From 436ad0903096c8dc06abbf6d8f0aef758d7af22c Mon Sep 17 00:00:00 2001 From: markj Date: Fri, 18 Oct 2019 14:05:13 +0000 Subject: [PATCH] Tighten mapping protections on preloaded files on amd64. - We load the kernel at 0x200000. Memory below that address need not be executable, so do not map it as such. - Remove references to .ldata and related sections in the kernel linker script. They come from ld.bfd's default linker script, but are not used, and we now use ld.lld to link the amd64 kernel. lld does not contain a default linker script. - Pad the .bss to a 2MB as we do between .text and .data. This forces the loader to load additional files starting in the following 2MB page, preserving the use of superpage mappings for kernel data. - Map memory above the kernel image with NX. The kernel linker now upgrades protections as needed, and other preloaded file types (e.g., entropy, microcode) need not be mapped with execute permissions in the first place. Reviewed by: kib MFC after: 1 month Sponsored by: Netflix Differential Revision: https://reviews.freebsd.org/D21859 --- sys/amd64/amd64/pmap.c | 17 +++++++++-------- sys/conf/ldscript.amd64 | 32 +++----------------------------- 2 files changed, 12 insertions(+), 37 deletions(-) diff --git a/sys/amd64/amd64/pmap.c b/sys/amd64/amd64/pmap.c index 938d92b0a988..de8a47417a1a 100644 --- a/sys/amd64/amd64/pmap.c +++ b/sys/amd64/amd64/pmap.c @@ -1422,22 +1422,22 @@ nkpt_init(vm_paddr_t addr) * * This function operates on 2M pages, since we map the kernel space that * way. - * - * Note that this doesn't currently provide any protection for modules. */ static inline pt_entry_t bootaddr_rwx(vm_paddr_t pa) { /* - * Everything in the same 2M page as the start of the kernel - * should be static. On the other hand, things in the same 2M - * page as the end of the kernel could be read-write/executable, - * as the kernel image is not guaranteed to end on a 2M boundary. + * The kernel is loaded at a 2MB-aligned address, and memory below that + * need not be executable. The .bss section is padded to a 2MB + * boundary, so memory following the kernel need not be executable + * either. Preloaded kernel modules have their mapping permissions + * fixed up by the linker. */ if (pa < trunc_2mpage(btext - KERNBASE) || - pa >= trunc_2mpage(_end - KERNBASE)) - return (X86_PG_RW); + pa >= trunc_2mpage(_end - KERNBASE)) + return (X86_PG_RW | pg_nx); + /* * The linker should ensure that the read-only and read-write * portions don't share the same 2M page, so this shouldn't @@ -1446,6 +1446,7 @@ bootaddr_rwx(vm_paddr_t pa) */ if (pa >= trunc_2mpage(brwsection - KERNBASE)) return (X86_PG_RW | pg_nx); + /* * Mark any 2M page containing kernel text as read-only. Mark * other pages with read-only data as read-only and not executable. diff --git a/sys/conf/ldscript.amd64 b/sys/conf/ldscript.amd64 index 91088722b2e4..d62b531454ec 100644 --- a/sys/conf/ldscript.amd64 +++ b/sys/conf/ldscript.amd64 @@ -45,12 +45,6 @@ SECTIONS .rela.got : { *(.rela.got) } .rel.bss : { *(.rel.bss .rel.bss.* .rel.gnu.linkonce.b.*) } .rela.bss : { *(.rela.bss .rela.bss.* .rela.gnu.linkonce.b.*) } - .rel.ldata : { *(.rel.ldata .rel.ldata.* .rel.gnu.linkonce.l.*) } - .rela.ldata : { *(.rela.ldata .rela.ldata.* .rela.gnu.linkonce.l.*) } - .rel.lbss : { *(.rel.lbss .rel.lbss.* .rel.gnu.linkonce.lb.*) } - .rela.lbss : { *(.rela.lbss .rela.lbss.* .rela.gnu.linkonce.lb.*) } - .rel.lrodata : { *(.rel.lrodata .rel.lrodata.* .rel.gnu.linkonce.lr.*) } - .rela.lrodata : { *(.rela.lrodata .rela.lrodata.* .rela.gnu.linkonce.lr.*) } .rel.plt : { *(.rel.plt) } .rela.plt : { *(.rela.plt) } .init : @@ -179,30 +173,10 @@ SECTIONS *(.dynbss) *(.bss .bss.* .gnu.linkonce.b.*) *(COMMON) - /* Align here to ensure that the .bss section occupies space up to - _end. Align after .bss to ensure correct alignment even if the - .bss section disappears because there are no input sections. - FIXME: Why do we need it? When there is no .bss section, we don't - pad the .data section. */ - . = ALIGN(. != 0 ? 64 / 8 : 1); + /* Ensure that the .bss section ends at a superpage boundary. + This way it can be mapped using non-executable large pages. */ + . = ALIGN(0x200000); } - .lbss : - { - *(.dynlbss) - *(.lbss .lbss.* .gnu.linkonce.lb.*) - *(LARGE_COMMON) - } - . = ALIGN(64 / 8); - .lrodata ALIGN(CONSTANT (MAXPAGESIZE)) + (. & (CONSTANT (MAXPAGESIZE) - 1)) : - { - *(.lrodata .lrodata.* .gnu.linkonce.lr.*) - } - .ldata ALIGN(CONSTANT (MAXPAGESIZE)) + (. & (CONSTANT (MAXPAGESIZE) - 1)) : - { - *(.ldata .ldata.* .gnu.linkonce.l.*) - . = ALIGN(. != 0 ? 64 / 8 : 1); - } - . = ALIGN(64 / 8); _end = .; PROVIDE (end = .); . = DATA_SEGMENT_END (.); /* Stabs debugging sections. */