Skip to content

[Sanitizers][Apple] Fix logic bugs that break RestrictMemoryToMaxAddress #124712

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jan 30, 2025
Merged

[Sanitizers][Apple] Fix logic bugs that break RestrictMemoryToMaxAddress #124712

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 11 additions & 5 deletions compiler-rt/lib/sanitizer_common/sanitizer_mac.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1203,13 +1203,14 @@ uptr MapDynamicShadow(uptr shadow_size_bytes, uptr shadow_scale,
const uptr left_padding =
Max<uptr>(granularity, 1ULL << min_shadow_base_alignment);

uptr space_size = shadow_size_bytes + left_padding;
uptr space_size = shadow_size_bytes;

uptr largest_gap_found = 0;
uptr max_occupied_addr = 0;

VReport(2, "FindDynamicShadowStart, space_size = %p\n", (void *)space_size);
uptr shadow_start =
FindAvailableMemoryRange(space_size, alignment, granularity,
FindAvailableMemoryRange(space_size, alignment, left_padding,
&largest_gap_found, &max_occupied_addr);
// If the shadow doesn't fit, restrict the address space to make it fit.
if (shadow_start == 0) {
Expand All @@ -1229,9 +1230,9 @@ uptr MapDynamicShadow(uptr shadow_size_bytes, uptr shadow_scale,
}
RestrictMemoryToMaxAddress(new_max_vm);
high_mem_end = new_max_vm - 1;
space_size = (high_mem_end >> shadow_scale) + left_padding;
space_size = (high_mem_end >> shadow_scale);
VReport(2, "FindDynamicShadowStart, space_size = %p\n", (void *)space_size);
shadow_start = FindAvailableMemoryRange(space_size, alignment, granularity,
shadow_start = FindAvailableMemoryRange(space_size, alignment, left_padding,
nullptr, nullptr);
if (shadow_start == 0) {
Report("Unable to find a memory range after restricting VM.\n");
Expand Down Expand Up @@ -1272,10 +1273,15 @@ uptr FindAvailableMemoryRange(uptr size, uptr alignment, uptr left_padding,
mach_msg_type_number_t count = kRegionInfoSize;
kr = mach_vm_region_recurse(mach_task_self(), &address, &vmsize, &depth,
(vm_region_info_t)&vminfo, &count);
if (kr == KERN_INVALID_ADDRESS) {

// There are cases where going beyond the processes' max vm does
// not return KERN_INVALID_ADDRESS so we check for going beyond that
// max address as well.
if (kr == KERN_INVALID_ADDRESS || address > max_vm_address) {
Copy link
Collaborator

@yln yln Jan 29, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Currently, if mach_vm_region_recurse() ever returns an error different from KERN_INVALID_ADDRESS (I don't think it currently does except if the arguments are wrong, then it returns KERN_INVALID_ARGUMENT), then we would still treat the result "successful enough" to update max_occupied_addr. Is that the intention or should it be kr != KERN_SUCCESS?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

+1 to changing the condition to kr != KERN_SUCCESS .

Copy link
Contributor Author

@thetruestblue thetruestblue Jan 29, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I dont know if this is the case. For this condition we simply finishing checking that iteration. A different return code would probably be more properly handled by not doing that and potentially throwing an error. I also don't know if we would expect different kernel codes in other cases (entitlements?) I would rather do a follow up PR with this change to keep this change more limited.

// No more regions beyond "address", consider the gap at the end of VM.
address = max_vm_address;
vmsize = 0;
kr = -1; // break after this iteration.
} else {
if (max_occupied_addr) *max_occupied_addr = address + vmsize;
}
Expand Down