mbuf: provide mc_uiotomc() a function to copy from uio(9) to mchain

[glebius]

Summary

Implement m_uiotombuf() as a wrapper around mc_uiotomc(). The M_EXTPG is
left untouched. The m_uiotombuf() is left as a compat KPI. New code
should use either mc_uiotomc() or m_uiotombuf_nomap().

[markj]

sys/kern/uipc_mbuf.c

2014

It would be nice to explain what "len" and "lspace" are.

2022–2023

Can you get rid of progress and just use mc->mc_len?

2029

What's the purpose of trying to handle the overflow case? We are silently truncating the uio and creating a 4GB chain, which presumably isn't what the caller wants...?

[glebius]

Review from Mark.

[markj]

sys/kern/uipc_mbuf.c

2000

I'd also assert that uio->uio_segflg == UIO_WRITE.

[glebius]

Utilize mc_first/etc.
Add assertion uio->uio_segflg == UIO_WRITE.

[glebius]

Correct UIO assertion.

[markj]

sys/kern/uipc_mbuf.c

2000

Sorry, I meant uio_rw of course. But why do you assert segflg == UIO_USERSPACE? It should be valid for kernel UIOs as well.

[glebius]

Don't assert uio is from userspace.