dri-devel January 2016

dri-devel@lists.freedesktop.org

194 participants
1200 discussions

[PATCH libdrm] Fix memory leak with drmModeGetConnectorCurrent()
by ville.syrjala＠linux.intel.com 07 Jan '16

07 Jan '16

From: Ville Syrjälä <ville.syrjala(a)linux.intel.com> drmModeGetConnectorCurrent() must provide temporary storage for the kernel to fill in at least one mode (asking for !=0 modes is how you prevent the heavyweight probe in the kernel). Currently we malloc that temp storage but we fail to free it before overwriting the pointer with the address of the actual storage we use to store the real mode list we get from the kernel in the second ioctl call. Let's just keep the temporary storage on the stack and thus we avoid the leak and also eliminate some pointless mallocs. Cc: Chris Wilson <chris(a)chris-wilson.co.uk> Fixes: 5ed5fa10600f ("mode: Retrieve only the current information for a Connector") Signed-off-by: Ville Syrjälä <ville.syrjala(a)linux.intel.com> --- xf86drmMode.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) diff --git a/xf86drmMode.c b/xf86drmMode.c index ab6b5195e8d3..7710061865ee 100644 --- a/xf86drmMode.c +++ b/xf86drmMode.c @@ -475,12 +475,13 @@ _drmModeGetConnector(int fd, uint32_t connector_id, int probe) { struct drm_mode_get_connector conn, counts; drmModeConnectorPtr r = NULL; + struct drm_mode_modeinfo stack_mode; memclear(conn); conn.connector_id = connector_id; if (!probe) { conn.count_modes = 1; - conn.modes_ptr = VOID2U64(drmMalloc(sizeof(struct drm_mode_modeinfo))); + conn.modes_ptr = VOID2U64(&stack_mode); } if (drmIoctl(fd, DRM_IOCTL_MODE_GETCONNECTOR, &conn)) @@ -504,7 +505,7 @@ retry: goto err_allocs; } else { conn.count_modes = 1; - conn.modes_ptr = VOID2U64(drmMalloc(sizeof(struct drm_mode_modeinfo))); + conn.modes_ptr = VOID2U64(&stack_mode); } if (conn.count_encoders) { @@ -525,7 +526,8 @@ retry: counts.count_encoders < conn.count_encoders) { drmFree(U642VOID(conn.props_ptr)); drmFree(U642VOID(conn.prop_values_ptr)); - drmFree(U642VOID(conn.modes_ptr)); + if (U642VOID(conn.modes_ptr) != &stack_mode) + drmFree(U642VOID(conn.modes_ptr)); drmFree(U642VOID(conn.encoders_ptr)); goto retry; @@ -567,7 +569,8 @@ retry: err_allocs: drmFree(U642VOID(conn.prop_values_ptr)); drmFree(U642VOID(conn.props_ptr)); - drmFree(U642VOID(conn.modes_ptr)); + if (U642VOID(conn.modes_ptr) != &stack_mode) + drmFree(U642VOID(conn.modes_ptr)); drmFree(U642VOID(conn.encoders_ptr)); return r; -- 2.4.10

3 3

[GIT PULL] drm-etnaviv-fixes for current -next
by Lucas Stach 07 Jan '16

07 Jan '16

Hi Dave, 2 small fixes for the etnaviv driver, both noticed by the smatch checker and reported/fixed by Dan Carpenter. Regards, Lucas The following changes since commit c11b8989635166c5a1e6aac1853a847bd664f8db: Merge tag 'omapdrm-4.5-resolved' of git://git.kernel.org/pub/scm/linux/kernel/git/tomba/linux into drm-next (2016-01-01 07:41:52 +1000) are available in the git repository at: git://git.pengutronix.de/git/lst/linux.git drm-etnaviv-fixes for you to fetch changes up to c33246d793b5bb9d8be7c67918136c310185c23d: drm/etnaviv: fix workaround for GC500 (2016-01-07 11:57:57 +0100) ---------------------------------------------------------------- Dan Carpenter (1): drm/etnaviv: unlock on error in etnaviv_gem_get_iova() Lucas Stach (1): drm/etnaviv: fix workaround for GC500 drivers/gpu/drm/etnaviv/etnaviv_gem.c | 6 ++++-- drivers/gpu/drm/etnaviv/etnaviv_gpu.c | 7 +++++-- 2 files changed, 9 insertions(+), 4 deletions(-)

1 0

[PATCH] drm/i915: intel_hpd_init(): Don't check connectors for an encoder
by Lyude 07 Jan '16

07 Jan '16

There's no reason to check for an active encoder when setting up hpd, especially since a disconnected connector isn't going to have an encoder in the first place. This fixes reprobing display connectors on resume. Signed-off-by: Lyude <cpaul(a)redhat.com> --- After some talking with danvet on IRC, I learned that calling drm_helper_hpd_irq_event() does actually trigger a full reprobe of each connector's status. It turns out this is the actual reason reprobing on resume hasn't been working (this was observed on a T440s): - We call hpd_init() - We check each connector for a couple of things before marking connector->polled with DRM_CONNECTOR_POLL_HPD, one of which is an active encoder. Of course, a disconnected port won't have an active encoder, so we don't add the flag to any of the connector's. - We call drm_helper_hpd_irq_event() - drm_helper_irq_event() checks each connector for the DRM_CONNECTOR_POLL_HPD flag. The only one that has it is eDP-1, so we skip reprobing each connector except that one. drivers/gpu/drm/i915/intel_hotplug.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/intel_hotplug.c b/drivers/gpu/drm/i915/intel_hotplug.c index b177857..18797bc 100644 --- a/drivers/gpu/drm/i915/intel_hotplug.c +++ b/drivers/gpu/drm/i915/intel_hotplug.c @@ -468,7 +468,7 @@ void intel_hpd_init(struct drm_i915_private *dev_priv) list_for_each_entry(connector, &mode_config->connector_list, head) { struct intel_connector *intel_connector = to_intel_connector(connector); connector->polled = intel_connector->polled; - if (connector->encoder && !connector->polled && I915_HAS_HOTPLUG(dev) && intel_connector->encoder->hpd_pin > HPD_NONE) + if (!connector->polled && I915_HAS_HOTPLUG(dev) && intel_connector->encoder->hpd_pin > HPD_NONE) connector->polled = DRM_CONNECTOR_POLL_HPD; if (intel_connector->mst_port) connector->polled = DRM_CONNECTOR_POLL_HPD; -- 2.5.0

2 6

[PATCH 4/5] drm: sti: fix cursor coordinates
by Benjamin Gaignard 07 Jan '16

07 Jan '16

fix x/y typo while setting cursor coordinates Signed-off-by: Fabien Dessenne <fabien.dessenne(a)st.com> --- drivers/gpu/drm/sti/sti_cursor.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/sti/sti_cursor.c b/drivers/gpu/drm/sti/sti_cursor.c index 8078631..a19693a 100644 --- a/drivers/gpu/drm/sti/sti_cursor.c +++ b/drivers/gpu/drm/sti/sti_cursor.c @@ -205,7 +205,7 @@ static void sti_cursor_atomic_update(struct drm_plane *drm_plane, writel(cursor->height << 16 | cursor->width, cursor->regs + CUR_SIZE); y = sti_vtg_get_line_number(*mode, dst_y); - x = sti_vtg_get_pixel_number(*mode, dst_y); + x = sti_vtg_get_pixel_number(*mode, dst_x); writel((y << 16) | x, cursor->regs + CUR_VPO); plane->status = STI_PLANE_UPDATED; -- 1.9.1

1 0

bisected: i915 modeset broken in ac9b8236551d1177fd07b56aef9b565d1864420d
by Meelis Roos 07 Jan '16

07 Jan '16

Between 4.4-rc3 and 4.4-rc4, i915 modesetting broke on my i5-2400 PC. Instead of seeing the new dense graphics mode, I see the last VGA text lines and no X appears either. I saw something similar on I865G but have not had time to check if it is the same issue. ac9b8236551d1177fd07b56aef9b565d1864420d is the first bad commit commit ac9b8236551d1177fd07b56aef9b565d1864420d Author: Ville Syrjälä <ville.syrjala(a)linux.intel.com> Date: Fri Nov 27 18:55:26 2015 +0200 drm/i915: Introduce a gmbus power domain Currently the gmbus code uses intel_aux_display_runtime_get/put in an effort to make sure the hardware is powered up sufficiently for gmbus. That function only takes the runtime PM reference which on VLV/CHV/BXT is not enough. We need the disp2d/pipe-a well on VLV/CHV and power well 2 on BXT. So add a new power domnain for gmbus and kill off the now unused intel_aux_display_runtime_get/put. And change intel_hdmi_set_edid() to use the gmbus power domain too since that's all we need there. Also toss in a BUILD_BUG_ON() to catch problems if we run out of bits for power domains. We're already really close to the limit... [Patrik: Add gmbus string to debugfs output] Signed-off-by: Ville Syrjälä <ville.syrjala(a)linux.intel.com> Reviewed-by: Patrik Jakobsson <patrik.jakobsson(a)linux.intel.com> [Cherry-picked from drm-intel-next-queued f0ab43e6 (Imre)] Signed-off-by: Imre Deak <imre.deak(a)intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1448643329-18675-3-git-send-em… Signed-off-by: Jani Nikula <jani.nikula(a)intel.com> :040000 040000 39379146d7e6dda8a4d5f8781ee3d307cce8c47e f4f09fae0485ad6263d31d425296fa9cd7de343b M drivers -- Meelis Roos (mroos(a)linux.ee)

3 5

[Bug 80868] New: Support screen scaling modes for external monitors
by bugzilla-daemon＠freedesktop.org 07 Jan '16

07 Jan '16

https://bugs.freedesktop.org/show_bug.cgi?id=80868 Priority: medium Bug ID: 80868 Assignee: dri-devel(a)lists.freedesktop.org Summary: Support screen scaling modes for external monitors Severity: enhancement Classification: Unclassified OS: All Reporter: kamil.paral(a)gmail.com Hardware: Other Status: NEW Version: XOrg CVS Component: DRM/Radeon Product: DRI As we steadily progress into the year of Linux gaming, there will be many requests regarding games. This is one of them :) Currently, it is possible to set scaling mode for a monitor only if the monitor is internal (LVDS, eDP). This is done through xrandr: $ xrandr --prop Screen 0: minimum 320 x 200, current 1680 x 1050, maximum 8192 x 8192 LVDS1 connected primary 1680x1050+0+0 (normal left inverted right x axis y axis) 331mm x 207mm <snip> scaling mode: Full aspect supported: None, Full, Center, Full aspect <snip> $ xrandr --output LVDS1 --set "scaling mode" "Center" However, it is not possible to do this for external panels (VGA, DVI, HDMI, DP). The "scaling mode" property is not exported for these panels (this was confirmed to me by ckoenig and agd5f on the #radeon irc channel, many thanks). Please allow users to set scaling mode even for external panels. There are many use cases for it, mainly related to gaming. Note: Many panels provide options to do the scaling on their own, however, only the very expensive ones provide good options. For example, the majority of "mainstream" panels don't provide the "center" scaling mode - however, that is very useful when playing older lower-resolution games when you prefer smaller and sharp image instead of larger and blurry. Second example are certain panels which support "full aspect" mode only for a small selection or resolutions, otherwise they simply scale to "full" ignoring image aspect ratio. For these panels, GPU scaling is essential if the user doesn't want to see distorted aspect ratio image. My personal use case is buying BenQ BL2411PT 1920x1200 panel, which *can not* display 1920x1080 resolution with correct aspect ratio - it always stretches it vertically. Yes, it's very dumb, yet that's how modern panels commonly work (and not just the cheap ones). This hasn't been a large issue in the past, because we had no games and opensource drivers were hardly able to run them anyway. Both things are changing rapidly. The proprietary AMD and NVIDIA drivers have been offering GPU scaling functionality for a long time, both on Linux and Windows. Here's an example of their GUI configuration: https://www.codeweavers.com/support/wiki/linux/faq/43_game_stretch Please allow us to use GPU scaling even with radeon driver. Thank you. There has been some technical details on the IRC, it is linked here: http://paste.fedoraproject.org/115407/ It seems to me that this functionality could be implemented in a simple and straightforward way: a) provide a single configuration option - "scaling mode" b) on internal panels default to "Full aspect" (which you already do) - that is reasonable default, because these panels have no control buttons c) on external panels default to "None" - that allows the user to easily configure scaling through the panel. Only if the user is dissatisfied, he/she can enable GPU scaling through xrandr. And here's one more user seconding my thoughts on the IRC: AbortRetryFail: 1:1 unscaled output for LCDs would be awesome. AbortRetryFail: 1280x720 looks horrible scaled up to fit a 1366x768 LCD Thanks. -- You are receiving this mail because: You are the assignee for the bug.

1 21

[PATCH 00/12] dw-hdmi development
by Russell King - ARM Linux 07 Jan '16

07 Jan '16

This sub-series is a mixture of development: * Removing the incorrect pixel repetition configuration code * Preventing pixel-doubled modes from being used * Adding interlaced video support * Implementing the sink_is_hdmi/sink_has_audio flags I suggested a few months ago * Only enabling audio support if the sink indicates it has audio * Avoiding double-enabling the HDMI interface * Fixing the mis-leading name of "dw_hdmi_phy_enable_power" * Adding connector mode forcing (important if your monitor bounces RXSENSE and HPD signals while in low-power mode.) * Improving the HDMI enable/disabling on sink status For review (and testing if people feel like it). Acks/tested-bys etc welcome. It applies on top of my drm-dwhdmi-devel branch, which is waiting for David Airlie to pull (see pull request on dri-devel, 15th July.) drivers/gpu/drm/bridge/dw_hdmi.c | 275 ++++++++++++++++++++++++++++++--------- drivers/gpu/ipu-v3/ipu-dc.c | 18 ++- drivers/gpu/ipu-v3/ipu-di.c | 129 +++++++++--------- 3 files changed, 291 insertions(+), 131 deletions(-) -- FTTC broadband for 0.8mile line: currently at 10.5Mbps down 400kbps up according to speedtest.net.

10 81

[RFC PATCH v1 0/2] Add RK3229 HDMI support
by Yakir Yang 07 Jan '16

07 Jan '16

RK3229 have integrated an DesignedWare HDMI controller and an INNO HDMI phy, so we can still reuse the dw-hdmi driver for RK3229 HDMI controller, but we need to create an separate driver for RK3229 HDMI PHY. Yakir Yang (2): drm: rockchip: hdmi: add RK3229 HDMI support dt-bindings: add document for rk3229-hdmi .../bindings/display/rockchip/dw_hdmi-rockchip.txt | 4 +- drivers/gpu/drm/bridge/dw-hdmi.c | 33 +- drivers/gpu/drm/rockchip/dw_hdmi-rockchip.c | 380 +++++++++++++++++++-- drivers/gpu/drm/rockchip/dw_hdmi-rockchip.h | 137 ++++++++ include/drm/bridge/dw_hdmi.h | 5 +- 5 files changed, 519 insertions(+), 40 deletions(-) create mode 100644 drivers/gpu/drm/rockchip/dw_hdmi-rockchip.h -- 2.1.2

1 3

[PATCH 1/2] Add device enumeration interface (v2)
by Jammy Zhou 06 Jan '16

06 Jan '16

drmGetDevices interface is added to enumernate GPU devices on the system v2: rebase the code and some improvement for the coding style Signed-off-by: Frank Min <Frank.Min(a)amd.com> Signed-off-by: Jammy Zhou <Jammy.Zhou(a)amd.com> (v2) --- Makefile.am | 3 ++- xf86drm.c | 48 ++++++++++++++++++++++++++++++++++++++++++++++++ xf86drm.h | 18 ++++++++++++++++++ 3 files changed, 68 insertions(+), 1 deletion(-) diff --git a/Makefile.am b/Makefile.am index 42d3d7f..8236ed8 100644 --- a/Makefile.am +++ b/Makefile.am @@ -89,7 +89,8 @@ SUBDIRS = \ libdrm_la_LTLIBRARIES = libdrm.la libdrm_ladir = $(libdir) libdrm_la_LDFLAGS = -version-number 2:4:0 -no-undefined -libdrm_la_LIBADD = @CLOCK_LIB@ +libdrm_la_LIBADD = @CLOCK_LIB@ \ + @PCIACCESS_LIBS@ libdrm_la_CPPFLAGS = -I$(top_srcdir)/include/drm AM_CFLAGS = \ diff --git a/xf86drm.c b/xf86drm.c index ffc53b8..4d67861 100644 --- a/xf86drm.c +++ b/xf86drm.c @@ -63,6 +63,7 @@ #include "xf86drm.h" #include "libdrm.h" +#include <pciaccess.h> #if defined(__FreeBSD__) || defined(__FreeBSD_kernel__) || defined(__DragonFly__) #define DRM_MAJOR 145 @@ -2817,3 +2818,50 @@ char *drmGetRenderDeviceNameFromFd(int fd) { return drmGetMinorNameForFD(fd, DRM_NODE_RENDER); } + +/** + * Enumerate the GPU devices on the system + * + * \param devs device array set to return the device information + * (if NULL, the number of device is returned) + * \param vendor the vendor ID for GPU devices to list + * (optional, if not specified, all GPU devices are returned) + * + * \return the number of GPU devices + */ +int drmGetDevices(drmDevicePtr devs, uint16_t vendor) +{ + struct pci_device_iterator * iter; + struct pci_device * dev; + uint32_t count = 0; + + if (pci_system_init()) + return -EINVAL; + + iter = pci_slot_match_iterator_create(NULL); + if (!iter) + return -EINVAL; + + while ((dev = pci_device_next(iter))) { + if (((dev->device_class == 0x30000) || + (dev->device_class == 0x38000)) && + ((vendor == 0) || (dev->vendor_id == vendor))){ + if (devs) { + devs[count].domain = dev->domain; + devs[count].bus = dev->bus; + devs[count].dev = dev->dev; + devs[count].func = dev->func; + devs[count].vendor_id = dev->vendor_id; + devs[count].device_id = dev->device_id; + devs[count].subvendor_id = dev->subvendor_id; + devs[count].subdevice_id = dev->subdevice_id; + devs[count].revision_id = dev->revision; + } + count++; + } + } + + pci_iterator_destroy(iter); + pci_system_cleanup(); + return count; +} diff --git a/xf86drm.h b/xf86drm.h index 40c55c9..6150e71 100644 --- a/xf86drm.h +++ b/xf86drm.h @@ -752,6 +752,24 @@ extern int drmPrimeFDToHandle(int fd, int prime_fd, uint32_t *handle); extern char *drmGetPrimaryDeviceNameFromFd(int fd); extern char *drmGetRenderDeviceNameFromFd(int fd); +/** + * Structure for a general pcie gpu device + */ +typedef struct _drmDevice { + uint16_t domain; + uint8_t bus; + uint8_t dev; + uint8_t func; + uint8_t revision_id; + uint16_t vendor_id; + uint16_t device_id; + uint16_t subvendor_id; + uint16_t subdevice_id; +} drmDevice, *drmDevicePtr; + + +extern int drmGetDevices(drmDevicePtr devs, uint16_t vendor); + #if defined(__cplusplus) || defined(c_plusplus) } #endif -- 1.9.1

4 9

[PATCH v2 1/6] drm/i915: Set connector_state->connector using the helper.
by Maarten Lankhorst 06 Jan '16

06 Jan '16

The atomic helper sets connector_state->connector, which the i915 code didn't. This will become a problem when we start using it. Signed-off-by: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com> Acked-by: Thierry Reding <treding(a)nvidia.com> --- drivers/gpu/drm/i915/intel_display.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 7dd7200d3ba9..1e42309ec40a 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -6477,13 +6477,11 @@ static void intel_connector_check_state(struct intel_connector *connector) int intel_connector_init(struct intel_connector *connector) { - struct drm_connector_state *connector_state; + drm_atomic_helper_connector_reset(&connector->base); - connector_state = kzalloc(sizeof *connector_state, GFP_KERNEL); - if (!connector_state) + if (!connector->base.state) return -ENOMEM; - connector->base.state = connector_state; return 0; } -- 2.1.0

2 13

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

dri-devel January 2016