svn commit: r278479 - in head: etc sys/kern

[Slawa Olhovchenkov]

On Tue, Feb 10, 2015 at 06:55:28PM +0000, Rui Paulo wrote:

> On Feb 10, 2015, at 10:52 AM, Slawa Olhovchenkov <slw at zxy.spb.ru> wrote:
> For common case (I am not talk about current devd implementation -- I
> am don't have any inforamtion/metrics/etc) routing and processing
> events may be sensitive to delay and ordering: may be exist
> requirement 'delay not more then 700ns', may be exist requirement
> 'next event process only after complete process previos event'. And
> some event handling may be very CPU/disk/etc consumption. Need to good
> think over and design API and architecure.
>  
> routing events do not come over devd (or any other /dev device):
> there's a routing socket specifically tailored for that purpose.

I am talk abot this:

nomatch 32 {
        match "bus" "uhub[0-9]+";
        match "mode" "host";
        match "vendor" "0x03eb";
        match "product" "0x2109";
        action "kldload -n uftdi";
};

>  For that type of latency, you probably don't want to be crossing
> the kernel/userland barrier too often (or never).

We have action for coredump. And have action for automount (when
inserting USB disk).

Must be coredump wait for complete mounting USB disk?
For flapping USB disk need to start multiple mount script?