Same problem

Deleted: net-internals-log.json 2.9 MB

net-internals-log.json 2.9 MB View Download

The original report and the report from #12 are both from Chrome 56, where the CopyTextureCHROMIUM upgrade hasn't started yet.

Re #15, these errors make me think that something is failing to allocate / produce a texture, and we're then trying to draw with this texture.

Can someone with a good feel for how the PDF viewer plugin works provide a bit more info? Does the plugin produce textures and hand them off to the compositor? Or does it (somehow) produce some other intermediate which we then rasterize?

thestig@ is in a better place to answer then me, but as far as I understand PDFium is producing bitmaps which get set into a ImageData class. We then take that and call pp::Graphics2D::PaintImageData.  What pepper does with it after that point, I don't know.

ericrk: Painting has never been my forte, so I don't know either. The PPAPI goes out to PepperGraphics2DHost::OnHostMsgPaintImageData(). Try tracing from there and see where it takes you?

I'm not seeing anything too obvious from looking at the the PDF drawing code. It doesn't appear to hit glCopySubTextureCHROMIUM. That code is instead hit by the chrome-produced UI around the PDF plugin (page counter, zoom toolbar).

Additionally, the about:gpu that produced these errors had GPU rasterization disabled, so it's not likely related to GPU raster.

I think it's going to be hard to make headway without a local repro or more information.

Can someone who is experiencing this issue answer the following:
- Does this reproduce reliably? Or is it just every so often?
- If it's reliable, does restarting Chrome help?

Also, if you are a user experiencing this issue, can you please also include your "about:gpu" page with your report? Thanks!

HI - The problem is not intermittent and happens each time you start chrome - how does one get the about gpu page?

The "about gpu page" is chrome://gpu or alternatively about:gpu

here is the gpu page is this what you require 

Deleted: chrome___gpu.pdf 1.3 MB

chrome___gpu.pdf 1.3 MB Download

continual - never doesn't come up with blank screen

Deleted: GPU.pdf 561 KB

GPU.pdf 561 KB Download

#26: it seems you have a dual GPU (Intel + NVidia) on your machine.  Can you switch your machine to use the NVidia GPU for Chrome (In NVidia Control Panel) and see if the problem is still there?  That way we will know if it's a driver specific issue or not.

#27 is also an Intel.

Note that we've seen one other instance of this (comment #15 on  bug 661596 ), which appeared to have similar corruption (dark grey screen), but wasn't restricted to PDFs. That user was also a single-GPU NVidia system.

re #15, note that the about:gpu pages seen don't all include the "GL ERROR :GL_INVALID_VALUE" errors. Additionally, some that do show this error have it coming from a different context (Offscreen-For-WebGL), so these errors are likely unrelated or a symptom of a larger issue that has already occurred.

Also, note that the background color seen on these blank pages is a dark-grey, the same color that's displayed on any PDF load while we wait for the actual content to fill in. I wonder if the PDF plugin could be getting hung up somewhere and is just failing to deliver content?

#28 - it works fine in Incognito does this rule out a driver issue?

I can't reproduce on a Win8 machine on Intel, but on that machine Chrome stable is already 57.

Does this issue reproducible also in Chrome 57?

#31: it can still be a driver issue, just Incognito mode doesn't trigger it.

I found this defect as I was having this problem.
In one of my many tabs, I had a pdf that was opened prior to applying a chrome update, and was now in a bad state.
Totally unresponsive.
I closed that tab, and now I am able to open and view pdfs.

Currently running Version 56.0.2924.87
Windows 7.

#28 -how do i switch the machine to use the NVidia GPU for Chrome (In NVidia Control Panel)? 

 Issue 689841  has been merged into this issue.

I got this on my work Linux workstation. If I go to the task manager, kill the Chrome PDF viewer process, and try to reload a PDF, no process appears. So it's a problem of the browser not calling the plugin at all, not a problem of the plugin hanging.

Let's wait and see if this is still an issue in M57...

Hi, I am having the same issue, with both pdf's on my desktop or pdf's i download . They come up but are completely black. If I open them up in acrobat, I can see them. Chrome Version 57.0.2987.98 (64-bit)OS Yosemite 10.10.5

npm: I can't reproduce the behavior in comment 36. Did something happen to cause Chrome to get wedged like that?

simondbailey@gmail.com: try the instructions from this link: https://support.lenovo.com/us/en/solutions/ht078706

thestig: no, it just randomly happened when trying to open some PDF (but none worked after). And it only caused problems when opening a PDF on my Chromium profile. But it got fixed after restarting.

I am having the same proble. Black Pdf

Deleted: net-internals-log.json 2.5 MB

net-internals-log.json 2.5 MB View Download

#39 - done that and applied the changes - no change PDFs still black

Ï had the same problem when trying to open any pdf (local or on web), and came to this page looking for a solution. It turns out that, like the user from comment 33, I also had a tab with a pdf open from a long time ago (which displayed the pdf fine), and when I closed that tab and restarted the browser, pdfs started working again. 
I am on Windows 10, and Chrome 56.0.2924.87

I wonder if we can dig further into the PDF plugin not loading issue reported by npm@ in #36. It sounds like users experiencing PDF problems can check if they're hitting the same PDF plugin loading issue npm@ saw by:

1. Check the chrome task manager for the PDF plugin (click the "..." to the right of the address bar, select "More tools", select Task Manager.
2. Check if the PDF plugin is running (should show up as "Plugin: Chrome PDF Viewer").
3. Kill the PDF plugin by selecting it in the list and clicking "End Process"
4. Refresh a tab with a PDF and see if the plugin re-starts successfully (shows up in the task manager again).
5. If not, this is likely the same issue that npm@ saw in #36.

Updating to Chrome 57 on Mac Pro worked for me.

 Issue 702970  has been merged into this issue.

chrome://gpu info for me. All PDF urls load blank grey screen.

Deleted: PDF_BLANK_gpu.htm 4.4 KB

PDF_BLANK_gpu.htm 4.4 KB View Download

Update:  PDFs load correctly with upgrade to Version 57.0.2987.110 (64-bit)

Do we know if this is fixed now that M57 is released? Are we still seeing reports of this happening?

#49 - still an issue at Version 57.0.2987.110

After upgrading to 57.0.2987.133, this issue was resolved for me.

Windows x64 10 Enterprise, Version 1511
Intel HD Graphics, 10/1/2015, 20.19.15.4300

A restart temporarily allowed PDFs to load but the problem returned. Had the issue on the previous version of Chrome and have it on the current version.

Chrome/57.0.2987.133
Operating system	Windows NT 10.0.14393
Software rendering list version	12.13
Driver bug list version	9.29
ANGLE commit id	c1a5d16e964a
2D graphics backend	Skia/57 ae9cc5d3588d52f4b371b55845704b25d88cf06d

ANGLE (AMD Radeon HD 8650G + HD 8750M Dual Graphics Direct3D11 vs_5_0 ps_5_0)

Log Messages
[6164:19408:0404/040439.241:ERROR:mf_helpers.cc(12)] : Error in dxva_video_decode_accelerator_win.cc on line 1528
[6164:19408:0404/040439.241:ERROR:mf_helpers.cc(12)] : Error in dxva_video_decode_accelerator_win.cc on line 610
[6164:19408:0404/040439.276:ERROR:gpu_video_decode_accelerator.cc(375)] : HW video decode not available for profile vp9 profile0
...
* a page fulll of the preceding 3 lines *
...
GpuProcessHostUIShim: The GPU process exited normally. Everything is okay.
[6164:19408:0405/143736.811:ERROR:interface_registry.cc(203)] : Failed to locate a binder for interface: chrome::mojom::ResourceUsageReporter requested by: exposed by: via InterfaceProviderSpec "service_manager:connector".
[6164:19408:0405/143736.812:WARNING:interface_registry.cc(210)] : InterfaceRegistry(service_manager:connector): --> InterfaceRegistry is not yet bound to a pipe.

I killed the process "Plugin: Chrome PDF Viewer" in the task manager and was able to view PDFs again without completely reloading Chrome. The keyboard shortcut for the task manager was not working and I had to go to the vertical ellipsis menu->More tools->Task manager to get to it.

I'm experiencing this black screen problem. I can preview a PDF in the Box.com UI, but when I download and try to open/view it, all I get is a black screen. I experienced the same when trying to open a PDF on myuhc.com. 

OS and Device details: Asus Chromebook

Deleted: net-internals-log.json 14.4 MB

net-internals-log.json 14.4 MB View Download

Deleted: Detailsfromchromeversionpage.pdf 33.6 KB

Detailsfromchromeversionpage.pdf 33.6 KB Download

Same problem with Chrome on Macos. Details below:

--- chrome://version ---
Google Chrome	57.0.2987.98 (Official Build) (64-bit)
Revision	f87f641e0af5bfed98578e340f1d5ca79651bf82-refs/branch-heads/2987@{#802}
OS	Mac OS X 
JavaScript	V8 5.7.492.63
Flash	25.0.0.127 /Users/nsonawane/Library/Application Support/Google/Chrome/PepperFlash/25.0.0.127/PepperFlashPlayer.plugin
User Agent	Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_3) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/57.0.2987.98 Safari/537.36
Command Line	/Applications/Google Chrome.app/Contents/MacOS/Google Chrome --enable-avfoundation --flag-switches-begin --flag-switches-end
Executable Path	/Applications/Google Chrome.app/Contents/MacOS/Google Chrome
Profile Path	/Users/nsonawane/Library/Application Support/Google/Chrome/Default
Variations	3095aa95-3f4a17df
8364a5c2-ca7d8d80
7c1bc906-f55a7974
2a33b90e-e7b70917
ba3f87da-b4a760c3
cf558fa6-ca7d8d80
58aac55e-3f4a17df
f3499283-f94608c8
31362330-3f4a17df
c70841c8-a2567007
9e201a2b-803f8fc4
6eb432aa-ca7d8d80
5274eb09-3f4a17df
684d1cdf-51126808
9773d3bd-ca7d8d80
2e109477-ca7d8d80
9e5c75f1-c16ec2e6
6b121ae7-ca7d8d80
f79cb77b-3d47f4f4
b7786474-d93a0620
23a898eb-ca7d8d80
4ea303a6-ecbb250e
9736de91-ca7d8d80
64005e71-4ad60575
7e9929f5-e9431caf
69bf80fa-91c810ef
867c4c68-3f4a17df
b2f0086-93053e47
7fc902e8-ca7d8d80
f11cb941-d6606b56
f47ae82a-86f22ee5
3ac60855-486e2a9c
f296190c-fae5d535
4442aae2-e1cc0f14
ed1d377-e1cc0f14
75f0f0a0-a5822863
e2b18481-d7f6b13c
e7e71889-e1cc0f14
828a5926-ca7d8d80
------------------------

chrome://net-internals attached

Hope this helps.

Deleted: net-internals-log.json 863 KB

net-internals-log.json 863 KB View Download

Re #49 - dsinclair@, sounds like this is still happening. Can you take a look?

I am having the same issue for a couple of months. I am on OSX 10.12.4, on a MacBook Pro (Retina, 13-inch, Late 2013)
Chrome://version attached

Deleted: chrome-About Version.html 5.1 KB

chrome-About Version.html 5.1 KB View Download

I'm not sure what I can do at this point, we don't have a reliable way to repro this issue. Without that, I'm at a loss as to how to track the issue down ...

Hi, I've been experiencing this issue for the past two weeks or so. There are multiple PCs at my company with this issue, and all of them started around the same time. 

Chrome version: Version 57.0.2987.133 (64-bit)
OS: Windows 10 Version 1703 (OS Build 15063.138)

PC Specs: Dell Optiplex 7040, Intel Core i7-6700T @ 2.80GHz, 16GB RAM

I've got personal computers running the same versions of Windows and Chrome where the PDFs are behaving normally. It's only the Dell hardware that's an issue. 

This is 100% reproducible for me.  I'm on a Thinkpad T430s running Windows 7.  I see this with 59.0.3067.6 dev and 60.0.3074.0 canary.  It survives across reboots.

FWIW, I see the tab as white instead of black.

Let me know if there's any debugging or tracing I can give you.

dbloch@ do you have hardware acceleration enabled? If so, can you disable and try to repro?

If it still repros, can you try opening the chrome task manager (under the ... > More Tools > Task Manager menu) and killing the 'Plugin: Chrome PDF Viewer' and reloading the PDF? (I'm curious to see if the pdf plugin gets reloaded or if we're getting stuck somewhere in the browser and the pdf plugin isn't seeing anything).

One other thing to try is, when you have the blank screen, can you open the print preview, does the preview content appear or is it also blank?

Is this one PDF or all PDFs you're seeing it with?

#63 Hi

1 - there is no plug in to kill (see attached)

2- print preview is also blank 

3  - it is all PDFs

Deleted: Capture.JPG 134 KB

	Capture.JPG 134 KB View Download

thestig@ I don't know how the plugin gets started and the PPAPI stuff around there. Do you have any ideas for things we could test to see why the pdf renderer wouldn't even be running?

Same answers as Simon.  He types faster than I do.

Re hardware acceleration, I'm not sure.  How do I check this?

Are there any crash reports in chrome://crashes ? One theory might be that the PDF plugin crashes and can't recover somehow.

The other theory is the plugin failed to start. Not sure why that would happen.

I found the hardware acceleration setting.  It was on, and when I turned it off it did fix the PDF viewer.  W00t!

#68 how did you switch it off - I have looked and it is greyed out  ta

#69  It just worked.  Maybe whether the control is live or not depends on the hardware.

dbloch@ just to verify, if you re-enable hardware acceleration does the PDF issue re-appear?

Wow, good question.  Today, when I re-enabled hardware acceleration, the issue doesn't reappear.

But note, today the issue is also gone on my dev channel issue, which I never touched.  Yesterday after I turned off hardware acceleration, it was working on my canary instance but not on my dev instance.  I wish I'd done this check then.  Anyway, I can no longer reproduce the problem.  Sorry.

simonbailey@ is the issue still happening for you? Does restarting Chrome cause it to go away or does it happen right after restart?

Did you happen to have a PDF open for a really long time before it started the first time?

This issue is still present for me regardless of hardware acceleration setting, restarting, safe mode, etc. 

Ten days ago I opened a topic on the Google help forums; they directed me here. 

https://productforums.google.com/forum/?utm_medium=email&utm_source=footer#!msg/chrome/gpPiiUQC7lo/jtaXqnzkAQAJ

josephluhn@ I'm not sure if it's the same issue or not, but it sounds similar. When you open the PDFs initially, do you leave them open (so, have PDFs open for long periods of time) and then PDFs start to fail to load?

It's back.  I
  - disabled hardware acceleration on my dev channel instance and relaunched.
  - confirmed that PDFs loaded.
  - re-enabled hardware acceleration and relaunched.
  - confirmed that PDFs  no longer load.

Let me know if there's anything else I can do while it's reproducible.

I never have PDFs open for long.

@Dsinclair, No. PDFs will fail to load even if it's the first page visited after starting Chrome. 

I click on the link, see a progress bar make it about 25-50% through, then the error message pops up. Doesn't matter which websites. 

Sorry if this isn't the right place for this. 

kbr@ per #75, it seems like Hardware Acceleration maybe playing a role here. Would someone from the GPU team be able to suggest what information we should grab or how to proceed?

Dbloch@, can you try this once or twice more just to make sure we've identified the right thing? (Sorry for the hassle). If so, can you paste the contents of about:gpu after an unsuccessful load (especially any error messages at the end)? Thanks!

It isn't completely reproducible, though hardware acceleration is strongly correlated.

disable again -> fixed
enable -> broken ("about:gpu" results from here)
disable -> *still broken*
enable -> ""
disable -> fixed

"about:gpu" output attached.  No error messages though :-(

I also saved it as a Google Doc, http://docs/document/d/1DxLI_njKGkdxr36-waOi_OkAtEvgxYW41PI77x7oDuI/edit, in case that's more useful.

Deleted: about_gpu.txt 9.3 KB

about_gpu.txt 9.3 KB View Download

dsinclair@ about:gpu information from any affected machine is important. But more important is some set of repro steps.

From a clean user profile on Chrome Canary I repeatedly tried turning on and off hardware acceleration via about:settings, restarting the browser, and navigating from https://www.khronos.org/registry/EGL/ back and forth to one of the linked specs. No failure to load any of the PDFs.

Here is about:gpu from my machine (Retina MacBook Pro, dual Intel and NVIDIA GPUs.)

Graphics Feature Status
Canvas: Hardware accelerated
Flash: Hardware accelerated
Flash Stage3D: Hardware accelerated
Flash Stage3D Baseline profile: Hardware accelerated
Compositing: Hardware accelerated
Multiple Raster Threads: Enabled
Native GpuMemoryBuffers: Hardware accelerated
Rasterization: Hardware accelerated
Video Decode: Hardware accelerated
Video Encode: Hardware accelerated
VPx Video Decode: Hardware accelerated
WebGL: Hardware accelerated
WebGL2: Hardware accelerated
Driver Bug Workarounds
add_and_true_to_loop_condition
adjust_src_dst_region_for_blitframebuffer
decode_encode_srgb_for_generatemipmap
disable_framebuffer_cmaa
disable_multimonitor_multisampling
disable_webgl_rgb_multisampling_usage
emulate_abs_int_function
get_frag_data_info_bug
init_two_cube_map_levels_before_copyteximage
msaa_is_slow
pack_parameters_workaround_with_pack_buffer
rebind_transform_feedback_before_resume
regenerate_struct_names
remove_invariant_and_centroid_for_essl3
rewrite_texelfetchoffset_to_texelfetch
scalarize_vec_and_mat_constructor_args
set_zero_level_before_generating_mipmap
unfold_short_circuit_as_ternary_operation
unpack_alignment_workaround_with_unpack_buffer
unpack_image_height_workaround_with_unpack_buffer
use_intermediary_for_copy_texture_image
use_unused_standard_shared_blocks
Problems Detected
Multisampling is buggy on OSX when multiple monitors are connected: 237931
Applied Workarounds: disable_multimonitor_multisampling
Unfold short circuit on Mac OS X: 307751
Applied Workarounds: unfold_short_circuit_as_ternary_operation
Always rewrite vec/mat constructors to be consistent: 398694
Applied Workarounds: scalarize_vec_and_mat_constructor_args
Mac drivers handle struct scopes incorrectly: 403957
Applied Workarounds: regenerate_struct_names
On Intel GPUs MSAA performance is not acceptable for GPU rasterization: 527565
Applied Workarounds: msaa_is_slow
glGenerateMipmap fails if the zero texture level is not set on some Mac drivers: 560499
Applied Workarounds: set_zero_level_before_generating_mipmap
Pack parameters work incorrectly with pack buffer bound: 563714
Applied Workarounds: pack_parameters_workaround_with_pack_buffer
Alignment works incorrectly with unpack buffer bound: 563714
Applied Workarounds: unpack_alignment_workaround_with_unpack_buffer
copyTexImage2D fails when reading from IOSurface on multiple GPU types.: 581777
Applied Workarounds: use_intermediary_for_copy_texture_image
Multisample renderbuffers with format GL_RGB8 have performance issues on Intel GPUs.: 607130
Applied Workarounds: disable_webgl_rgb_multisampling_usage
Limited enabling of Chromium GL_INTEL_framebuffer_CMAA: 535198
Applied Workarounds: disable_framebuffer_cmaa
glGetFragData{Location|Index} works incorrectly on Max: 638340
Applied Workarounds: get_frag_data_info_bug
glResumeTransformFeedback works incorrectly on Intel GPUs: 638514
Applied Workarounds: rebind_transform_feedback_before_resume
Result of abs(i) where i is an integer in vertex shader is wrong: 642227
Applied Workarounds: emulate_abs_int_function
Rewrite texelFetchOffset to texelFetch for Intel Mac: 642605
Applied Workarounds: rewrite_texelfetchoffset_to_texelfetch
Rewrite condition in for and while loops for Intel Mac: 644669
Applied Workarounds: add_and_true_to_loop_condition
Decode and encode before generateMipmap for srgb format textures on macosx: 634519
Applied Workarounds: decode_encode_srgb_for_generatemipmap
Init first two levels before CopyTexImage2D for cube map texture on Intel Mac 10.12: 648197
Applied Workarounds: init_two_cube_map_levels_before_copyteximage
Insert statements to reference all members in unused std140/shared blocks on Mac: 618464
Applied Workarounds: use_unused_standard_shared_blocks
Tex(Sub)Image3D performs incorrectly when uploading from unpack buffer with GL_UNPACK_IMAGE_HEIGHT greater than zero on Intel Macs: 654258
Applied Workarounds: unpack_image_height_workaround_with_unpack_buffer
adjust src/dst region if blitting pixels outside read framebuffer on Mac: 644740
Applied Workarounds: adjust_src_dst_region_for_blitframebuffer
Mac driver GL 4.1 requires invariant and centroid to match between shaders: 639760, 641129
Applied Workarounds: remove_invariant_and_centroid_for_essl3
Disable KHR_blend_equation_advanced until cc shaders are updated: 661715
Version Information
Data exported	4/20/2017, 10:44:08 AM
Chrome version	Chrome/60.0.3076.0
Operating system	Mac OS X 10.12.4
Software rendering list version	13.3
Driver bug list version	10.3
ANGLE commit id	b8126691cf41
2D graphics backend	Skia/60 da16434928b4c721b462d780d5140f9fc68d1413-
Command Line Args	Chrome Canary.app/Contents/MacOS/Google Chrome Canary --user-data-dir=/tmp/c2 --flag-switches-begin --flag-switches-end
Driver Information
Initialization time	51
In-process GPU	false
Passthrough Command Decoder	false
Supports overlays	false
Sandboxed	true
GPU0	VENDOR = 0x10de, DEVICE= 0x0fe9
GPU1	VENDOR = 0x8086, DEVICE= 0x0d26 *ACTIVE*
Optimus	true
Optimus	true
AMD switchable	false
Driver vendor	
Driver version	10.24.45
Driver date	
Pixel shader version	4.10
Vertex shader version	4.10
Max. MSAA samples	8
Machine model name	MacBookPro
Machine model version	11.3
GL_VENDOR	Intel Inc.
GL_RENDERER	Intel Iris Pro OpenGL Engine
GL_VERSION	4.1 INTEL-10.24.45
GL_EXTENSIONS	GL_ARB_blend_func_extended GL_ARB_draw_buffers_blend GL_ARB_draw_indirect GL_ARB_ES2_compatibility GL_ARB_explicit_attrib_location GL_ARB_gpu_shader_fp64 GL_ARB_gpu_shader5 GL_ARB_instanced_arrays GL_ARB_internalformat_query GL_ARB_occlusion_query2 GL_ARB_sample_shading GL_ARB_sampler_objects GL_ARB_separate_shader_objects GL_ARB_shader_bit_encoding GL_ARB_shader_subroutine GL_ARB_shading_language_include GL_ARB_tessellation_shader GL_ARB_texture_buffer_object_rgb32 GL_ARB_texture_cube_map_array GL_ARB_texture_gather GL_ARB_texture_query_lod GL_ARB_texture_rgb10_a2ui GL_ARB_texture_storage GL_ARB_texture_swizzle GL_ARB_timer_query GL_ARB_transform_feedback2 GL_ARB_transform_feedback3 GL_ARB_vertex_attrib_64bit GL_ARB_vertex_type_2_10_10_10_rev GL_ARB_viewport_array GL_EXT_debug_label GL_EXT_debug_marker GL_EXT_framebuffer_multisample_blit_scaled GL_EXT_texture_compression_s3tc GL_EXT_texture_filter_anisotropic GL_EXT_texture_sRGB_decode GL_APPLE_client_storage GL_APPLE_container_object_shareable GL_APPLE_flush_render GL_APPLE_object_purgeable GL_APPLE_rgb_422 GL_APPLE_row_bytes GL_APPLE_texture_range GL_ATI_texture_mirror_once GL_NV_texture_barrier
Disabled Extensions	GL_KHR_blend_equation_advanced GL_KHR_blend_equation_advanced_coherent
Window system binding vendor	
Window system binding version	
Window system binding extensions	
Direct rendering	Yes
Reset notification strategy	0x0000
GPU process crash count	0
Compositor Information
Tile Update Mode	Zero-copy
Partial Raster	Enabled
GpuMemoryBuffers Status
ATC	Software only
ATCIA	Software only
DXT1	Software only
DXT5	Software only
ETC1	Software only
R_8	GPU_READ_CPU_READ_WRITE, GPU_READ_CPU_READ_WRITE_PERSISTENT
RG_88	Software only
BGR_565	Software only
RGBA_4444	Software only
RGBX_8888	Software only
RGBA_8888	GPU_READ, SCANOUT, SCANOUT_CPU_READ_WRITE
BGRX_8888	GPU_READ, SCANOUT, SCANOUT_CPU_READ_WRITE
BGRA_8888	GPU_READ, SCANOUT, SCANOUT_CPU_READ_WRITE, GPU_READ_CPU_READ_WRITE, GPU_READ_CPU_READ_WRITE_PERSISTENT
RGBA_F16	GPU_READ_CPU_READ_WRITE, GPU_READ_CPU_READ_WRITE_PERSISTENT
YVU_420	Software only
YUV_420_BIPLANAR	GPU_READ_CPU_READ_WRITE, GPU_READ_CPU_READ_WRITE_PERSISTENT
UYVY_422	GPU_READ_CPU_READ_WRITE, GPU_READ_CPU_READ_WRITE_PERSISTENT

Adding Internals>Plugins>Pepper since this may be specific to PPAPI plugins' rendering path.

kbr@ we have not found any consistent repro, it happens on some machines, and it happens randomly. Would it possible for someone in MTV to take a look at dbloch@'s machine to see if anything can be found (dblock@ assuming that works for you)?

Okay by me.  It's a Thinkpad T430s, which is about four years old.  Maybe it affects older machines more?

I looked at dbloch@'s machine and the symptom is that the browser doesn't look like it's attempting to load the PDF at all. From e.g. https://www.khronos.org/registry/EGL/ , clicking any of the links simply doesn't cause the navigation to occur at all.

Opening any of the PDF links in a new tab results in the URL of the PDF in the URL bar, but a blank tab, and I'm quite sure the navigation didn't occur.

The bug is caused by some Finch experiment. I created two new user-data-dir directories and in one of them toggling HW acceleration off and then back on didn't provoke the bug, but in the second one it did. I'm attaching three variations. One is from dbloch's current user-data-dir, where PDF loads fail. The second is from a new user-data-dir where they were succeeding. The third is from a second new user-data-dir where after toggling hardware accelerating on and off they fail.

Deleted: variations-dbloch-failing.txt 1.1 KB

variations-dbloch-failing.txt 1.1 KB View Download

Deleted: variations-dbloch-new-data-dir-not-failing.txt 1.1 KB

variations-dbloch-new-data-dir-not-failing.txt 1.1 KB View Download

Deleted: variations-dbloch-new-data-dir-2-failing.txt 1.1 KB

variations-dbloch-new-data-dir-2-failing.txt 1.1 KB View Download

I symbolized the variations and looked for those that were the same between the two failing user profiles, and different in the working user profile. Here are the results:

Failing:
LoadingWithMojo-Control
NTPTilesInInstantService-Control
PersistentHistograms-EnabledOnDiskNoStability3
SafeBrowsingV4LocalDatabaseManagerEnabled-V4Only2
SocketReadIfReady-Control
WebFontsInterventionV2-Default
YieldBetweenContentScriptRuns-Enabled

Working:
LoadingWithMojo-Enabled
NTPTilesInInstantService-Enabled
PersistentHistograms-NotEnabled3
SafeBrowsingV4LocalDatabaseManagerEnabled-Control2
SocketReadIfReady-Enabled
WebFontsInterventionV2-Enabled-2g-1
YieldBetweenContentScriptRuns-Control

So, if I'm reading that right, the 3 experiments on the failing case that are enabled are:

 * PersistentHistograms-EnabledOnDiskNoStability3
 * SafeBrowsingV4LocalDatabaseManagerEnabled-V4Only2
 * YieldBetweenContentScriptRuns-Enabled

The other ones, I'm guessing Control means it's the Control group and are using the code from previously? Adding SafeBrowsing label, I'm not sure what the other two would fall under.

+ksakamoto for YieldBetweenContentScriptRuns ( bug 636655 ) -- that sounds like the most likely cause to me.

Note: I tried launching the browser with --disable-features=YieldBetweenContentScriptRuns and unless I misspelled the flag the bug still reproduced. Basically, toggling several of those settings with either --enable-features or --disable-features didn't work.

When the bug stops happening, toggling the "Use hardware acceleration when available" flag and restarting the browser a couple of times makes it happen again -- but it also happens when that flag is turned off. Removing Internals>GPU.

japhet@ indicates there are some bugs filed against Canary where PlzNavigate causes PDF links to not load.

For the record: here is the browser version from dbloch's machine that repro'd these issues.

Google Chrome	59.0.3071.9 (Official Build) dev (64-bit) (cohort: Dev)
Revision	2ae1b9cb3ac16142bb504b4529323d14d0eb6f93-refs/branch-heads/3071@{#28}
OS	Windows
JavaScript	V8 5.9.211.4
Flash	24.0.0.189 internal-not-yet-present
User Agent	Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/59.0.3071.9 Safari/537.36

japhet@ suspects this is the same as  Issue 709771 .

I don't think PlzNavigate is the cause of all these issues as they've been reported on M56 and M57 as well. Which means there are probably a few different issues in here that have a similar outcome.

YieldBetweenContentScriptRuns use old-type flags, so it's enabled/disabled by
 --force-fieldtrials='YieldBetweenContentScriptRuns/Enabled/' or
 --force-fieldtrials='YieldBetweenContentScriptRuns/Control/'.

That said, this is Canary/Dev only experiment so I don't think this is the cause of this issue.

FYI SafeBrowsingV4LocalDatabaseManagerEnabled has been enabled 50/50 on dev/canary/beta and 1% on stable for some months now.

It's definitely possible that I misunderstood or mis-diagnosed which experiments could have had an effect. A second pair of eyes would be appreciated. At the same time, given that it reproduces with some user profiles and not with others, I'm 99% sure it's caused by a Finch experiment. Can do more work with dbloch@ if others have suggestions on what else to try.

dbloch@, kbr@ if this is still an issue on dbloch@'s machine, can we try out the flags mentioned in  crbug.com/709771  to see if this root cause is the same?

From comment #40 on that issue:
A little more poking and I've found that this reproduces the problem:

--enable-features=browser-side-navigation,SafeBrowsingV4LocalDatabaseManagerEnabled,SafeBrowsingV4OnlyEnabled

And this does not repro:

--enable-features=browser-side-navigation --disable-features=SafeBrowsingV4LocalDatabaseManagerEnabled,SafeBrowsingV4OnlyEnabled

dbloch@ and I confirmed that it's the same root cause on his machine. The only confusion was that a fresh --user-data-dir had to be chosen to avoid interacting with the already-running Chrome instance.

Awesome, thanks for the confirmation. I'm going to mark this as a dupe of 709771. Once that's fixed, if folks are still seeing issues I think we should open new bugs as they're separate problems.

The way  issue 709771  manifests is that there is no navigation being performed at all. It seems the original report is that the PDF won't render, which means we navigated to it, but there was an issue displaying the contents. I don't think the two are exactly the same.

I think there are several things entwined in this bug, some fixed, some 709771 and possibly others, it's hard to tell. There are reports of users having this fixed as of M57 that were seeing it in M56. There are users who see it fixed with the flags flipped that also trip  issue 709771 . I'm not sure if there are another class of user in there or not.

I think the best route forward to is dupe this in favour of  issue 709771 . If, after that issue is resolved, there are users seeing problems similar to this bug we can open a new issue to get a fresh start and untangle the fixed from unfixed in here.

Seeing same issue still, PLEASE FIX:


Google Inc.
Copyright 2017 Google Inc. All rights reserved.
Google Chrome	57.0.2987.133 (Official Build) (64-bit)
Revision	ec33cd0c06881d919ac0de419d829ad914e0be8f-refs/branch-heads/2987@{#887}
OS	Mac OS X 
JavaScript	V8 5.7.492.71
Flash	25.0.0.148 /Users/cballenger/Library/Application Support/Google/Chrome/PepperFlash/25.0.0.148/PepperFlashPlayer.plugin
User Agent	Mozilla/5.0 (Macintosh; Intel Mac OS X 10_11_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/57.0.2987.133 Safari/537.36
Command Line	/Applications/Google Chrome.app/Contents/MacOS/Google Chrome --flag-switches-begin --flag-switches-end
Executable Path	/Applications/Google Chrome.app/Contents/MacOS/Google Chrome
Profile Path	/Users/cballenger/Library/Application Support/Google/Chrome/Default
Variations	ea8deb27-3d47f4f4
3095aa95-3f4a17df
7c1bc906-f55a7974
2a33b90e-e7b70917
ba3f87da-b4a760c3
cf558fa6-48a16532
58aac55e-3f4a17df
f3499283-f94608c8
31362330-3f4a17df
c70841c8-a2567007
349d561b-3f4a17df
9e201a2b-26bbc519
6eb432aa-ca7d8d80
5274eb09-3f4a17df
684d1cdf-51126808
9773d3bd-ca7d8d80
b22b3d54-3d47f4f4
2e109477-ca7d8d80
165e16d1-3f4a17df
9e5c75f1-c16ec2e6
6b121ae7-ca7d8d80
f79cb77b-3d47f4f4
b7786474-d93a0620
23a898eb-ca7d8d80
48bd06b2-f23d1dea
4ea303a6-fea2917d
3d7e3f6a-2eb01455
9736de91-ca7d8d80
64005e71-4ad60575
f069f10b-803f8fc4
69bf80fa-91c810ef
867c4c68-3f4a17df
b2f0086-93053e47
7fc902e8-ca7d8d80
f11cb941-11910166
f47ae82a-86f22ee5
3ac60855-486e2a9c
f296190c-f9f7acb5
4442aae2-a5822863
ed1d377-e1cc0f14
75f0f0a0-a5822863
e2b18481-bd104136
e7e71889-e1cc0f14
828a5926-ca7d8d80

Deleted: net-internals-log.json 237 KB

net-internals-log.json 237 KB View Download

Version 59.0.3071.71
Windows 8.1 Enterprise
Dell Latitude

Is the google chrome development team still working on this?

Logs:
https://drive.google.com/file/d/0B_XTSb-2rgCeTDB5RFBkZmRaWHc/view?usp=sharing

This issue was duplicated into  issue 709771  which we believe is fixed. It has been verified in version 60.0.3088.3.

The PDF loading issue is still present for me in Canary 61.0.3115.0

Two PDFs that have issues: 
http://hoffmannursery.com/assets/files/files/Hoffman_Nursery_Availability.pdf
http://www.northcreeknurseries.com/_ccLib/downloads/North_Creek_Availability_All_Plugs.pdf

Can you please file a new bug so we can work out what's going on?

Sure thing. 

I am on chromeos 59.0.3071.91 on a Toshiba. Just in the last week or so I notice that if I go to a PDF in a tab and then move to another tab and return, I have to re-load the PDF. It is very annoying if I have two PDF files open, as I am continually re-loading. This did not happen a few weeks ago - is there a switch I can flip to get the multi-pdf behaviour back, or is this the same bug as presented here (blank screen until I re-load).

Re #109: it is an issue but it sounds different to this one. Please file a new bug.

I have found this sort of issue for sometime with both Chrome and Chrome Canary. Some pdfs work as expected, others are either entirely blank or only start showing text at some point part way through the document. If downloaded and read in Acrobat reader etc the pdf is fine. My other solution to date has been to use IE to access. 

It appears to be a problem with some pdf documents online as part of an online planning system used by a number of councils in England. 

Having had the problem today with a number of files I decided to check further and so found this ongoing bug report.

Chrome details:
Google Chrome	59.0.3071.115 (Official Build) (64-bit) (cohort: Stable)
Revision	3cf8514bb1239453fd15ff1f7efee389ac9df8ba-refs/branch-heads/3071@{#820}
OS	Windows
JavaScript	V8 5.9.211.38
Flash	26.0.0.137 C:\Users\Clerk\AppData\Local\Google\Chrome\User Data\PepperFlash\26.0.0.137\pepflashplayer.dll
User Agent	Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/59.0.3071.115 Safari/537.36
Command Line	"C:\Program Files (x86)\Google\Chrome\Application\chrome.exe" --no-startup-window /prefetch:5 --flag-switches-begin --flag-switches-end
Executable Path	C:\Program Files (x86)\Google\Chrome\Application\chrome.exe

Windows 10 Pro version 1703 OS Build 15063.483 
Dell Optiplex 3020

chrome://net-export log file attached

pdf file path http://edrms.milton-keynes.gov.uk/NorthgatePublicDocs/00299511.pdf - file attached.

I have also reproduced the issue with the same pdf file using Canary 62.0.3171.0 on Windows 10 Home on a HP laptop  

Deleted: 00299511.pdf 210 KB

00299511.pdf 210 KB Download

Deleted: chrome-net-export-log.json 573 KB

chrome-net-export-log.json 573 KB View Download

re Comment #111, this sounds like a different issue, can you please file a new bug?

Ive tried all the same steps and its not working either. I have not added nor deleted any extensions that I have used for the last year and for some reason for the last 2 weeks I havent been able to view pdfs. I do remember a recent update that I did with Chrome and come to think about it the problems started happening shortly after it was installed. 

"Whoops,

There was a problem loading this page"

Is what I get, sometimes on one page or another its not always the same and some times it works. Any help would be appreciated as this is putting a big kink in my work flows. 

Hi, jkarren@, are you experiencing similar problem on a recent version of Chrome? If so, pls file a new bug at https://bugs.chromium.org/p/chromium/issues/entry
Commenting on a closed bug may not work.

pdf displays but when you click on it the browser window goes blank (white).

Please file new bugs, instead of adding comments to old bugs with similar symptoms. As of this writing, the newest bug with PDFs and blank screens is  bug 870404 .