Re: Question: handling early hotplug interrupts

2017-08-31 Thread Michael Ellerman
Nathan Fontenot writes: > On 08/30/2017 01:09 AM, Michael Ellerman wrote: ... >> >> So this is simply a case of the init calls being out of order. >> >> We either need to create the pseries_hp_wq earlier, or register the >> event sources IRQs later. I'm not sure which is better. > > Perhaps I'm

Re: Question: handling early hotplug interrupts

2017-08-30 Thread Nathan Fontenot
On 08/30/2017 01:09 AM, Michael Ellerman wrote: > Daniel Henrique Barboza writes: > >> Hi Ben, >> >> On 08/29/2017 06:55 PM, Benjamin Herrenschmidt wrote: >>> On Tue, 2017-08-29 at 17:43 -0300, Daniel Henrique Barboza wrote: Hi, This is a scenario I've been facing when working in e

Re: Question: handling early hotplug interrupts

2017-08-29 Thread Michael Ellerman
Daniel Henrique Barboza writes: > Hi Ben, > > On 08/29/2017 06:55 PM, Benjamin Herrenschmidt wrote: >> On Tue, 2017-08-29 at 17:43 -0300, Daniel Henrique Barboza wrote: >>> Hi, >>> >>> This is a scenario I've been facing when working in early device >>> hotplugs in QEMU. When a device is added, a

Re: Question: handling early hotplug interrupts

2017-08-29 Thread Michael Ellerman
Daniel Henrique Barboza writes: > Hi, > > This is a scenario I've been facing when working in early device > hotplugs in QEMU. When a device is added, a IRQ pulse is fired to warn > the guest of the event, then the kernel fetches it by calling > 'check_exception' and handles it. If the hotplug

Re: Question: handling early hotplug interrupts

2017-08-29 Thread Daniel Henrique Barboza
Hi Ben, On 08/29/2017 06:55 PM, Benjamin Herrenschmidt wrote: On Tue, 2017-08-29 at 17:43 -0300, Daniel Henrique Barboza wrote: Hi, This is a scenario I've been facing when working in early device hotplugs in QEMU. When a device is added, a IRQ pulse is fired to warn the guest of the event, th

Re: Question: handling early hotplug interrupts

2017-08-29 Thread Benjamin Herrenschmidt
On Tue, 2017-08-29 at 17:43 -0300, Daniel Henrique Barboza wrote: > Hi, > > This is a scenario I've been facing when working in early device > hotplugs in QEMU. When a device is added, a IRQ pulse is fired to warn > the guest of the event, then the kernel fetches it by calling > 'check_exceptio

Question: handling early hotplug interrupts

2017-08-29 Thread Daniel Henrique Barboza
Hi, This is a scenario I've been facing when working in early device hotplugs in QEMU. When a device is added, a IRQ pulse is fired to warn the guest of the event, then the kernel fetches it by calling 'check_exception' and handles it. If the hotplug is done too early (before SLOF, for exampl