Skip to contentSkip to conversation

fpga manager framework with configfs interface

RFC PATCH H. Peter Anvin 2 participants 2 messages May 13, 2014 – Jun 25, 2014 13 messages not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

#5 lore raw
Quoted 9 lines · Pavel Machek

Hi!

Quoted 7 lines
Quoted 4 lines

Echoing filenames into files in configfs is strange at the very least. Think chroot, and what about permissions...?

It would be much better to just cat bitstream into configfs...

No binary attributes supported for configfs. Apparently this is done on purpose.

Quoted 9 lines
Quoted 5 lines

In fact it was my first idea about the configfs interface for DT overlays but decided to go with firmware file names instead.

If this is valid use case (which seems to be) would patches adding binary attributes to configfs be accepted?

I'd say this is very valid use, but I'm not configfs maintainer. It is certainly better then echoing filenames there.

Almost anything is better...

-hpa

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): CANk1AXQcqzu16OaT3LziWaroeyyuUoXU1CQy4e6BoGKLOk6fSg@mail.gmail.com

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): CANk1AXS+JZCv7CMJW956PvnJ_YvebyyLVm_TkCPZr5TX1Wb4Yw@mail.gmail.com

A message that is not in the archive (sent off-list or not yet received): [email protected]

#14 lore raw
Quoted 8 lines · H. Peter Anvin
Quoted 6 lines · Alan Tull

Yes, that's why I've recently felt more love for the firmware interface. Since I need resume, it gives me a straightforward path to be able to reprogram the FPGA without holding ram for the bitfile. Also it is scalable.

Quoted 7 lines

The firmware interface is definitely pretty much ideal for the "production driver" use case... after all, the fact that the firmware happens to be an FPGA is to some degree an implementation detail (well, modulo fun things like showing up on the bus with one device ID and then disconnecting and reconnecting with another device ID after firmware load.)

There are *so* many use cases for FPGA bitstreams...

Agreed!

Philip

Quoted 5 lines

-hpa