Jika/redox-kernel
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Investigate why user heap isn't mapped

Browse Source 
Took me way too long to spot this :D
This commit is contained in:
jD91mZM2
2020-08-15 17:22:34 +02:00
parent e3814c1ca2
commit da7b813fa9
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
src/syscall/process.rs
View File

@@ -455,14 +455,14 @@ pub fn clone(flags: CloneFlags, stack_base: usize) -> Result<ContextId> {
// Copy user heap mapping, if found
if let Some(heap_shared) = heap_opt {
//TODO: find out why this gets unmapped (perhaps it is never mapped)
// This will not always be mapped. `context.heap` starts off
// with a size of zero, and then any call to `brk` increases
// it. But, if `brk` doesn't increase it, it's zero.
if let Some(frame) = active_table.p4()[crate::USER_HEAP_PML4].pointed_frame() {
let flags = active_table.p4()[crate::USER_HEAP_PML4].flags();
active_table.with(&mut new_table, &mut temporary_page, |mapper| {
mapper.p4_mut()[crate::USER_HEAP_PML4].set(frame, flags);
});
} else {
println!("clone: user heap not mapped for {:X?}", heap_shared);
}
context.heap = Some(heap_shared);
}