From: Alex Deucher <alexdeucher@gmail.com>
To: Jerome Glisse <jglisse@redhat.com>
Cc: airlied@gmail.com, linux-kernel@vger.kernel.org, dri-devel@lists.sf.net
Subject: Re: [PATCH 1/2] drm/radeon/kms: Fallback to non AGP when acceleration fails to initialize
Date: Tue, 6 Oct 2009 12:08:50 -0400 [thread overview]
Message-ID: <a728f9f90910060908i6c4cb7d9yd011722e0b216282@mail.gmail.com> (raw)
In-Reply-To: <1254842860-5492-1-git-send-email-jglisse@redhat.com>
On Tue, Oct 6, 2009 at 11:27 AM, Jerome Glisse <jglisse@redhat.com> wrote:
> When GPU acceleration is not working with AGP try to fallback to non
> AGP GART (either PCI or PCIE GART). This should make KMS failure on
> AGP less painfull. We still need to find out what is wrong when AGP
> fails but at least user have a lot of more chances to get a working
> configuration with acceleration. This patch also cleanup R600/RV770
> fallback path so they use same code as others asics.
>
> Signed-off-by: Jerome Glisse <jglisse@redhat.com>
> ---
> drivers/gpu/drm/radeon/r600.c | 18 +-----------------
> drivers/gpu/drm/radeon/radeon_device.c | 25 ++++++++++++++++++++++++-
> drivers/gpu/drm/radeon/rv770.c | 18 +-----------------
> 3 files changed, 26 insertions(+), 35 deletions(-)
>
> diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
> index 9f7eb08..2cef638 100644
> --- a/drivers/gpu/drm/radeon/r600.c
> +++ b/drivers/gpu/drm/radeon/r600.c
> @@ -1560,15 +1560,8 @@ int r600_init(struct radeon_device *rdev)
> if (r)
> return r;
> r = r600_mc_init(rdev);
> - if (r) {
> - if (rdev->flags & RADEON_IS_AGP) {
> - /* Retry with disabling AGP */
> - r600_fini(rdev);
> - rdev->flags &= ~RADEON_IS_AGP;
> - return r600_init(rdev);
> - }
> + if (r)
> return r;
> - }
> /* Memory manager */
> r = radeon_object_init(rdev);
> if (r)
> @@ -1597,15 +1590,8 @@ int r600_init(struct radeon_device *rdev)
>
> r = r600_startup(rdev);
> if (r) {
> - if (rdev->flags & RADEON_IS_AGP) {
> - /* Retry with disabling AGP */
> - r600_fini(rdev);
> - rdev->flags &= ~RADEON_IS_AGP;
> - return r600_init(rdev);
> - }
> r600_suspend(rdev);
> r600_wb_fini(rdev);
> - radeon_ib_pool_fini(rdev);
> radeon_ring_fini(rdev);
> r600_pcie_gart_fini(rdev);
> rdev->accel_working = false;
> @@ -1637,10 +1623,8 @@ void r600_fini(struct radeon_device *rdev)
> radeon_gem_fini(rdev);
> radeon_fence_driver_fini(rdev);
> radeon_clocks_fini(rdev);
> -#if __OS_HAS_AGP
> if (rdev->flags & RADEON_IS_AGP)
> radeon_agp_fini(rdev);
> -#endif
> radeon_object_fini(rdev);
> radeon_atombios_fini(rdev);
> kfree(rdev->bios);
> diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
> index 2d07ccc..ec02006 100644
> --- a/drivers/gpu/drm/radeon/radeon_device.c
> +++ b/drivers/gpu/drm/radeon/radeon_device.c
> @@ -565,8 +565,31 @@ int radeon_device_init(struct radeon_device *rdev,
> DRM_INFO("register mmio size: %u\n", (unsigned)rdev->rmmio_size);
>
> r = radeon_init(rdev);
> - if (r) {
> + if (r)
> return r;
> + if (rdev->flags & RADEON_IS_AGP && !rdev->accel_working) {
> + /* Acceleration not working on AGP card try again
> + * with fallback to PCI or PCIE GART
> + */
> + radeon_fini(rdev);
> + rdev->flags &= ~RADEON_IS_AGP;
> + if (rdev->family >= CHIP_RV515 ||
> + rdev->family == CHIP_RV380 ||
> + rdev->family == CHIP_RV410 ||
> + rdev->family == CHIP_R423) {
> + DRM_INFO("Forcing AGP to PCIE mode\n");
> + rdev->flags |= RADEON_IS_PCIE;
> + rdev->asic->gart_tlb_flush = &rv370_pcie_gart_tlb_flush;
> + rdev->asic->gart_set_page = &rv370_pcie_gart_set_page;
> + } else {
> + DRM_INFO("Forcing AGP to PCI mode\n");
> + rdev->flags |= RADEON_IS_PCI;
> + rdev->asic->gart_tlb_flush = &r100_pci_gart_tlb_flush;
> + rdev->asic->gart_set_page = &r100_pci_gart_set_page;
> + }
The above code is wrong for r6xx/r7xx. should be:
if (rdev->family >= CHIP_R600) {
DRM_INFO("Forcing AGP to PCIE mode\n");
rdev->flags |= RADEON_IS_PCIE;
} else if (rdev->family >= CHIP_RV515 ||
rdev->family == CHIP_RV380 ||
rdev->family == CHIP_RV410 ||
rdev->family == CHIP_R423) {
DRM_INFO("Forcing AGP to PCIE mode\n");
rdev->flags |= RADEON_IS_PCIE;
rdev->asic->gart_tlb_flush = &rv370_pcie_gart_tlb_flush;
rdev->asic->gart_set_page = &rv370_pcie_gart_set_page;
} else {
DRM_INFO("Forcing AGP to PCI mode\n");
rdev->flags |= RADEON_IS_PCI;
rdev->asic->gart_tlb_flush = &r100_pci_gart_tlb_flush;
rdev->asic->gart_set_page = &r100_pci_gart_set_page;
}
Otherwise you end up with rv370 pcie gart functions.
> + r = radeon_init(rdev);
> + if (r)
> + return r;
> }
> if (radeon_testing) {
> radeon_test_moves(rdev);
> diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c
> index 295cf14..5c597df 100644
> --- a/drivers/gpu/drm/radeon/rv770.c
> +++ b/drivers/gpu/drm/radeon/rv770.c
> @@ -923,15 +923,8 @@ int rv770_init(struct radeon_device *rdev)
> if (r)
> return r;
> r = rv770_mc_init(rdev);
> - if (r) {
> - if (rdev->flags & RADEON_IS_AGP) {
> - /* Retry with disabling AGP */
> - rv770_fini(rdev);
> - rdev->flags &= ~RADEON_IS_AGP;
> - return rv770_init(rdev);
> - }
> + if (r)
> return r;
> - }
> /* Memory manager */
> r = radeon_object_init(rdev);
> if (r)
> @@ -960,15 +953,8 @@ int rv770_init(struct radeon_device *rdev)
>
> r = rv770_startup(rdev);
> if (r) {
> - if (rdev->flags & RADEON_IS_AGP) {
> - /* Retry with disabling AGP */
> - rv770_fini(rdev);
> - rdev->flags &= ~RADEON_IS_AGP;
> - return rv770_init(rdev);
> - }
> rv770_suspend(rdev);
> r600_wb_fini(rdev);
> - radeon_ib_pool_fini(rdev);
> radeon_ring_fini(rdev);
> rv770_pcie_gart_fini(rdev);
> rdev->accel_working = false;
> @@ -999,10 +985,8 @@ void rv770_fini(struct radeon_device *rdev)
> radeon_gem_fini(rdev);
> radeon_fence_driver_fini(rdev);
> radeon_clocks_fini(rdev);
> -#if __OS_HAS_AGP
> if (rdev->flags & RADEON_IS_AGP)
> radeon_agp_fini(rdev);
> -#endif
> radeon_object_fini(rdev);
> radeon_atombios_fini(rdev);
> kfree(rdev->bios);
> --
> 1.6.4.4
>
>
> ------------------------------------------------------------------------------
> Come build with us! The BlackBerry® Developer Conference in SF, CA
> is the only developer event you need to attend this year. Jumpstart your
> developing skills, take BlackBerry mobile applications to market and stay
> ahead of the curve. Join us from November 9-12, 2009. Register now!
> http://p.sf.net/sfu/devconf
> --
> _______________________________________________
> Dri-devel mailing list
> Dri-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/dri-devel
>
prev parent reply other threads:[~2009-10-06 16:09 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-10-06 15:27 Jerome Glisse
2009-10-06 15:27 ` [PATCH 2/2] drm/radeon/kms: Fix AGP support for R600/RV770 family Jerome Glisse
2009-10-06 16:08 ` Alex Deucher [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=a728f9f90910060908i6c4cb7d9yd011722e0b216282@mail.gmail.com \
--to=alexdeucher@gmail.com \
--cc=airlied@gmail.com \
--cc=dri-devel@lists.sf.net \
--cc=jglisse@redhat.com \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®