Replying to my own email to close this out...
This problem was solved by using appropriately sized usb transfer
buffer, with this commit:
https://github.com/sigrokproject/libsigrok/commit/e847645b7be832c1d62d1ad2312e3a3d4de467ce
The PV nightly builds for linux and windows are working now for Kingst
LA1016/LA2016. macOS not yet tested AFAIK.
On 2021-09-14 18:59, Kevin Grant wrote:
Hello,
Replying to my own email here.
It looks like libusb folks have discussed enabling RAW_IO but have so
far decided not to (Jan2021), as it seemingly adds some complexity to
the driver operation:
https://github.com/libusb/libusb/issues/490
Furthermore, some quotes from here:
https://docs.microsoft.com/en-us/windows-hardware/drivers/usbcon/winusb-functions-for-pipe-policy-modification
AUTO_CLEAR_STALL This policy is useful only when you have multiple
pending read requests to the endpoint when RAW_IO is disabled.
RAW_IO imposes certain restrictions on the buffer that is passed by the
caller:
The buffer length must be a multiple of the maximum endpoint packet
size.
The length must be less than or equal to the value of
MAXIMUM_TRANSFER_SIZE retrieved by WinUsb_GetPipePolicy.
So, AUTO_CLEAR_STALL is no longer effective due to RAW_IO, and that
perhaps causes the Kingst EP6IN read to stall / fail.
I propose that we remove RAW_IO pipe policy from mxe-fixes.patch
Comments welcome.
Regards,
Kevin
On 2021-09-14 16:54, Kevin Grant wrote:
Hello,
I made some changes to the Kingst LA2016 logic analyser driver which
were pulled in the last few days.
https://github.com/sigrokproject/libsigrok/pull/131
I have tested the nightly build of PulseView for Linux and both the
LA2016 and LA1016 are working, which is good.
However, the Windows build does not work. At the point of reading
capture data from EP6IN, PulseView stalls and crashes. I can fix this
problem by removing part of the libusb patch:
https://github.com/sigrokproject/sigrok-util/blob/master/cross-compile/mingw/mxe_fixes.patch
Specifically, commenting out these lines fixes the problem:
++ if (!WinUSBX[sub_api].SetPipePolicy(winusb_handle,
endpoint_address,
++ RAW_IO, sizeof(UCHAR), &policy)) {
++ usbi_dbg("failed to enable RAW_IO for endpoint %02X",
endpoint_address);
++ }
So, setting RAW_IO policy causes problems for me. Does anybody know
what this patch was intended to fix and if there is a Kingst LA driver
change required to support it?
For reference, the Kingst driver code is here:
https://github.com/sigrokproject/libsigrok/tree/master/src/hardware/kingst-la2016
Thanks,
Kevin
_______________________________________________
sigrok-devel mailing list
sigrok-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sigrok-devel
_______________________________________________
sigrok-devel mailing list
sigrok-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sigrok-devel
_______________________________________________
sigrok-devel mailing list
sigrok-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sigrok-devel
