Project: chromium Issues People Development process History Sign in
New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.
Starred by 16 users
Status: Fixed
Owner:
Closed: Mar 2014
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux
Pri: 1
Type: Bug

Blocking:
issue 516679



Sign in to add a comment
Shrinking the window doesn't redraw/re-fit all of the elements on Intel drivers, likely related to partial swaps
Reported by saiarcot...@gmail.com, Jan 30 2014 Back to list
UserAgent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Ubuntu Chromium/34.0.1809.0 Chrome/34.0.1809.0 Safari/537.36

Steps to reproduce the problem:
1. Make sure the browser is in windowed mode.
2. Drag the window inward either from the left side or the right side, so that the width of the window decreases

What is the expected behavior?
The address bar and any extension buttons are resized and moved to fit the new width, and the close buttons (which are on the right side, but should be on the left, but that's a different bug) 

What went wrong?
The address bar, extension buttons, and close buttons remain in the same relative position and size. Hovering over where the extension and close buttons should be causes the buttons to reappear, but they still look out of place.

Did this work before? N/A 

Chrome version: 34.0.1809.0  Channel: dev
OS Version: Ubuntu 13.10
Flash Version: Shockwave Flash 12.0 r0

This occurs in Aura, but works fine in GTK.
 
Also, increasing the width doesn't cause this problem, and switching to another window and then back to Chromium fixes the appearance.

I've attached a video of this.
Chromium Resize.avi
4.1 MB Download
Labels: Proj-DesktopAura
Comment 3 by e...@chromium.org, Feb 4 2014
Cc: sadrul@chromium.org piman@chromium.org
Can't reproduce this locally. Adding gpu/x11 experts.
Comment 4 by piman@chromium.org, Feb 4 2014
Cc: jbau...@chromium.org
Could you give us the output of the about:gpu page?

This looks like missing damage at the aura/views level, but I wonder if it could be related to partial swaps and/or damage rects in the compositor.
about:gpu is below.

Graphics Feature Status
Canvas: Software only, hardware acceleration unavailable
3D CSS: Hardware accelerated
Compositing: Hardware accelerated on all pages and threaded
CSS Animation: Accelerated and threaded
Flash 3D: Hardware accelerated
Flash Stage3D: Unavailable. Hardware acceleration unavailable
Flash Stage3D Baseline profile: Unavailable. Hardware acceleration unavailable
WebGL multisampling: Hardware accelerated
Video: Hardware accelerated
Video Decode: Software only, hardware acceleration unavailable
Video Encode: Hardware accelerated
WebGL: Hardware accelerated
Problems Detected
Accelerated 2d canvas is unstable in Linux at the moment
Stage3D is not supported on Linux: 129848
Accelerated video decode is unavailable on Mac and Linux: 137247, 133828
Driver Bug Workarounds
clear_alpha_in_readpixels
disable_ext_occlusion_query
set_texture_filter_before_generating_mipmap
Version Information
Data exported	2/3/2014 8:53:22 PM
Chrome version	Chrome/34.0.1809.0
Operating system	Linux 3.11.0-15-generic
Software rendering list version	6.25
Driver bug list version	3.8
ANGLE revision	2453
2D graphics backend	Skia
Command Line Args	--use-system-title-bar --ppapi-flash-path=/usr/lib/pepflashplugin-installer/libpepflashplayer.so --ppapi-flash-version=12.0.0.41 --incognito --flag-switches-begin --disable-ntp-other-sessions-menu --enable-deferred-image-decoding --enable-download-resumption --enable-encrypted-media --enable-offline-cache-access --enable-panels --disable-password-generation --enable-scroll-prediction --enable-smooth-scrolling --save-page-as-mhtml --flag-switches-end
Driver Information
Initialization time	2531
Sandboxed	true
GPU0	VENDOR = 0x8086, DEVICE= 0x0116
Optimus	false
AMD switchable	false
Driver vendor	Mesa
Driver version	9.2.1
Driver date	
Pixel shader version	1.30
Vertex shader version	1.30
Machine model	
GL version	3.0
GL_VENDOR	Intel Open Source Technology Center
GL_RENDERER	Mesa DRI Intel(R) Sandybridge Mobile
GL_VERSION	3.0 Mesa 9.2.1
GL_EXTENSIONS	GL_ARB_multisample GL_EXT_abgr GL_EXT_bgra GL_EXT_blend_color GL_EXT_blend_minmax GL_EXT_blend_subtract GL_EXT_copy_texture GL_EXT_polygon_offset GL_EXT_subtexture GL_EXT_texture_object GL_EXT_vertex_array GL_EXT_compiled_vertex_array GL_EXT_texture GL_EXT_texture3D GL_IBM_rasterpos_clip GL_ARB_point_parameters GL_EXT_draw_range_elements GL_EXT_packed_pixels GL_EXT_point_parameters GL_EXT_rescale_normal GL_EXT_separate_specular_color GL_EXT_texture_edge_clamp GL_SGIS_generate_mipmap GL_SGIS_texture_border_clamp GL_SGIS_texture_edge_clamp GL_SGIS_texture_lod GL_ARB_framebuffer_sRGB GL_ARB_multitexture GL_EXT_framebuffer_sRGB GL_IBM_multimode_draw_arrays GL_IBM_texture_mirrored_repeat GL_3DFX_texture_compression_FXT1 GL_ARB_texture_cube_map GL_ARB_texture_env_add GL_ARB_transpose_matrix GL_EXT_blend_func_separate GL_EXT_fog_coord GL_EXT_multi_draw_arrays GL_EXT_secondary_color GL_EXT_texture_env_add GL_EXT_texture_filter_anisotropic GL_EXT_texture_lod_bias GL_INGR_blend_func_separate GL_NV_blend_square GL_NV_light_max_exponent GL_NV_texgen_reflection GL_NV_texture_env_combine4 GL_S3_s3tc GL_SUN_multi_draw_arrays GL_ARB_texture_border_clamp GL_ARB_texture_compression GL_EXT_framebuffer_object GL_EXT_texture_compression_s3tc GL_EXT_texture_env_combine GL_EXT_texture_env_dot3 GL_MESA_window_pos GL_NV_packed_depth_stencil GL_NV_texture_rectangle GL_ARB_depth_texture GL_ARB_occlusion_query GL_ARB_shadow GL_ARB_texture_env_combine GL_ARB_texture_env_crossbar GL_ARB_texture_env_dot3 GL_ARB_texture_mirrored_repeat GL_ARB_window_pos GL_ATI_envmap_bumpmap GL_EXT_stencil_two_side GL_EXT_texture_cube_map GL_NV_depth_clamp GL_APPLE_packed_pixels GL_APPLE_vertex_array_object GL_ARB_draw_buffers GL_ARB_fragment_program GL_ARB_fragment_shader GL_ARB_shader_objects GL_ARB_vertex_program GL_ARB_vertex_shader GL_ATI_draw_buffers GL_ATI_texture_env_combine3 GL_ATI_texture_float GL_EXT_shadow_funcs GL_EXT_stencil_wrap GL_MESA_pack_invert GL_MESA_ycbcr_texture GL_NV_primitive_restart GL_ARB_depth_clamp GL_ARB_fragment_program_shadow GL_ARB_half_float_pixel GL_ARB_occlusion_query2 GL_ARB_point_sprite GL_ARB_shading_language_100 GL_ARB_sync GL_ARB_texture_non_power_of_two GL_ARB_vertex_buffer_object GL_ATI_blend_equation_separate GL_EXT_blend_equation_separate GL_OES_read_format GL_ARB_color_buffer_float GL_ARB_pixel_buffer_object GL_ARB_texture_compression_rgtc GL_ARB_texture_float GL_ARB_texture_rectangle GL_EXT_packed_float GL_EXT_pixel_buffer_object GL_EXT_texture_compression_dxt1 GL_EXT_texture_compression_rgtc GL_EXT_texture_rectangle GL_EXT_texture_sRGB GL_EXT_texture_shared_exponent GL_ARB_framebuffer_object GL_EXT_framebuffer_blit GL_EXT_framebuffer_multisample GL_EXT_packed_depth_stencil GL_APPLE_object_purgeable GL_ARB_vertex_array_object GL_ATI_separate_stencil GL_EXT_draw_buffers2 GL_EXT_draw_instanced GL_EXT_gpu_program_parameters GL_EXT_texture_array GL_EXT_texture_integer GL_EXT_texture_sRGB_decode GL_EXT_timer_query GL_OES_EGL_image GL_MESA_texture_array GL_ARB_copy_buffer GL_ARB_depth_buffer_float GL_ARB_draw_instanced GL_ARB_half_float_vertex GL_ARB_instanced_arrays GL_ARB_map_buffer_range GL_ARB_texture_rg GL_ARB_texture_swizzle GL_ARB_vertex_array_bgra GL_EXT_separate_shader_objects GL_EXT_texture_swizzle GL_EXT_vertex_array_bgra GL_NV_conditional_render GL_AMD_draw_buffers_blend GL_ARB_ES2_compatibility GL_ARB_debug_output GL_ARB_draw_buffers_blend GL_ARB_draw_elements_base_vertex GL_ARB_explicit_attrib_location GL_ARB_fragment_coord_conventions GL_ARB_provoking_vertex GL_ARB_sampler_objects GL_ARB_seamless_cube_map GL_ARB_shader_texture_lod GL_ARB_texture_cube_map_array GL_ARB_texture_multisample GL_ARB_texture_query_lod GL_ARB_texture_rgb10_a2ui GL_ARB_uniform_buffer_object GL_ARB_vertex_type_2_10_10_10_rev GL_EXT_provoking_vertex GL_EXT_texture_snorm GL_MESA_texture_signed_rgba GL_ARB_get_program_binary GL_ARB_robustness GL_ARB_shader_bit_encoding GL_ARB_timer_query GL_ANGLE_texture_compression_dxt3 GL_ANGLE_texture_compression_dxt5 GL_ARB_internalformat_query GL_ARB_shading_language_420pack GL_ARB_shading_language_packing GL_ARB_texture_storage GL_EXT_framebuffer_multisample_blit_scaled GL_EXT_transform_feedback GL_ARB_ES3_compatibility GL_ARB_invalidate_subdata GL_ARB_texture_storage_multisample
Window system binding vendor	SGI
Window system binding version	1.4
Window system binding extensions	GLX_ARB_create_context GLX_ARB_create_context_profile GLX_ARB_multisample GLX_EXT_create_context_es2_profile GLX_EXT_import_context GLX_EXT_texture_from_pixmap GLX_EXT_visual_info GLX_EXT_visual_rating GLX_MESA_copy_sub_buffer GLX_OML_swap_method GLX_SGI_swap_control GLX_SGIS_multisample GLX_SGIX_fbconfig GLX_SGIX_pbuffer GLX_SGIX_visual_select_group GLX_INTEL_swap_event
Reset notification strategy	0x8261

Comment 6 by piman@chromium.org, Feb 4 2014
Ok, on this driver we would use partial swaps (since GLX_MESA_copy_sub_buffer is supported) whereas on NVIDIA we typically do not, that may be why erg@ can't reproduce.

@saiarcot895: can you try running chrome while passing the --ui-disable-partial-swap command-line parameter and see if that helps (you'd have to quit chrome entirely and restart with this parameter). If it fixes the problem, then that'd be a definite lead.
Yes, that fixed it. Elements are now resizing appropriately.

Also, you may have hit two birds with one stone, because it seems that with the option, bus 317516 doesn't occur.
Typo:  bug 317516 .
Comment 9 by piman@chromium.org, Feb 4 2014
Owner: piman@chromium.org
Status: Assigned
Summary: Shrinking the window doesn't redraw/re-fit all of the elements on Intel drivers, likely related to partial swaps (was: Shrinking the window doesn't redraw/re-fit all of the elements)
@#7: thanks!

I will try to repro to see if we're doing something iffy or if the driver is buggy.
Comment 10 by kbr@chromium.org, Feb 11 2014
Another data point: vivianz@ and I reproduced the corrupted rendering upon resizing the window smaller, on a Lenovo ThinkPad T420s running Ubuntu 12.04. We had to disable NVIDIA Optimus in the BIOS and force the machine onto the integrated GPU. Running with --ui-disable-partial-swap worked around the problem.

Attached is about:gpu output from this machine.

about-gpu.txt
5.6 KB View Download
Comment 11 by kbr@chromium.org, Feb 11 2014
Cc: vivianz@chromium.org kbr@chromium.org
Comment 12 by piman@chromium.org, Feb 11 2014
Cc: e...@chromium.org wiltzius@chromium.org rponnada@chromium.org varkha@chromium.org srsridhar@chromium.org
 Issue 317595  has been merged into this issue.
Comment 13 by e...@chromium.org, Feb 20 2014
Hey piman, has there been any progress on this?
It's odd that copy_sub_buffer would matter, because when resizing the window we should be redrawing the entire window and using swapbuffers instead of glxCopySubBufferMESA.
Cc: ccameron@chromium.org
Comment 16 by e...@chromium.org, Feb 24 2014
Labels: -Pri-2 Pri-1 ReleaseBlock-Beta M-35
 Issue 347303  has been merged into this issue.
The same issue is not repro on a amd graphic card linux box either, this is Intel GPU card ONLY.
Can reproduce this behavior on on Lenovo T410s (Intel Arrandale HD) using Chromium aura 34.0.1847.14 (253219) on Gentoo. --ui-disable-partial-swap workaround the problem for me, too. 

Attached the gpu:about. The implication of this bug are more drastic with Awesome WM, e.g. doesn't redraw the whole screen when changing tags.
about_gpu.txt
6.4 KB View Download
Cc: marc...@chromium.org
@#14: it's because of the implementation of the intel driver. Upon resize, the driver sizes the back buffer on SwapBuffer (through the GetBuffers DRI2 API), which doesn't happen when using glXCopySubBufferMESA (It wouldn't happen at the right time anyway - at the last swap rather than on Resize).

We used to have a workaround for this on Chrome OS (needed for screen resolution changes), which involved a "dummy" SwapBuffers on Resize, but that can cause a flash, and I think that'll be too visible here.

Short term, we should just blacklist partial swaps on Intel, and maybe look for a workaround (e.g. maybe use glCopyPixels or something before the dummy SwapBuffers on resize? Or maybe there's a way to hint the driver to force a GetBuffers?).
Comment 21 by kbr@chromium.org, Mar 4 2014
Blacklisting partial swaps on the Intel driver sounds like the best short-term solution. Should we do this only on non-Chrome OS Linux platforms?

We definitely don't want to blacklist it on Chrome OS :)

We could simply blacklist it always on Linux for now.
Comment 23 by kbr@chromium.org, Mar 4 2014
Cc: zmo@chromium.org
+zmo

OK, I didn't know whether a "linux" blacklist entry would apply to Chrome OS.

Mo, do you think you might be able to help with adding a driver bug workaround for disabling partial swap (--ui-disable-partial-swap)?

Comment 24 by zmo@chromium.org, Mar 11 2014
linux and chromeos are separated in blacklist, so yes, we definitely can blacklist one without affecting the other.
Comment 25 by e...@chromium.org, Mar 11 2014
 Issue 351433  has been merged into this issue.
Comment 26 by e...@chromium.org, Mar 20 2014
Ping. Has partial swaps been blacklisted yet? If so, could you please remove the beta blocker tag?
Project Member Comment 28 by bugdroid1@chromium.org, Mar 23 2014
------------------------------------------------------------------
r258740 | piman@chromium.org | 2014-03-22T02:43:43.124103Z

Changed paths:
   M http://src.chromium.org/viewvc/chrome/trunk/src/ui/gl/gl_surface_egl.cc?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/gpu/config/gpu_driver_bug_workaround_type.h?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/ui/gl/gl_surface_egl.h?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/ui/gl/gl_surface_x11.cc?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/content/common/gpu/image_transport_surface_mac.cc?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/content/common/gpu/texture_image_transport_surface.cc?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/ui/gl/gl_surface_glx.cc?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/content/common/gpu/texture_image_transport_surface.h?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/ui/gl/gl_surface_glx.h?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/gpu/command_buffer/service/gles2_cmd_decoder.cc?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/ui/gl/gl_surface_win.cc?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/gpu/command_buffer/service/gl_surface_mock.h?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/gpu/config/gpu_driver_bug_list_json.cc?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/ui/gl/gl_surface.cc?r1=258740&r2=258739&pathrev=258740
   M http://src.chromium.org/viewvc/chrome/trunk/src/ui/gl/gl_surface.h?r1=258740&r2=258739&pathrev=258740

Blacklist partial swaps on linux

BUG= 339493 

Review URL: https://codereview.chromium.org/207443007
-----------------------------------------------------------------
Comment 29 by piman@chromium.org, Mar 24 2014
Status: Fixed
Project Member Comment 30 by bugdroid1@chromium.org, Oct 6 2014
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/16e7fae640eb254acd54533d865b3a01645607ac

commit 16e7fae640eb254acd54533d865b3a01645607ac
Author: sl.ostapenko <sl.ostapenko@samsung.com>
Date: Mon Oct 06 18:20:59 2014

Apply disable_post_sub_buffers_for_onscreen_surfaces only to mesa gl.

Since this bug appears only on mesa for intel, limit workaround only to the mesa driver.

BUG= 339493 

Review URL: https://codereview.chromium.org/612483004

Cr-Commit-Position: refs/heads/master@{#298280}

[modify] https://chromium.googlesource.com/chromium/src.git/+/16e7fae640eb254acd54533d865b3a01645607ac/gpu/config/gpu_driver_bug_list_json.cc

Do you have a way to limit this workaround for Mesa on Intel, leaving Mesa on AMD alone?
Comment 32 by piman@chromium.org, Jan 13 2015
I suspect AMD is also affected on Mesa - this is really due to how DRI2 works, see #20.
I can't redproduce this bug after disabling the GPU blacklist and passing --disable-gpu-driver-bug-workarounds on the command line. (So this workaround is no longer listed in about:gpu.) Chromium appears to be much much smoother now, when comparing to the "blacklist enabled, workaround enabled" state.

What is the mesa bug report ID for this bug so I can double-check if there is another way to reproduce?

AMD Radeon HD8670,
GL_VENDOR	X.Org
GL_RENDERER	Gallium 0.4 on AMD BARTS
GL_VERSION	3.0 Mesa 10.3.2

Chrome version	Chrome/40.0.2214.111
Operating system	Linux 3.16.0-30-generic

(Installed chromium from Ubuntu 14.04 repositories…)
The bug pretty much still exists for me on latest version of Chrome with Intel Drivers:

Data exported	8/2/2015, 8:42:10 AM
Chrome version	Chrome/44.0.2403.125
Operating system	Linux 4.1.3-1-ARCH
Software rendering list version	0
Driver bug list version	8.19
ANGLE commit id	fa9744b09e24
2D graphics backend	Skia
Command Line Args	--ui-disable-partial-swap --flag-switches-begin --ignore-gpu-blacklist --flag-switches-end
Driver Information
Initialization time	52
Sandboxed	true
GPU0	VENDOR = 0x8086, DEVICE= 0x1616
Optimus	false
AMD switchable	false
Driver vendor	Mesa
Driver version	10.6.3
Driver date	
Pixel shader version	1.30
Vertex shader version	1.30
Max. MSAA samples	8
Machine model name	
Machine model version	
GL_VENDOR	Intel Open Source Technology Center
GL_RENDERER	Mesa DRI Intel(R) HD Graphics 5500 (Broadwell GT2)
GL_VERSION	3.0 Mesa 10.6.3
Comment 35 by kbr@chromium.org, Aug 4 2015
Blocking: chromium:516679
Comment 36 by e...@chromium.org, Aug 4 2015
Cc: -e...@chromium.org
Comment 37 Deleted
>> Do you have a way to limit this workaround for Mesa on Intel, leaving Mesa on
>> AMD alone?

> I suspect AMD is also affected on Mesa - this is really due to how DRI2 works,
> see #20.

I have not looked into detail but since I saw "glX"CopySubBufferMESA in the discussion, could we add "gl_type": "gl", in the entry:
     {
      "id": 68,
      "description": "Disable partial swaps on linux drivers",
      "cr_bugs": [339493],
      "os": {
        "type": "linux"
      },
      "driver_vendor": "Mesa",
      "features": [
        "disable_post_sub_buffers_for_onscreen_surfaces"
      ]
    },
in order to enable post sub suffers with --use-gl=egl ?

Indeed I compared implementation of eglPostSubBufferNV vs glXCopySubBufferMESA
in mesa and the former is a wrapper around "xcb_dri2_copy" whereas the later is a wrapper around pipe->resource_copy_region which uses libdrm directly.

An easy way to reproduce the problem is to maximize and then un-maximize the window, and the address glitch appears and remains visible.
I can reproduce with nouveau driver with use-gl=desktop (i.e. glx), but it looks good with --use-gl=egl

So it looks to me that we can add "gl_type": "gl" to that entry 68.
What do you think ?





Comment 39 by piman@chromium.org, Jun 10 2016
If it works with egl, I don't necessarily object to expose tentatively, but make sure to test well (I won't be able to, I don't have a relevant configuration),  with heavy resize, and also with things that draw only once (e.g. context menus, tooltips), if possible under a variety of configs (with/without a compositing manager).

Separately, I have thoughts about how to force the GetBuffers, but I haven't had time to explore yet.
I created  http://crbug.com/619514  for that.
Project Member Comment 41 by bugdroid1@chromium.org, Jul 19 2016
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/6bcbbd2886e61691f69ec14c27e62b69f81a271b

commit 6bcbbd2886e61691f69ec14c27e62b69f81a271b
Author: cwallez <cwallez@chromium.org>
Date: Tue Jul 19 02:45:20 2016

gpu driver bug list: correctly backlist partial swap on Mesa

This entry previously used "gl_vendor": "Mesa" when the GL_VENDOR string
is "Intel Open Source Technology Center". Try to match "Mesa" in the
GL_RENDERER string instead.

BUG= 339493 
BUG= 628168 
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2156993004
Cr-Commit-Position: refs/heads/master@{#406194}

[modify] https://crrev.com/6bcbbd2886e61691f69ec14c27e62b69f81a271b/gpu/config/gpu_driver_bug_list_json.cc

Sign in to add a comment