dri-devel February 2022

dri-devel@lists.freedesktop.org

394 participants
865 discussions

fbdev git tree update
by Helge Deller 02 Feb '22

02 Feb '22

Hello DRI and fbdev developers, I've now mostly checked all queued-up patches on the fbdev mailing list: https://patchwork.kernel.org/project/linux-fbdev/list/ and applied the ones which seemed appropriate. IMHO there is nothing really critical/important/conflicting in there. Shortlog is below, the git "fbdev for-next" branch is here: https://git.kernel.org/pub/scm/linux/kernel/git/deller/linux-fbdev.git/log/… If you see critical things in there or want me to drop (or add) something, just please let me know. If everything goes well I plan to send a pull request when the merge window for kernel v5.18 opens. Helge ---------------------------------------------------------------- Arnd Bergmann (1): agp: define proper stubs for empty helpers Cai Huoqing (11): video: fbdev: au1200fb: Make use of dma_mmap_coherent() video: fbdev: omapfb: panel-lgphilips-lb035q02: Make use of the helper function dev_err_probe() video: fbdev: omapfb: lcd_ams_delta: Make use of the helper function dev_err_probe() video: fbdev: omapfb: panel-sharp-ls037v7dw01: Make use of the helper function dev_err_probe() video: fbdev: omapfb: panel-tpo-td043mtea1: Make use of the helper function dev_err_probe() video: fbdev: da8xx-fb: Make use of the helper function dev_err_probe() video: fbdev: pxa168fb: Make use of the helper function dev_err_probe() video: fbdev: pxa3xx-gcu: Make use of the helper function dev_err_probe() video: fbdev: ssd1307fb: Make use of the helper function dev_err_probe() video: fbdev: s3c-fb: Make use of the helper function dev_err_probe() video: fbdev: mmp: Make use of the helper function dev_err_probe() Changcheng Deng (1): video: fbmem: use swap() to make code cleaner in fb_rotate_logo() Chunyang Zhong (1): video: fbdev: ocfb: add const to of_device_id Colin Ian King (6): video: fbdev: aty128fb: make some arrays static const video: fbdev: mb862xx: remove redundant assignment to pointer ptr video: fbdev: via: Fix spelling mistake "bellow" -> "below" video: fbdev: atyfb: Remove assigned but never used variable statements video: fbdev: asiliantfb: remove redundant assignment to variable Ftarget video: fbdev: pxa168fb: Initialize pointers with NULL and not plain integer 0 Dan Carpenter (3): video: fbdev: savagefb: make a variable local video: fbdev: atmel_lcdfb: fix an error code in atmel_lcdfb_probe() video: fbdev: fbcvt.c: fix printing in fb_cvt_print_name() Evgeny Novikov (1): video: fbdev: w100fb: Reset global state George Kennedy (1): video: fbdev: cirrusfb: check pixclock to avoid divide by zero Greg Kroah-Hartman (1): video: fbdev: omapfb: use default_groups in kobj_type Jiasheng Jiang (1): video: fbdev: imxfb: Check for null res pointer Krzysztof Kozlowski (1): video: fbdev: s3c-fb: drop unneeded MODULE_ALIAS Luca Weiss (2): backlight: qcom-wled: Add PM6150L compatible dt-bindings: simple-framebuffer: allow standalone compatible Minghao Chi (1): video: fbdev: mach64_ct: remove redundant res variable Tim Gardner (1): video: fbdev: nvidiafb: Use strscpy() to prevent buffer overflow Wang Hai (1): video: fbdev: smscufx: Fix null-ptr-deref in ufx_usb_probe() Xu Wang (2): backlight: lm3630a_bl: Remove redundant 'flush_workqueue()' calls video: fbdev: omapfb: Remove redundant 'flush_workqueue()' calls Yang Guang (2): video: fbdev: sis: use swap() to make code cleaner video: fbdev: omapfb: acx565akm: replace snprintf with sysfs_emit YueHaibing (1): video: fbdev: controlfb: Fix COMPILE_TEST build Z. Liu (1): video: fbdev: matroxfb: set maxvram of vbG200eW to the same as vbG200 to avoid black screen

2 4

[PATCH v3 0/3] drm/i915: Fix header test and log spam on !x86
by Lucas De Marchi 02 Feb '22

02 Feb '22

Some minor fixes and changes to help porting i915 to arm64, or even anything !x86. v3: No changes, just submit to the right mailing list. Lucas De Marchi (3): drm: Stop spamming log with drm_cache message drm/i915: Fix header test for !CONFIG_X86 drm/i915: Do not spam log with missing arch support drivers/gpu/drm/drm_cache.c | 9 +++------ drivers/gpu/drm/i915/i915_mm.h | 4 ++-- 2 files changed, 5 insertions(+), 8 deletions(-) -- 2.35.0

4 9

[PATCH v12 0/9] MIPS: JZ4780 and CI20 HDMI
by H. Nikolaus Schaller 02 Feb '22

02 Feb '22

PATCH V12 2022-01-31 13:26:54: This version reworks how hdmi ddc power is controlled by connector and not by ddc/hdmi bridge driver. Also some patches of the previous version of this series have been removed since they are already applied to mips-next/linux/next/v5.17-rc1. Fixes and changes: - repair interworking of dw-hdmi with connector-hdmi (by hns(a)goldelico.com) - fix JZ_REG_LCD_OSDC setup for jz4780 (by hns(a)goldelico.com and paul(a)crapouillou.net) - adjustments for ci20.dts to use connector gpio for +5v (suggested by several) - to add control of "ddc-en-gpios" to hdmi-connector driver (by hns(a)goldelico.com) - regulator code removed because we now use the "ddc-en-gpios" of the connector driver (suggested by paul(a)crapouillou.net) - bindings: addition of "ddc-i2c-bus" and "hdmi-5v-supply" removed (suggested by robh+dt(a)kernel.org) PATCH V11 2021-12-02 19:39:52: - patch 4/8: change devm_regulator_get_optional to devm_regulator_get and remove NULL check (requested by broonie(a)kernel.org) - patch 3/8: make hdmi-5v-supply required (requested by broonie(a)kernel.org) PATCH V10 2021-11-30 22:26:41: - patch 3/8: fix $id and $ref paths (found by robh(a)kernel.org) PATCH V9 2021-11-24 22:29:14: - patch 6/8: remove optional <0> for assigned-clocks and unintentionally included "unwedge" setup (found by paul(a)crapouillou.net) - patch 4/8: some cosmetics make regulator enable/disable only if not NULL (found by paul(a)crapouillou.net) simplify/fix error handling and driver cleanup on remove (proposed by paul(a)crapouillou.net) - patch 3/8: fix #include path in example (found by paul(a)crapouillou.net) fix missing "i" in unevaluatedProperties (found by robh(a)kernel.org) fix 4 spaces indentation for required: property (found by robh(a)kernel.org) PATCH V8 2021-11-23 19:14:00: - fix a bad editing result from patch 2/8 (found by paul(a)crapouillou.net) PATCH V7 2021-11-23 18:46:23: - changed gpio polarity of hdmi_power to 0 (suggested by paul(a)crapouillou.net) - fixed LCD1 irq number (bug found by paul(a)crapouillou.net) - removed "- 4" for calculating max_register (suggested by paul(a)crapouillou.net) - use unevaluatedPropertes instead of additionalProperties (suggested by robh(a)kernel.org) - moved and renamed ingenic,jz4780-hdmi.yaml (suggested by robh(a)kernel.org) - adjusted assigned-clocks changes to upstream which added some for SSI (by hns(a)goldelico.com) - rebased and tested with v5.16-rc2 + patch set drm/ingenic by paul(a)crapouillou.net (by hns(a)goldelico.com) PATCH V6 2021-11-10 20:43:33: - changed CONFIG_DRM_INGENIC_DW_HDMI to "m" (by hns(a)goldelico.com) - made ingenic-dw-hdmi an independent platform driver which can be compiled as module and removed error patch fixes for IPU (suggested by paul(a)crapouillou.net) - moved assigned-clocks from jz4780.dtsi to ci20.dts (suggested by paul(a)crapouillou.net) - fixed reg property in jz4780.dtsi to cover all registers incl. gamma and vee (by hns(a)goldelico.com) - added a base patch to calculate regmap size from DTS reg property (requested by paul(a)crapouillou.net) - restored resetting all bits except one in LCDOSDC (requested by paul(a)crapouillou.net) - clarified setting of cpos (suggested by paul(a)crapouillou.net) - moved bindings definition for ddc-i2c-bus (suggested by paul(a)crapouillou.net) - simplified mask definitions for JZ_LCD_DESSIZE (requested by paul(a)crapouillou.net) - removed setting alpha premultiplication (suggested by paul(a)crapouillou.net) - removed some comments (suggested by paul(a)crapouillou.net) PATCH V5 2021-10-05 14:28:44: - dropped mode_fixup and timings support in dw-hdmi as it is no longer needed in this V5 (by hns(a)goldelico.com) - dropped "drm/ingenic: add some jz4780 specific features" (stimulated by paul(a)crapouillou.net) - fixed typo in commit subject: "synopsis" -> "synopsys" (by hns(a)goldelico.com) - swapped clocks in jz4780.dtsi to match synopsys,dw-hdmi.yaml (by hns(a)goldelico.com) - improved, simplified, fixed, dtbschecked ingenic-jz4780-hdmi.yaml and made dependent of bridge/synopsys,dw-hdmi.yaml (based on suggestions by maxime(a)cerno.tech) - fixed binding vs. driver&DTS use of hdmi-5v regulator (suggested by maxime(a)cerno.tech) - dropped "drm/bridge: synopsis: Fix to properly handle HPD" - was a no longer needed workaround for a previous version (suggested by maxime(a)cerno.tech) PATCH V4 2021-09-27 18:44:38: - fix setting output_port = 1 (issue found by paul(a)crapouillou.net) - ci20.dts: convert to use hdmi-connector (by hns(a)goldelico.com) - add a hdmi-regulator to control +5V power (by hns(a)goldelico.com) - added a fix to dw-hdmi to call drm_kms_helper_hotplug_event on plugin event detection (by hns(a)goldelico.com) - always allocate extended descriptor but initialize only for jz4780 (by hns(a)goldelico.com) - updated to work on top of "[PATCH v3 0/6] drm/ingenic: Various improvements v3" (by paul(a)crapouillou.net) - rebased to v5.13-rc3 PATCH V3 2021-08-08 07:10:50: This series adds HDMI support for JZ4780 and CI20 board (and fixes one IPU related issue in registration error path) - [patch 1/8] switched from mode_fixup to atomic_check (suggested by robert.foss(a)linaro.org) - the call to the dw-hdmi specialization is still called mode_fixup - [patch 3/8] diverse fixes for ingenic-drm-drv (suggested by paul(a)crapouillou.net) - factor out some non-HDMI features of the jz4780 into a separate patch - multiple fixes around max height - do not change regmap config but a copy on stack - define some constants - factor out fixing of drm_init error path for IPU into separate patch - use FIELD_PREP() - [patch 8/8] conversion to component framework dropped (suggested by Laurent.pinchart(a)ideasonboard.com and paul(a)crapouillou.net) PATCH V2 2021-08-05 16:08:05: - code and commit messages revisited for checkpatch warnings - rebased on v5.14-rc4 - include (failed, hence RFC 8/8) attempt to convert to component framework (was suggested by Paul Cercueil <paul(a)crapouillou.net> a while ago) This series adds HDMI support for JZ4780 and CI20 board H. Nikolaus Schaller (5): drm/ingenic: prepare ingenic drm for later addition of JZ4780 drm/synopsys+ingenic: repair hot plug detection dw-hdmi/ingenic-dw-hdmi: repair interworking with hdmi-connector drm/bridge: display-connector: add ddc-en gpio support MIPS: DTS: CI20: fix how ddc power is enabled Paul Boddie (3): drm/ingenic: Add support for JZ4780 and HDMI output drm/ingenic: Add dw-hdmi driver specialization for jz4780 [RFC] drm/ingenic: add some more features specific to jz4780 Sam Ravnborg (1): dt-bindings: display: Add ingenic,jz4780-dw-hdmi DT Schema .../display/bridge/ingenic,jz4780-hdmi.yaml | 83 ++++++++++++++ arch/mips/boot/dts/ingenic/ci20.dts | 15 +-- drivers/gpu/drm/bridge/display-connector.c | 17 +++ drivers/gpu/drm/bridge/synopsys/dw-hdmi.c | 13 ++- drivers/gpu/drm/ingenic/Kconfig | 9 ++ drivers/gpu/drm/ingenic/Makefile | 1 + drivers/gpu/drm/ingenic/ingenic-drm-drv.c | 104 ++++++++++++++++- drivers/gpu/drm/ingenic/ingenic-drm.h | 38 +++++++ drivers/gpu/drm/ingenic/ingenic-dw-hdmi.c | 106 ++++++++++++++++++ include/drm/bridge/dw_hdmi.h | 1 + 10 files changed, 368 insertions(+), 19 deletions(-) create mode 100644 Documentation/devicetree/bindings/display/bridge/ingenic,jz4780-hdmi.yaml create mode 100644 drivers/gpu/drm/ingenic/ingenic-dw-hdmi.c -- 2.33.0

4 25

[next] arm: panel-edp.c:(.text+0xb74): undefined reference to `drm_panel_dp_aux_backlight'
by Naresh Kamboju 02 Feb '22

02 Feb '22

Following builds failed on Linux next 20220202 arm architecture. - arm-gcc-10-omap2plus_defconfig - arm-clang-nightly-shmobile_defconfig - arm-gcc-8-lpc32xx_defconfig - arm-clang-13-shmobile_defconfig - arm-gcc-10-shmobile_defconfig - arm-clang-11-shmobile_defconfig - arm-clang-11-omap2plus_defconfig - arm-clang-13-omap2plus_defconfig - arm-clang-11-lpc32xx_defconfig - arm-gcc-8-omap2plus_defconfig - arm-gcc-9-vexpress_defconfig - arm-clang-nightly-lpc32xx_defconfig - arm-gcc-9-shmobile_defconfig - arm-clang-13-lpc32xx_defconfig - arm-gcc-10-sama5_defconfig - arm-clang-11-vexpress_defconfig - arm-clang-11-sama5_defconfig - arm-gcc-9-omap2plus_defconfig - arm-clang-nightly-sama5_defconfig - arm-clang-13-vexpress_defconfig - arm-clang-nightly-vexpress_defconfig - arm-gcc-9-lpc32xx_defconfig - arm-clang-12-vexpress_defconfig - arm-gcc-10-vexpress_defconfig - arm-clang-12-shmobile_defconfig - arm-gcc-11-omap2plus_defconfig - arm-gcc-9-sama5_defconfig - arm-gcc-8-shmobile_defconfig - arm-gcc-10-lpc32xx_defconfig - arm-clang-12-omap2plus_defconfig - arm-gcc-8-vexpress_defconfig - arm-clang-12-sama5_defconfig - arm-clang-nightly-omap2plus_defconfig - arm-gcc-11-lpc32xx_defconfig - arm-gcc-11-sama5_defconfig - arm-gcc-11-shmobile_defconfig - arm-gcc-11-vexpress_defconfig - arm-gcc-8-sama5_defconfig - arm-clang-13-sama5_defconfig - arm-clang-12-lpc32xx_defconfig make --silent --keep-going --jobs=8 ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- 'CC=sccache arm-linux-gnueabihf-gcc' 'HOSTCC=sccache gcc' vexpress_defconfig make --silent --keep-going --jobs=8 ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- 'CC=sccache arm-linux-gnueabihf-gcc' 'HOSTCC=sccache gcc' arm-linux-gnueabihf-ld: drivers/gpu/drm/panel/panel-edp.o: in function `panel_edp_probe': panel-edp.c:(.text+0xb74): undefined reference to `drm_panel_dp_aux_backlight' make[1]: *** [/builds/linux/Makefile:1222: vmlinux] Error 1 Reported-by: Linux Kernel Functional Testing <lkft(a)linaro.org> meta data: ----------- git describe: next-20220202 git_repo: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git target_arch: arm kconfig: vexpress_defconfig toolchain: gcc-11 Build log: ------------- https://builds.tuxbuild.com/24XRim72vFXix6l6MdAJNENy6je/ Steps to reproduce: -------------------- # To install tuxmake on your system globally: # sudo pip3 install -U tuxmake # # See https://docs.tuxmake.org/ for complete documentation. # Original tuxmake command with fragments listed below. tuxmake --runtime podman --target-arch arm --toolchain gcc-11 --kconfig vexpress_defconfig tuxmake --runtime podman --target-arch arm --toolchain gcc-11 --kconfig https://builds.tuxbuild.com/24XRim72vFXix6l6MdAJNENy6je/config -- Linaro LKFT https://lkft.linaro.org

4 5

Re: [PATCH v4 00/10] Add MEMORY_DEVICE_COHERENT for coherent device memory mapping
by Jason Gunthorpe 02 Feb '22

02 Feb '22

On Wed, Feb 02, 2022 at 03:57:50PM +0100, Christoph Hellwig wrote: > On Thu, Jan 27, 2022 at 02:32:58PM -0800, Andrew Morton wrote: > > On Wed, 26 Jan 2022 21:09:39 -0600 Alex Sierra <alex.sierra(a)amd.com> wrote: > > > > > This patch series introduces MEMORY_DEVICE_COHERENT, a type of memory > > > owned by a device that can be mapped into CPU page tables like > > > MEMORY_DEVICE_GENERIC and can also be migrated like > > > MEMORY_DEVICE_PRIVATE. > > > > Some more reviewer input appears to be desirable here. > > > > I was going to tentatively add it to -mm and -next, but problems. > > 5.17-rc1's mm/migrate.c:migrate_vma_check_page() is rather different > > from the tree you patched. Please redo, refresh and resend? > > I really hate adding more types with the weird one off page refcount. > We need to clean that mess up first. Is there anyone who could give an outline of what is needed to make fsdax use compound pages/folios for its PMD stuff? I already suggested removing that as a way forward, and was shot down, but nobody is standing up to maintain this code and fix it :( We got devdax and the DRM stuff fixed now, so FSDAX is the next blocker on this work. The people who want this to advance have no idea about FSs or what to do, unfortunately. Jason

1 0

[PATCH v4] phy: dphy: Correct clk_pre parameter
by Liu Ying 02 Feb '22

02 Feb '22

The D-PHY specification (v1.2) explicitly mentions that the T-CLK-PRE parameter's unit is Unit Interval(UI) and the minimum value is 8. Also, kernel doc of the 'clk_pre' member of struct phy_configure_opts_mipi_dphy mentions that it should be in UI. However, the dphy core driver wrongly sets 'clk_pre' to 8000, which seems to hint that it's in picoseconds. So, let's fix the dphy core driver to correctly reflect the T-CLK-PRE parameter's minimum value according to the D-PHY specification. I'm assuming that all impacted custom drivers shall program values in TxByteClkHS cycles into hardware for the T-CLK-PRE parameter. The D-PHY specification mentions that the frequency of TxByteClkHS is exactly 1/8 the High-Speed(HS) bit rate(each HS bit consumes one UI). So, relevant custom driver code is changed to program those values as DIV_ROUND_UP(cfg->clk_pre, BITS_PER_BYTE), then. Note that I've only tested the patch with RM67191 DSI panel on i.MX8mq EVK. Help is needed to test with other i.MX8mq, Meson and Rockchip platforms, as I don't have the hardwares. Fixes: 2ed869990e14 ("phy: Add MIPI D-PHY configuration options") Cc: Andrzej Hajda <andrzej.hajda(a)intel.com> Cc: Neil Armstrong <narmstrong(a)baylibre.com> Cc: Robert Foss <robert.foss(a)linaro.org> Cc: Laurent Pinchart <Laurent.pinchart(a)ideasonboard.com> Cc: Jonas Karlman <jonas(a)kwiboo.se> Cc: Jernej Skrabec <jernej.skrabec(a)gmail.com> Cc: David Airlie <airlied(a)linux.ie> Cc: Daniel Vetter <daniel(a)ffwll.ch> Cc: Kishon Vijay Abraham I <kishon(a)ti.com> Cc: Vinod Koul <vkoul(a)kernel.org> Cc: Kevin Hilman <khilman(a)baylibre.com> Cc: Jerome Brunet <jbrunet(a)baylibre.com> Cc: Martin Blumenstingl <martin.blumenstingl(a)googlemail.com> Cc: Heiko Stuebner <heiko(a)sntech.de> Cc: Maxime Ripard <mripard(a)kernel.org> Cc: Guido Günther <agx(a)sigxcpu.org> Cc: Wyon Bi <bivvy.bi(a)rock-chips.com> Tested-by: Liu Ying <victor.liu(a)nxp.com> # RM67191 DSI panel on i.MX8mq EVK Reviewed-by: Andrzej Hajda <andrzej.hajda(a)intel.com> Reviewed-by: Neil Armstrong <narmstrong(a)baylibre.com> # for phy-meson-axg-mipi-dphy.c Tested-by: Neil Armstrong <narmstrong(a)baylibre.com> # for phy-meson-axg-mipi-dphy.c Tested-by: Guido Günther <agx(a)sigxcpu.org> # Librem 5 (imx8mq) with it's rather picky panel Reviewed-by: Laurent Pinchart <laurent.pinchart(a)ideasonboard.com> Signed-off-by: Liu Ying <victor.liu(a)nxp.com> --- v3->v4: * Fix commit message as the patch doesn't fix D-PHY documentation. v2->v3: * Drop D-PHY documentation change. (Laurent) * Collect R-b tags and T-b tags. * Cc Wyon. v1->v2: * Use BITS_PER_BYTE macro. (Andrzej) * Drop dsi argument from ui2bc() in nwl-dsi.c. drivers/gpu/drm/bridge/nwl-dsi.c | 12 +++++------- drivers/phy/amlogic/phy-meson-axg-mipi-dphy.c | 3 ++- drivers/phy/phy-core-mipi-dphy.c | 4 ++-- drivers/phy/rockchip/phy-rockchip-inno-dsidphy.c | 3 ++- 4 files changed, 11 insertions(+), 11 deletions(-) diff --git a/drivers/gpu/drm/bridge/nwl-dsi.c b/drivers/gpu/drm/bridge/nwl-dsi.c index a7389a0facfb..af07eeb47ca0 100644 --- a/drivers/gpu/drm/bridge/nwl-dsi.c +++ b/drivers/gpu/drm/bridge/nwl-dsi.c @@ -7,6 +7,7 @@ */ #include <linux/bitfield.h> +#include <linux/bits.h> #include <linux/clk.h> #include <linux/irq.h> #include <linux/math64.h> @@ -196,12 +197,9 @@ static u32 ps2bc(struct nwl_dsi *dsi, unsigned long long ps) /* * ui2bc - UI time periods to byte clock cycles */ -static u32 ui2bc(struct nwl_dsi *dsi, unsigned long long ui) +static u32 ui2bc(unsigned int ui) { - u32 bpp = mipi_dsi_pixel_format_to_bpp(dsi->format); - - return DIV64_U64_ROUND_UP(ui * dsi->lanes, - dsi->mode.clock * 1000 * bpp); + return DIV_ROUND_UP(ui, BITS_PER_BYTE); } /* @@ -232,12 +230,12 @@ static int nwl_dsi_config_host(struct nwl_dsi *dsi) } /* values in byte clock cycles */ - cycles = ui2bc(dsi, cfg->clk_pre); + cycles = ui2bc(cfg->clk_pre); DRM_DEV_DEBUG_DRIVER(dsi->dev, "cfg_t_pre: 0x%x\n", cycles); nwl_dsi_write(dsi, NWL_DSI_CFG_T_PRE, cycles); cycles = ps2bc(dsi, cfg->lpx + cfg->clk_prepare + cfg->clk_zero); DRM_DEV_DEBUG_DRIVER(dsi->dev, "cfg_tx_gap (pre): 0x%x\n", cycles); - cycles += ui2bc(dsi, cfg->clk_pre); + cycles += ui2bc(cfg->clk_pre); DRM_DEV_DEBUG_DRIVER(dsi->dev, "cfg_t_post: 0x%x\n", cycles); nwl_dsi_write(dsi, NWL_DSI_CFG_T_POST, cycles); cycles = ps2bc(dsi, cfg->hs_exit); diff --git a/drivers/phy/amlogic/phy-meson-axg-mipi-dphy.c b/drivers/phy/amlogic/phy-meson-axg-mipi-dphy.c index cd2332bf0e31..fdbd64c03e12 100644 --- a/drivers/phy/amlogic/phy-meson-axg-mipi-dphy.c +++ b/drivers/phy/amlogic/phy-meson-axg-mipi-dphy.c @@ -9,6 +9,7 @@ #include <linux/bitfield.h> #include <linux/bitops.h> +#include <linux/bits.h> #include <linux/clk.h> #include <linux/delay.h> #include <linux/io.h> @@ -250,7 +251,7 @@ static int phy_meson_axg_mipi_dphy_power_on(struct phy *phy) (DIV_ROUND_UP(priv->config.clk_zero, temp) << 16) | (DIV_ROUND_UP(priv->config.clk_prepare, temp) << 24)); regmap_write(priv->regmap, MIPI_DSI_CLK_TIM1, - DIV_ROUND_UP(priv->config.clk_pre, temp)); + DIV_ROUND_UP(priv->config.clk_pre, BITS_PER_BYTE)); regmap_write(priv->regmap, MIPI_DSI_HS_TIM, DIV_ROUND_UP(priv->config.hs_exit, temp) | diff --git a/drivers/phy/phy-core-mipi-dphy.c b/drivers/phy/phy-core-mipi-dphy.c index 288c9c67aa74..ccb4045685cd 100644 --- a/drivers/phy/phy-core-mipi-dphy.c +++ b/drivers/phy/phy-core-mipi-dphy.c @@ -36,7 +36,7 @@ int phy_mipi_dphy_get_default_config(unsigned long pixel_clock, cfg->clk_miss = 0; cfg->clk_post = 60000 + 52 * ui; - cfg->clk_pre = 8000; + cfg->clk_pre = 8; cfg->clk_prepare = 38000; cfg->clk_settle = 95000; cfg->clk_term_en = 0; @@ -97,7 +97,7 @@ int phy_mipi_dphy_config_validate(struct phy_configure_opts_mipi_dphy *cfg) if (cfg->clk_post < (60000 + 52 * ui)) return -EINVAL; - if (cfg->clk_pre < 8000) + if (cfg->clk_pre < 8) return -EINVAL; if (cfg->clk_prepare < 38000 || cfg->clk_prepare > 95000) diff --git a/drivers/phy/rockchip/phy-rockchip-inno-dsidphy.c b/drivers/phy/rockchip/phy-rockchip-inno-dsidphy.c index 347dc79a18c1..630e01b5c19b 100644 --- a/drivers/phy/rockchip/phy-rockchip-inno-dsidphy.c +++ b/drivers/phy/rockchip/phy-rockchip-inno-dsidphy.c @@ -5,6 +5,7 @@ * Author: Wyon Bi <bivvy.bi(a)rock-chips.com> */ +#include <linux/bits.h> #include <linux/kernel.h> #include <linux/clk.h> #include <linux/iopoll.h> @@ -364,7 +365,7 @@ static void inno_dsidphy_mipi_mode_enable(struct inno_dsidphy *inno) * The value of counter for HS Tclk-pre * Tclk-pre = Tpin_txbyteclkhs * value */ - clk_pre = DIV_ROUND_UP(cfg->clk_pre, t_txbyteclkhs); + clk_pre = DIV_ROUND_UP(cfg->clk_pre, BITS_PER_BYTE); /* * The value of counter for HS Tlpx Time -- 2.25.1

2 1

[PATCH v3 1/2] dt-bindings: display: Turn lvds.yaml into a generic schema
by Maxime Ripard 02 Feb '22

02 Feb '22

The lvds.yaml file so far was both defining the generic LVDS properties (such as data-mapping) that could be used for any LVDS sink, but also the panel-lvds binding. That last binding was to describe LVDS panels simple enough, and had a number of other bindings using it as a base to specialise it further. However, this situation makes it fairly hard to extend and reuse both the generic parts, and the panel-lvds itself. Let's remove the panel-lvds parts and leave only the generic LVDS properties. Reviewed-by: Rob Herring <robh(a)kernel.org> Signed-off-by: Maxime Ripard <maxime(a)cerno.tech> --- Changes from v2: - Fix references to that file Changes from v1: - Moved the schema out of panel --- .../bindings/display/bridge/lvds-codec.yaml | 2 +- .../bindings/display/{panel => }/lvds.yaml | 31 ++----------------- .../display/panel/advantech,idk-1110wr.yaml | 19 ++++++++++-- .../display/panel/innolux,ee101ia-01d.yaml | 23 ++++++++++++-- .../display/panel/mitsubishi,aa104xd12.yaml | 19 ++++++++++-- .../display/panel/mitsubishi,aa121td01.yaml | 19 ++++++++++-- .../display/panel/sgd,gktw70sdae4se.yaml | 19 ++++++++++-- MAINTAINERS | 2 +- 8 files changed, 93 insertions(+), 41 deletions(-) rename Documentation/devicetree/bindings/display/{panel => }/lvds.yaml (86%) diff --git a/Documentation/devicetree/bindings/display/bridge/lvds-codec.yaml b/Documentation/devicetree/bindings/display/bridge/lvds-codec.yaml index 5079c1cc337b..27b905b81b12 100644 --- a/Documentation/devicetree/bindings/display/bridge/lvds-codec.yaml +++ b/Documentation/devicetree/bindings/display/bridge/lvds-codec.yaml @@ -67,7 +67,7 @@ properties: - vesa-24 description: | The color signals mapping order. See details in - Documentation/devicetree/bindings/display/panel/lvds.yaml + Documentation/devicetree/bindings/display/lvds.yaml port@1: $ref: /schemas/graph.yaml#/properties/port diff --git a/Documentation/devicetree/bindings/display/panel/lvds.yaml b/Documentation/devicetree/bindings/display/lvds.yaml similarity index 86% rename from Documentation/devicetree/bindings/display/panel/lvds.yaml rename to Documentation/devicetree/bindings/display/lvds.yaml index 49460c9dceea..55751402fb13 100644 --- a/Documentation/devicetree/bindings/display/panel/lvds.yaml +++ b/Documentation/devicetree/bindings/display/lvds.yaml @@ -1,10 +1,10 @@ # SPDX-License-Identifier: GPL-2.0 %YAML 1.2 --- -$id: http://devicetree.org/schemas/display/panel/lvds.yaml# +$id: http://devicetree.org/schemas/display/lvds.yaml# $schema: http://devicetree.org/meta-schemas/core.yaml# -title: LVDS Display Panel +title: LVDS Display Common Properties maintainers: - Laurent Pinchart <laurent.pinchart+renesas(a)ideasonboard.com> @@ -26,18 +26,7 @@ description: |+ Device compatible with those specifications have been marketed under the FPD-Link and FlatLink brands. -allOf: - - $ref: panel-common.yaml# - properties: - compatible: - contains: - const: panel-lvds - description: - Shall contain "panel-lvds" in addition to a mandatory panel-specific - compatible string defined in individual panel bindings. The "panel-lvds" - value shall never be used on its own. - data-mapping: enum: - jeida-18 @@ -96,22 +85,6 @@ properties: If set, reverse the bit order described in the data mappings below on all data lanes, transmitting bits for slots 6 to 0 instead of 0 to 6. - port: true - ports: true - -required: - - compatible - - data-mapping - - width-mm - - height-mm - - panel-timing - -oneOf: - - required: - - port - - required: - - ports - additionalProperties: true ... diff --git a/Documentation/devicetree/bindings/display/panel/advantech,idk-1110wr.yaml b/Documentation/devicetree/bindings/display/panel/advantech,idk-1110wr.yaml index 93878c2cd370..3a8c2c11f9bd 100644 --- a/Documentation/devicetree/bindings/display/panel/advantech,idk-1110wr.yaml +++ b/Documentation/devicetree/bindings/display/panel/advantech,idk-1110wr.yaml @@ -11,13 +11,23 @@ maintainers: - Thierry Reding <thierry.reding(a)gmail.com> allOf: - - $ref: lvds.yaml# + - $ref: panel-common.yaml# + - $ref: /schemas/display/lvds.yaml/# + +select: + properties: + compatible: + contains: + const: advantech,idk-1110wr + + required: + - compatible properties: compatible: items: - const: advantech,idk-1110wr - - {} # panel-lvds, but not listed here to avoid false select + - const: panel-lvds data-mapping: const: jeida-24 @@ -35,6 +45,11 @@ additionalProperties: false required: - compatible + - data-mapping + - width-mm + - height-mm + - panel-timing + - port examples: - |+ diff --git a/Documentation/devicetree/bindings/display/panel/innolux,ee101ia-01d.yaml b/Documentation/devicetree/bindings/display/panel/innolux,ee101ia-01d.yaml index a69681e724cb..566e11f6bfc0 100644 --- a/Documentation/devicetree/bindings/display/panel/innolux,ee101ia-01d.yaml +++ b/Documentation/devicetree/bindings/display/panel/innolux,ee101ia-01d.yaml @@ -11,15 +11,26 @@ maintainers: - Thierry Reding <thierry.reding(a)gmail.com> allOf: - - $ref: lvds.yaml# + - $ref: panel-common.yaml# + - $ref: /schemas/display/lvds.yaml/# + +select: + properties: + compatible: + contains: + const: innolux,ee101ia-01d + + required: + - compatible properties: compatible: items: - const: innolux,ee101ia-01d - - {} # panel-lvds, but not listed here to avoid false select + - const: panel-lvds backlight: true + data-mapping: true enable-gpios: true power-supply: true width-mm: true @@ -27,5 +38,13 @@ properties: panel-timing: true port: true +required: + - compatible + - data-mapping + - width-mm + - height-mm + - panel-timing + - port + additionalProperties: false ... diff --git a/Documentation/devicetree/bindings/display/panel/mitsubishi,aa104xd12.yaml b/Documentation/devicetree/bindings/display/panel/mitsubishi,aa104xd12.yaml index b5e7ee230fa6..5cf3c588f46d 100644 --- a/Documentation/devicetree/bindings/display/panel/mitsubishi,aa104xd12.yaml +++ b/Documentation/devicetree/bindings/display/panel/mitsubishi,aa104xd12.yaml @@ -11,13 +11,23 @@ maintainers: - Thierry Reding <thierry.reding(a)gmail.com> allOf: - - $ref: lvds.yaml# + - $ref: panel-common.yaml# + - $ref: /schemas/display/lvds.yaml/# + +select: + properties: + compatible: + contains: + const: mitsubishi,aa104xd12 + + required: + - compatible properties: compatible: items: - const: mitsubishi,aa104xd12 - - {} # panel-lvds, but not listed here to avoid false select + - const: panel-lvds vcc-supply: description: Reference to the regulator powering the panel VCC pins. @@ -39,6 +49,11 @@ additionalProperties: false required: - compatible - vcc-supply + - data-mapping + - width-mm + - height-mm + - panel-timing + - port examples: - |+ diff --git a/Documentation/devicetree/bindings/display/panel/mitsubishi,aa121td01.yaml b/Documentation/devicetree/bindings/display/panel/mitsubishi,aa121td01.yaml index 977c50a85b67..54750cc5440d 100644 --- a/Documentation/devicetree/bindings/display/panel/mitsubishi,aa121td01.yaml +++ b/Documentation/devicetree/bindings/display/panel/mitsubishi,aa121td01.yaml @@ -11,13 +11,23 @@ maintainers: - Thierry Reding <thierry.reding(a)gmail.com> allOf: - - $ref: lvds.yaml# + - $ref: panel-common.yaml# + - $ref: /schemas/display/lvds.yaml/# + +select: + properties: + compatible: + contains: + const: mitsubishi,aa121td01 + + required: + - compatible properties: compatible: items: - const: mitsubishi,aa121td01 - - {} # panel-lvds, but not listed here to avoid false select + - const: panel-lvds vcc-supply: description: Reference to the regulator powering the panel VCC pins. @@ -39,6 +49,11 @@ additionalProperties: false required: - compatible - vcc-supply + - data-mapping + - width-mm + - height-mm + - panel-timing + - port examples: - |+ diff --git a/Documentation/devicetree/bindings/display/panel/sgd,gktw70sdae4se.yaml b/Documentation/devicetree/bindings/display/panel/sgd,gktw70sdae4se.yaml index e63a570ae59d..44e02decdf3a 100644 --- a/Documentation/devicetree/bindings/display/panel/sgd,gktw70sdae4se.yaml +++ b/Documentation/devicetree/bindings/display/panel/sgd,gktw70sdae4se.yaml @@ -11,13 +11,23 @@ maintainers: - Thierry Reding <thierry.reding(a)gmail.com> allOf: - - $ref: lvds.yaml# + - $ref: panel-common.yaml# + - $ref: /schemas/display/lvds.yaml/# + +select: + properties: + compatible: + contains: + const: sgd,gktw70sdae4se + + required: + - compatible properties: compatible: items: - const: sgd,gktw70sdae4se - - {} # panel-lvds, but not listed here to avoid false select + - const: panel-lvds data-mapping: const: jeida-18 @@ -35,6 +45,11 @@ additionalProperties: false required: - compatible + - port + - data-mapping + - width-mm + - height-mm + - panel-timing examples: - |+ diff --git a/MAINTAINERS b/MAINTAINERS index 41d2916600e0..368072da0a05 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -6079,7 +6079,7 @@ L: dri-devel(a)lists.freedesktop.org T: git git://anongit.freedesktop.org/drm/drm-misc S: Maintained F: drivers/gpu/drm/panel/panel-lvds.c -F: Documentation/devicetree/bindings/display/panel/lvds.yaml +F: Documentation/devicetree/bindings/display/lvds.yaml DRM DRIVER FOR MANTIX MLAF057WE51 PANELS M: Guido Günther <agx(a)sigxcpu.org> -- 2.34.1

2 11

[PATCH 1/9] drm/ttm: add common accounting to the resource mgr v3
by Christian König 02 Feb '22

02 Feb '22

It makes sense to have this in the common manager for debugging and accounting of how much resources are used. v2: cleanup kerneldoc a bit v3: drop the atomic, update counter under lock instead Signed-off-by: Christian König <christian.koenig(a)amd.com> Reviewed-by: Huang Rui <ray.huang(a)amd.com> --- drivers/gpu/drm/ttm/ttm_resource.c | 30 ++++++++++++++++++++++++++++++ include/drm/ttm/ttm_resource.h | 11 +++++++++-- 2 files changed, 39 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/ttm/ttm_resource.c b/drivers/gpu/drm/ttm/ttm_resource.c index 68344c90549b..7f9ec64ebaba 100644 --- a/drivers/gpu/drm/ttm/ttm_resource.c +++ b/drivers/gpu/drm/ttm/ttm_resource.c @@ -41,6 +41,8 @@ void ttm_resource_init(struct ttm_buffer_object *bo, const struct ttm_place *place, struct ttm_resource *res) { + struct ttm_resource_manager *man; + res->start = 0; res->num_pages = PFN_UP(bo->base.size); res->mem_type = place->mem_type; @@ -50,6 +52,11 @@ void ttm_resource_init(struct ttm_buffer_object *bo, res->bus.is_iomem = false; res->bus.caching = ttm_cached; res->bo = bo; + + man = ttm_manager_type(bo->bdev, place->mem_type); + spin_lock(&bo->bdev->lru_lock); + man->usage += bo->base.size; + spin_unlock(&bo->bdev->lru_lock); } EXPORT_SYMBOL(ttm_resource_init); @@ -65,6 +72,9 @@ EXPORT_SYMBOL(ttm_resource_init); void ttm_resource_fini(struct ttm_resource_manager *man, struct ttm_resource *res) { + spin_lock(&man->bdev->lru_lock); + man->usage -= res->bo->base.size; + spin_unlock(&man->bdev->lru_lock); } EXPORT_SYMBOL(ttm_resource_fini); @@ -166,6 +176,7 @@ void ttm_resource_manager_init(struct ttm_resource_manager *man, spin_lock_init(&man->move_lock); man->bdev = bdev; man->size = p_size; + man->usage = 0; for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) INIT_LIST_HEAD(&man->lru[i]); @@ -226,6 +237,24 @@ int ttm_resource_manager_evict_all(struct ttm_device *bdev, } EXPORT_SYMBOL(ttm_resource_manager_evict_all); +/** + * ttm_resource_manager_usage + * + * @man: A memory manager object. + * + * Return how many resources are currently used. + */ +uint64_t ttm_resource_manager_usage(struct ttm_resource_manager *man) +{ + uint64_t usage; + + spin_lock(&man->bdev->lru_lock); + usage = man->usage; + spin_unlock(&man->bdev->lru_lock); + return usage; +} +EXPORT_SYMBOL(ttm_resource_manager_usage); + /** * ttm_resource_manager_debug * @@ -238,6 +267,7 @@ void ttm_resource_manager_debug(struct ttm_resource_manager *man, drm_printf(p, " use_type: %d\n", man->use_type); drm_printf(p, " use_tt: %d\n", man->use_tt); drm_printf(p, " size: %llu\n", man->size); + drm_printf(p, " usage: %llu\n", ttm_resource_manager_usage(man)); if (man->func->debug) man->func->debug(man, p); } diff --git a/include/drm/ttm/ttm_resource.h b/include/drm/ttm/ttm_resource.h index 69eea9d6399b..5516d6340aa7 100644 --- a/include/drm/ttm/ttm_resource.h +++ b/include/drm/ttm/ttm_resource.h @@ -27,6 +27,7 @@ #include <linux/types.h> #include <linux/mutex.h> +#include <linux/atomic.h> #include <linux/dma-buf-map.h> #include <linux/dma-fence.h> #include <drm/drm_print.h> @@ -130,10 +131,15 @@ struct ttm_resource_manager { struct dma_fence *move; /* - * Protected by the global->lru_lock. + * Protected by the bdev->lru_lock. */ - struct list_head lru[TTM_MAX_BO_PRIORITY]; + + /** + * @usage: How much of the resources are used, protected by the + * bdev->lru_lock. + */ + uint64_t usage; }; /** @@ -283,6 +289,7 @@ void ttm_resource_manager_init(struct ttm_resource_manager *man, int ttm_resource_manager_evict_all(struct ttm_device *bdev, struct ttm_resource_manager *man); +uint64_t ttm_resource_manager_usage(struct ttm_resource_manager *man); void ttm_resource_manager_debug(struct ttm_resource_manager *man, struct drm_printer *p); -- 2.25.1

1 8

[PATCH v3 0/3] Fix regression introduced by disabling accelerated scrolling in fbcon
by Helge Deller 02 Feb '22

02 Feb '22

This series reverts two patches which disabled scrolling acceleration in fbcon/fbdev. Those patches introduced a regression for fbdev-supported graphic cards because of the performance penalty by doing screen scrolling by software instead of using existing 2D hardware acceleration. The third patch introduces a new config option CONFIG_FRAMEBUFFER_CONSOLE_LEGACY_ACCELERATION, which -if enabled- switches fbcon to utilize the bitblt and fillrect hardware acceleration in the framebuffer console. If disabled, such acceleration will not be used, even if it is supported by the graphics hardware driver. This series is being discussed on linux-fbdev and dri-devel mailing lists. Helge Helge Deller (3): Revert "fbdev: Garbage collect fbdev scrolling acceleration, part 1 (from TODO list)" Revert "fbcon: Disable accelerated scrolling" fbcon: Add option to enable legacy hardware acceleration Documentation/gpu/todo.rst | 24 - drivers/video/console/Kconfig | 11 + drivers/video/fbdev/core/bitblit.c | 16 + drivers/video/fbdev/core/fbcon.c | 557 +++++++++++++++++++++++- drivers/video/fbdev/core/fbcon.h | 72 +++ drivers/video/fbdev/core/fbcon_ccw.c | 28 +- drivers/video/fbdev/core/fbcon_cw.c | 28 +- drivers/video/fbdev/core/fbcon_rotate.h | 9 + drivers/video/fbdev/core/fbcon_ud.c | 37 +- drivers/video/fbdev/core/tileblit.c | 16 + drivers/video/fbdev/skeletonfb.c | 12 +- include/linux/fb.h | 2 +- 12 files changed, 744 insertions(+), 68 deletions(-) -- 2.34.1

2 7

[RFC PATCH] drm/i915: Fix vma lifetime
by Thomas Hellström 02 Feb '22

02 Feb '22

It's unclear what reference the initial vma refcount refers to. A vma can have multiple weak references, the object vma list, the vm's bound list and the gt's closed_list, and that initial vma refcount can be put from lookups of all these lists. The problem is that if another user is also holding another vma refcount, removal of other weak references will not happen and that initial vma refcount may be put again, causing refcount underflow. In addition, putting the initial reference will not stop anyone from calling i915_vma_close() which may put the vma back on the gt's closed list, from which it is never removed so when the vma is finally freed, the list becomes corrupted. From what I can tell this never happens in the current code, because __i915_vm_close() is never called from gem context close due to a vm open-count inconsistency which saves us here. Introduce a selftest that illustrates the problems mentioned above. Also fix this by re-introducing i915_vma_destroy() which removes all weak references of the vma and *then* puts the initial vma refcount. Perhaps a better name would be i915_vma_revoke() or i915_vma_zombify(), since other callers may still hold a refcount, but with the prospect of being able to replace that with the object lock in the near future, let's stick with i915_vma_destroy(). The newly introduced selftest produces the following output without the introduced fixes: [ 316.077020] i915: Running i915_vma_mock_selftests/igt_vma_lifetime [ 316.083261] Refcount = 2. Now closing vm. [ 316.087708] Refcount = 1. Now closing vma. [ 316.091858] i915_vma_close() caused list corruption. [ 316.096824] Refcount = 1. Now freeing obj. [ 316.101073] Refcount = 0. Now freeing vma. [ 316.105180] ------------[ cut here ]------------ [ 316.109819] refcount_t: underflow; use-after-free. [ 316.114632] WARNING: CPU: 1 PID: 7701 at lib/refcount.c:28 refcount_warn_saturate+0xa6/0xf0 Fixes: 76f9764cc3d5 ("drm/i915: Introduce a vma.kref") Cc: Chris Wilson <chris(a)chris-wilson.co.uk> Cc: Imre Deak <imre.deak(a)intel.com> Signed-off-by: Thomas Hellström <thomas.hellstrom(a)linux.intel.com> --- drivers/gpu/drm/i915/gem/i915_gem_object.c | 14 +--- .../drm/i915/gem/selftests/i915_gem_mman.c | 4 +- drivers/gpu/drm/i915/gt/intel_gtt.c | 17 ++-- drivers/gpu/drm/i915/i915_vma.c | 79 ++++++++++++++++--- drivers/gpu/drm/i915/i915_vma.h | 3 + drivers/gpu/drm/i915/selftests/i915_vma.c | 69 ++++++++++++++++ 6 files changed, 152 insertions(+), 34 deletions(-) diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c b/drivers/gpu/drm/i915/gem/i915_gem_object.c index e03e362d320b..78c4cbe82031 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c @@ -267,12 +267,6 @@ void __i915_gem_object_pages_fini(struct drm_i915_gem_object *obj) if (!list_empty(&obj->vma.list)) { struct i915_vma *vma; - /* - * Note that the vma keeps an object reference while - * it is active, so it *should* not sleep while we - * destroy it. Our debug code errs insits it *might*. - * For the moment, play along. - */ spin_lock(&obj->vma.lock); while ((vma = list_first_entry_or_null(&obj->vma.list, struct i915_vma, @@ -280,13 +274,7 @@ void __i915_gem_object_pages_fini(struct drm_i915_gem_object *obj) GEM_BUG_ON(vma->obj != obj); spin_unlock(&obj->vma.lock); - /* Verify that the vma is unbound under the vm mutex. */ - mutex_lock(&vma->vm->mutex); - atomic_and(~I915_VMA_PIN_MASK, &vma->flags); - __i915_vma_unbind(vma); - mutex_unlock(&vma->vm->mutex); - - __i915_vma_put(vma); + i915_vma_destroy(vma); spin_lock(&obj->vma.lock); } diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c index ba29767348be..af36bffd064b 100644 --- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c +++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c @@ -167,7 +167,7 @@ static int check_partial_mapping(struct drm_i915_gem_object *obj, out: i915_gem_object_lock(obj, NULL); - __i915_vma_put(vma); + i915_vma_destroy(vma); i915_gem_object_unlock(obj); return err; } @@ -264,7 +264,7 @@ static int check_partial_mappings(struct drm_i915_gem_object *obj, return err; i915_gem_object_lock(obj, NULL); - __i915_vma_put(vma); + i915_vma_destroy(vma); i915_gem_object_unlock(obj); if (igt_timeout(end_time, diff --git a/drivers/gpu/drm/i915/gt/intel_gtt.c b/drivers/gpu/drm/i915/gt/intel_gtt.c index 46be4197b93f..c9a24b021a7e 100644 --- a/drivers/gpu/drm/i915/gt/intel_gtt.c +++ b/drivers/gpu/drm/i915/gt/intel_gtt.c @@ -105,14 +105,19 @@ void __i915_vm_close(struct i915_address_space *vm) list_for_each_entry_safe(vma, vn, &vm->bound_list, vm_link) { struct drm_i915_gem_object *obj = vma->obj; - /* Keep the obj (and hence the vma) alive as _we_ destroy it */ - if (!kref_get_unless_zero(&obj->base.refcount)) + if (!kref_get_unless_zero(&obj->base.refcount)) { + /* + * Unbind the dying vma to ensure the bound_list + * is completely drained. We leave the destruction to + * the object destructor. + */ + atomic_and(~I915_VMA_PIN_MASK, &vma->flags); + WARN_ON(__i915_vma_unbind(vma)); continue; + } - atomic_and(~I915_VMA_PIN_MASK, &vma->flags); - WARN_ON(__i915_vma_unbind(vma)); - __i915_vma_put(vma); - + /* Keep the obj (and hence the vma) alive as _we_ destroy it */ + i915_vma_destroy_locked(vma); i915_gem_object_put(obj); } GEM_BUG_ON(!list_empty(&vm->bound_list)); diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index 7815850b644e..689a10286a9e 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -1606,15 +1606,27 @@ void i915_vma_reopen(struct i915_vma *vma) void i915_vma_release(struct kref *ref) { struct i915_vma *vma = container_of(ref, typeof(*vma), ref); + + i915_vm_put(vma->vm); + i915_active_fini(&vma->active); + GEM_WARN_ON(vma->resource); + i915_vma_free(vma); +} + +static inline void force_unbind(struct i915_vma *vma) +{ + if (!drm_mm_node_allocated(&vma->node)) + return; + + atomic_and(~I915_VMA_PIN_MASK, &vma->flags); + WARN_ON(__i915_vma_unbind(vma)); + GEM_BUG_ON(drm_mm_node_allocated(&vma->node)); +} + +static inline void release_references(struct i915_vma *vma) +{ struct drm_i915_gem_object *obj = vma->obj; - if (drm_mm_node_allocated(&vma->node)) { - mutex_lock(&vma->vm->mutex); - atomic_and(~I915_VMA_PIN_MASK, &vma->flags); - WARN_ON(__i915_vma_unbind(vma)); - mutex_unlock(&vma->vm->mutex); - GEM_BUG_ON(drm_mm_node_allocated(&vma->node)); - } GEM_BUG_ON(i915_vma_is_active(vma)); spin_lock(&obj->vma.lock); @@ -1624,11 +1636,49 @@ void i915_vma_release(struct kref *ref) spin_unlock(&obj->vma.lock); __i915_vma_remove_closed(vma); - i915_vm_put(vma->vm); - i915_active_fini(&vma->active); - GEM_WARN_ON(vma->resource); - i915_vma_free(vma); + __i915_vma_put(vma); +} + +/** + * i915_vma_destroy_locked - Remove all weak reference to the vma and put + * the initial reference. + * + * This function should be called when it's decided the vma isn't needed + * anymore. The caller must assure that it doesn't race with another lookup + * plus destroy, typically by taking an appropriate reference. + * + * Current callsites are + * - __i915_gem_object_pages_fini() + * - __i915_vm_close() - Blocks the above function by taking a reference on + * the object. + * - __i915_vma_parked() - Blocks the above functions by taking an open-count on + * the vm and a reference on the object. + * + * Because of locks taken during destruction, a vma is also guaranteed to + * stay alive while the following locks are held if it was looked up while + * holding one of the locks: + * - vm->mutex + * - obj->vma.lock + * - gt->closed_lock + * + * A vma user can also temporarily keep the vma alive while holding a vma + * reference. + */ +void i915_vma_destroy_locked(struct i915_vma *vma) +{ + lockdep_assert_held(&vma->vm->mutex); + + force_unbind(vma); + release_references(vma); +} + +void i915_vma_destroy(struct i915_vma *vma) +{ + mutex_lock(&vma->vm->mutex); + force_unbind(vma); + mutex_unlock(&vma->vm->mutex); + release_references(vma); } void i915_vma_parked(struct intel_gt *gt) @@ -1643,10 +1693,13 @@ void i915_vma_parked(struct intel_gt *gt) /* XXX All to avoid keeping a reference on i915_vma itself */ - if (!kref_get_unless_zero(&obj->base.refcount)) + if (!kref_get_unless_zero(&obj->base.refcount)) { + list_del_init(&vma->closed_link); continue; + } if (!i915_vm_tryopen(vm)) { + list_del_init(&vma->closed_link); i915_gem_object_put(obj); continue; } @@ -1662,7 +1715,7 @@ void i915_vma_parked(struct intel_gt *gt) if (i915_gem_object_trylock(obj, NULL)) { INIT_LIST_HEAD(&vma->closed_link); - __i915_vma_put(vma); + i915_vma_destroy(vma); i915_gem_object_unlock(obj); } else { /* back you go.. */ diff --git a/drivers/gpu/drm/i915/i915_vma.h b/drivers/gpu/drm/i915/i915_vma.h index 011af044ad4f..67ae7341c7e0 100644 --- a/drivers/gpu/drm/i915/i915_vma.h +++ b/drivers/gpu/drm/i915/i915_vma.h @@ -236,6 +236,9 @@ static inline void __i915_vma_put(struct i915_vma *vma) kref_put(&vma->ref, i915_vma_release); } +void i915_vma_destroy_locked(struct i915_vma *vma); +void i915_vma_destroy(struct i915_vma *vma); + #define assert_vma_held(vma) dma_resv_assert_held((vma)->obj->base.resv) static inline void i915_vma_lock(struct i915_vma *vma) diff --git a/drivers/gpu/drm/i915/selftests/i915_vma.c b/drivers/gpu/drm/i915/selftests/i915_vma.c index 0280605a2673..94aba0c48793 100644 --- a/drivers/gpu/drm/i915/selftests/i915_vma.c +++ b/drivers/gpu/drm/i915/selftests/i915_vma.c @@ -913,6 +913,74 @@ static int igt_vma_partial(void *arg) return err; } +static int igt_vma_lifetime(void *arg) +{ + struct i915_ggtt *ggtt = arg; + struct drm_i915_private *i915 = ggtt->vm.i915; + struct drm_i915_gem_object *obj; + struct i915_vma *vma; + struct i915_ppgtt *ppgtt = mock_ppgtt(i915, "refcount"); + struct i915_address_space *vm; + bool refctd = 0; + int err = 0; + + if (!ppgtt) + return -ENOMEM; + + vm = &ppgtt->vm; + obj = i915_gem_object_create_internal(i915, PAGE_SIZE); + if (IS_ERR(obj)) { + err = PTR_ERR(obj); + goto out_obj; + } + + vma = i915_vma_instance(obj, vm, NULL); + if (IS_ERR(vma)) { + err = PTR_ERR(vma); + goto out_vma; + } + + atomic_set(&vma->open_count, 1); + err = i915_vma_pin(vma, 0, 0, PIN_USER); + + __i915_vma_get(vma); + refctd = true; + + /* + * Now run vma close, vm close and obj unref, all wanting to + * destroy the vma. The vma should survive until we put it + * without any refcount underflow warning. Moving forward we might + * be using object lock instead of the vma refcount, but then we'd + * have to put the vma before object put ofc. + */ + + pr_info("Refcount = %d. Now closing vm.\n", kref_read(&vma->ref)); + __i915_vm_close(vm); + pr_info("Refcount = %d. Now closing vma.\n", kref_read(&vma->ref)); + i915_vma_close(vma); + i915_vma_parked(vma->vm->gt); + + /* Did i915_vma_parked forget to remove the vma from the list? */ + if (i915_vma_is_closed(vma)) { + pr_err("i915_vma_close() caused list corruption.\n"); + list_del_init(&vma->closed_link); + err = -EINVAL; + } + + pr_info("Refcount = %d. Now freeing obj.\n", kref_read(&vma->ref)); +out_vma: + i915_gem_object_put(obj); + i915_gem_flush_free_objects(i915); + if (refctd) { + pr_info("Refcount = %d. Now freeing vma.\n", kref_read(&vma->ref)); + __i915_vma_put(vma); + } +out_obj: + i915_vm_put(vm); + + return err; +} + int i915_vma_mock_selftests(void) { static const struct i915_subtest tests[] = { @@ -920,6 +988,7 @@ int i915_vma_mock_selftests(void) SUBTEST(igt_vma_pin1), SUBTEST(igt_vma_rotate_remap), SUBTEST(igt_vma_partial), + SUBTEST(igt_vma_lifetime), }; struct drm_i915_private *i915; struct intel_gt *gt; -- 2.34.1

1 0

← Newer
1
...
80
81
82
83
84
85
86
87
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

dri-devel February 2022