Keyboard Shortcuts

File
u :up to issue
m :publish + mail comments
M :edit review message
j / k :jump to file after / before current file
J / K :jump to next file with a comment after / before current file
Side-by-side diff
i :toggle intra-line diffs
e :expand all comments
c :collapse all comments
s :toggle showing all comments
n / p :next / previous diff chunk or comment
N / P :next / previous comment
<Up> / <Down> :next / previous line
<Enter> :respond to / edit current comment
d :mark current comment as done
Issue
u :up to list of issues
m :publish + mail comments
j / k :jump to patch after / before current patch
o / <Enter> :open current patch in side-by-side view
i :open current patch in unified diff view
Issue List
j / k :jump to issue after / before current issue
o / <Enter> :open current issue
# : close issue
Comment/message editing
<Ctrl> + s or <Ctrl> + Enter :save comment
<Esc> :cancel edit
Rietveld Code Review Tool
Help | Bug tracker | Discussion group | Source code | Sign in
(244)
Issues Repositories Search
Open Issues | Closed Issues | All Issues | Sign in with your Google Account to create issues and add comments

Issue 1316041: fast path memory state machine

Can't Edit
Can't Publish+Mail
Start Review
Created:
15 years, 7 months ago by kcc
Modified:
15 years, 7 months ago
Reviewers:
Timur Iskhodzhanov
Base URL:
http://data-race-test.googlecode.com/svn/trunk/tsan/
Visibility:
Public.

Patch Set 1 #

Patch Set 2 : fix 80 char #

Total comments: 16

Patch Set 3 : more comments, changed name of stats #

Patch Set 4 : less indents #

Patch Set 5 : more comments #

Patch Set 6 : x #

Created: 15 years, 7 months ago
Download [raw] [tar.bz2]
Unified diffs Side-by-side diffs Delta from patch set Stats (+146 lines, -52 lines) Patch
M thread_sanitizer.cc View 1 2 3 4 5 17 chunks +141 lines, -52 lines 0 comments Download
M ts_stats.h View 1 2 3 4 5 2 chunks +5 lines, -0 lines 0 comments Download
Total messages: 2
|
kcc
15 years, 7 months ago (2010年05月26日 07:41:14 UTC) #1
Sign in to reply to this message.
Timur Iskhodzhanov
http://codereview.appspot.com/1316041/diff/2001/3001 File thread_sanitizer.cc (right): http://codereview.appspot.com/1316041/diff/2001/3001#newcode193 thread_sanitizer.cc:193: LID Singleton() const { return LID(raw()); } GetSingleton() ? ...
15 years, 7 months ago (2010年05月26日 09:12:29 UTC) #2
http://codereview.appspot.com/1316041/diff/2001/3001
File thread_sanitizer.cc (right):
http://codereview.appspot.com/1316041/diff/2001/3001#newcode193
thread_sanitizer.cc:193: LID Singleton() const { return LID(raw()); }
GetSingleton() ?
http://codereview.appspot.com/1316041/diff/2001/3001#newcode212
thread_sanitizer.cc:212: SID Singleton() const {
GetSingleton() ?
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5656
thread_sanitizer.cc:5656: // Tuple segment sets and check for race.
// Returns true iff bla-bla-bla, writes the result to *new_sval
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5661
thread_sanitizer.cc:5661: //#define MSM_STAT(i) G_stats->msm[i]++;
do { if (DEBUG_MODE) ...; } while(0) ??
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5665
thread_sanitizer.cc:5665: if (rd_ssid.IsEmpty()) {
I think I'd be good to insert the following to the beginning of the function:
if (!rd_ssid.IsSingleton() || !wr_ssid.IsSingleton())
 return false;
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5670
thread_sanitizer.cc:5670: MSM_STAT(0);
FastPath1 vs msm[0] ?
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5673
thread_sanitizer.cc:5673: // wr_ssid != 0
DCHECK ?
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5674
thread_sanitizer.cc:5674: if (wr_ssid.IsSingleton()) {
if (!wr_ssid.IsSingleton())
 return false;
... to avoid extra indent
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5683
thread_sanitizer.cc:5683: // same thread, but different segments.
.. segment.
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5698
thread_sanitizer.cc:5698: } else {
DCHECK(!rd_ssid.IsEmpty());
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5699
thread_sanitizer.cc:5699: if (wr_ssid.IsEmpty()) {
invert the condition to get rid of extra indent
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5701
thread_sanitizer.cc:5701: if (rd_ssid.IsSingleton()) {
if (!rd_ssid.IsSingleton())
 return false;
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5750
thread_sanitizer.cc:5750: ShadowValue *sval_p = NULL;
new_sval_p
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5757
thread_sanitizer.cc:5757: }
what do you think about "GetOrAddSvalAtOffset" ?
http://codereview.appspot.com/1316041/diff/2001/3001#newcode5762
thread_sanitizer.cc:5762: if (!MemoryStateMachineSameThread(is_w, old_sval, tid,
hm... is it publish-friendly?
http://codereview.appspot.com/1316041/diff/2001/3002
File ts_stats.h (right):
http://codereview.appspot.com/1316041/diff/2001/3002#newcode216
ts_stats.h:216: uintptr_t msm[16];
more descriptive name / comment what these stat means.
Sign in to reply to this message.
|
Powered by Google App Engine
RSS Feeds Recent Issues | This issue
This is Rietveld f62528b

AltStyle によって変換されたページ (->オリジナル) /