How to Distribute? (TrueCrypt, Cifs.ko, ntfs.ko, mount.cifs) - Page 17 - maemo.org

Active Topics

The N900 is now 15 years old! (7)
to Nokia N900 by Danct12 - 1 day ago
Sailfish OS for the Motorola Moto G7 Power (XT1955-5) - (ocean) (4)
to SailfishOS by Maemish - 1 day, 19 hrs ago
F(x)tec Pro1x (QWERTY) for sale (0)
to Buy & Sell by aln00000 - 4 days, 4 hrs ago
more...

Page 17 of 22

Thread Tools

Nathan	2010-02-12 , 15:50
Posts: 452 \| Thanked: 522 times \| Joined on Nov 2007	#161

Originally Posted by slewis1972

Hi Guys

Last night I created a folder called NAS under media and mounted it via:

mount -tcifs //xxx.xxx.xxx.xxx/media/Music/Music /media/NAS -ouser=xxxx,pass=xxxxxx,ip=xxxxxx

1. Do a 'lsmod' from the terminal and make sure that cifs shows up in the modules list.
2. If it does then try your mount command above, and then do a 'dmesg' and see what the error message is.

Nathan

Quote & Reply |

Nathan	2010-02-12 , 18:27
Posts: 452 \| Thanked: 522 times \| Joined on Nov 2007	#162

Originally Posted by luigi

@Nathan: I've tried the ",direct" option in an attempt to increase the speed of the cifs mount, but for me it turns out there is no noticeable difference with or without that option.

The average time it takes to copy 100Mbit from my server is nearly 30sec. - compared with the 12 sec. it takes to scp it. Browsing through a folder with images or viewing a video is hardly possible.

I'm curious to know experiences of others.

On a connection that shows quality of about 27% in wifiinfo -- Yeah, I'm a ways from my router. So I'm sure this actually slows things down a bit. But I did all my measurements multiple times from the same place.

Well, in my tests this is what occurred:
- scp was almost twice as fast as cifs
- cifs with "direct" was twice as fast as cifs without it.

Direct numbers
- I can copy a 30 meg file in 60 seconds (Cifs, direct mode)
- I can copy the same 30 meg file in 33 seconds, via scp.

Doing some google research, it does appear that cifs is known to be slower than ftp/nfs/scp. So this is probably not going to improve any, as it is a protocol issue.

Nathan

Quote & Reply |

Nathan	2010-02-12 , 18:39
Posts: 452 \| Thanked: 522 times \| Joined on Nov 2007	#163

I also dropped a new version of TrueCrypt into the repository with the missing control files and the sudoers addition. Please vote on it so that we can move it out of "testing" and into extras. (Truecrypt itself is unchanged -- just install related stuff).

Nathan

Quote & Reply |

maartenmk	2010-02-12 , 19:11
Posts: 182 \| Thanked: 69 times \| Joined on Nov 2009 @ Netherlands	#164

For me the direct mode actually seems to be slower. 500KB/s would be plenty for streaming, but I reach nowhere near that speed, more like 50. Any ideas?

I did manage to get it working without the script of GameboyRMH now. Nathan, could it be your instructions in the 1st post are incorrect? Looking at the script, I tried this:
mount -t cifs //<server>/<path> /<your>/<local>/<path> -o user=<user>,pass=<pass>,ip=<ip of server>,direct

And it worked, while the original instructions didn't.

Quote & Reply |

Nathan	2010-02-12 , 22:19
Posts: 452 \| Thanked: 522 times \| Joined on Nov 2007	#165

Originally Posted by maartenmk

For me the direct mode actually seems to be slower. 500KB/s would be plenty for streaming, but I reach nowhere near that speed, more like 50. Any ideas?

I did manage to get it working without the script of GameboyRMH now. Nathan, could it be your instructions in the 1st post are incorrect? Looking at the script, I tried this:
mount -t cifs //<server>/<path> /<your>/<local>/<path> -o user=<user>,pass=<pass>,ip=<ip of server>,direct

And it worked, while the original instructions didn't.

You have to have the original "mount.cifs" installed (see the attachment on post #1), to use it. I haven't used the script that others developed -- so I have no idea how it works.

Now, you can use either "mount.cifs" or "mount -t cifs" -- the rest of the command line is identical. I just ran the tests again using both commands to verify "direct" worked with both of them and was getting <400k (no direct) vs >600k (direct). It does fluctuates quite a bit on large files without direct, where it is consistent on large files with "direct". I also tried my next set of tests close to my router and my non-direct was a bit faster than before. But using "direct" on the command line with mount.cifs and mount -t cifs was still much better.

This is my hardware setup.
A Server which is physically wired to the Linksys 54g Access Point, using only wireless G on Channel 1. I have the "larger" linksys antennas, and am also running running ddt-wrt on the AP. I tried messing with the xmit power but that didn't help a bit. I do have the access point set to use one antenna as Transmit and the other as Receive (rather than auto).

On my n900 I have cifs & ntfs all auto loading on startup. I have
- wireless power savings set to on (intermediate)
- wifi transmissions power to 100mw
- wpa encryption (not wpa2)
- scanning = never.

Nathan

Quote & Reply |

The Following User Says Thank You to Nathan For This Useful Post:
maartenmk

wmarone	2010-02-12 , 22:21
Posts: 1,746 \| Thanked: 2,100 times \| Joined on Sep 2009	#166

For the next release could you include nls_utf8.ko?

I've got a bunch of files with Japanese filenames shared from my desktop that don't appear properly with the default ANSI encoding. I built and loaded the module myself and confirmed it worked, but it'd be nice to have it as part of an official, managed package.

Quote & Reply |

Nathan	2010-02-12 , 22:47
Posts: 452 \| Thanked: 522 times \| Joined on Nov 2007	#167

Originally Posted by wmarone

For the next release could you include nls_utf8.ko?

I've got a bunch of files with Japanese filenames shared from my desktop that don't appear properly with the default ANSI encoding. I built and loaded the module myself and confirmed it worked, but it'd be nice to have it as part of an official, managed package.

Actually it is already also in extras-testing, I stuck it their a while back since someone else had the same issue as you. "kernel-modules-nls-utf8" I think is its name. ;-)

Out of curiosity, do you have to manually load the utf8 module; or does the os auto load it when it detects a utf8 name?

I've been meaning to test this; but haven't had time. I have a script in the kernel module that loads it on startup; but if the os is smart enough to load it when it needs it then I don't need to auto-load it on startup.

Nathan

Quote & Reply |

wmarone	2010-02-12 , 22:58
Posts: 1,746 \| Thanked: 2,100 times \| Joined on Sep 2009	#168

Doh, didn't think to look for it. And there it is. I was too busy trying to figure out why the "user" option in my fstab was being ignored (apparently busybox doesn't acknowlege it?)

I had to manually insert it into the modules.dep and move it into /lib/modules/`uname -r`, after which I modprobe'd it. I suspect it would probably have automatically loaded it, but didn't try. I can this evening however, after I clean out my existing changes and install the package.

Quote & Reply |

	qwazix	2010-02-17 , 12:28
	Moderator \| Posts: 2,622 \| Thanked: 5,447 times \| Joined on Jan 2010	#169

as I suspect I got the wifi bug, the weird thing is that it started only after installing the cifs module.
I just updated to pr1.1.1 and waiting to see if it is resolved (cifs installed)
________
Kinky_lola

Last edited by qwazix; 2011-08-21 at 10:21.

Quote & Reply |

wickedgtr1	2010-02-18 , 18:30
Posts: 17 \| Thanked: 3 times \| Joined on Feb 2010	#170

After installing the Cifs Kernel module, it breaks the built in screen calibration utility -- tested on my very used nokia, and a friends brand new one.

I have just done a complete flash/emmc flash using the command line flasher and will re-test.

Other then that - it works great, thank you.