binder: validate alloc->mm in ->mmap() handler
PATCH 1/1 binder: validate alloc->mm in ->mmap() handler
commit 3ce00bb7e91cf57d723905371507af57182c37ef upstream.
Since commit 1da52815d5f1 ("binder: fix alloc->vma_vm_mm null-ptr dereference") binder caches a pointer to the current->mm during open(). This fixes a null-ptr dereference reported by syzkaller. Unfortunately, it also opens the door for a process to update its mm after the open(), (e.g. via execve) making the cached alloc->mm pointer invalid.
Things get worse when the process continues to mmap() a vma. From this point forward, binder will attempt to find this vma using an obsolete alloc->mm reference. Such as in binder_update_page_range(), where the wrong vma is obtained via vma_lookup(), yet binder proceeds to happily insert new pages into it.
To avoid this issue fail the ->mmap() callback if we detect a mismatch between the vma->vm_mm and the original alloc->mm pointer. This prevents alloc->vm_addr from getting set, so that any subsequent vma_lookup() calls fail as expected.
--- drivers/android/binder_alloc.c | 7 +++++++ 1 file changed, 7 insertions(+)
drivers/android/binder_alloc.c +7 -0
@@ -760,6 +760,12 @@ int binder_alloc_mmap_handler(struct binder_alloc *alloc,const char *failure_string;struct binder_buffer *buffer; +if (unlikely (vma->vm_mm != alloc->vma_vm_mm)) { + ret = -EINVAL ; + failure_string ="invalid vma->vm_mm" ; + goto err_invalid_mm; + } +mutex_lock (&binder_alloc_mmap_lock);if (alloc->buffer_size) { ret = -EBUSY ;
@@ -806,6 +812,7 @@ int binder_alloc_mmap_handler(struct binder_alloc *alloc, alloc->buffer_size =0 ; err_already_mapped:mutex_unlock (&binder_alloc_mmap_lock); +err_invalid_mm: binder_alloc_debug(BINDER_DEBUG_USER_ERROR, "%s: %d %lx-%lx %s failed %d\n", __func__, alloc->pid, vma->vm_start, vma->vm_end,
Signature · 1 line
2.38.1.584.g0f3c55d4c2-goog
Greg Kroah-Hartman
maintainerQuoted 28 lines · Carlos Llamas
commit
3ce00bb7e91cf57d723905371507af57182c37efupstream.Since commit
1da52815d5f1("binder: fixalloc->vma_vm_mmnull-ptr dereference") binder caches a pointer to thecurrent->mmduringopen(). This fixes a null-ptr dereference reported by syzkaller. Unfortunately, it also opens the door for a process to update its mm after theopen(), (e.g. via execve) making the cachedalloc->mmpointer invalid.Things get worse when the process continues to
mmap()a vma. From this point forward, binder will attempt to find this vma using an obsoletealloc->mmreference. Such as inbinder_update_page_range(), where the wrong vma is obtained viavma_lookup(), yet binder proceeds to happily insert new pages into it.To avoid this issue fail the ->mmap() callback if we detect a mismatch between the
vma->vm_mmand the originalalloc->mmpointer. This preventsalloc->vm_addrfrom getting set, so that any subsequentvma_lookup()calls fail as expected.Fixes 1da52815d5f1 ("binder: fix alloc->vma_vm_mm null-ptr dereference")Reported-by Jann Horn <[email protected]>Cc <[email protected]> # 5.15+Signed-off-by Carlos Llamas <[email protected]>Acked-by Todd Kjos <[email protected]>Signed-off-by Greg Kroah-Hartman <[email protected]>[cmllamas: renamed alloc->mm since missing e66b77e50522]Signed-off-by Carlos Llamas <[email protected]>
Quoted 3 lines
--- drivers/android/binder_alloc.c | 7 +++++++ 1 file changed, 7 insertions(+)
This is already in the 6.0 queue, is this a different version?
thanks,
greg k-h
Quoted 33 lines · Greg Kroah-Hartman
Quoted 28 lines · Carlos Llamas
commit
3ce00bb7e91cf57d723905371507af57182c37efupstream.Since commit
1da52815d5f1("binder: fixalloc->vma_vm_mmnull-ptr dereference") binder caches a pointer to thecurrent->mmduringopen(). This fixes a null-ptr dereference reported by syzkaller. Unfortunately, it also opens the door for a process to update its mm after theopen(), (e.g. via execve) making the cachedalloc->mmpointer invalid.Things get worse when the process continues to
mmap()a vma. From this point forward, binder will attempt to find this vma using an obsoletealloc->mmreference. Such as inbinder_update_page_range(), where the wrong vma is obtained viavma_lookup(), yet binder proceeds to happily insert new pages into it.To avoid this issue fail the ->mmap() callback if we detect a mismatch between the
vma->vm_mmand the originalalloc->mmpointer. This preventsalloc->vm_addrfrom getting set, so that any subsequentvma_lookup()calls fail as expected.Fixes 1da52815d5f1 ("binder: fix alloc->vma_vm_mm null-ptr dereference")Reported-by Jann Horn <[email protected]>Cc <[email protected]> # 5.15+Signed-off-by Carlos Llamas <[email protected]>Acked-by Todd Kjos <[email protected]>Signed-off-by Greg Kroah-Hartman <[email protected]>[cmllamas: renamed alloc->mm since missing e66b77e50522]Signed-off-by Carlos Llamas <[email protected]>Quoted 3 lines
--- drivers/android/binder_alloc.c | 7 +++++++ 1 file changed, 7 insertions(+)
Quoted 5 lines
This is already in the 6.0 queue, is this a different version?
thanks,
greg k-h
Oh, please ignore this. It seems Sasha backported the change correctly.