Re: [PATCH dovetail 00/11] riscv: Add dovetail support

Philippe Gerum <[email protected]>
Newsgroups dev.linux.lists.xenomai
Message-ID <87pl3ii35l.fsf@2a01cb0804738000b42250ae26cae50b.ipv6.abo.wanadoo.fr>
Tobias Schaffner <[email protected]> writes:

>
> Changes since v5:
>  * Use mark_trap_entry/exit_raw for kernel mode exception paths which
>    use irq_nmi_entry/exit for now, as the nmi irq handlers expect hard
>    irqs to stay off. This may be be subject to further optimization at
>    a later stage, but I do not see any negative impact latency wise in
>    my tests.
>

Agreed. All traps branching to the nmi_entry code seem to denote some
kernel breakage or a debugger stop, so latency may be the least of our
problem if taken anyway.

However, we may still have to mirror the hardware irq state into the
stall bit when running in-band, to pass sanity checks like
WARN_ON(irqs_disabled()) if any.

-- 
Philippe.
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.