[blocked roll] webkit_tests fail on wasm memory allocations |
||||||||
Issue descriptionSince 0cd7468b86d0958fdc6ae1d7f69da1569e05b5f8 ("[wasm] Always enable guard regions on 64-bit platforms"), some webkit_tests related to wasm memory fail: https://build.chromium.org/p/client.v8.fyi/builders/V8-Blink%20Linux%2064 This blocks the roll of v8 into chromium. Please revert or fix as soon as possible.
,
Apr 3 2018
,
Apr 4 2018
,
Apr 4 2018
,
Apr 4 2018
Here's a link to one of the failing builds, for future reference: https://build.chromium.org/p/client.v8.fyi/builders/V8-Blink%20Linux%2064/builds/22678
,
Apr 4 2018
I was just able to repro this with the original patch. After trying again with my new patch (https://crrev.com/c/996466), this did not repro. I'm going to call it fixed.
,
Apr 6 2018
Reopening, since this appeared again on the bots last night after relanding the original change.
,
Apr 7 2018
dziwka
,
Apr 13 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/927dffa49808aa941f76b405430e450019cfe1d5 commit 927dffa49808aa941f76b405430e450019cfe1d5 Author: Eric Holk <eholk@chromium.org> Date: Fri Apr 13 20:46:58 2018 Use higher dynamic address space limit unconditionally. We are working to enable full guard regions for WebAssembly memories on all 64-bit platforms, whereas previously they were only used when the WebAssembly trap handler feature was used. Since each memory + guard region requires 8 GiB of address space, the original 16 GiB limit is no longer sufficient. This CL makes the behavior of having a higher hard limit on the address space and a lower soft limit that is raised when needed to allocate a WebAssembly memory available even when not using trap handlers. Bug: chromium:828499 , v8:7619 Change-Id: I31a279176b91bf18be140a15bf2f64ae62debba1 Reviewed-on: https://chromium-review.googlesource.com/1000334 Reviewed-by: Kentaro Hara <haraken@chromium.org> Reviewed-by: Jorge Lucangeli Obes <jorgelo@chromium.org> Reviewed-by: Chris Palmer <palmer@chromium.org> Commit-Queue: Eric Holk <eholk@chromium.org> Cr-Commit-Position: refs/heads/master@{#550754} [modify] https://crrev.com/927dffa49808aa941f76b405430e450019cfe1d5/content/renderer/renderer_main_platform_delegate_linux.cc [modify] https://crrev.com/927dffa49808aa941f76b405430e450019cfe1d5/services/service_manager/sandbox/linux/sandbox_linux.cc [modify] https://crrev.com/927dffa49808aa941f76b405430e450019cfe1d5/services/service_manager/sandbox/linux/sandbox_seccomp_bpf_linux.h
,
Apr 17 2018
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/927dffa49808aa941f76b405430e450019cfe1d5 commit 927dffa49808aa941f76b405430e450019cfe1d5 Author: Eric Holk <eholk@chromium.org> Date: Fri Apr 13 20:46:58 2018 Use higher dynamic address space limit unconditionally. We are working to enable full guard regions for WebAssembly memories on all 64-bit platforms, whereas previously they were only used when the WebAssembly trap handler feature was used. Since each memory + guard region requires 8 GiB of address space, the original 16 GiB limit is no longer sufficient. This CL makes the behavior of having a higher hard limit on the address space and a lower soft limit that is raised when needed to allocate a WebAssembly memory available even when not using trap handlers. Bug: chromium:828499 , v8:7619 Change-Id: I31a279176b91bf18be140a15bf2f64ae62debba1 Reviewed-on: https://chromium-review.googlesource.com/1000334 Reviewed-by: Kentaro Hara <haraken@chromium.org> Reviewed-by: Jorge Lucangeli Obes <jorgelo@chromium.org> Reviewed-by: Chris Palmer <palmer@chromium.org> Commit-Queue: Eric Holk <eholk@chromium.org> Cr-Commit-Position: refs/heads/master@{#550754} [modify] https://crrev.com/927dffa49808aa941f76b405430e450019cfe1d5/content/renderer/renderer_main_platform_delegate_linux.cc [modify] https://crrev.com/927dffa49808aa941f76b405430e450019cfe1d5/services/service_manager/sandbox/linux/sandbox_linux.cc [modify] https://crrev.com/927dffa49808aa941f76b405430e450019cfe1d5/services/service_manager/sandbox/linux/sandbox_seccomp_bpf_linux.h
,
Dec 12
|
||||||||
►
Sign in to add a comment |
||||||||
Comment 1 by eholk@chromium.org
, Apr 3 2018