On 11/22/2024 09:59, Alex Deucher wrote:
On Fri, Nov 22, 2024 at 10:57 AM Mario Limonciello
wrote:
On 11/22/2024 08:28, Lazar, Lijo wrote:
On 11/19/2024 1:33 AM, Mario Limonciello wrote:
As part of the suspend sequence VRAM needs to be evicted on dGPUs.
In order to make suspend/resume more
On Fri, Nov 22, 2024 at 10:57 AM Mario Limonciello
wrote:
>
> On 11/22/2024 08:28, Lazar, Lijo wrote:
> >
> >
> > On 11/19/2024 1:33 AM, Mario Limonciello wrote:
> >> As part of the suspend sequence VRAM needs to be evicted on dGPUs.
> >> In order to make suspend/resume more reliable we moved this
On 11/19/2024 1:33 AM, Mario Limonciello wrote:
> As part of the suspend sequence VRAM needs to be evicted on dGPUs.
> In order to make suspend/resume more reliable we moved this into
> the pmops prepare() callback so that the suspend sequence would fail
> but the system could remain operational
On 11/22/2024 08:28, Lazar, Lijo wrote:
On 11/19/2024 1:33 AM, Mario Limonciello wrote:
As part of the suspend sequence VRAM needs to be evicted on dGPUs.
In order to make suspend/resume more reliable we moved this into
the pmops prepare() callback so that the suspend sequence would fail
but t
On Fri, Nov 22, 2024 at 5:03 AM Christian König
wrote:
>
> Am 18.11.24 um 21:03 schrieb Mario Limonciello:
> > As part of the suspend sequence VRAM needs to be evicted on dGPUs.
> > In order to make suspend/resume more reliable we moved this into
> > the pmops prepare() callback so that the suspen
Am 18.11.24 um 21:03 schrieb Mario Limonciello:
As part of the suspend sequence VRAM needs to be evicted on dGPUs.
In order to make suspend/resume more reliable we moved this into
the pmops prepare() callback so that the suspend sequence would fail
but the system could remain operational under hi
On 11/18/2024 14:03, Mario Limonciello wrote:
As part of the suspend sequence VRAM needs to be evicted on dGPUs.
In order to make suspend/resume more reliable we moved this into
the pmops prepare() callback so that the suspend sequence would fail
but the system could remain operational under high
As part of the suspend sequence VRAM needs to be evicted on dGPUs.
In order to make suspend/resume more reliable we moved this into
the pmops prepare() callback so that the suspend sequence would fail
but the system could remain operational under high memory usage suspend.
Another class of issues
