Weak interior handles by davidwrighton · Pull Request #100446 · dotnet/runtime

davidwrighton · 2024-03-29T17:09:53Z

Add a new type of handle (HNDTYPE_WEAK_INTERIOR_POINTER) and a use of it to make storage of managed class objects consistent between collectible and non-collectible variants

…sistent between collectible and non-collectible types

dotnet-policy-service · 2024-03-29T17:10:30Z

Tagging subscribers to this area: @mangod9
See info in area-owners.md if you want to be subscribed.

davidwrighton · 2024-03-29T17:10:32Z

@Maoni0 I'd like it if you could take a look at this sometime. I've not managed to test it ... but this is the sort of thing I want to do

teo-tsirpanis · 2024-03-29T19:27:08Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in

Maoni0 · 2024-04-01T23:23:47Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

otherwise the change looks ok to me.

…terior_handle

Don't forget to set the typehandle into the RuntimeType object

…ch point into the middle of the specified object

davidwrighton · 2024-04-15T22:42:07Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

Ok, done.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

Unfortunately my understanding is that api can only be called when the runtime is suspended due to race conditions, so I can't use this in any of the cases I'm working with.

jkotas · 2024-04-15T22:59:43Z

+ // if we did then copy the value
+ if (pUserData)
+ {
+ TADDR pObjectInteriorPointer = **dac_cast<DPTR(DPTR(TADDR))>(pUserData);


Nit: None of the surrounding code is DACized so dac_cast is unnecessary complication.

Yeah, I thought that VerifyHeap worked in the DAC, but it turns out that it is an entirely parallel managed implementation these days. I'll pull that out if you'd like.

There is a build break introduced by the dac_cast macro. My guess is that the easiest way to fix it is to delete the use here.

jkotas

There is a potential follow up codegen improvement for collectible assemblies:

We can change getRuntimeTypePointer to return InfoAccessType in addition to the pointer itself and allow JIT to fetch the Type object by a simple dereference instead of calling the helper in collectible code.

cc @EgorBo

jkotas · 2024-04-18T00:03:02Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in
// 2. We hold the m_crstLoaderAllocator lock, and double check to ensure that the data is ready to be filled in
Nit

jkotas · 2024-04-18T00:16:10Z

+ CONTRACTL_END;
+
+ const RUNTIMETYPEHANDLE handle = m_hExposedClassObject;
+ OBJECTREF retVal = ObjectToOBJECTREF((Object*)handle);


Suggested change
OBJECTREFretVal=ObjectToOBJECTREF((Object*)handle);
OBJECTREFretVal=ObjectToOBJECTREF(handle);
RUNTIMETYPEHANDLE can be defined as Object* to save casts like this. It is actually Object* now.

- Implementation of weak interior handle - Use new weak interior handle to make managed class object storage consistent between collectible and non-collectible types

davidwrighton · 2024-03-29T17:09:53Z

Add a new type of handle (HNDTYPE_WEAK_INTERIOR_POINTER) and a use of it to make storage of managed class objects consistent between collectible and non-collectible variants

…sistent between collectible and non-collectible types

dotnet-policy-service · 2024-03-29T17:10:30Z

Tagging subscribers to this area: @mangod9
See info in area-owners.md if you want to be subscribed.

davidwrighton · 2024-03-29T17:10:32Z

@Maoni0 I'd like it if you could take a look at this sometime. I've not managed to test it ... but this is the sort of thing I want to do

teo-tsirpanis · 2024-03-29T19:27:08Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in

Maoni0 · 2024-04-01T23:23:47Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

otherwise the change looks ok to me.

…terior_handle

Don't forget to set the typehandle into the RuntimeType object

…ch point into the middle of the specified object

davidwrighton · 2024-04-15T22:42:07Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

Ok, done.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

Unfortunately my understanding is that api can only be called when the runtime is suspended due to race conditions, so I can't use this in any of the cases I'm working with.

jkotas · 2024-04-15T22:59:43Z

+ // if we did then copy the value
+ if (pUserData)
+ {
+ TADDR pObjectInteriorPointer = **dac_cast<DPTR(DPTR(TADDR))>(pUserData);


Nit: None of the surrounding code is DACized so dac_cast is unnecessary complication.

Yeah, I thought that VerifyHeap worked in the DAC, but it turns out that it is an entirely parallel managed implementation these days. I'll pull that out if you'd like.

There is a build break introduced by the dac_cast macro. My guess is that the easiest way to fix it is to delete the use here.

jkotas

There is a potential follow up codegen improvement for collectible assemblies:

We can change getRuntimeTypePointer to return InfoAccessType in addition to the pointer itself and allow JIT to fetch the Type object by a simple dereference instead of calling the helper in collectible code.

cc @EgorBo

jkotas · 2024-04-18T00:03:02Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in
// 2. We hold the m_crstLoaderAllocator lock, and double check to ensure that the data is ready to be filled in
Nit

jkotas · 2024-04-18T00:16:10Z

+ CONTRACTL_END;
+
+ const RUNTIMETYPEHANDLE handle = m_hExposedClassObject;
+ OBJECTREF retVal = ObjectToOBJECTREF((Object*)handle);


Suggested change
OBJECTREFretVal=ObjectToOBJECTREF((Object*)handle);
OBJECTREFretVal=ObjectToOBJECTREF(handle);
RUNTIMETYPEHANDLE can be defined as Object* to save casts like this. It is actually Object* now.

- Implementation of weak interior handle - Use new weak interior handle to make managed class object storage consistent between collectible and non-collectible types

davidwrighton · 2024-03-29T17:09:53Z

Add a new type of handle (HNDTYPE_WEAK_INTERIOR_POINTER) and a use of it to make storage of managed class objects consistent between collectible and non-collectible variants

…sistent between collectible and non-collectible types

dotnet-policy-service · 2024-03-29T17:10:30Z

Tagging subscribers to this area: @mangod9
See info in area-owners.md if you want to be subscribed.

davidwrighton · 2024-03-29T17:10:32Z

@Maoni0 I'd like it if you could take a look at this sometime. I've not managed to test it ... but this is the sort of thing I want to do

teo-tsirpanis · 2024-03-29T19:27:08Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in

Maoni0 · 2024-04-01T23:23:47Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

otherwise the change looks ok to me.

…terior_handle

Don't forget to set the typehandle into the RuntimeType object

…ch point into the middle of the specified object

davidwrighton · 2024-04-15T22:42:07Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

Ok, done.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

Unfortunately my understanding is that api can only be called when the runtime is suspended due to race conditions, so I can't use this in any of the cases I'm working with.

jkotas · 2024-04-15T22:59:43Z

+ // if we did then copy the value
+ if (pUserData)
+ {
+ TADDR pObjectInteriorPointer = **dac_cast<DPTR(DPTR(TADDR))>(pUserData);


Nit: None of the surrounding code is DACized so dac_cast is unnecessary complication.

Yeah, I thought that VerifyHeap worked in the DAC, but it turns out that it is an entirely parallel managed implementation these days. I'll pull that out if you'd like.

There is a build break introduced by the dac_cast macro. My guess is that the easiest way to fix it is to delete the use here.

jkotas

There is a potential follow up codegen improvement for collectible assemblies:

We can change getRuntimeTypePointer to return InfoAccessType in addition to the pointer itself and allow JIT to fetch the Type object by a simple dereference instead of calling the helper in collectible code.

cc @EgorBo

jkotas · 2024-04-18T00:03:02Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in
// 2. We hold the m_crstLoaderAllocator lock, and double check to ensure that the data is ready to be filled in
Nit

jkotas · 2024-04-18T00:16:10Z

+ CONTRACTL_END;
+
+ const RUNTIMETYPEHANDLE handle = m_hExposedClassObject;
+ OBJECTREF retVal = ObjectToOBJECTREF((Object*)handle);


Suggested change
OBJECTREFretVal=ObjectToOBJECTREF((Object*)handle);
OBJECTREFretVal=ObjectToOBJECTREF(handle);
RUNTIMETYPEHANDLE can be defined as Object* to save casts like this. It is actually Object* now.

- Implementation of weak interior handle - Use new weak interior handle to make managed class object storage consistent between collectible and non-collectible types

davidwrighton · 2024-03-29T17:09:53Z

Add a new type of handle (HNDTYPE_WEAK_INTERIOR_POINTER) and a use of it to make storage of managed class objects consistent between collectible and non-collectible variants

…sistent between collectible and non-collectible types

dotnet-policy-service · 2024-03-29T17:10:30Z

Tagging subscribers to this area: @mangod9
See info in area-owners.md if you want to be subscribed.

davidwrighton · 2024-03-29T17:10:32Z

@Maoni0 I'd like it if you could take a look at this sometime. I've not managed to test it ... but this is the sort of thing I want to do

teo-tsirpanis · 2024-03-29T19:27:08Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in

Maoni0 · 2024-04-01T23:23:47Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

otherwise the change looks ok to me.

…terior_handle

Don't forget to set the typehandle into the RuntimeType object

…ch point into the middle of the specified object

davidwrighton · 2024-04-15T22:42:07Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

Ok, done.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

Unfortunately my understanding is that api can only be called when the runtime is suspended due to race conditions, so I can't use this in any of the cases I'm working with.

jkotas · 2024-04-15T22:59:43Z

+ // if we did then copy the value
+ if (pUserData)
+ {
+ TADDR pObjectInteriorPointer = **dac_cast<DPTR(DPTR(TADDR))>(pUserData);


Nit: None of the surrounding code is DACized so dac_cast is unnecessary complication.

Yeah, I thought that VerifyHeap worked in the DAC, but it turns out that it is an entirely parallel managed implementation these days. I'll pull that out if you'd like.

There is a build break introduced by the dac_cast macro. My guess is that the easiest way to fix it is to delete the use here.

jkotas

There is a potential follow up codegen improvement for collectible assemblies:

We can change getRuntimeTypePointer to return InfoAccessType in addition to the pointer itself and allow JIT to fetch the Type object by a simple dereference instead of calling the helper in collectible code.

cc @EgorBo

jkotas · 2024-04-18T00:03:02Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in
// 2. We hold the m_crstLoaderAllocator lock, and double check to ensure that the data is ready to be filled in
Nit

jkotas · 2024-04-18T00:16:10Z

+ CONTRACTL_END;
+
+ const RUNTIMETYPEHANDLE handle = m_hExposedClassObject;
+ OBJECTREF retVal = ObjectToOBJECTREF((Object*)handle);


Suggested change
OBJECTREFretVal=ObjectToOBJECTREF((Object*)handle);
OBJECTREFretVal=ObjectToOBJECTREF(handle);
RUNTIMETYPEHANDLE can be defined as Object* to save casts like this. It is actually Object* now.

- Implementation of weak interior handle - Use new weak interior handle to make managed class object storage consistent between collectible and non-collectible types

davidwrighton · 2024-03-29T17:09:53Z

Add a new type of handle (HNDTYPE_WEAK_INTERIOR_POINTER) and a use of it to make storage of managed class objects consistent between collectible and non-collectible variants

…sistent between collectible and non-collectible types

dotnet-policy-service · 2024-03-29T17:10:30Z

Tagging subscribers to this area: @mangod9
See info in area-owners.md if you want to be subscribed.

davidwrighton · 2024-03-29T17:10:32Z

@Maoni0 I'd like it if you could take a look at this sometime. I've not managed to test it ... but this is the sort of thing I want to do

teo-tsirpanis · 2024-03-29T19:27:08Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in

Maoni0 · 2024-04-01T23:23:47Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

otherwise the change looks ok to me.

…terior_handle

Don't forget to set the typehandle into the RuntimeType object

…ch point into the middle of the specified object

davidwrighton · 2024-04-15T22:42:07Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

Ok, done.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

Unfortunately my understanding is that api can only be called when the runtime is suspended due to race conditions, so I can't use this in any of the cases I'm working with.

jkotas · 2024-04-15T22:59:43Z

+ // if we did then copy the value
+ if (pUserData)
+ {
+ TADDR pObjectInteriorPointer = **dac_cast<DPTR(DPTR(TADDR))>(pUserData);


Nit: None of the surrounding code is DACized so dac_cast is unnecessary complication.

Yeah, I thought that VerifyHeap worked in the DAC, but it turns out that it is an entirely parallel managed implementation these days. I'll pull that out if you'd like.

There is a build break introduced by the dac_cast macro. My guess is that the easiest way to fix it is to delete the use here.

jkotas

There is a potential follow up codegen improvement for collectible assemblies:

We can change getRuntimeTypePointer to return InfoAccessType in addition to the pointer itself and allow JIT to fetch the Type object by a simple dereference instead of calling the helper in collectible code.

cc @EgorBo

jkotas · 2024-04-18T00:03:02Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in
// 2. We hold the m_crstLoaderAllocator lock, and double check to ensure that the data is ready to be filled in
Nit

jkotas · 2024-04-18T00:16:10Z

+ CONTRACTL_END;
+
+ const RUNTIMETYPEHANDLE handle = m_hExposedClassObject;
+ OBJECTREF retVal = ObjectToOBJECTREF((Object*)handle);


Suggested change
OBJECTREFretVal=ObjectToOBJECTREF((Object*)handle);
OBJECTREFretVal=ObjectToOBJECTREF(handle);
RUNTIMETYPEHANDLE can be defined as Object* to save casts like this. It is actually Object* now.

- Implementation of weak interior handle - Use new weak interior handle to make managed class object storage consistent between collectible and non-collectible types

davidwrighton · 2024-03-29T17:09:53Z

Add a new type of handle (HNDTYPE_WEAK_INTERIOR_POINTER) and a use of it to make storage of managed class objects consistent between collectible and non-collectible variants

…sistent between collectible and non-collectible types

dotnet-policy-service · 2024-03-29T17:10:30Z

Tagging subscribers to this area: @mangod9
See info in area-owners.md if you want to be subscribed.

davidwrighton · 2024-03-29T17:10:32Z

@Maoni0 I'd like it if you could take a look at this sometime. I've not managed to test it ... but this is the sort of thing I want to do

teo-tsirpanis · 2024-03-29T19:27:08Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in

Maoni0 · 2024-04-01T23:23:47Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

otherwise the change looks ok to me.

…terior_handle

Don't forget to set the typehandle into the RuntimeType object

…ch point into the middle of the specified object

davidwrighton · 2024-04-15T22:42:07Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

Ok, done.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

Unfortunately my understanding is that api can only be called when the runtime is suspended due to race conditions, so I can't use this in any of the cases I'm working with.

jkotas · 2024-04-15T22:59:43Z

+ // if we did then copy the value
+ if (pUserData)
+ {
+ TADDR pObjectInteriorPointer = **dac_cast<DPTR(DPTR(TADDR))>(pUserData);


Nit: None of the surrounding code is DACized so dac_cast is unnecessary complication.

Yeah, I thought that VerifyHeap worked in the DAC, but it turns out that it is an entirely parallel managed implementation these days. I'll pull that out if you'd like.

There is a build break introduced by the dac_cast macro. My guess is that the easiest way to fix it is to delete the use here.

jkotas

There is a potential follow up codegen improvement for collectible assemblies:

We can change getRuntimeTypePointer to return InfoAccessType in addition to the pointer itself and allow JIT to fetch the Type object by a simple dereference instead of calling the helper in collectible code.

cc @EgorBo

jkotas · 2024-04-18T00:03:02Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in
// 2. We hold the m_crstLoaderAllocator lock, and double check to ensure that the data is ready to be filled in
Nit

jkotas · 2024-04-18T00:16:10Z

+ CONTRACTL_END;
+
+ const RUNTIMETYPEHANDLE handle = m_hExposedClassObject;
+ OBJECTREF retVal = ObjectToOBJECTREF((Object*)handle);


Suggested change
OBJECTREFretVal=ObjectToOBJECTREF((Object*)handle);
OBJECTREFretVal=ObjectToOBJECTREF(handle);
RUNTIMETYPEHANDLE can be defined as Object* to save casts like this. It is actually Object* now.

- Implementation of weak interior handle - Use new weak interior handle to make managed class object storage consistent between collectible and non-collectible types

davidwrighton · 2024-03-29T17:09:53Z

Add a new type of handle (HNDTYPE_WEAK_INTERIOR_POINTER) and a use of it to make storage of managed class objects consistent between collectible and non-collectible variants

…sistent between collectible and non-collectible types

dotnet-policy-service · 2024-03-29T17:10:30Z

Tagging subscribers to this area: @mangod9
See info in area-owners.md if you want to be subscribed.

davidwrighton · 2024-03-29T17:10:32Z

@Maoni0 I'd like it if you could take a look at this sometime. I've not managed to test it ... but this is the sort of thing I want to do

teo-tsirpanis · 2024-03-29T19:27:08Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in

Maoni0 · 2024-04-01T23:23:47Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

otherwise the change looks ok to me.

…terior_handle

Don't forget to set the typehandle into the RuntimeType object

…ch point into the middle of the specified object

davidwrighton · 2024-04-15T22:42:07Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

Ok, done.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

Unfortunately my understanding is that api can only be called when the runtime is suspended due to race conditions, so I can't use this in any of the cases I'm working with.

jkotas · 2024-04-15T22:59:43Z

+ // if we did then copy the value
+ if (pUserData)
+ {
+ TADDR pObjectInteriorPointer = **dac_cast<DPTR(DPTR(TADDR))>(pUserData);


Nit: None of the surrounding code is DACized so dac_cast is unnecessary complication.

Yeah, I thought that VerifyHeap worked in the DAC, but it turns out that it is an entirely parallel managed implementation these days. I'll pull that out if you'd like.

There is a build break introduced by the dac_cast macro. My guess is that the easiest way to fix it is to delete the use here.

jkotas

There is a potential follow up codegen improvement for collectible assemblies:

We can change getRuntimeTypePointer to return InfoAccessType in addition to the pointer itself and allow JIT to fetch the Type object by a simple dereference instead of calling the helper in collectible code.

cc @EgorBo

jkotas · 2024-04-18T00:03:02Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in
// 2. We hold the m_crstLoaderAllocator lock, and double check to ensure that the data is ready to be filled in
Nit

jkotas · 2024-04-18T00:16:10Z

+ CONTRACTL_END;
+
+ const RUNTIMETYPEHANDLE handle = m_hExposedClassObject;
+ OBJECTREF retVal = ObjectToOBJECTREF((Object*)handle);


Suggested change
OBJECTREFretVal=ObjectToOBJECTREF((Object*)handle);
OBJECTREFretVal=ObjectToOBJECTREF(handle);
RUNTIMETYPEHANDLE can be defined as Object* to save casts like this. It is actually Object* now.

- Implementation of weak interior handle - Use new weak interior handle to make managed class object storage consistent between collectible and non-collectible types

davidwrighton · 2024-03-29T17:09:53Z

Add a new type of handle (HNDTYPE_WEAK_INTERIOR_POINTER) and a use of it to make storage of managed class objects consistent between collectible and non-collectible variants

…sistent between collectible and non-collectible types

dotnet-policy-service · 2024-03-29T17:10:30Z

Tagging subscribers to this area: @mangod9
See info in area-owners.md if you want to be subscribed.

davidwrighton · 2024-03-29T17:10:32Z

@Maoni0 I'd like it if you could take a look at this sometime. I've not managed to test it ... but this is the sort of thing I want to do

teo-tsirpanis · 2024-03-29T19:27:08Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in

Maoni0 · 2024-04-01T23:23:47Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

otherwise the change looks ok to me.

…terior_handle

Don't forget to set the typehandle into the RuntimeType object

…ch point into the middle of the specified object

davidwrighton · 2024-04-15T22:42:07Z

I presume the extra info for these handles must always point to the middle of the primary target of this handle (it's not obvious from the caller I'm seeing in loaderallocator.cpp). we should add some verification code that this is the case.

Ok, done.

I thought I'd mention we do have a `GCHeap::GetContainingObject" that gets you from an interior pointer to its containing object, in case this is useful to you.

Unfortunately my understanding is that api can only be called when the runtime is suspended due to race conditions, so I can't use this in any of the cases I'm working with.

jkotas · 2024-04-15T22:59:43Z

+ // if we did then copy the value
+ if (pUserData)
+ {
+ TADDR pObjectInteriorPointer = **dac_cast<DPTR(DPTR(TADDR))>(pUserData);


Nit: None of the surrounding code is DACized so dac_cast is unnecessary complication.

Yeah, I thought that VerifyHeap worked in the DAC, but it turns out that it is an entirely parallel managed implementation these days. I'll pull that out if you'd like.

There is a build break introduced by the dac_cast macro. My guess is that the easiest way to fix it is to delete the use here.

jkotas

There is a potential follow up codegen improvement for collectible assemblies:

We can change getRuntimeTypePointer to return InfoAccessType in addition to the pointer itself and allow JIT to fetch the Type object by a simple dereference instead of calling the helper in collectible code.

cc @EgorBo

jkotas · 2024-04-18T00:03:02Z

+ //
+ // We work around these details by the following means
+ // 1. We use a LOADERHANDLE to keep the object alive until the LoaderAllocator is freed.
+ // 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in


Suggested change
// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in
// 2. We hold the m_crstLoaderAllocator lock, and double check to ensure that the data is ready to be filled in
Nit

jkotas · 2024-04-18T00:16:10Z

+ CONTRACTL_END;
+
+ const RUNTIMETYPEHANDLE handle = m_hExposedClassObject;
+ OBJECTREF retVal = ObjectToOBJECTREF((Object*)handle);


Suggested change
OBJECTREFretVal=ObjectToOBJECTREF((Object*)handle);
OBJECTREFretVal=ObjectToOBJECTREF(handle);
RUNTIMETYPEHANDLE can be defined as Object* to save casts like this. It is actually Object* now.

- Implementation of weak interior handle - Use new weak interior handle to make managed class object storage consistent between collectible and non-collectible types

davidwrighton added 2 commits March 29, 2024 07:49

Implementation of weak interior handle
5f36950

Use new weak interior handle to make managed class object storage con…
e19eaeb
…sistent between collectible and non-collectible types

ghost added the area-VM-coreclr label Mar 29, 2024

dotnet-policy-serviceBot assigned davidwrighton Mar 29, 2024

davidwrighton mentioned this pull request Mar 29, 2024
Make normal statics simpler #99183
Merged

teo-tsirpanis reviewed Mar 29, 2024
View reviewed changes

This was referenced Mar 29, 2024
Intermittend failures in ThreadPoolTests on Mono Linux ARM64 #85960
Closed
LibraryImportGenerator unit test segfault illegal memory access #89054
Closed
Segmentation fault in System.Text.RegularExpressions.Tests #93206
Closed

Merge branch 'main' ofhttps://github.com/dotnet/runtimeinto weak_in…
56da86b
…terior_handle

This was referenced Apr 15, 2024
ComInterfaceGenerator.Unit.Tests failed on linux-x64 running interpreter #100683
Closed
LibraryImportGenerator.Unit.Tests crashing on linux-x64 mono interpreter #100800
Open
CompileToAssembly_SimpleUseInDebug test failure on windows amd64 #100964
Closed

davidwrighton added 2 commits April 15, 2024 12:07

Adjust where the GC_NOTRIGGER section begins
1db7694
Don't forget to set the typehandle into the RuntimeType object

Add validation that weak interior pointers have interior pointers whi…
6a5b21f
…ch point into the middle of the specified object

davidwrighton marked this pull request as ready for review April 15, 2024 22:42

jkotas reviewed Apr 15, 2024
View reviewed changes

Remove DAC support from new verification routine
7cfd502

build-analysisBot mentioned this pull request Apr 16, 2024
The hosted runner encountered an error while running your job. (Error Type: Disconnect). dotnet/dnceng#1919
Closed
3 tasks

davidwrighton requested review from Maoni0, cshung and jkotas April 16, 2024 18:44

jkotas approved these changes Apr 18, 2024
View reviewed changes

Code review fixes
c1a3dfa

davidwrighton merged commit 56510f0 into dotnet:mainApr 18, 2024

github-actionsBot locked and limited conversation to collaborators May 19, 2024

davidwrighton added 2 commits March 29, 2024 07:49

Implementation of weak interior handle
5f36950

Use new weak interior handle to make managed class object storage con…
e19eaeb
…sistent between collectible and non-collectible types

ghost added the area-VM-coreclr label Mar 29, 2024

dotnet-policy-serviceBot assigned davidwrighton Mar 29, 2024

davidwrighton mentioned this pull request Mar 29, 2024
Make normal statics simpler #99183
Merged

teo-tsirpanis reviewed Mar 29, 2024
View reviewed changes

This was referenced Mar 29, 2024
Intermittend failures in ThreadPoolTests on Mono Linux ARM64 #85960
Closed
LibraryImportGenerator unit test segfault illegal memory access #89054
Closed
Segmentation fault in System.Text.RegularExpressions.Tests #93206
Closed

Merge branch 'main' ofhttps://github.com/dotnet/runtimeinto weak_in…
56da86b
…terior_handle

This was referenced Apr 15, 2024
ComInterfaceGenerator.Unit.Tests failed on linux-x64 running interpreter #100683
Closed
LibraryImportGenerator.Unit.Tests crashing on linux-x64 mono interpreter #100800
Open
CompileToAssembly_SimpleUseInDebug test failure on windows amd64 #100964
Closed

davidwrighton added 2 commits April 15, 2024 12:07

Adjust where the GC_NOTRIGGER section begins
1db7694
Don't forget to set the typehandle into the RuntimeType object

Add validation that weak interior pointers have interior pointers whi…
6a5b21f
…ch point into the middle of the specified object

davidwrighton marked this pull request as ready for review April 15, 2024 22:42

jkotas reviewed Apr 15, 2024
View reviewed changes

Remove DAC support from new verification routine
7cfd502

build-analysisBot mentioned this pull request Apr 16, 2024
The hosted runner encountered an error while running your job. (Error Type: Disconnect). dotnet/dnceng#1919
Closed
3 tasks

davidwrighton requested review from Maoni0, cshung and jkotas April 16, 2024 18:44

jkotas approved these changes Apr 18, 2024
View reviewed changes

Code review fixes
c1a3dfa

davidwrighton merged commit 56510f0 into dotnet:mainApr 18, 2024

github-actionsBot locked and limited conversation to collaborators May 19, 2024

davidwrighton added 2 commits March 29, 2024 07:49

Implementation of weak interior handle
5f36950

Use new weak interior handle to make managed class object storage con…
e19eaeb
…sistent between collectible and non-collectible types

ghost added the area-VM-coreclr label Mar 29, 2024

dotnet-policy-serviceBot assigned davidwrighton Mar 29, 2024

davidwrighton mentioned this pull request Mar 29, 2024
Make normal statics simpler #99183
Merged

teo-tsirpanis reviewed Mar 29, 2024
View reviewed changes

This was referenced Mar 29, 2024
Intermittend failures in ThreadPoolTests on Mono Linux ARM64 #85960
Closed
LibraryImportGenerator unit test segfault illegal memory access #89054
Closed
Segmentation fault in System.Text.RegularExpressions.Tests #93206
Closed

Merge branch 'main' ofhttps://github.com/dotnet/runtimeinto weak_in…
56da86b
…terior_handle

This was referenced Apr 15, 2024
ComInterfaceGenerator.Unit.Tests failed on linux-x64 running interpreter #100683
Closed
LibraryImportGenerator.Unit.Tests crashing on linux-x64 mono interpreter #100800
Open
CompileToAssembly_SimpleUseInDebug test failure on windows amd64 #100964
Closed

davidwrighton added 2 commits April 15, 2024 12:07

Adjust where the GC_NOTRIGGER section begins
1db7694
Don't forget to set the typehandle into the RuntimeType object

Add validation that weak interior pointers have interior pointers whi…
6a5b21f
…ch point into the middle of the specified object

davidwrighton marked this pull request as ready for review April 15, 2024 22:42

jkotas reviewed Apr 15, 2024
View reviewed changes

Remove DAC support from new verification routine
7cfd502

build-analysisBot mentioned this pull request Apr 16, 2024
The hosted runner encountered an error while running your job. (Error Type: Disconnect). dotnet/dnceng#1919
Closed
3 tasks

davidwrighton requested review from Maoni0, cshung and jkotas April 16, 2024 18:44

jkotas approved these changes Apr 18, 2024
View reviewed changes

Code review fixes
c1a3dfa

davidwrighton merged commit 56510f0 into dotnet:mainApr 18, 2024

github-actionsBot locked and limited conversation to collaborators May 19, 2024

davidwrighton added 2 commits March 29, 2024 07:49

Implementation of weak interior handle
5f36950

Use new weak interior handle to make managed class object storage con…
e19eaeb
…sistent between collectible and non-collectible types

ghost added the area-VM-coreclr label Mar 29, 2024

dotnet-policy-serviceBot assigned davidwrighton Mar 29, 2024

davidwrighton mentioned this pull request Mar 29, 2024
Make normal statics simpler #99183
Merged

teo-tsirpanis reviewed Mar 29, 2024
View reviewed changes

This was referenced Mar 29, 2024
Intermittend failures in ThreadPoolTests on Mono Linux ARM64 #85960
Closed
LibraryImportGenerator unit test segfault illegal memory access #89054
Closed
Segmentation fault in System.Text.RegularExpressions.Tests #93206
Closed

Merge branch 'main' ofhttps://github.com/dotnet/runtimeinto weak_in…
56da86b
…terior_handle

This was referenced Apr 15, 2024
ComInterfaceGenerator.Unit.Tests failed on linux-x64 running interpreter #100683
Closed
LibraryImportGenerator.Unit.Tests crashing on linux-x64 mono interpreter #100800
Open
CompileToAssembly_SimpleUseInDebug test failure on windows amd64 #100964
Closed

davidwrighton added 2 commits April 15, 2024 12:07

Adjust where the GC_NOTRIGGER section begins
1db7694
Don't forget to set the typehandle into the RuntimeType object

Add validation that weak interior pointers have interior pointers whi…
6a5b21f
…ch point into the middle of the specified object

davidwrighton marked this pull request as ready for review April 15, 2024 22:42

jkotas reviewed Apr 15, 2024
View reviewed changes

Remove DAC support from new verification routine
7cfd502

build-analysisBot mentioned this pull request Apr 16, 2024
The hosted runner encountered an error while running your job. (Error Type: Disconnect). dotnet/dnceng#1919
Closed
3 tasks

davidwrighton requested review from Maoni0, cshung and jkotas April 16, 2024 18:44

jkotas approved these changes Apr 18, 2024
View reviewed changes

Code review fixes
c1a3dfa

davidwrighton merged commit 56510f0 into dotnet:mainApr 18, 2024

github-actionsBot locked and limited conversation to collaborators May 19, 2024

davidwrighton added 2 commits March 29, 2024 07:49

Implementation of weak interior handle
5f36950

Use new weak interior handle to make managed class object storage con…
e19eaeb
…sistent between collectible and non-collectible types

ghost added the area-VM-coreclr label Mar 29, 2024

dotnet-policy-serviceBot assigned davidwrighton Mar 29, 2024

davidwrighton mentioned this pull request Mar 29, 2024
Make normal statics simpler #99183
Merged

teo-tsirpanis reviewed Mar 29, 2024
View reviewed changes

This was referenced Mar 29, 2024
Intermittend failures in ThreadPoolTests on Mono Linux ARM64 #85960
Closed
LibraryImportGenerator unit test segfault illegal memory access #89054
Closed
Segmentation fault in System.Text.RegularExpressions.Tests #93206
Closed

Merge branch 'main' ofhttps://github.com/dotnet/runtimeinto weak_in…
56da86b
…terior_handle

This was referenced Apr 15, 2024
ComInterfaceGenerator.Unit.Tests failed on linux-x64 running interpreter #100683
Closed
LibraryImportGenerator.Unit.Tests crashing on linux-x64 mono interpreter #100800
Open
CompileToAssembly_SimpleUseInDebug test failure on windows amd64 #100964
Closed

davidwrighton added 2 commits April 15, 2024 12:07

Adjust where the GC_NOTRIGGER section begins
1db7694
Don't forget to set the typehandle into the RuntimeType object

Add validation that weak interior pointers have interior pointers whi…
6a5b21f
…ch point into the middle of the specified object

davidwrighton marked this pull request as ready for review April 15, 2024 22:42

jkotas reviewed Apr 15, 2024
View reviewed changes

Remove DAC support from new verification routine
7cfd502

build-analysisBot mentioned this pull request Apr 16, 2024
The hosted runner encountered an error while running your job. (Error Type: Disconnect). dotnet/dnceng#1919
Closed
3 tasks

davidwrighton requested review from Maoni0, cshung and jkotas April 16, 2024 18:44

jkotas approved these changes Apr 18, 2024
View reviewed changes

Code review fixes
c1a3dfa

davidwrighton merged commit 56510f0 into dotnet:mainApr 18, 2024

github-actionsBot locked and limited conversation to collaborators May 19, 2024

davidwrighton added 2 commits March 29, 2024 07:49

Implementation of weak interior handle
5f36950

Use new weak interior handle to make managed class object storage con…
e19eaeb
…sistent between collectible and non-collectible types

ghost added the area-VM-coreclr label Mar 29, 2024

dotnet-policy-serviceBot assigned davidwrighton Mar 29, 2024

davidwrighton mentioned this pull request Mar 29, 2024
Make normal statics simpler #99183
Merged

teo-tsirpanis reviewed Mar 29, 2024
View reviewed changes

This was referenced Mar 29, 2024
Intermittend failures in ThreadPoolTests on Mono Linux ARM64 #85960
Closed
LibraryImportGenerator unit test segfault illegal memory access #89054
Closed
Segmentation fault in System.Text.RegularExpressions.Tests #93206
Closed

Merge branch 'main' ofhttps://github.com/dotnet/runtimeinto weak_in…
56da86b
…terior_handle

This was referenced Apr 15, 2024
ComInterfaceGenerator.Unit.Tests failed on linux-x64 running interpreter #100683
Closed
LibraryImportGenerator.Unit.Tests crashing on linux-x64 mono interpreter #100800
Open
CompileToAssembly_SimpleUseInDebug test failure on windows amd64 #100964
Closed

davidwrighton added 2 commits April 15, 2024 12:07

Adjust where the GC_NOTRIGGER section begins
1db7694
Don't forget to set the typehandle into the RuntimeType object

Add validation that weak interior pointers have interior pointers whi…
6a5b21f
…ch point into the middle of the specified object

davidwrighton marked this pull request as ready for review April 15, 2024 22:42

jkotas reviewed Apr 15, 2024
View reviewed changes

Remove DAC support from new verification routine
7cfd502

build-analysisBot mentioned this pull request Apr 16, 2024
The hosted runner encountered an error while running your job. (Error Type: Disconnect). dotnet/dnceng#1919
Closed
3 tasks

davidwrighton requested review from Maoni0, cshung and jkotas April 16, 2024 18:44

jkotas approved these changes Apr 18, 2024
View reviewed changes

Code review fixes
c1a3dfa

davidwrighton merged commit 56510f0 into dotnet:mainApr 18, 2024

github-actionsBot locked and limited conversation to collaborators May 19, 2024

davidwrighton added 2 commits March 29, 2024 07:49

Implementation of weak interior handle
5f36950

Use new weak interior handle to make managed class object storage con…
e19eaeb
…sistent between collectible and non-collectible types

ghost added the area-VM-coreclr label Mar 29, 2024

dotnet-policy-serviceBot assigned davidwrighton Mar 29, 2024

davidwrighton mentioned this pull request Mar 29, 2024
Make normal statics simpler #99183
Merged

teo-tsirpanis reviewed Mar 29, 2024
View reviewed changes

This was referenced Mar 29, 2024
Intermittend failures in ThreadPoolTests on Mono Linux ARM64 #85960
Closed
LibraryImportGenerator unit test segfault illegal memory access #89054
Closed
Segmentation fault in System.Text.RegularExpressions.Tests #93206
Closed

Merge branch 'main' ofhttps://github.com/dotnet/runtimeinto weak_in…
56da86b
…terior_handle

This was referenced Apr 15, 2024
ComInterfaceGenerator.Unit.Tests failed on linux-x64 running interpreter #100683
Closed
LibraryImportGenerator.Unit.Tests crashing on linux-x64 mono interpreter #100800
Open
CompileToAssembly_SimpleUseInDebug test failure on windows amd64 #100964
Closed

davidwrighton added 2 commits April 15, 2024 12:07

Adjust where the GC_NOTRIGGER section begins
1db7694
Don't forget to set the typehandle into the RuntimeType object

Add validation that weak interior pointers have interior pointers whi…
6a5b21f
…ch point into the middle of the specified object

davidwrighton marked this pull request as ready for review April 15, 2024 22:42

jkotas reviewed Apr 15, 2024
View reviewed changes

Remove DAC support from new verification routine
7cfd502

build-analysisBot mentioned this pull request Apr 16, 2024
The hosted runner encountered an error while running your job. (Error Type: Disconnect). dotnet/dnceng#1919
Closed
3 tasks

davidwrighton requested review from Maoni0, cshung and jkotas April 16, 2024 18:44

jkotas approved these changes Apr 18, 2024
View reviewed changes

Code review fixes
c1a3dfa

davidwrighton merged commit 56510f0 into dotnet:mainApr 18, 2024

github-actionsBot locked and limited conversation to collaborators May 19, 2024

davidwrighton added 2 commits March 29, 2024 07:49

Implementation of weak interior handle
5f36950

Use new weak interior handle to make managed class object storage con…
e19eaeb
…sistent between collectible and non-collectible types

ghost added the area-VM-coreclr label Mar 29, 2024

dotnet-policy-serviceBot assigned davidwrighton Mar 29, 2024

davidwrighton mentioned this pull request Mar 29, 2024
Make normal statics simpler #99183
Merged

teo-tsirpanis reviewed Mar 29, 2024
View reviewed changes

This was referenced Mar 29, 2024
Intermittend failures in ThreadPoolTests on Mono Linux ARM64 #85960
Closed
LibraryImportGenerator unit test segfault illegal memory access #89054
Closed
Segmentation fault in System.Text.RegularExpressions.Tests #93206
Closed

Merge branch 'main' ofhttps://github.com/dotnet/runtimeinto weak_in…
56da86b
…terior_handle

This was referenced Apr 15, 2024
ComInterfaceGenerator.Unit.Tests failed on linux-x64 running interpreter #100683
Closed
LibraryImportGenerator.Unit.Tests crashing on linux-x64 mono interpreter #100800
Open
CompileToAssembly_SimpleUseInDebug test failure on windows amd64 #100964
Closed

davidwrighton added 2 commits April 15, 2024 12:07

Adjust where the GC_NOTRIGGER section begins
1db7694
Don't forget to set the typehandle into the RuntimeType object

Add validation that weak interior pointers have interior pointers whi…
6a5b21f
…ch point into the middle of the specified object

davidwrighton marked this pull request as ready for review April 15, 2024 22:42

jkotas reviewed Apr 15, 2024
View reviewed changes

Remove DAC support from new verification routine
7cfd502

build-analysisBot mentioned this pull request Apr 16, 2024
The hosted runner encountered an error while running your job. (Error Type: Disconnect). dotnet/dnceng#1919
Closed
3 tasks

davidwrighton requested review from Maoni0, cshung and jkotas April 16, 2024 18:44

jkotas approved these changes Apr 18, 2024
View reviewed changes

Code review fixes
c1a3dfa

davidwrighton merged commit 56510f0 into dotnet:mainApr 18, 2024

github-actionsBot locked and limited conversation to collaborators May 19, 2024

	// 2. We hold the crstLoaderAllocatorLock, and double check to wnsure that the data is ready to be filled in
	// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in

	// 2. We hold the crstLoaderAllocatorLock, and double check to ensure that the data is ready to be filled in
	// 2. We hold the m_crstLoaderAllocator lock, and double check to ensure that the data is ready to be filled in

	OBJECTREFretVal=ObjectToOBJECTREF((Object*)handle);
	OBJECTREFretVal=ObjectToOBJECTREF(handle);

Uh oh!

Conversation

davidwrighton commented Mar 29, 2024

Uh oh!

dotnet-policy-serviceBot commented Mar 29, 2024

Uh oh!

davidwrighton commented Mar 29, 2024

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Maoni0 commented Apr 1, 2024

Uh oh!

davidwrighton commented Apr 15, 2024• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

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!

jkotas left a comment

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!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Uh oh!

Conversation

davidwrighton commented Mar 29, 2024

Uh oh!

dotnet-policy-serviceBot commented Mar 29, 2024

Uh oh!

davidwrighton commented Mar 29, 2024

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Maoni0 commented Apr 1, 2024

Uh oh!

davidwrighton commented Apr 15, 2024• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

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!

jkotas left a comment

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!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Uh oh!

Conversation

davidwrighton commented Mar 29, 2024

Uh oh!

davidwrighton commented Apr 15, 2024•
edited
Loading

davidwrighton commented Apr 15, 2024•
edited
Loading

davidwrighton commented Apr 15, 2024•
edited
Loading