Nouveau - Jan 2018 - [PATCH v3] drm/nouveau: Move irq setup/teardown to pci ctor/dtor

For a while we've been having issues with seemingly random interrupts
coming from nvidia cards when resuming them. Originally the fix for this
was thought to be just re-arming the MSI interrupt registers right after
re-allocating our IRQs, however it seems a lot of what we do is both
wrong and not even nessecary.

This was made apparent by what appeared to be a regression in the
mainline kernel that started introducing suspend/resume issues for
nouveau:

        a0c9259dc4e1 (irq/matrix: Spread interrupts on allocation)

After this commit was introduced, we started getting interrupts from the
GPU before we actually re-allocated our own IRQ (see references below)
and assigned the IRQ handler. Investigating this turned out that the
problem was not with the commit, but the fact that nouveau even
free/allocates it's irqs before and after suspend/resume.

For starters: drivers in the linux kernel haven't had to handle
freeing/re-allocating their IRQs during suspend/resume cycles for quite
a while now. Nouveau seems to be one of the few drivers left that still
does this, despite the fact there's no reason we actually need to since
disabling interrupts from the device side should be enough, as the
kernel is already smart enough to know to disable host-side interrupts
for us before going into suspend. Since we were tearing down our IRQs by
hand however, that means there was a short period during resume where
interrupts could be received before we re-allocated our IRQ which would
lead to us getting an unhandled IRQ. Since we never handle said IRQ and
re-arm the interrupt registers, this would cause us to miss all of the
interrupts from the GPU and cause our init process to start timing out
on anything requiring interrupts.

So, since this whole setup/teardown every suspend/resume cycle is
useless anyway, move irq setup/teardown into the pci subdev's ctor/dtor
functions instead so they're only called at driver load and driver
unload. This should fix most of the issues with pending interrupts on
resume, along with getting suspend/resume for nouveau to work again.

As well, this probably means we can also just remove the msi rearm call
inside nvkm_pci_init(). But since our main focus here is to fix
suspend/resume before 4.15, we'll save that for a later patch.

Signed-off-by: Lyude Paul <lyude at redhat.com>
Cc: Karol Herbst <kherbst at redhat.com>
Cc: Thomas Gleixner <tglx at linutronix.de>
Cc: Mike Galbraith <efault at gmx.de>
Cc: stable at vger.kernel.org
---
Changes since v2:
 - Remove teardown, just reuse pci->irq to indicate when we're tearing
   down the driver

 drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c | 44 +++++++++++++++++---------
 1 file changed, 29 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c
b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c
index deb96de54b00..3b2cad639388 100644
--- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c
+++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c
@@ -71,6 +71,10 @@ nvkm_pci_intr(int irq, void *arg)
 	struct nvkm_pci *pci = arg;
 	struct nvkm_device *device = pci->subdev.device;
 	bool handled = false;
+
+	if (pci->irq < 0)
+		return IRQ_HANDLED;
+
 	nvkm_mc_intr_unarm(device);
 	if (pci->msi)
 		pci->func->msi_rearm(pci);
@@ -84,11 +88,6 @@ nvkm_pci_fini(struct nvkm_subdev *subdev, bool suspend)
 {
 	struct nvkm_pci *pci = nvkm_pci(subdev);
 
-	if (pci->irq >= 0) {
-		free_irq(pci->irq, pci);
-		pci->irq = -1;
-	}
-
 	if (pci->agp.bridge)
 		nvkm_agp_fini(pci);
 
@@ -108,8 +107,20 @@ static int
 nvkm_pci_oneinit(struct nvkm_subdev *subdev)
 {
 	struct nvkm_pci *pci = nvkm_pci(subdev);
-	if (pci_is_pcie(pci->pdev))
-		return nvkm_pcie_oneinit(pci);
+	struct pci_dev *pdev = pci->pdev;
+	int ret;
+
+	if (pci_is_pcie(pci->pdev)) {
+		ret = nvkm_pcie_oneinit(pci);
+		if (ret)
+			return ret;
+	}
+
+	ret = request_irq(pdev->irq, nvkm_pci_intr, IRQF_SHARED, "nvkm",
pci);
+	if (ret)
+		return ret;
+	pci->irq = pdev->irq;
+
 	return 0;
 }
 
@@ -117,7 +128,6 @@ static int
 nvkm_pci_init(struct nvkm_subdev *subdev)
 {
 	struct nvkm_pci *pci = nvkm_pci(subdev);
-	struct pci_dev *pdev = pci->pdev;
 	int ret;
 
 	if (pci->agp.bridge) {
@@ -131,28 +141,32 @@ nvkm_pci_init(struct nvkm_subdev *subdev)
 	if (pci->func->init)
 		pci->func->init(pci);
 
-	ret = request_irq(pdev->irq, nvkm_pci_intr, IRQF_SHARED, "nvkm",
pci);
-	if (ret)
-		return ret;
-
-	pci->irq = pdev->irq;
-
 	/* Ensure MSI interrupts are armed, for the case where there are
 	 * already interrupts pending (for whatever reason) at load time.
 	 */
 	if (pci->msi)
 		pci->func->msi_rearm(pci);
 
-	return ret;
+	return 0;
 }
 
 static void *
 nvkm_pci_dtor(struct nvkm_subdev *subdev)
 {
 	struct nvkm_pci *pci = nvkm_pci(subdev);
+	int irq;
+
 	nvkm_agp_dtor(pci);
+
+	if (pci->irq >= 0) {
+		irq = pci->irq;
+		pci->irq = -1;
+		free_irq(irq, pci);
+	}
+
 	if (pci->msi)
 		pci_disable_msi(pci->pdev);
+
 	return nvkm_pci(subdev);
 }
 
-- 
2.14.3

On Thu, Jan 25, 2018 at 06:29:53PM -0500, Lyude Paul
wrote:
> This was made apparent by what appeared to be a regression in the
> mainline kernel that started introducing suspend/resume issues for
> nouveau:
> 
>         a0c9259dc4e1 (irq/matrix: Spread interrupts on allocation)
I'm just a dumb user here, but I confirm:
CPU: AMD Phenom II X6 1055T, GPU: GTX 560 Ti
100% fail to resume GPU on 4.15-rc*, 100% ok with your patch.

And that spinning rust hates being repeatedly powered off and on... :/
> ---
> Changes since v2:
>  - Remove teardown, just reuse pci->irq to indicate when we're
tearing
>    down the driver
> 
>  drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c | 44
+++++++++++++++++---------
>  1 file changed, 29 insertions(+), 15 deletions(-)
Meow!
-- 
⢀⣴⠾⠻⢶⣦⠀ 
⣾⠁⢰⠒⠀⣿⡁ Imagine there are bandits in your house, your kid is bleeding out,
⢿⡄⠘⠷⠚⠋⠀ the house is on fire, and seven big-ass trumpets are playing in the
⠈⠳⣄⠀⠀⠀⠀ sky.  Your cat demands food.  The priority should be obvious...

On Fri, 2018-01-26 at 02:20 +0100, Adam Borowski wrote:
> On Thu, Jan 25, 2018 at 06:29:53PM -0500, Lyude Paul wrote:
> > This was made apparent by what appeared to be a regression in the
> > mainline kernel that started introducing suspend/resume issues for
> > nouveau:
> > 
> >         a0c9259dc4e1 (irq/matrix: Spread interrupts on allocation)
> 
> I'm just a dumb user here, but I confirm:
> CPU: AMD Phenom II X6 1055T, GPU: GTX 560 Ti
> 100% fail to resume GPU on 4.15-rc*, 100% ok with your patch.
Ditto.. and my GTX 980 works again.

	-Mike