1088 - Necko: Netlib performance ->URLImpl Locking Overhead

Status: VERIFIED FIXED

(Core :: Networking, defect, P2)

Description

[troy]

Currently URLImpl uses NSPR cached monitors in order to be thread safe. The
problem is NSPR cached monitors have a very large performance overhead.

Kipp pointed out that NSPR 2.0 has simple locks, see PR_NewLock() which on
Windows are just a critical section.

We should have one NSPR lock for all of the URLImpl objects which makes the
resource overhead low

Comment 1

[leger]

rickg, would you consider this a ss: bug?

Comment 2

[leger]

Taking off ss: list per bug mtg today.

Comment 3

[don]

Re-assigned to [e-mail].

Gagan, is this a netlib bug?  Who should get this?

Comment 4

[Gagan]

I think that sounds feasible. Adding Warren for comments and accepting bug.

Comment 5

[leger]

Noting [Perf] in summary.  Starting to track bugs to review for performance
milestone.

Comment 6

[leger]

Setting all current Open/Normal to M4.

Comment 7

[leger]

Moving [Perf] from Status to Status Whiteboard.

Comment 8

[Paul MacQuiddy]

setting paulmac as QA contact for all gagan's bugs (sorry for the spam)

Comment 9

[Gagan]

Per DP's suggestion marking these till M8. Though Necko lands with M7, we will
be able to verify it for M8.

Comment 10

[James Everingham]

I'm moving this to target M9, Necko will be enabled somewhere during late M8 or
early M9.  We will need to get on this and it cannot be postponed past the M9
milestone.

Comment 11

[leger]

Changing all Networking Library/Browser bugs to Networking-Core component for
Browser.

Occasionally, Bugzilla will burp and cause Verified bugs to reopen when I do
this in a bulk change.  If this happens, I will fix. ;-)

Comment 12

[Gagan]

Probably not relevant with Necko. Pl. verify.

Comment 13

[Warren Harris]

I don't believe that the URL methods need to be thread-safe now. The DOM and
Netlib are no longer operating on the same object. Rick Potts should confirm
this.

Comment 14

[rpotts (gone)]

I'm resolving this bug because multiple threads are no longer sharing the same
objects.  It used to be that the underlying URL_Struct was manipulated by both
the netlib thread and the UI thread.

This is no longer the case...

-- rick

Comment 15

[leger]

Bulk move of all Networking-Core (to be deleted component) bugs to new
Networking component.