Skip to content

V
V8
Switch branch/tag
  1. 02 Jun, 2020 8 commits
    • Clemens Backes's avatar
      [liftoff] Use ZoneVectors during compilation · d47ab50b
      Clemens Backes authored 
      The allocation of heap storage for the vectors shows up in profiles, and
also the destruction of the contained elements.
Since we already have a compilation zone available, we can easily move
the vectors and all contained elements into the zone, where they will be
deallocated together with all other data at once.

R=thibaudm@chromium.org

Bug: v8:10576
Change-Id: I2c3d318f9b19b1b8c2e4e57e7523cd1c79192cc6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2224220Reviewed-by: 's avatarThibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68099}
      d47ab50b
    • Dominik Inführ's avatar
      [heap] Do not verify slots in concurrent sweeper · 37e3cac7
      Dominik Inführ authored 
      Cannot verify whether slots are cleared from a concurrent sweeping task,
the OLD_TO_NEW remembered set is owned by the main thread.

Bug: v8:10562
Change-Id: I0c880f7f96350aa35b44d2966c9621576825374c
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2215820
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Reviewed-by: 's avatarUlan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68098}
      37e3cac7
    • Clemens Backes's avatar
      [d8] Improve error message for illegal --trace-path · f8439175
      Clemens Backes authored 
      Unconditionally check for errors (instead of a DCHECK), and print a
better error message (including the file path) if the file cannot be
opened.

R=tebbi@chromium.org
CC=machenbach@chromium.org

Bug: chromium:1087039
Change-Id: Ia5cb76b309e78631ecf9462de6c7cd3eb4bf9e59
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2224226Reviewed-by: 's avatarTobias Tebbi <tebbi@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68097}
      f8439175
    • Clemens Backes's avatar
      [wasm][debug] Support multi-threaded breakpoints · 5fcb414a
      Clemens Backes authored 
      This adds support for multiple isolates sharing the same module but
setting different breakpoints. This is simulated by having a debugger
test that runs in the "--isolates" variant, i.e. two isolates running
the same test at the same time. Both isolates will set and remove
breakpoints.

The DebugInfo will keep a separate list of breakpoints per isolate, and
when recompiling a function for debugging it will respect all
breakpoints in all isolates.
In order to ensure consistency if multiple isolates are setting or
removing breakpoints simultaneously, we go back to a more coarse-grained
locking scheme, where the DebugInfo lock is held while re-compiling
Liftoff functions.

While recompilation will install the code in the module-global code
table and jump table (and hence all isolates will use it for future
calls), only the stack of the requesting isolate is rewritten to
immediately use new code. This is OK, because other isolates are not
interested in the new breakpoint(s) anyway.
On {SetBreakpoint}, we always need to rewrite the stack of the
requesting isolate though, even if the breakpoint was set before by
another isolate.

Drive-by: Some fixes in SharedFunctionInfo in order to support setting
breakpoints via the Debug mirror.

R=thibaudm@chromium.org

Bug: v8:10359
Change-Id: If659afb273260fc5e8124b4b617fb4322de473c7
Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_rel
Cq-Include-Trybots: luci.v8.try:v8_linux64_tsan_isolates_rel_ng
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2218059Reviewed-by: 's avatarBenedikt Meurer <bmeurer@chromium.org>
Reviewed-by: 's avatarThibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68096}
      5fcb414a
    • Georg Neis's avatar
      [turbofan] Fix a typer monotonicity bug · f3d46392
      Georg Neis authored 
      Bug: chromium:1085804
Change-Id: I98f12da97334bd5fd32bd01b1eca56be895dc0a8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2218286
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: 's avatarTobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68095}
      f3d46392
    • Leszek Swirski's avatar
      [offthread] Don't set marking bit on off-thread large pages · e4cc52bf
      Leszek Swirski authored 
      Bug: chromium:1088350
Change-Id: I9dc9f6b323ebe2e751167880acad0eed3584222a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2224227
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: 's avatarUlan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68094}
      e4cc52bf
    • Daniel Bevenius's avatar
      [tools] Add error handling to no_arg_cmd · fcac59ad
      Daniel Bevenius authored 
      Currently, it can be little difficult to understand why a command in
lldb-commands.py stops working. For example, at the moment running the
jlh command results in an empty line:

$ lldb --one-line "command script import ../../tools/lldb_commands.py" \
     v8_hello_world
(lldb) br s -f hello-world.cc -l 49
(lldb) jlh script

(lldb)

With this commit this would instead display the following error message:

(lldb) jlh script
Failed to evaluate command
_v8_internal_Print_Object(*(v8::internal::Object**)(*(void*)(script))) :
error: cannot cast from type 'v8::Local<v8::Script>' to pointer type
'void *'

The output is really only two lines but I've wrapped the lines here so
they don't exceed the 72 column width. I'll follow up with a commit to
fix the issue reported.

Change-Id: I634a412b616dad7cadd74dce36418d27c1997777
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2083477Reviewed-by: 's avatarLeszek Swirski <leszeks@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68093}
      fcac59ad
    • jiepan's avatar
      Fix printing order of opcode prefix and space · 1f5842ad
      jiepan authored 
      Leading whitespaces are used to indiate control depth,
opcode prefix should be printed after the whitespaces.

Change-Id: I0a22864d1d5a2e643b15a4c10909c0387922f8e3
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2224959Reviewed-by: 's avatarClemens Backes <clemensb@chromium.org>
Commit-Queue: Jie Pan <jie.pan@intel.com>
Cr-Commit-Position: refs/heads/master@{#68092}
      1f5842ad
  3. 01 Jun, 2020 5 commits
  5. 31 May, 2020 1 commit
  7. 30 May, 2020 2 commits
  9. 29 May, 2020 20 commits
  11. 28 May, 2020 4 commits