svn commit: r357293 - head/sys/net

[Jeff Roberson]

On Thu, 30 Jan 2020, Hans Petter Selasky wrote:

> On 2020-01-30 21:56, Gleb Smirnoff wrote:
>> On Thu, Jan 30, 2020 at 12:04:03PM +0000, Hans Petter Selasky wrote:
>> H> Author: hselasky
>> H> Date: Thu Jan 30 12:04:02 2020
>> H> New Revision: 357293
>> H> URL: https://svnweb.freebsd.org/changeset/base/357293
>> H>
>> H> Log:
>> H>   Widen EPOCH(9) usage in netisr.
>> H>
>> H>   Software interrupt handlers are allowed to sleep. In swi_net() there
>> H>   is a read lock behind NETISR_RLOCK() which in turn ends up calling
>> H>   msleep() which means the whole of swi_net() cannot be protected by an
>> H>   EPOCH(9) section. By default the NETISR_LOCKING feature is disabled.
>> H>
>> H>   This issue was introduced by r357004. This is a preparation step for
>> H>   replacing the functionality provided by r357004.
>> H>
>> H>   Found by:	kib@
>> H>   Sponsored by:	Mellanox Technologies
>> 
>> What?! NETISR_RLOCK() which in turn ends up calling msleep()? Can you 
>> please
>> explain this nonsense?
>> 
>
> See rms_rlock_fallback() in kern_rmlock.c .

The network stack uses rm_ not rms_.

Jeff

>
> --HPS
>