| Message ID | 20180210143501.30407-1-marc.zyngier@arm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hi Marc, Am Samstag, 10. Februar 2018, 15:35:01 CET schrieb Marc Zyngier: > The rockchip DRM driver is quite careful to disable interrupts > when taking a lock that is also taken in interrupt context, > which is a good thing. > > What is a bit over the top is to use spin_lock_irqsave when > already in interrupt context, as you cannot take another > interrupt again, and disabling interrupt is just pure > overhead. > > Switching to the non _irqsave version in interrupt context is > more logical, and less heavy handed. > > Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> please note, that we had a maintainer swap for the Rockchip drm-component, where Sandy replaced Mark [0] ... with me acting as sort-of (and not yet up to speed) backup. So I guess Rockchip drm patches should also include Sandy Huang <hjc@rock-chips.com> Heiko [0] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=b415b79529391abd29b6a30b93697821f01e082f
Hi Heiko, On 10/02/18 16:23, Heiko Stuebner wrote: > Hi Marc, > > Am Samstag, 10. Februar 2018, 15:35:01 CET schrieb Marc Zyngier: >> The rockchip DRM driver is quite careful to disable interrupts >> when taking a lock that is also taken in interrupt context, >> which is a good thing. >> >> What is a bit over the top is to use spin_lock_irqsave when >> already in interrupt context, as you cannot take another >> interrupt again, and disabling interrupt is just pure >> overhead. >> >> Switching to the non _irqsave version in interrupt context is >> more logical, and less heavy handed. >> >> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> > > please note, that we had a maintainer swap for the Rockchip drm-component, > where Sandy replaced Mark [0] ... with me acting as sort-of (and not yet > up to speed) backup. > > So I guess Rockchip drm patches should also include > Sandy Huang <hjc@rock-chips.com> I guess that update didn't make it into 4.15, which is why I didn't spot it. I'll repost the patches shortly including Sandy on Cc. Thanks, M.
diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c index 72554f404b7e..e23877e72c43 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c @@ -1159,15 +1159,14 @@ static void vop_handle_vblank(struct vop *vop) { struct drm_device *drm = vop->drm_dev; struct drm_crtc *crtc = &vop->crtc; - unsigned long flags; - spin_lock_irqsave(&drm->event_lock, flags); + spin_lock(&drm->event_lock); if (vop->event) { drm_crtc_send_vblank_event(crtc, vop->event); drm_crtc_vblank_put(crtc); vop->event = NULL; } - spin_unlock_irqrestore(&drm->event_lock, flags); + spin_unlock(&drm->event_lock); if (test_and_clear_bit(VOP_PENDING_FB_UNREF, &vop->pending)) drm_flip_work_commit(&vop->fb_unref_work, system_unbound_wq); @@ -1178,21 +1177,20 @@ static irqreturn_t vop_isr(int irq, void *data) struct vop *vop = data; struct drm_crtc *crtc = &vop->crtc; uint32_t active_irqs; - unsigned long flags; int ret = IRQ_NONE; /* * interrupt register has interrupt status, enable and clear bits, we * must hold irq_lock to avoid a race with enable/disable_vblank(). */ - spin_lock_irqsave(&vop->irq_lock, flags); + spin_lock(&vop->irq_lock); active_irqs = VOP_INTR_GET_TYPE(vop, status, INTR_MASK); /* Clear all active interrupt sources */ if (active_irqs) VOP_INTR_SET_TYPE(vop, clear, active_irqs, 1); - spin_unlock_irqrestore(&vop->irq_lock, flags); + spin_unlock(&vop->irq_lock); /* This is expected for vop iommu irqs, since the irq is shared */ if (!active_irqs)
The rockchip DRM driver is quite careful to disable interrupts when taking a lock that is also taken in interrupt context, which is a good thing. What is a bit over the top is to use spin_lock_irqsave when already in interrupt context, as you cannot take another interrupt again, and disabling interrupt is just pure overhead. Switching to the non _irqsave version in interrupt context is more logical, and less heavy handed. Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> --- drivers/gpu/drm/rockchip/rockchip_drm_vop.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-)