ChibiOS Free Embedded RTOS

See my second point, I think that is the problem.

The code looks OK but without trying it I cannot be sure. I wish to was able to tell if some code is OK by just looking at it

Giovanni

Tectu wrote:

Jeroen3 wrote:I''m at 4e11d82, "Fixed bug #514". (git)

As far as I know that repo state even uses the 'old' FatFS library. The new one is 0.10b.
Personally don't have issues to use SDHC cards. I didn't have issues with them in the past and with the latest master.


~ Tectu

I tried to use the 0.10b (ffrom original site) , but receive some errors during compilation, need 'patched' version? where i found?
or i must do ssomething (not only overwrite the old llibrary)?

Compiling fatfs_diskio.c
Compiling fatfs_syscall.c
Compiling ff.c
c:/ChibiStudio/chibios/os/various/fatfs_bindings/fatfs_diskio.c:111:9: error: conflicting types for 'disk_read'
In file included from c:/ChibiStudio/chibios/os/various/fatfs_bindings/fatfs_diskio.c:11:0:
c:/ChibiStudio/chibios/ext/fatfs/src/diskio.h:37:9: note: previous declaration of 'disk_read' was here
c:/ChibiStudio/chibios/os/various/fatfs_bindings/fatfs_diskio.c:152:9: error: conflicting types for 'disk_write'
In file included from c:/ChibiStudio/chibios/os/various/fatfs_bindings/fatfs_diskio.c:11:0:
c:/ChibiStudio/chibios/ext/fatfs/src/diskio.h:38:9: note: previous declaration of 'disk_write' was here
make: *** [build/obj/fatfs_diskio.o] Error 1
make: *** Waiting for unfinished jobs....

Please try using FatFS provided with ChibiOS, there could be patches.

Giovanni

with the fatfs-0.9-patched.zip (in the etc\ directory) not have problem with compilation, but thinked can use the 0.10b from the orginal site
but still have problem with mmcard, anyway today i have time to look what happen

thanks

I've got more card features running. And I wanted to ask if ~220 KB/s of file copy operation is reasonable?
Running with 3 open files. Copy A->B with files of 1 to 10 MB. And a continuous data logging file in the background.

Consider that FatFS is done to be small not necessarily fast, it lacks any kind of buffering so the library continuously reads/writes FAT blocks. It is unlikely that you will reach the transfer rate of the naked SPI port.

Giovanni

Correct, It doesn't have buffering. That would be waste of the already limited resources.
I don't need it to be ultra fast, I was just wondering if this is a reasonable speed. Seems like it is.

It also seems like the card itself is the limiting factor since the chRegistry says copying 10 MB uses 1 unit of cpu, 1 ms, of the copying thread.
Or everything is running in interrupt mode.

Everything runs on DMA and IRQs so very little CPU time is involed.

Giovanni

Meanwhile, I can confirm writing and reading from an 8GB SanDisk SDHC card works*.
However, getting metadata doesn't. It won't read the filesize correctly anymore. It does with a FILINFO.
It seems after f_open, the fsize isn't being set correctly if the file is beyond 4 GB.

That is probably a FatFS limitation.

Giovanni