Wi-Fi Combo Forum Discussions
Hi there, I am new to socket programming so wondering whether is it possible to port libcurl or rest interface to wicked sdk, so that which will ease developing client side interface. I need to send HTTP request messages and parse responses.
Show LessHi gangi,
Our customer has met 'bootloader' failed issue while downloading image into BCM94390WCD2.
Maybe, customer had met this issue while trying to test OTA function.
I have no idea to solve this issue. Could you let me know what I should do?
Test environments are as follows,
Module and EVB version : BCM94390WCD2 / BCM9WCD3EVAL1
SDK : SDK 3.1.1
application : snip.scan-BCM94390WCD2 download run
Compile and Download error log :
**** Build of configuration Default for project Wiced-SDK ****
C:\WICED-SDK-3.1.1\Wiced-SDK\make.exe snip.scan-BCM94390WCD2 download run
Making .gdbinit
Building Bootloader
Building Serial Flash Loader App
Finished Building Bootloader
................
Creating Filesystem
Downloading Bootloader ...
Building apps lookup table
platforms/BCM94390WCD2/BCM94390WCD2_targets.mk:111: recipe for target 'download_bootloader' failed
make.exe[1]: *** [download_bootloader] Error 1
make.exe[1]: *** Waiting for unfinished jobs....
make: *** [main_app] Error 2
Makefile:185: recipe for target 'main_app' failed
Thanks,
Sung-Mok Lee
Show LessThe JSON parser library does not seem to support negative "numbers", the "-" is cut of the object string before the user callback been
invoked.
This can be fixed by including the minus sign as part of the object strings, replacing the following tests in multiple places of the JSON parser lib
/* Move value_start token until we encounter a non-digit value */
- while ( ( ( *value_start >= '0' ) && ( *value_start <= '9') ) || ( *value_start == '.' ) )
by
while ( ( ( *value_start >= '0' ) && ( *value_start <= '9') ) || ( *value_start == '.' ) || ( *value_start == '-' )
Show LessHi @ all!
My HW (SN8000) try to join to an AP (LinkSys). Used mode is WPA2 and the HW is only activated as client.
Now I find a strange behavior, when I change the channel on the AP (including power cycle for clean startup)
- Channel 1-6 and 13 works fine (join takes 5-6s)
- On channel 7-10 I have a long time (~15s and more) until the join is complete, sometimes it fails
=> Is there any settings or software behavior for different channels?
=> Anyone see the same problems? (or similar)
Thanks for helping,
Jan Gottwald
Show LessHi,
We saw a major issue on calling wiced_tcp_disconnect with tis context that will lead to system crash.. This can be very easily reproduced on the demo code snip.https_client sample in the SDK.
Best Regards,
Aaron Lee
Show LessHi,
For our project, I have developed the DMA I2S API/driver in WICED SDK 2.4.1. We use the Inventek module ISM43362_M3G_L44 which uses the STM32F205 processor . We configure the SPI2 as the I2S master. When I tried to port it from SDK 2.4.1 to SDK 3.1.1, I found it doesn’t work and my application is locked up and keeps rebooting. Can you provide followings
1. How to register my I2S_dma_irq() in SDK3.1.1. We use DMA1_stream3. I can’t find any ISR registration example code in SDK 3.1.1.
2. I also found that SPI2 is already used for wifi gSPI bus in SDK 3.1.1. Maybe this will cause the conflict to our i2s pin configuration. How can I disable this wifi gSPI bus because we will use spi2 as i2s bus interface and we always the sdio bus for wifi bus?
Thanks,
Wenjian
Show LessUsing SDK 3.1.2 and the Macronix MX25Lxxxx series of flash, you may have noticed that write operation is very slow.
Bytes are in fact written one by one, with a lots of overhead on the SPI, here is how to speed it up:
spi_flash.c in sflash_write()
Line 163 change:
max_write_size = (unsigned int) 1; /* TODO: this should be 256, but that causes write errors */
To:
max_write_size = (unsigned int) 256;
and later, line 197 replace:
write_size = ( size > max_write_size )? max_write_size : size;
By:
if (device_address % max_write_size) {
/* Not on a page boundary, limit write within the page*/
unsigned int nbBytesLeftOnPage = max_write_size - (device_address % max_write_size);
write_size = (size > nbBytesLeftOnPage) ? nbBytesLeftOnPage : size;
} else {
/* Page boundary, write as many bytes as we can */
write_size = size > max_write_size ? max_write_size : size;
}
Message was edited by: Guillaume Brivet Fix logic when write is not page-aligned
Show LessHello,
I'm currently using SDK_3.1.1 and likely find a bug in "wicked_apps_erase" when calling "wiced_framework_app_erase".
The api is used to erase specified application depended on app header location was stored in DCT, if I understand the code correctly, the problem is using "app_header.sectors[ index ].start" as maximum sectors need to be erased, it's not quite right, but should be app_header.sectors[ index ].count instead, otherwise unexpected sectors will be erased.
It's currently modification I think:
wiced_result_t wiced_apps_erase( const image_location_t* app_header_location )
{
sflash_handle_t sflash_handle;
app_header_t app_header;
uint8_t index;
/* Loop on each entry and erase the sectors*/
init_sflash( &sflash_handle, PLATFORM_SFLASH_PERIPHERAL_ID, SFLASH_WRITE_ALLOWED );
sflash_read( &sflash_handle, app_header_location->detail.external_fixed.location, &app_header, sizeof(app_header_t) );
for ( index = 0; index < app_header.count; index++ )
{
unsigned long sector;
#if 0 //Original
for ( sector = 0; sector < app_header.sectors[ index ].start; sector++ )
#else //Fix
for ( sector = 0; sector < app_header.sectors[ index ].count; sector++ )
#endif
{
sflash_sector_erase( &sflash_handle, ( app_header.sectors[ index ].start + sector ) * SFLASH_SECTOR_SIZE );
}
}
return WICED_SUCCESS;
}
--
Andy
Show Less