RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-26 20:17:44 +00:00
Code Issues Activity Actions

Kernel: Don't hog VMObject lock when remapping a region page

Browse source 
We really only need the VMObject lock when accessing the physical pages
array, so once we have a strong pointer to the physical page we want to
remap, we can give up the VMObject lock.

This fixes a deadlock I encountered while building DOOM on SMP.
This commit is contained in:
Andreas Kling 2022-08-18 18:54:36 +02:00
parent 10399a258f
commit b560442fe1
2 changed files with 53 additions and 42 deletions
Download patch file Download diff file Expand all files Collapse all files

3
Kernel/Memory/Region.h
View file

@ -199,7 +199,7 @@ private:
Region(NonnullRefPtr<VMObject>, size_t offset_in_vmobject, OwnPtr<KString>, Region::Access access, Cacheable, bool shared);
Region(VirtualRange const&, NonnullRefPtr<VMObject>, size_t offset_in_vmobject, OwnPtr<KString>, Region::Access access, Cacheable, bool shared);
[[nodiscard]] bool remap_vmobject_page(size_t page_index, bool with_flush = true);
[[nodiscard]] bool remap_vmobject_page(size_t page_index, NonnullRefPtr<PhysicalPage>);
void set_access_bit(Access access, bool b)
{
@ -214,6 +214,7 @@ private:
[[nodiscard]] PageFaultResponse handle_zero_fault(size_t page_index);
[[nodiscard]] bool map_individual_page_impl(size_t page_index);
[[nodiscard]] bool map_individual_page_impl(size_t page_index, RefPtr<PhysicalPage>);
RefPtr<PageDirectory> m_page_directory;
VirtualRange m_range;