Dmitry

XDMA data corruption issue (0xFFFFFFFF every other read) was fixed yet was not mentioned anywhere

5

**TL;DR: if you have the same issue, you should upgrade to Vivado 2019.1 or newer. Or if you are a Xilinx/AMD employee, then you should write your documentation better** I...

Fix ERESTARTSYS leak to userspace

3

Fix ERESTARTSYS leak to userspace by replacing it with the appropriate for userspace error codes. Fixes #320

ERESTARTSYS is intended for internal kernel purposes and should not leak to userspace. The driver should return error codes that are standard in userspace. This is EAGAIN (Try Again) for...

In [commit 748af10](https://github.com/Xilinx/dma_ip_drivers/commit/748af10d85aedeec5baf076dfa8715060aceb404) a fix for wait_event_interruptible_timeout was introduced to handle ERESTARTSYS error that causes premature return from wait and thus false timeout errors and data losses. The same fix...

In XDMA in Channel Identifier registers there is version field that is documented like that: ``` Version 8'h01: 2015.3 and 2015.4 8'h02: 2016.1 8'h03: 2016.2 8'h04: 2016.3 8'h05: 2016.4 8'h06:...

Always (i.e. don't require a build with LIBXDMA_DEBUG flag) print "Engine was not running!!! Clearing status\n" message as it simplifies diagnostics. Fixes #312

Fix buffer overflow in engine_status_dump() function

1

fixes #314, also fixes a typo (INVLIAD_LEN => INVALID_LEN)

In function engine_status_dump() there is char buffer[256]. If engine status is 0xFFFFFFFF, which means all possible flags, calling engine_status_dump() causes buffer overflow and kernel crash. This is impossible in normal...

Add wrappers to handle ERESTARTSYS return code from both wait_* and swait_* functions, fixes #311

In current code, message "Engine was not running!!! Clearing status\n" is printed with dbg_tfr macro that only prints when module is built with __LIBXDMA_DEBUG__ flag. However, building with this flag...