ZED-bare metal SD card write-using drivers-debugging

Folks really need your help with this,

Basically I am attempting to write to a 16 GB sandisk microSD card. following the instructions in this link

http://zedboard.org/content/standalone-sd-card-sdio

Here's a brief of the setup:

h/w:
1) 16GB uSD (San disk), fat32 formatted, default allocation size. Zedboard

2) Vivado:
- zynqPS with SD0 and UART1 enabled. MIO 46-CD, 47-WP, and 38( power) is this necessary?, 'Fast speed' and pullups disabled (ON ALL PINS).

-None of the jumpers on my zedboard are set to boot from SD (i.e. J6 or JP7-JP11).

s/w:

Using Yaro's(from the link) C files with the xsdps drivers that it calls.

Here's my debug attempt so far- I basically stepped through the code and a status failure shows up from:

" XSdps_Get_BusWidth() "

stepping into this code I find that, an error flag is actually set in here:

if (StatusReg & XSDPS_INTR_ERR_MASK) {.....

StatusReg is being set just above this if() statement by reading from the XSDPS_NORM_INTR_STS_OFFSET and is being compared with XSDPS_INTR_ERR_MASK (x8000).

I'm wondering if the BlkCnt of 1 that is passed to the address XSdPs_BLK_SIZE_OFFSET is a problem. Because before this it's always been a value of 0.

.....

Thats about as far as I could get before really being at the end of understanding whats going on and how to fix it.

Please..any directions on how to further debug would be HUGELY appreciated.

0 zedhed over 10 years ago

Hi omkarpradhan,

I don't think that using debug mode should impact what you are trying to do.

What version of tools are you using?

Take a look at this thread here:

http://microzed.org/content/read-txt-matrices-sd-ddr-zedboard

In that thread, the XILFFS was used to access the SD card. On my end, I was at least able to use 2014.4 to build a design where I could run the example code for the XILFFS library on ZedBoard which reads and writes to the SD card.

Regards,

-Kevin
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 Former Member over 10 years ago in reply to zedhed

Kevin,

Like the title says I am using 2014.2 Are you sure the version difference causes some changes?

I noted in the link that I had referenced there was something about disabling data chaches. Any idea how to do this?

Also does my h/w config look correct? Do you have the same setup?

Omkar
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 Former Member over 10 years ago in reply to zedhed

Holy ***, it works now. Man thanks for that tip about the example.

Also don't know if changing the h/w config a bit did it. I'm going to run around the block celebrate and then come back to figure the rest of it out.

I was using a very similar code in another discussion on this forum.

By the way have you been able to read/write text files? If so any experiments with how fast you can read write?

Omkar
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 Former Member over 10 years ago in reply to zedhed

Kevin,

It seems like the default example is set up to write NOT in English. I can't understand the language that it is writing in. I looks to be a mix of chinese/korean script.

Can you tell me where to change this?
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 zedhed over 10 years ago

Hi Omkar,

I am not sure that the xilffs_polled_example.c is setup to write in any specific language.

If I look inside of FfsSdPolledExample() for SDK 2014.4, I can see the following test pattern get setup within a buffer:

#define TEST 7

...

for(BuffCnt = 0; BuffCnt < FileSize; BuffCnt++){
SourceAddress[BuffCnt] = TEST + BuffCnt;
}

That ends up being a bunch of sequential 32-bit word data so it just so happens when you open the file it appears to be NOT in English. If you open the file with a HEX editor, you should see a data pattern starting at 0x00000007 and counting upward from there.

Regards,

-Kevin
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel