SYSTIMER device polling

Matthew Dillon dillon at apollo.backplane.com
Sun Oct 9 09:10:32 PDT 2005


:On 11:10, Fri 07 Oct 05, Sepherosa Ziehau wrote:
:> Hi all,
:>=20
:> Following patch(against HEAD) make device polling utilize seperate
:> systimer, instead of hard clock:
:> http://leaf.dragonflybsd.org/~sephe/systimer_poll.diff
:>=20
:> Tested with fxp(4)
:>=20
:> Please review it.
:>=20
:> Cheers,
:> sephe
:>=20
:> --
:> Live Free or Die
:
:I like it. But maybe struct systimer for device polling should be moved to
:struct globaldata the others gd_*clocks are?
:
:By the way, is it possible to pull for device a on cpu0 and device b on
:cpu1? I mean destributing the device pulling work amoung multiple cpus.
:
:--=20
:Alexander 'alxl' Lobachov	mailto:alxl at xxxxxxxxx
:KeyID:			0x0ACA53C6
:Key fingerprint:	8C07 4350 C429 C4E6 B4A9  BF59 1103 9E47 0ACA 53C6

    The current polling implementation is basically just on cpu 0, so for
    this first stage we should probably keep it on cpu 0.  In that light,
    the patch is fine and I will be committing it.
 
					-Matt
					Matthew Dillon 
					<dillon at xxxxxxxxxxxxx>





More information about the Submit mailing list