Documentation/mm/vmalloced-kernel-stacks.rst

1 .. SPDX-License-Identifier: GPL-2.0
12 --------
19 ------------
23 it difficult to isolate and root-cause.
25 Virtually-mapped kernel stacks with guard pages causes kernel stack
30 support for virtually mapped stacks with guard pages. This feature
40 --------------------
45 - vmalloc space must be large enough to hold many kernel stacks. This
46   may rule out many 32-bit architectures.
47 - Stacks in vmalloc space need to work reliably.  For example, if
53 - If the stack overflows into a guard page, something reasonable
58 ----------
63 - Enable this if you want the use virtually-mapped kernel stacks
64   with guard pages. This causes kernel stack overflows to be caught
65   immediately rather than causing difficult-to-diagnose corruption.
84 -----------
87 virtually contiguous memory pages from the page level allocator. These
88 pages are mapped into contiguous kernel virtual space with PAGE_KERNEL
94 - Allocated stacks are cached and later reused by new threads, so memcg
97 - vm_struct is cached to be able to find when thread free is initiated
100 - On arm64, all VMAP's stacks need to have the same alignment to ensure
103 - This does not address interrupt stacks - according to the original patch
113 stack and a non-null stack_vm_area pointer serves as a indication that the
121 -----------------------
123 Leading and trailing guard pages help detect stack overflows. When stack
124 overflows into the guard pages, handlers have to be careful not overflow
129 stack overflow on the double-fault stack.
131 Testing VMAP allocation with guard pages
132 ----------------------------------------
135 and trailing guard page? The following lkdtm tests can help detect any
144 -----------
146 - A percpu cache of vmalloced stacks appears to be a bit faster than a
147   high-order stack allocation, at least when the cache hits.
148 - THREAD_INFO_IN_TASK gets rid of arch-specific thread_info entirely and
151 - The thread stack can be free'ed as soon as the task is dead (without
152   waiting for RCU) and then, if vmapped stacks are in use, cache the