown process. But, it might be getting stuck on the mmcblk
device that really isn't there (but that the mapper application might be trying
to use). And, it is patiently waiting - forever - for that "device" to unmount.

But I may be wrong.

In any case, things "seem" to work a bit more nicely without that line there.

I say "seems" becuase I'm still having problems with mounting cards. (I'll
continue this in the other thread I started on potential hardware problems.)

Ed
 
Reply


 
Forum Jump


All times are GMT. The time now is 11:59.