Cleanup most usages of Windows file mapping #121896
Conversation
github-actions
bot
added
the
needs-area-label
dotnet-policy-service
bot
added
the
community-contribution
huoyaoyuan
added
area-PAL-coreclr
and removed
needs-area-label
| #ifdef HOST_WINDOWS | ||
| #define MEMORY_MAPPED_STRESSLOG_BASE_ADDRESS (void*)0x400000000000 | ||
| #else | ||
| #define MEMORY_MAPPED_STRESSLOG_BASE_ADDRESS nullptr |
There was a problem hiding this comment.
Is this removing support for memory mapped stresslog on non-Windows?
There was a problem hiding this comment.
Yes. NativeAOT copy of stresslog removes memory map support on all platforms. I haven't measured the impact.
| void* m_address; | ||
|
|
||
| #ifdef TARGET_WINDOWS | ||
| HANDLE m_hFile; |
There was a problem hiding this comment.
I do not think these Windows-specific need to be kept open and stored here. It should be fine to close them at the end of Open.
There was a problem hiding this comment.
I'm not sure how serious should we design this. The helper is only used through ilasm/ildasm, and we may get rid of the native version at all in foreseeable future.
Or, just hold off for this cleanup and rewrite ilasm/ildasm first.
There was a problem hiding this comment.
Alternatively, we can just avoid mmap at all in them and read the whole file into array. I don't expect the I/O performance to matter here.
2 participants
The remaining cases are superpmi (included in #121367) and PEImage, which involves LoadLibrary and we would want more future-proof abstraction.
All the cases included in this PR are transitively used by ilasm/ildasm. For stgio, I'm doing simple ifdef's on OS api, since it's too complicated around different I/O models, and we would get rid of it with new metadata library.