[spi_device] fifo2sram_adapter enhancement

[eunchan]

#12053 revises the spid_upload module to create an overflow event and to track the pointer of the FIFO.

It would be easier if spid_fifo2sram_adapter (eventually prim_fifo2sram_adapter) provides the overflow event at the same cycle when the producer side pushes an entry when the FIFO is full.

Issue here is:

• fifo2sram only writes data to SRAM no way to track the consumer side.
• SPI_DEVICE upload logic requires to keep update the data in SRAM even in full case.

Plan:

• At lease, for now, adds clear port to reset the pointer (currently rst_ni does that)
• Make the module configurable to be able to write when full (currently this is the default behavior)

[andreaskurth]

Triaged for spi_device. I think this is not relevant for this release because we have a mechanism that works, though this issue proposes a cleaner implementation IIUC. Thus keeping Type:FutureRelease Not relevant to currently planned releases/milestones .

[msfschaffner]

CC @a-will

[a-will]

Currently, spid_fifo2sram_adapter is only used in contexts where SPI-side overflow is not possible. For flash, wrapping is the correct behavior, and for TPM, we stop writing once the requested amount of data has been received.

For the TPM Read FIFO, overflow would cause bizarre behavior, but software has an explicit length that only lasts for the current command. We probably don't need to go the extra mile to protect software from this specific case.

Closing, as there is no planned use case. Please re-open if I've made an error.