Fix secondary stub cache and add a test for the stub cache lookups.

Fix secondary stub cache and add a test for the stub cache lookups.
Committed: https://code.google.com/p/v8/source/detail?r=10864

[Sven Panne, 2012-02-29 09:37:42]

LGTM with tiny nits.

https://chromiumcodereview.appspot.com/9496010/diff/1/src/arm/stub-cache-arm.cc
File src/arm/stub-cache-arm.cc (right):

https://chromiumcodereview.appspot.com/9496010/diff/1/src/arm/stub-cache-arm....
src/arm/stub-cache-arm.cc:199: // entry size being 12.
Hmmm, I can see multiplication by 3 and a left shift by one, but that leaves a
factor of 2. To be honest, I don't fully understand the calculations in
ProbeTable, some comments in prose would be helpful...

https://chromiumcodereview.appspot.com/9496010/diff/1/src/flag-definitions.h
File src/flag-definitions.h (right):

https://chromiumcodereview.appspot.com/9496010/diff/1/src/flag-definitions.h#...
src/flag-definitions.h:566: DEFINE_bool(test_secondary_stub_cache,
These should probably be debug-only flags.

https://chromiumcodereview.appspot.com/9496010/diff/1/test/cctest/test-api.cc
File test/cctest/test-api.cc (right):

https://chromiumcodereview.appspot.com/9496010/diff/1/test/cctest/test-api.cc...
test/cctest/test-api.cc:16112: static int* LookupCounter(const char* name) {
It seems that all those statics here are only used in DEBUG mode, so I guess
that gcc will complain in release mode.

https://chromiumcodereview.appspot.com/9496010/diff/1/test/cctest/test-api.cc...
test/cctest/test-api.cc:16138: TEST(SecondaryStubCache) {
This differs from TEST(PrimaryStubCache) only in a single line, so a separate
method for the common code might be appropriate.

Just out of curiosity: Where/how do we set/reset the i::FLAG_foo flags for each
test?

[Erik Corry, 2012-02-29 10:45:59]

https://chromiumcodereview.appspot.com/9496010/diff/1/src/arm/stub-cache-arm.cc
File src/arm/stub-cache-arm.cc (right):

https://chromiumcodereview.appspot.com/9496010/diff/1/src/arm/stub-cache-arm....
src/arm/stub-cache-arm.cc:199: // entry size being 12.
On 2012/02/29 09:37:42, Sven Panne wrote:
> Hmmm, I can see multiplication by 3 and a left shift by one, but that leaves a
> factor of 2. To be honest, I don't fully understand the calculations in
> ProbeTable, some comments in prose would be helpful...

Done.

https://chromiumcodereview.appspot.com/9496010/diff/1/src/flag-definitions.h
File src/flag-definitions.h (right):

https://chromiumcodereview.appspot.com/9496010/diff/1/src/flag-definitions.h#...
src/flag-definitions.h:566: DEFINE_bool(test_secondary_stub_cache,
On 2012/02/29 09:37:42, Sven Panne wrote:
> These should probably be debug-only flags.

That is already the case.

https://chromiumcodereview.appspot.com/9496010/diff/1/test/cctest/test-api.cc
File test/cctest/test-api.cc (right):

https://chromiumcodereview.appspot.com/9496010/diff/1/test/cctest/test-api.cc...
test/cctest/test-api.cc:16112: static int* LookupCounter(const char* name) {
On 2012/02/29 09:37:42, Sven Panne wrote:
> It seems that all those statics here are only used in DEBUG mode, so I guess
> that gcc will complain in release mode.

It does not complain about probes_counter etc. but it does complain about
kMegaMorphicTestProgram.  Go figure.

https://chromiumcodereview.appspot.com/9496010/diff/1/test/cctest/test-api.cc...
test/cctest/test-api.cc:16138: TEST(SecondaryStubCache) {
On 2012/02/29 09:37:42, Sven Panne wrote:
> This differs from TEST(PrimaryStubCache) only in a single line, so a separate
> method for the common code might be appropriate.

Done
 
> Just out of curiosity: Where/how do we set/reset the i::FLAG_foo flags for
each
> test?

Each test is run in its own process, which terminates, so no need to reset
anything.  This is why these tests are not THREADED_TEST tests.