Re: [lwip-devel] PPPOS and PBUF_POOL

[Joel Cunningham]

On Aug 31, 2017, at 1:50 PM, address@hidden wrote:

Axel Lin wrote:

I think it's to avoid memory fragment.
The way pppos_write/pppos_netif_output using the PBUF_POOL is different from
other places because the pbuf is freed immediately by pppos_output_last().

Hmm, allocating from the heap and directly freeing it again does not produce fragmentation: in fact, it doesn't change anything.
I don't know why it is that way. I should probably be changed.

I studied the PPPOS usage a little more and I’m wondering if PBUF_POOL was chosen because it’s not known how much the payload will expand once escaped and so the implementer used PBUF_POOL as a means of determining the “largest packet we can send”.  See pppos_output_append()

I’ll file a bug for this issue.

Joel