[lkml]   [2007]   [Dec]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    SubjectRe: [patch] Make MMCONFIG space (extended PCI config space) a driver opt-in issue

    > Are you prepared to guarantee that freely mixing mmconfig and type1
    > config accesses at runtime will always work, on all chipsets? I'm
    > talking about silicon here, not kernel software.

    We spinlock config space accesses. So the silicon will never see
    simultaneous accesses, which is what I strongly suspect would be broken
    in chipsets (asking for livelocks here).

    Now there is the question of whether the silicon implements MMCONFIG
    writes in some asynchronous way, in which case, there might be an issue
    if you go bang the port right away but I very much doubt it.


     \ /
      Last update: 2007-12-23 22:17    [W:0.021 / U:1.916 seconds]
    ©2003-2017 Jasper Spaans. hosted at Digital OceanAdvertise on this site