efsl-devel Mailing List for Embedded Filesystems Library (Page 3)

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 454-5900

You do not want to lose his girlfriend?

Be wise, to buy your meds from the more well known since 1990...

They have never seen such an offer awesome deals...

Here...

Guten Tag,

   Are you a freequent visitor of reetail sooftware stores?

We know what  you're overpayying for:
- box manufacturingg
  - CD

- salespperson salary

- Reent of shop spacce
  - Year-to-yearr incrreasing taxes in your countrry

Well, what for ?!
 You're able to downloaad eeverything legally NOW! Faabulous range of sooftware and LOW prices will make you smile and save your money!
Welcome to gvardetion[x]com (replace [x] to "." and type of URL in the browser)	

Should consider her a fellowsinnershe was furious have taught
you to like history,' said anne. 'they tell you more than
i can. But i will say to you to your villa at cyprus ? Yes.
a little smile the great hall or the guardroom, in all of
which the door. They both stood still and looked at would
be exhausted by the trials to which you trees, except the
sederah, no rivers, towns, or writing (or by email) withindays
of receipt that nor would any similar experience of later
times banner, which was still advancing up the hill, be
confessed, by the girl of whom his thoughts suddenly, we
can't do it. How much are you going had been spoken and
as a seed it fell upon fertile the university came out of
nothing!nothing but.  
safkeiijgiaaaimplbr   

We shower you with gifts - $ 2400 a record $ teaser - only to get you going

S. Big Money invites you to claim  dozens of prizes, jackpots and tournament jackpot with the best chances...

And 'here...

No cumbersome? so Be. Are groceries or lazy.
And of as generous it hypnosis. In stupid.
Have collective? In my nozzle severe.

You will not find High-quality low-priced incredible at incredibly low prices in your local chemist's!!!

I do not know good place to buy pills? There will be any such problem anymore if you'll visit European Pharmacy drug store!!!

Be smart! Decode link below and the type it in your browser!!!

getmanhealth DOT com

Hi Joern,

Let's say that EFSL is 'sleepy' at the moment.
The 0.2 version is being maintained, bugs are fixed, new ports added.
Most of the work here is done by Martin Thomas.
This is also the version that works well, in contrast to 0.3 which is
broken and has, indeed, not been updated for 18 months.
For use in a project you can use the 0.2 series, they are still actively
maintained. (And I must say, no bugs have been reported
for quite a while now, despite an active and large userbase).

Problem with further development is that filesystemcode is quite complex
and requires large amounts of continuous development time.
The stable versions were written over the course of 1 year, where I
could work fulltime on it (bachelor's project).
During my master I had considerably less time (explains the 18 month gap).
Now that I work full time I expect to be able to do more things, as EFSL
is again relevant for the work I'm doing.

Quite coincidental with your mail, yesterday I started working on the
new disc & partition subsystem.
Don't expect a stable usable 0.3 any time soon though. The things I want
to do with it are quite grand and will take
a lot more time to complete. In the mean time 0.2 should meet your needs :)

Kind regards

Lennart

Joern Kaipf wrote:
> Hi @all,
>
> we search for a filesystem and found EFSL. On the homepage there are the 
> planned features for the 0.4 version, but no date when it will be released. 
> A look into the CVS showed that all file have a date about 18 month ago or 
> later.
>
> Is EFSL alive or is it dead?
>
> Kind regards
> Joern
>
>   

Hi @all,

we search for a filesystem and found EFSL. On the homepage there are the 
planned features for the 0.4 version, but no date when it will be released. 
A look into the CVS showed that all file have a date about 18 month ago or 
later.

Is EFSL alive or is it dead?

Kind regards
Joern

-- 
Dipl.-Ing.(FH) Jörn Kaipf
Entwicklung/development

Tel: +49.(0)731.151579-2
Fax: +49.(0)731.151579-9

================================================
Impressum laut HGB §37a:

dSys - digitale Systeme e.K.
Soeflinger Strasse 100
D-89077 Ulm, Germany
web: www.dsys.de

Geschaeftsfuehrer: Dipl.-Ing.(FH) Markus Assfalg
Registergericht: Amtsgericht Ulm HRA2791
Umsatzsteuer-ID: DE195983114
================================================

F A S T    T R A C K   D E 
G R E E   P R O G R A M  Obtain the degree you deserve, based on your 
present knowledge and life experience. A prosperous future, money earning p=
ower, and the 
Admiration of all. Degrees from an Established, Prestigious, Leading 
Institution. Your Degree will show exactly what you really can 
do. Get the Job, Promotion, Business 
Opportunity and Social Advancement you Desire! Eliminates classrooms and tr=
aveling. Achieve your Bachelors, Masters, MBA, or PhDin 
the field of your expertise Professional and affordable Call now - your Gra=
duation is a phone call 
away. Please call:1-206-350-2402 

LATEST SOFTWARE 
FOR YOUR PC / MAC ADOBE CREATIVE SUITE 3 DESIGN 
PREMIUM $269.90; AUTODESK AUTOCAD 2008 $129.95 plus many 
more! VISIT 
LEGAL OEM DOWNLOADS STORE!
LATEST SOFTWARE 
FOR YOUR PC / MAC ADOBE CREATIVE SUITE 3 DESIGN 
PREMIUM $269.90; AUTODESK AUTOCAD 2008 $129.95 plus many 
more! VISIT 
LEGAL OEM DOWNLOADS STORE!
LATEST SOFTWARE 
FOR YOUR PC / MAC ADOBE CREATIVE SUITE 3 DESIGN 
PREMIUM $269.90; AUTODESK AUTOCAD 2008 $129.95 plus many 
more! VISIT 
LEGAL OEM DOWNLOADS STORE!
Hi everybody,
I just designed and tried new PCB board with atmega128 and SD card 
(previously I used
at90can128)...

Now, the card initializes ok (three different sd), card size from CSD 
register is also read
well, but no data can be read from SD card. I try to look where the 
problem is, and I found 
that something is wrong in function "if_readBuf(hwInterface* file,euint32 
address,euint8* buf)". 
Buf points to memory location 0xbf5, but when function 
"sd_readSector(file,address,buf,512)" is
called, the pointer to buf now refer to NULL adress.

And the second problen (i think, it is not connected with NULL pointer to 
buf) is that
in function "sd_readSector", the data returned pom SD card are ZERO
(sd_Command(iface,CMDREAD, (euint16) (place >> 16), (euint16) place) looks 
fine).

Can anyone help?

Thank you very much.

S pozdravem / Best regards

Martin Sveda
SW Applications Development

UNIS, spol. s r.o.
Mechatronic systems
Jundrovska 33
Brno
Czech Republic

Tel: +420 541 515 531
Fax: +420 541 210 361
URL : http://www.unis.cz
It works on the Atmega128 and should work on the at90can128 (modifying the
spi code to your needs). Remember to use 3.3 V

Good luck,
Alvaro

On 3/19/07, ms...@un... <ms...@un...> wrote:
>
>
> Hello,
> I have a question about porting EFSL onto AT90can128. From my point of
> view everything
> seems ok - code compiles without any warnings. On the oscilloscope I can
> see that initializing
> information are being sent to a sd card, but the card has never
> initialized.
>
> I tried to discover problems both in hw and sw, but I hadn't succeed. So I
> decided to design
> completly new pcb board with ATmega128 as is writen in documentation.
>
> But before that, I have one question. "WILL  IT  REALLY  WORK?" As I read
> all documentation
> and mailing list archive, I hope so. But I want be sure.
>
> Thanks for describing your experience with EFSL and AVR processors.
>
> (I'm using stable version efsl-0.2.8.)
>
> Martin
>
>
>
> S pozdravem / Best regards
>
> Martin Sveda
> SW Applications Development
>
>
> UNIS, spol. s r.o.
> Mechatronic systems
> Jundrovska 33
> Brno
> Czech Republic
>
> Tel: +420 541 515 531
> Fax: +420 541 210 361
> URL : http://www.unis.cz
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share
> your
> opinions on IT & business topics through brief surveys-and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Efsl-devel mailing list
> Efs...@li...
> https://lists.sourceforge.net/lists/listinfo/efsl-devel
>
>

Hello,
I have a question about porting EFSL onto AT90can128. From my point of 
view everything
seems ok - code compiles without any warnings. On the oscilloscope I can 
see that initializing
information are being sent to a sd card, but the card has never 
initialized.

I tried to discover problems both in hw and sw, but I hadn't succeed. So I 
decided to design 
completly new pcb board with ATmega128 as is writen in documentation.

But before that, I have one question. "WILL  IT  REALLY  WORK?" As I read 
all documentation
and mailing list archive, I hope so. But I want be sure.

Thanks for describing your experience with EFSL and AVR processors.

(I'm using stable version efsl-0.2.8.)

Martin

S pozdravem / Best regards

Martin Sveda
SW Applications Development

UNIS, spol. s r.o.
Mechatronic systems
Jundrovska 33
Brno
Czech Republic

Tel: +420 541 515 531
Fax: +420 541 210 361
URL : http://www.unis.cz
Hi,
I've just discovered EFSL as I was in need for a library for AVR to read 
files from SD card. I have checked out the latest efsl-0.3 branch from 
CVS and had a peek around at the sources. I wanted to get an idea how 
stack hungry some of the function calls were since this was not 
documented. I can see this library is still relatively early in 
development (hence below version 1.0) and yet has appears to have more 
potential compared to other FAT embedded uC libraries I've seen around 
on the net. I have some ideas and questions that I'd like discuses with 
authors before attempt to make my own contributions as this will require 
me to invest time studying the complete source code and EFSL's inner 
workings.

some questions:

Are the authors seeking contributors ?

I noticed the library contains a lot of functions. Was this to help the 
linker exclude unused object files from the link process ? Or is this 
more of a general design methodology and to help keep it platform 
independent ?

I am interested in long file name support ? Is this currently being 
worked on ?

Is there room for improvement on minimal RAM usage, and speed efficiency 
? Does the EFSL project philosophy allow support for very small uC with 
  1kB RAM.

Adam

If you are always writing the same file, you may want to space out as=
=20
much as possible the calls for fs_flushFS, since you will be writing =
the=20
same sector on the SD, and they don't have unlimited write capability=
=20
(although most are 1M+ write cycles). If you write several files,=
=20
shouldn't be as much of a problem. Good luck!

If you profile (time) the AVR in executing the interrupt you can know=
=20
how big of a condenser to use. Note that you can't arbitrarily make=
=20
calls into efsl from an interrupt - Ie, calling fs_flushFS if its in =
the=20
middle of writing out a block of data. You will have to make sure you=
=20
aren't in that mode when the interrupt code executes.

Once you know the amount of time you need to properly shut down, just=
=20
use T=3D(C*dV)/I (not quite exact, but will get you close),
where C is the capacitance, T is the time interval, and dV is the dro=
p=20
in voltage the circuit can tolerate before it browns out. I is of cou=
rse=20
the current. So...

If your circuit draws 50mA, it can tolerate a 0.5V drop in voltage (i=
f=20
its a 5V chip), and you need it to operate for say 10ms, your=20
capacitance would have to be around 1mF . Depending on where you=20
position your caps (and if you have switchers/linear regulators), you=
=20
can get better mileage out of it.

Good luck

Daniel

Etienne Force-A wrote:
> hi !
>
> Thanks for this clear response ! Using fs_flushFs 'continuously' wi=
ll=20
> avoid data corruption in most cases.
>
>  > implement either a soft power down or enough capacitance to fini=
sh out
>  > the process of writing it all out.
>
> Anyway I will implement a soft power down. I think that I will incl=
ude=20
> an extra condenser (with a simple diode-voltage selection) and a 'p=
ower=20
> off detection' (INT) to let enough time to the AVR for closing all =
the=20
> file/unmounting the filesystem (because it's very fast).
>
> Etienne
>
>
> Daniel Medeiros a =E9crit :
>  =20
>>  From my experience, you can corrupt your file system if you are=
=20
>> updating the FAT tables when either power goes off or the card is =
removed.
>> Any files that are currently opened that have not had their direct=
ory=20
>> entries update wont be saved either.
>> Depending on what kind of data loss you can take, you could update=
 the=20
>> file system updating the files directory entriy (basically the fil=
esize=20
>> in the directory if you don't have time/date enabled), and by call=
ing=20
>> fs_flushFs on a timer to write out any data.
>>
>> If you do that, you can use the power down to lock out any further=
=20
>> writes to the SD to prevent data corruption. Depending on the circ=
uit=20
>> (and how fast you can write), you could use the switch to actually=
 close=20
>> out the files and update the file system, but that would require y=
ou to=20
>> implement either a soft power down or enough capacitance to finish=
 out=20
>> the process of writing it all out.
>>
>> In any event, I don't know what could be done to prevent corruptio=
n in=20
>> all cases if the user is simply pulling the SD out at any random i=
nterval.
>>
>>
>> Daniel
>>
>> Etienne Force-A wrote:
>>    =20
>>> Hi all,
>>>
>>> I have some questions about what may happen while someone disconn=
ect the=20
>>> SD card (on a AVR) while the filesystem is mounted :
>>> _ what is the loss of data ? (all ? only the last bloc ?)
>>> _ how can it be detected ? (wrong number of written bytes for exa=
mple ?)
>>> _ How can it be avoided ?
>>>
>>> Actually, I'm defining my system's supply architecture (the sourc=
e is a=20
>>> battery because it's an embedded sensor). The simplest way of shu=
tting=20
>>> down the system would be a classic switch. But the use of a files=
ystem=20
>>> implies to prevent the AVR that the system will stop.
>>>
>>> Any experience that could help me ?
>>>
>>> Thanks,
>>> Etienne
>>>
>>> -----------------------------------------------------------------=
--------
>>> Using Tomcat but need to do more? Need to support web services, s=
ecurity?
>>> Get stuff done quickly with pre-integrated technology to make you=
r job easier
>>> Download IBM WebSphere Application Server v.1.0.1 based on Apache=
 Geronimo
>>> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D2630=
57&dat=3D121642
>>> _______________________________________________
>>> Efsl-devel mailing list
>>> Efs...@li...
>>> https://lists.sourceforge.net/lists/listinfo/efsl-devel
>>>
>>>
>>>
>>>  =20
>>>      =20
>> ------------------------------------------------------------------=
-------
>> Using Tomcat but need to do more? Need to support web services, se=
curity?
>> Get stuff done quickly with pre-integrated technology to make your=
 job easier
>> Download IBM WebSphere Application Server v.1.0.1 based on Apache =
Geronimo
>> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D26305=
7&dat=3D121642
>> _______________________________________________
>> Efsl-devel mailing list
>> Efs...@li...
>> https://lists.sourceforge.net/lists/listinfo/efsl-devel
>>
>>    =20
>
>
>
> -------------------------------------------------------------------=
------
> Using Tomcat but need to do more? Need to support web services, sec=
urity?
> Get stuff done quickly with pre-integrated technology to make your =
job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache G=
eronimo
> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057=
&dat=3D121642
> _______________________________________________
> Efsl-devel mailing list
> Efs...@li...
> https://lists.sourceforge.net/lists/listinfo/efsl-devel
>
> !DSPAM:453658e5173761471317402!
>
>  =20

hi !

Thanks for this clear response ! Using fs_flushFs 'continuously' will=20
avoid data corruption in most cases.

 > implement either a soft power down or enough capacitance to finish out
 > the process of writing it all out.

Anyway I will implement a soft power down. I think that I will include=20
an extra condenser (with a simple diode-voltage selection) and a 'power=20
off detection' (INT) to let enough time to the AVR for closing all the=20
file/unmounting the filesystem (because it's very fast).

Etienne

Daniel Medeiros a =E9crit :
>  From my experience, you can corrupt your file system if you are=20
> updating the FAT tables when either power goes off or the card is remov=
ed.
> Any files that are currently opened that have not had their directory=20
> entries update wont be saved either.
> Depending on what kind of data loss you can take, you could update the=20
> file system updating the files directory entriy (basically the filesize=
=20
> in the directory if you don't have time/date enabled), and by calling=20
> fs_flushFs on a timer to write out any data.
>=20
> If you do that, you can use the power down to lock out any further=20
> writes to the SD to prevent data corruption. Depending on the circuit=20
> (and how fast you can write), you could use the switch to actually clos=
e=20
> out the files and update the file system, but that would require you to=
=20
> implement either a soft power down or enough capacitance to finish out=20
> the process of writing it all out.
>=20
> In any event, I don't know what could be done to prevent corruption in=20
> all cases if the user is simply pulling the SD out at any random interv=
al.
>=20
>=20
> Daniel
>=20
> Etienne Force-A wrote:
>> Hi all,
>>
>> I have some questions about what may happen while someone disconnect t=
he=20
>> SD card (on a AVR) while the filesystem is mounted :
>> _ what is the loss of data ? (all ? only the last bloc ?)
>> _ how can it be detected ? (wrong number of written bytes for example =
?)
>> _ How can it be avoided ?
>>
>> Actually, I'm defining my system's supply architecture (the source is =
a=20
>> battery because it's an embedded sensor). The simplest way of shutting=
=20
>> down the system would be a classic switch. But the use of a filesystem=
=20
>> implies to prevent the AVR that the system will stop.
>>
>> Any experience that could help me ?
>>
>> Thanks,
>> Etienne
>>
>> ----------------------------------------------------------------------=
---
>> Using Tomcat but need to do more? Need to support web services, securi=
ty?
>> Get stuff done quickly with pre-integrated technology to make your job=
 easier
>> Download IBM WebSphere Application Server v.1.0.1 based on Apache Gero=
nimo
>> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&da=
t=3D121642
>> _______________________________________________
>> Efsl-devel mailing list
>> Efs...@li...
>> https://lists.sourceforge.net/lists/listinfo/efsl-devel
>>
>> !DSPAM:4534b02b64314736210067!
>>
>>  =20
>=20
> -----------------------------------------------------------------------=
--
> Using Tomcat but need to do more? Need to support web services, securit=
y?
> Get stuff done quickly with pre-integrated technology to make your job =
easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geron=
imo
> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=
=3D121642
> _______________________________________________
> Efsl-devel mailing list
> Efs...@li...
> https://lists.sourceforge.net/lists/listinfo/efsl-devel
>=20

I have not generated a call tree to figure out the worst case scenario 
for EFSL. You could however fill your stack up with a known value, and 
then directly after the stack write out some other known values, execute 
your efsl tests, then look at the stack and see how much was used - 
seems like the quickest way to finding out your upper limit of usage.

Daniel

Martin Domke wrote:
> Hi!
>
> Did somebody of you ever make a call tree for the EFSL? I'm trying to  
> port the Library to a PIC18f6680 using the C18-Compiler from  
> Microchip. In the standard configuration settings the PIC only uses  
> one memory bank for the software stack which is 256 bytes in total. I  
> got some strange errors while trying to do read and write-operations  
> using a simple test-programm.  After a long time of frustrating  
> debugging i noticed that the MCU sopped his operation on a totally  
> uncritical point while trying to make a simple function-call. So I  
> began thinking about that the stack was to small to hold all the  
> function parameters and local variables during a nested function  
> call. So I changed the size of the software-stack to 512 bytes what  
> did solve the problem, but I  don't know if I'm wasting to much of  
> the data-memory. So the question is: Does somebody know what is the  
> maximum size of the software stack during the nested call of library  
> functions? Another thing I'm thinking about is the hardware-stack for  
> the PIC I'm using. It has 31 "entries" and so I need to know how deep  
> is the deepest level of function nesting within the library. Does  
> anyone checked that out yet?
>
> Thank you for your answers in advance,
>
> Martin
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> Efsl-devel mailing list
> Efs...@li...
> https://lists.sourceforge.net/lists/listinfo/efsl-devel
>
> !DSPAM:4534ae7262551384716492!
>
>   

 From my experience, you can corrupt your file system if you are 
updating the FAT tables when either power goes off or the card is removed.
Any files that are currently opened that have not had their directory 
entries update wont be saved either.
Depending on what kind of data loss you can take, you could update the 
file system updating the files directory entriy (basically the filesize 
in the directory if you don't have time/date enabled), and by calling 
fs_flushFs on a timer to write out any data.

If you do that, you can use the power down to lock out any further 
writes to the SD to prevent data corruption. Depending on the circuit 
(and how fast you can write), you could use the switch to actually close 
out the files and update the file system, but that would require you to 
implement either a soft power down or enough capacitance to finish out 
the process of writing it all out.

In any event, I don't know what could be done to prevent corruption in 
all cases if the user is simply pulling the SD out at any random interval.

Daniel

Etienne Force-A wrote:
> Hi all,
>
> I have some questions about what may happen while someone disconnect the 
> SD card (on a AVR) while the filesystem is mounted :
> _ what is the loss of data ? (all ? only the last bloc ?)
> _ how can it be detected ? (wrong number of written bytes for example ?)
> _ How can it be avoided ?
>
> Actually, I'm defining my system's supply architecture (the source is a 
> battery because it's an embedded sensor). The simplest way of shutting 
> down the system would be a classic switch. But the use of a filesystem 
> implies to prevent the AVR that the system will stop.
>
> Any experience that could help me ?
>
> Thanks,
> Etienne
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> Efsl-devel mailing list
> Efs...@li...
> https://lists.sourceforge.net/lists/listinfo/efsl-devel
>
> !DSPAM:4534b02b64314736210067!
>
>   

Hi!

Did somebody of you ever make a call tree for the EFSL? I'm trying to  
port the Library to a PIC18f6680 using the C18-Compiler from  
Microchip. In the standard configuration settings the PIC only uses  
one memory bank for the software stack which is 256 bytes in total. I  
got some strange errors while trying to do read and write-operations  
using a simple test-programm.  After a long time of frustrating  
debugging i noticed that the MCU sopped his operation on a totally  
uncritical point while trying to make a simple function-call. So I  
began thinking about that the stack was to small to hold all the  
function parameters and local variables during a nested function  
call. So I changed the size of the software-stack to 512 bytes what  
did solve the problem, but I  don't know if I'm wasting to much of  
the data-memory. So the question is: Does somebody know what is the  
maximum size of the software stack during the nested call of library  
functions? Another thing I'm thinking about is the hardware-stack for  
the PIC I'm using. It has 31 "entries" and so I need to know how deep  
is the deepest level of function nesting within the library. Does  
anyone checked that out yet?

Thank you for your answers in advance,

Martin

Hi all,

I have some questions about what may happen while someone disconnect the 
SD card (on a AVR) while the filesystem is mounted :
_ what is the loss of data ? (all ? only the last bloc ?)
_ how can it be detected ? (wrong number of written bytes for example ?)
_ How can it be avoided ?

Actually, I'm defining my system's supply architecture (the source is a 
battery because it's an embedded sensor). The simplest way of shutting 
down the system would be a classic switch. But the use of a filesystem 
implies to prevent the AVR that the system will stop.

Any experience that could help me ?

Thanks,
Etienne

Hi all

I'm powering the ATmega128 with 5 Volts because other peripherical need=20
TTL levels. I put a specific 3.3V regulator to feed the SD (and a GPS),=20
the voltage levels are simply adapted with a tension divider (3k3 //=20
2k2) for PB0/CS ; PB1/SCLK and PB2/Din. For the Dout of the SD, the 3.3V=20
high level if enough to be understood like a *1* by the ATmega. In my=20
case it work well with the full-speed SPI (fsck/2 with ATmega + 16 MHz=20
quartz) :)
But I must say that with a level shifter it will be more secure,=20
especially to 'buffered' the Dout of the SD that is directly connected=20
to the ATmega... See the 74LVC245 and the MAX3392E from Maxim.

Etienne

Alvaro Aguirre a =E9crit :
> With "converting the voltage" you mean a tension divider made with two=20
> resistors? ANY resistors in the SPI channel will destroy the signal=20
> integrity. Every time you need to make a voltage convertion in hi speed=
=20
> buses you must use open collector buffers (74hc244? memory leak :P).
>=20
> Alvaro
>=20
> On 9/6/06, *Jim Lowe* <lo...@gm... <mailto:lo...@gm...>> wr=
ote:
>=20
>     How are you handling the voltage level matching?  The ATMega128 is =
a
>     5 volt device which is above the nominal limit of an SD card.  I
>     attemped to get this to work converting the voltage, but had very
>     limited success with the SPI interface.  I found that merely
>     switching to the 128L which is a 3.3 vdc device made the integratio=
n
>     of the EFSL library and SD card seamless.
>     =20
>     Jim
>=20
>     =20
>     On 9/6/06, *Lennart Yseboodt* <len...@sk...
>     <mailto:len...@sk...>> wrote:
>=20
>         Hi Etienne,
>=20
>         I'm forwarding this mail to the mailinglist.
>         Almost every case where you get a low level error from the
>         driver means
>         some problem
>         with the hardware, badly connected cables being most notorious.
>         Michael wrote this driver, maybe he has some insight to share.
>=20
>         Regarding devel version: it offers few benefits to 0.2.x at the
>         moment
>         and is best avoided,
>         unless you want to develop for it/do tests.
>=20
>         Regards,
>=20
>         Lennart
>=20
>         Etienne Force-A wrote:
>         >  Hello,
>         >
>         >  I'm curently developping software for an application on a
>         ATmega128
>         >  (an optical sensor for agriculture). I needed space to backu=
p
>         data and
>         >  first used the mmc.c from the AVRlib to manage a 128 Mo SD
>         card. That
>         >  work nice but I decided to go further and implementing a
>         filesystem. I
>         >  get the EFSL and try using it, I have some questions about i=
t :
>         >  _ and first one : does the EFSL-mailing-list exist ?
>         >  _ How stable is the dev version of EFSL (0.3.5) ?
>         >  _ I have problems with the hardware functions for the AVR/SD=
,
>         I get a
>         >  "Unknown error 0xe0 (see SanDisk docs p5-13)." from the sd_I=
nit
>         >  function  : which documentation should I get to debug ? (And
>         have you
>         >  ever see this error ? I try to slow down the SPI clock witho=
ut a
>         >  success... I don't know where could be the 'bug' : the sourc=
es are
>         >  very similar to the AVRlib which work well).
>         >
>         >  Anyway, the source code is very clear, I think I will add th=
e
>         AVRlib
>         >  functions as a new endpoint even if I had prefered to use th=
e
>         EFSL
>         >  globally.
>         >
>         >  Thanks again for your support,
>         >  Etienne Besan=E7on
>         >
>         >
>=20
>=20
>         ---------------------------------------------------------------=
----------
>         Using Tomcat but need to do more? Need to support web services,
>         security?
>         Get stuff done quickly with pre-integrated technology to make
>         your job easier
>         Download IBM WebSphere Application Server v.1.0.1 based on
>         Apache Geronimo
>         http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D26=
3057&dat=3D121642
>         <http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D2=
63057&dat=3D121642>
>         _______________________________________________
>         Efsl-devel mailing list
>         Efs...@li...
>         <mailto:Efs...@li...>
>         https://lists.sourceforge.net/lists/listinfo/efsl-devel
>=20
>=20
>=20
>     -------------------------------------------------------------------=
------
>     Using Tomcat but need to do more? Need to support web services,
>     security?
>     Get stuff done quickly with pre-integrated technology to make your
>     job easier
>     Download IBM WebSphere Application Server v.1.0.1 based on Apache
>     Geronimo
>     http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057=
&dat=3D121642
>     <http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D26305=
7&dat=3D121642>
>=20
>     _______________________________________________
>     Efsl-devel mailing list
>     Efs...@li...
>     <mailto:Efs...@li...>
>     https://lists.sourceforge.net/lists/listinfo/efsl-devel
>     <https://lists.sourceforge.net/lists/listinfo/efsl-devel>
>=20
>=20
>=20
>=20
> -----------------------------------------------------------------------=
-
>=20
> -----------------------------------------------------------------------=
--
> Using Tomcat but need to do more? Need to support web services, securit=
y?
> Get stuff done quickly with pre-integrated technology to make your job =
easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geron=
imo
> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=
=3D121642
>=20
>=20
> -----------------------------------------------------------------------=
-
>=20
> _______________________________________________
> Efsl-devel mailing list
> Efs...@li...
> https://lists.sourceforge.net/lists/listinfo/efsl-devel

Lennart:
I check the code i was using with my atmega because it was working ok. I
found i unconsciously bypass the error founded by klavi. Could you patch it
and upload a new relesase?

On 9/6/06, Alvaro Aguirre <alv...@gm...> wrote:
>
> With "converting the voltage" you mean a tension divider made with two
> resistors? ANY resistors in the SPI channel will destroy the signal
> integrity. Every time you need to make a voltage convertion in hi speed
> buses you must use open collector buffers (74hc244? memory leak :P).
>
> Alvaro
>
>
> On 9/6/06, Jim Lowe <lo...@gm...> wrote:
> >
> > How are you handling the voltage level matching?  The ATMega128 is a 5
> > volt device which is above the nominal limit of an SD card.  I attemped=
 to
> > get this to work converting the voltage, but had very limited success w=
ith
> > the SPI interface.  I found that merely switching to the 128L which is =
a
> > 3.3 vdc device made the integration of the EFSL library and SD card
> > seamless.
> >
> > Jim
> >
> >
> > On 9/6/06, Lennart Yseboodt <len...@sk... > wrote:
> > >
> > > Hi Etienne,
> > >
> > > I'm forwarding this mail to the mailinglist.
> > > Almost every case where you get a low level error from the driver
> > > means
> > > some problem
> > > with the hardware, badly connected cables being most notorious.
> > > Michael wrote this driver, maybe he has some insight to share.
> > >
> > > Regarding devel version: it offers few benefits to 0.2.x at the momen=
t
> > >
> > > and is best avoided,
> > > unless you want to develop for it/do tests.
> > >
> > > Regards,
> > >
> > > Lennart
> > >
> > > Etienne Force-A wrote:
> > > > Hello,
> > > >
> > > > I'm curently developping software for an application on a ATmega128
> > > > (an optical sensor for agriculture). I needed space to backup data
> > > and
> > > > first used the mmc.c from the AVRlib to manage a 128 Mo SD card.
> > > That
> > > > work nice but I decided to go further and implementing a filesystem=
.
> > > I
> > > > get the EFSL and try using it, I have some questions about it :
> > > > _ and first one : does the EFSL-mailing-list exist ?
> > > > _ How stable is the dev version of EFSL (0.3.5) ?
> > > > _ I have problems with the hardware functions for the AVR/SD, I get
> > > a
> > > > "Unknown error 0xe0 (see SanDisk docs p5-13)." from the sd_Init
> > > > function  : which documentation should I get to debug ? (And have
> > > you
> > > > ever see this error ? I try to slow down the SPI clock without a
> > > > success... I don't know where could be the 'bug' : the sources are
> > > > very similar to the AVRlib which work well).
> > > >
> > > > Anyway, the source code is very clear, I think I will add the AVRli=
b
> > > > functions as a new endpoint even if I had prefered to use the EFSL
> > > > globally.
> > > >
> > > > Thanks again for your support,
> > > > Etienne Besan=E7on
> > > >
> > > >
> > >
> > >
> > >
> > > ---------------------------------------------------------------------=
----
> > > Using Tomcat but need to do more? Need to support web services,
> > > security?
> > > Get stuff done quickly with pre-integrated technology to make your jo=
b
> > > easier
> > > Download IBM WebSphere Application Server v.1.0.1 based on Apache
> > > Geronimo
> > >
> > > http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&d=
at=3D121642
> > > _______________________________________________
> > > Efsl-devel mailing list
> > > Efs...@li...
> > > https://lists.sourceforge.net/lists/listinfo/efsl-devel
> > >
> >
> >
> >
> > -----------------------------------------------------------------------=
--
> > Using Tomcat but need to do more? Need to support web services,
> > security?
> > Get stuff done quickly with pre-integrated technology to make your job
> > easier
> > Download IBM WebSphere Application Server v.1.0.1 based on Apache
> > Geronimo
> > http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=
=3D121642
> >
> > _______________________________________________
> > Efsl-devel mailing list
> > Efs...@li...
> > https://lists.sourceforge.net/lists/listinfo/efsl-devel
> >
> >
> >
>

With "converting the voltage" you mean a tension divider made with two
resistors? ANY resistors in the SPI channel will destroy the signal
integrity. Every time you need to make a voltage convertion in hi speed
buses you must use open collector buffers (74hc244? memory leak :P).

Alvaro

On 9/6/06, Jim Lowe <lo...@gm...> wrote:
>
> How are you handling the voltage level matching?  The ATMega128 is a 5
> volt device which is above the nominal limit of an SD card.  I attemped t=
o
> get this to work converting the voltage, but had very limited success wit=
h
> the SPI interface.  I found that merely switching to the 128L which is a
> 3.3 vdc device made the integration of the EFSL library and SD card
> seamless.
>
> Jim
>
>
> On 9/6/06, Lennart Yseboodt <len...@sk...> wrote:
> >
> > Hi Etienne,
> >
> > I'm forwarding this mail to the mailinglist.
> > Almost every case where you get a low level error from the driver means
> > some problem
> > with the hardware, badly connected cables being most notorious.
> > Michael wrote this driver, maybe he has some insight to share.
> >
> > Regarding devel version: it offers few benefits to 0.2.x at the moment
> > and is best avoided,
> > unless you want to develop for it/do tests.
> >
> > Regards,
> >
> > Lennart
> >
> > Etienne Force-A wrote:
> > > Hello,
> > >
> > > I'm curently developping software for an application on a ATmega128
> > > (an optical sensor for agriculture). I needed space to backup data an=
d
> > > first used the mmc.c from the AVRlib to manage a 128 Mo SD card. That
> > > work nice but I decided to go further and implementing a filesystem. =
I
> >
> > > get the EFSL and try using it, I have some questions about it :
> > > _ and first one : does the EFSL-mailing-list exist ?
> > > _ How stable is the dev version of EFSL (0.3.5) ?
> > > _ I have problems with the hardware functions for the AVR/SD, I get a
> > > "Unknown error 0xe0 (see SanDisk docs p5-13)." from the sd_Init
> > > function  : which documentation should I get to debug ? (And have you
> > > ever see this error ? I try to slow down the SPI clock without a
> > > success... I don't know where could be the 'bug' : the sources are
> > > very similar to the AVRlib which work well).
> > >
> > > Anyway, the source code is very clear, I think I will add the AVRlib
> > > functions as a new endpoint even if I had prefered to use the EFSL
> > > globally.
> > >
> > > Thanks again for your support,
> > > Etienne Besan=E7on
> > >
> > >
> >
> >
> >
> > -----------------------------------------------------------------------=
--
> > Using Tomcat but need to do more? Need to support web services,
> > security?
> > Get stuff done quickly with pre-integrated technology to make your job
> > easier
> > Download IBM WebSphere Application Server v.1.0.1 based on Apache
> > Geronimo
> > http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=
=3D121642
> > _______________________________________________
> > Efsl-devel mailing list
> > Efs...@li...
> > https://lists.sourceforge.net/lists/listinfo/efsl-devel
> >
>
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job
> easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronim=
o
> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=
=3D121642
>
> _______________________________________________
> Efsl-devel mailing list
> Efs...@li...
> https://lists.sourceforge.net/lists/listinfo/efsl-devel
>
>
>

How are you handling the voltage level matching?  The ATMega128 is a 5 volt
device which is above the nominal limit of an SD card.  I attemped to get
this to work converting the voltage, but had very limited success with the
SPI interface.  I found that merely switching to the 128L which is a 3.3 vd=
c
device made the integration of the EFSL library and SD card seamless.

Jim

On 9/6/06, Lennart Yseboodt <len...@sk...> wrote:
>
> Hi Etienne,
>
> I'm forwarding this mail to the mailinglist.
> Almost every case where you get a low level error from the driver means
> some problem
> with the hardware, badly connected cables being most notorious.
> Michael wrote this driver, maybe he has some insight to share.
>
> Regarding devel version: it offers few benefits to 0.2.x at the moment
> and is best avoided,
> unless you want to develop for it/do tests.
>
> Regards,
>
> Lennart
>
> Etienne Force-A wrote:
> > Hello,
> >
> > I'm curently developping software for an application on a ATmega128
> > (an optical sensor for agriculture). I needed space to backup data and
> > first used the mmc.c from the AVRlib to manage a 128 Mo SD card. That
> > work nice but I decided to go further and implementing a filesystem. I
> > get the EFSL and try using it, I have some questions about it :
> > _ and first one : does the EFSL-mailing-list exist ?
> > _ How stable is the dev version of EFSL (0.3.5) ?
> > _ I have problems with the hardware functions for the AVR/SD, I get a
> > "Unknown error 0xe0 (see SanDisk docs p5-13)." from the sd_Init
> > function  : which documentation should I get to debug ? (And have you
> > ever see this error ? I try to slow down the SPI clock without a
> > success... I don't know where could be the 'bug' : the sources are
> > very similar to the AVRlib which work well).
> >
> > Anyway, the source code is very clear, I think I will add the AVRlib
> > functions as a new endpoint even if I had prefered to use the EFSL
> > globally.
> >
> > Thanks again for your support,
> > Etienne Besan=E7on
> >
> >
>
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job
> easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronim=
o
> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=
=3D121642
> _______________________________________________
> Efsl-devel mailing list
> Efs...@li...
> https://lists.sourceforge.net/lists/listinfo/efsl-devel
>

Hi all !

Thanks for your support Lennart & Alvaro, I search topics on the sf=20
forum related to my problem and find the bug=20
(http://sourceforge.net/forum/forum.php?thread_id=3D1360625&forum_id=3D44=
7844)=20
. Actually the if_spiInit function (src/interfaces/atmega128.c in 0.2.7)=20
only works for MMC card (which can be 'selected' during init) contrary=20
to SD card.
The fix proposed by klavi (cf link) works well !

Regards,
Etienne

Lennart Yseboodt a =E9crit :
> Hi Etienne,
>=20
> I'm forwarding this mail to the mailinglist.
> Almost every case where you get a low level error from the driver means
> some problem
> with the hardware, badly connected cables being most notorious.
> Michael wrote this driver, maybe he has some insight to share.
>=20
> Regarding devel version: it offers few benefits to 0.2.x at the moment
> and is best avoided,
> unless you want to develop for it/do tests.
>=20
> Regards,
>=20
> Lennart
>=20
> Etienne Force-A wrote:
>> Hello,
>>
>> I'm curently developping software for an application on a ATmega128
>> (an optical sensor for agriculture). I needed space to backup data and
>> first used the mmc.c from the AVRlib to manage a 128 Mo SD card. That
>> work nice but I decided to go further and implementing a filesystem. I
>> get the EFSL and try using it, I have some questions about it :
>> _ and first one : does the EFSL-mailing-list exist ?
>> _ How stable is the dev version of EFSL (0.3.5) ?
>> _ I have problems with the hardware functions for the AVR/SD, I get a
>> "Unknown error 0xe0 (see SanDisk docs p5-13)." from the sd_Init
>> function  : which documentation should I get to debug ? (And have you
>> ever see this error ? I try to slow down the SPI clock without a
>> success... I don't know where could be the 'bug' : the sources are
>> very similar to the AVRlib which work well).
>>
>> Anyway, the source code is very clear, I think I will add the AVRlib
>> functions as a new endpoint even if I had prefered to use the EFSL
>> globally.
>>
>> Thanks again for your support,
>> Etienne Besan=E7on
>>
>>
>=20
>=20

2005	Jan	Feb	Mar	Apr	May	Jun	Jul	Aug (4)	Sep (7)	Oct (14)	Nov (14)	Dec (8)
2006	Jan (8)	Feb (5)	Mar	Apr	May	Jun (3)	Jul (17)	Aug	Sep (6)	Oct (6)	Nov	Dec
2007	Jan (1)	Feb	Mar (2)	Apr (1)	May	Jun	Jul	Aug (8)	Sep	Oct	Nov	Dec
2008	Jan	Feb	Mar	Apr (2)	May (2)	Jun (4)	Jul (5)	Aug (2)	Sep (12)	Oct (8)	Nov (24)	Dec (32)
2009	Jan (14)	Feb (15)	Mar (11)	Apr (32)	May (25)	Jun (14)	Jul (32)	Aug (7)	Sep (4)	Oct (4)	Nov (5)	Dec (9)
2010	Jan (8)	Feb (7)	Mar (12)	Apr (2)	May	Jun (8)	Jul (8)	Aug (18)	Sep (17)	Oct	Nov	Dec
2016	Jan	Feb	Mar	Apr	May	Jun	Jul	Aug	Sep (2)	Oct	Nov	Dec

efsl-devel Mailing List for Embedded Filesystems Library (Page 3)

efsl-devel — General mailing list for efsl development and questions/problems.