ll_prof: Reduce profiling hooks overhead from >400% to 25%.

ll_prof: Reduce profiling hooks overhead from >400% to 25%.

Switched to using binary low-level log instead of the textual log used
by the ticks processor. The binary log contains code-related events,
code object names, and their bodies. When writing to the log we ask
glibc to use a larger buffer. To avoid complex processing of the
snapshot log (which is still textual) the serializer emits final
snapshot position to code name mappings that can be quickly be read
without replaying the snapshot log. (This might be useful for the
ticks processor.)

Committed: http://code.google.com/p/v8/source/detail?r=7729

[mnaganov (inactive), 2011-04-29 14:20:28]

LGTM with comments

http://codereview.chromium.org/6904127/diff/1/src/log.cc
File src/log.cc (right):

http://codereview.chromium.org/6904127/diff/1/src/log.cc#newcode1098
src/log.cc:1098: HashMap::Entry* entry = address_to_name_map_->Lookup(
This code is repeated several times. Worth extracting into a function? Or even
better -- wrap the hashmap into a class.

http://codereview.chromium.org/6904127/diff/1/src/utils.h
File src/utils.h (right):

http://codereview.chromium.org/6904127/diff/1/src/utils.h#newcode254
src/utils.h:254: static inline uint32_t ComputePointerHash(void* ptr) {
As we intend to deal with Addresses, maybe it's a good idea to use an Address
type here?

[Vitaly Repeshko, 2011-04-29 15:50:39]

Thanks for quick review!

http://codereview.chromium.org/6904127/diff/1/src/log.cc
File src/log.cc (right):

http://codereview.chromium.org/6904127/diff/1/src/log.cc#newcode1098
src/log.cc:1098: HashMap::Entry* entry = address_to_name_map_->Lookup(
On 2011/04/29 14:20:28, Mikhail Naganov (Chromium) wrote:
> This code is repeated several times. Worth extracting into a function? Or even
> better -- wrap the hashmap into a class.

Created a class. It's +20 lines overall, but the code is hopefully more
readable.

http://codereview.chromium.org/6904127/diff/1/src/utils.h
File src/utils.h (right):

http://codereview.chromium.org/6904127/diff/1/src/utils.h#newcode254
src/utils.h:254: static inline uint32_t ComputePointerHash(void* ptr) {
On 2011/04/29 14:20:28, Mikhail Naganov (Chromium) wrote:
> As we intend to deal with Addresses, maybe it's a good idea to use an Address
> type here?

Don't think so --- it'll require casting from everything that is not byte*.