profiler changes for tiered compilation #14612

davmason · 2017-10-20T07:14:28Z

Adds new profiler APIs for tiered compilation support. Also as part of this change I updated ProfToEEInterfaceImpl::GetILToNativeMapping2 to properly respect the requested rejit ID.

noahfalk · 2017-10-20T08:20:11Z

src/vm/proftoeeinterfaceimpl.cpp


-            // Now that tiered compilation can create more than one jitted code version for the same rejit id
-            // we are arbitrarily choosing the first one to return. To return all of them we'd presumably need
-            // a new profiler API.


This comment is still correct - we are arbitrarily choosing an implementation. The comment could also mention the new APIs you've added rather than refer to hypothetical new APIs.

noahfalk · 2017-10-20T08:24:27Z

src/vm/proftoeeinterfaceimpl.cpp

+
+                pCodeVersionManager->GetILCodeVersion(pMD, reJitId);
+            }
+


Lets add another comment to indicate this is an arbitrarily chosen version and profilers are recommended to use GetILToNativeMapping3 to avoid that ambiguity.

Ok, sounds good

noahfalk · 2017-10-20T08:25:11Z

src/vm/proftoeeinterfaceimpl.cpp


+/*
+ * GetNativeCodeStartAddresses
+ * 


Whoops! Taken care of now

noahfalk · 2017-10-20T08:43:49Z

src/vm/proftoeeinterfaceimpl.cpp

+                                                           ReJITID reJitId, 
+                                                           ULONG32 cCodeStartAddresses, 
+                                                           ULONG32 *pcCodeStartAddresses, 
+                                                           UINT_PTR codeStartAddresses[])


Check out GetModuleInfo2 for example of how it deals with the variable size character buffer and the associated size in/out parameters. I know there aren't a ton of total examples in the profiler API and we already aren't strictly consistent which adds to the confusion, but I think GetModuleInfo2's behavior is a reasonable standard to emulate.

Ok, I made it so that pcCodeStartAddresses and codeStartAddresses are optional. If that's not what you meant, let me know.

I was thinking of the different ways you can call GetModuleInfo2 that don't work here yet:

You can call with only pcchName/pcCodeStartAddresses non-null and receive the size of the buffer that would be needed (here this only works if the caller also allocated a sufficiently large buffer)

You can call with pcchName/pcCodeStartAddresses null if you don't care about the size (looks like you still go to E_INVALIDARG in this case)

Ok, I see what you're saying now. I missed it the first time. I just updated it to be able to call with just with either pcCodeAddresses or codeStartAddresses null now and the other will be filled out appropriately

noahfalk · 2017-10-20T08:46:01Z

src/vm/proftoeeinterfaceimpl.cpp

+        CodeVersionManager *pCodeVersionManager = pMD->GetCodeVersionManager();
+
+        ILCodeVersion ilCodeVersion = NULL;
+        {


I'd put the entire enumeration under the lock. Did anything lead you to avoid putting it under the lock (we might want to change whatever that was too)

I forgot to add the lock initially, then when I ran the tests only GetILCodeversion had an assert about the lock so I just put the it around that one place.

I have changed it to be around the whole enumeration now though

noahfalk · 2017-10-20T08:48:00Z

src/vm/proftoeeinterfaceimpl.cpp

+
+/*
+ * GetILToNativeMapping3
+ * 


Another todo to do

noahfalk · 2017-10-20T08:49:38Z

src/vm/proftoeeinterfaceimpl.h


    // end ICorProfilerInfo8

+    // beging ICorProfiler9


begin ICorProfilerInfo9

thanks, done

noahfalk · 2017-10-20T08:50:17Z

src/vm/proftoeeinterfaceimpl.h

+        ULONG32* pcCodeInfos, 
+        COR_PRF_CODE_INFO codeInfos[]);
+
+    // end ICorProfiler9


ICorProfilerInfo9

noahfalk · 2017-10-20T08:54:03Z

src/vm/proftoeeinterfaceimpl.cpp

        hr = ValidateParametersForGetCodeInfo(pMethodDesc, cCodeInfos, codeInfos);
        if (SUCCEEDED(hr))
        {
            CodeVersionManager* pCodeVersionManager = pMethodDesc->GetCodeVersionManager();


I notice a pre-existing bug - this code path didn't grab the code version manager lock and it needs to

If you run the tests with a debug build it should give you asserts anywhere you are missing the lock.

Thanks. I haven't run the existing profiler tests yet, but in general I am running with debug coreclr so I should catch these once I do.

noahfalk · 2017-10-20T08:56:18Z

src/vm/proftoeeinterfaceimpl.cpp

+                pCodeVersionManager->GetILCodeVersion(pMD, reJitId);
+            }
+
+            NativeCodeVersionCollection nativeCodeVersions = ilCodeVersion.GetNativeCodeVersions(pMD);


The whole enumeration should be inside the lock (but you can leave the call to GetILtoNativeMapping3 outside of it)

noahfalk · 2017-10-23T23:31:53Z

src/vm/proftoeeinterfaceimpl.cpp

+            else
+            {
+                for(ULONG32 i = 0; i < cCodeStartAddresses; ++i)
+                {


You've got a buffer overrun here. cCodeAddresses can be larger than trueLen but addresses only has length trueLen.

noahfalk · 2017-10-23T23:33:55Z

LGTM, modulo that one remaining buffer overrun

davmason added the area-Diagnostics label Oct 20, 2017

davmason self-assigned this Oct 20, 2017

davmason requested a review from noahfalk October 20, 2017 07:14

dnfclas added the cla-already-signed label Oct 20, 2017

noahfalk reviewed Oct 20, 2017

View reviewed changes

davmason added 4 commits October 20, 2017 14:03

profiler changes

0bccef1

code review feedback

22f32db

missed one thing on code review feedback

8412d0e

more code review feedback

2138ef9

noahfalk reviewed Oct 23, 2017

View reviewed changes

fix potential buffer overrun

e9b9896

davmason merged commit 5d66791 into dotnet:master Oct 24, 2017

davmason deleted the new_apis branch October 24, 2017 03:24

profiler changes for tiered compilation #14612

profiler changes for tiered compilation #14612

Uh oh!

Conversation

davmason commented Oct 20, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

noahfalk Oct 20, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

noahfalk Oct 20, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

noahfalk commented Oct 23, 2017

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

noahfalk Oct 20, 2017 •

edited

Loading

noahfalk Oct 20, 2017 •

edited

Loading