Re: [RFC PATCH 04/11] kvm tools: console: unconditionally output to any console
On Tue, May 7, 2013 at 2:34 AM, Sasha Levin wrote: > On 05/03/2013 12:09 PM, Will Deacon wrote: >> On Fri, May 03, 2013 at 05:02:14PM +0100, Sasha Levin wrote: >>> On 05/03/2013 05:19 AM, Pekka Enberg wrote: On Wed, May 1, 2013 at 6:50 PM, Will Deacon wrote: > From: Marc Zyngier > > Kvmtool suppresses any output to a console that has not been elected > as *the* console. > > While this makes sense on the input side (we want the input to be sent > to one console driver only), it seems to be the wrong thing to do on > the output side, as it effectively prevents the guest from switching > from one console to another (think earlyprintk using 8250 to virtio > console). > > After all, the guest *does* poke this device and outputs something > there. > > Just remove the kvm->cfg.active_console test from the output paths. > > Signed-off-by: Marc Zyngier > Signed-off-by: Will Deacon Seems reasonable. Asias, Sasha? >>> >>> I remember at trying it some time ago but dropped it for a reason I don't >>> remember at the moment. >>> >>> Can I have the weekend to play with it to try and figure out why? >> >> There's no rush from my point of view (hence the RFC) so take as long as you >> need! > > Looks good to me! > > > Thanks, > Sasha > I am fine with having 8250 emulated by KVMTOOL, but I am more inclined towards having a full para-virtualized (PV) machine emulated by KVMTOOL. Best Regards, Anup -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [RFC PATCH 04/11] kvm tools: console: unconditionally output to any console
On 05/03/2013 12:09 PM, Will Deacon wrote: > On Fri, May 03, 2013 at 05:02:14PM +0100, Sasha Levin wrote: >> On 05/03/2013 05:19 AM, Pekka Enberg wrote: >>> On Wed, May 1, 2013 at 6:50 PM, Will Deacon wrote: From: Marc Zyngier Kvmtool suppresses any output to a console that has not been elected as *the* console. While this makes sense on the input side (we want the input to be sent to one console driver only), it seems to be the wrong thing to do on the output side, as it effectively prevents the guest from switching from one console to another (think earlyprintk using 8250 to virtio console). After all, the guest *does* poke this device and outputs something there. Just remove the kvm->cfg.active_console test from the output paths. Signed-off-by: Marc Zyngier Signed-off-by: Will Deacon >>> >>> Seems reasonable. Asias, Sasha? >>> >> >> I remember at trying it some time ago but dropped it for a reason I don't >> remember at the moment. >> >> Can I have the weekend to play with it to try and figure out why? > > There's no rush from my point of view (hence the RFC) so take as long as you > need! Looks good to me! Thanks, Sasha -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [RFC PATCH 04/11] kvm tools: console: unconditionally output to any console
On Fri, May 3, 2013 at 5:19 PM, Pekka Enberg wrote: > On Wed, May 1, 2013 at 6:50 PM, Will Deacon wrote: >> From: Marc Zyngier >> >> Kvmtool suppresses any output to a console that has not been elected >> as *the* console. >> >> While this makes sense on the input side (we want the input to be sent >> to one console driver only), it seems to be the wrong thing to do on >> the output side, as it effectively prevents the guest from switching >> from one console to another (think earlyprintk using 8250 to virtio >> console). >> >> After all, the guest *does* poke this device and outputs something >> there. >> >> Just remove the kvm->cfg.active_console test from the output paths. >> >> Signed-off-by: Marc Zyngier >> Signed-off-by: Will Deacon > > Seems reasonable. Asias, Sasha? This patch itself looks good to me. But we have more issues for the console devices and termials with regard to multiple console support: 1) All the console outputs (spapr_hvcons.c, spapr_rtas.c virtio/console.c) are redirected to term 0. 2) With multiple console support, the cfg.active_console logic is not very useful at all. 3) Four serial devices ttyS0-3 are initialized unconditionally and mapped to term 0-3. 4) Using --tty option, we can map a term to /dev/pts/N on host. I think we can merge --tty option to --console option. I have something like this in my mind: --console type=serial,backend=stdio --console type=virtio,backend=pts --console type=hv,backend=pts e.g to add two serial consoles ttyS0 and ttyS1 and one virtio console hvc0, ttyS0 is mapped the stdio and ttyS1 and hvc0 are mapped to pts, we use this: --console type=serial,backend=stdio --console type=serial,backend=pts --console type=virtio,backend=pts > -- > To unsubscribe from this list: send the line "unsubscribe kvm" in > the body of a message to majord...@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- Asias -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [RFC PATCH 04/11] kvm tools: console: unconditionally output to any console
On Fri, May 03, 2013 at 05:02:14PM +0100, Sasha Levin wrote: > On 05/03/2013 05:19 AM, Pekka Enberg wrote: > > On Wed, May 1, 2013 at 6:50 PM, Will Deacon wrote: > >> From: Marc Zyngier > >> > >> Kvmtool suppresses any output to a console that has not been elected > >> as *the* console. > >> > >> While this makes sense on the input side (we want the input to be sent > >> to one console driver only), it seems to be the wrong thing to do on > >> the output side, as it effectively prevents the guest from switching > >> from one console to another (think earlyprintk using 8250 to virtio > >> console). > >> > >> After all, the guest *does* poke this device and outputs something > >> there. > >> > >> Just remove the kvm->cfg.active_console test from the output paths. > >> > >> Signed-off-by: Marc Zyngier > >> Signed-off-by: Will Deacon > > > > Seems reasonable. Asias, Sasha? > > > > I remember at trying it some time ago but dropped it for a reason I don't > remember at the moment. > > Can I have the weekend to play with it to try and figure out why? There's no rush from my point of view (hence the RFC) so take as long as you need! Will -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [RFC PATCH 04/11] kvm tools: console: unconditionally output to any console
On 05/03/2013 05:19 AM, Pekka Enberg wrote: > On Wed, May 1, 2013 at 6:50 PM, Will Deacon wrote: >> From: Marc Zyngier >> >> Kvmtool suppresses any output to a console that has not been elected >> as *the* console. >> >> While this makes sense on the input side (we want the input to be sent >> to one console driver only), it seems to be the wrong thing to do on >> the output side, as it effectively prevents the guest from switching >> from one console to another (think earlyprintk using 8250 to virtio >> console). >> >> After all, the guest *does* poke this device and outputs something >> there. >> >> Just remove the kvm->cfg.active_console test from the output paths. >> >> Signed-off-by: Marc Zyngier >> Signed-off-by: Will Deacon > > Seems reasonable. Asias, Sasha? > I remember at trying it some time ago but dropped it for a reason I don't remember at the moment. Can I have the weekend to play with it to try and figure out why? Thanks, Sasha -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [RFC PATCH 04/11] kvm tools: console: unconditionally output to any console
On Wed, May 1, 2013 at 6:50 PM, Will Deacon wrote: > From: Marc Zyngier > > Kvmtool suppresses any output to a console that has not been elected > as *the* console. > > While this makes sense on the input side (we want the input to be sent > to one console driver only), it seems to be the wrong thing to do on > the output side, as it effectively prevents the guest from switching > from one console to another (think earlyprintk using 8250 to virtio > console). > > After all, the guest *does* poke this device and outputs something > there. > > Just remove the kvm->cfg.active_console test from the output paths. > > Signed-off-by: Marc Zyngier > Signed-off-by: Will Deacon Seems reasonable. Asias, Sasha? -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
