s390: Add stackprotector support

Stackprotector support was previously unavailable on s390 because by
default compilers generate code which is not suitable for the kernel:
the canary value is accessed via thread local storage, where the address
of thread local storage is within access registers 0 and 1.

Using those registers also for the kernel would come with a significant
performance impact and more complicated kernel entry/exit code, since
access registers contents would have to be exchanged on every kernel entry
and exit.

With the upcoming gcc 16 release new compiler options will become available
which allow to generate code suitable for the kernel. [1]

Compiler option -mstack-protector-guard=global instructs gcc to generate
stackprotector code that refers to a global stackprotector canary value via
symbol __stack_chk_guard. Access to this value is guaranteed to occur via
larl and lgrl instructions.

Furthermore, compiler option -mstack-protector-guard-record generates a
section containing all code addresses that reference the canary value.

To allow for per task canary values the instructions which load the address
of __stack_chk_guard are patched so they access a lowcore field instead: a
per task canary value is available within the task_struct of each task, and
is written to the per-cpu lowcore location on each context switch.

Also add sanity checks and debugging option to be consistent with other
kernel code patching mechanisms.

Full debugging output can be enabled with the following kernel command line
options:

debug_stackprotector
bootdebug
ignore_loglevel
earlyprintk
dyndbg="file stackprotector.c +p"

Example debug output:

stackprot: 0000021e402d4eda: c010005a9ae3 -> c01f00070240

where "<insn address>: <old insn> -> <new insn>".

[1] gcc commit 0cd1f03939d5 ("s390: Support global stack protector")

Reviewed-by: Sven Schnelle <svens@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>

1 files changed, 156 insertions, 0 deletions

diff --git a/arch/s390/kernel/stackprotector.c b/arch/s390/kernel/stackprotector.c
new file mode 100644
index 000000000000..d4e40483f008
--- /dev/null
+++ b/arch/s390/kernel/stackprotector.c
@@ -0,0 +1,156 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#ifndef pr_fmt
+#define pr_fmt(fmt)	"stackprot: " fmt
+#endif
+
+#include <linux/export.h>
+#include <linux/uaccess.h>
+#include <linux/printk.h>
+#include <asm/abs_lowcore.h>
+#include <asm/sections.h>
+#include <asm/machine.h>
+#include <asm/asm-offsets.h>
+#include <asm/arch-stackprotector.h>
+
+#ifdef __DECOMPRESSOR
+
+#define DEBUGP		boot_debug
+#define EMERGP		boot_emerg
+#define PANIC		boot_panic
+
+#else /* __DECOMPRESSOR */
+
+#define DEBUGP		pr_debug
+#define EMERGP		pr_emerg
+#define PANIC		panic
+
+#endif /* __DECOMPRESSOR */
+
+int __bootdata_preserved(stack_protector_debug);
+
+unsigned long __stack_chk_guard;
+EXPORT_SYMBOL(__stack_chk_guard);
+
+struct insn_ril {
+	u8 opc1 : 8;
+	u8 r1	: 4;
+	u8 opc2 : 4;
+	u32 imm;
+} __packed;
+
+/*
+ * Convert a virtual instruction address to a real instruction address. The
+ * decompressor needs to patch instructions within the kernel image based on
+ * their virtual addresses, while dynamic address translation is still
+ * disabled. Therefore a translation from virtual kernel image addresses to
+ * the corresponding physical addresses is required.
+ *
+ * After dynamic address translation is enabled and when the kernel needs to
+ * patch instructions such a translation is not required since the addresses
+ * are identical.
+ */
+static struct insn_ril *vaddress_to_insn(unsigned long vaddress)
+{
+#ifdef __DECOMPRESSOR
+	return (struct insn_ril *)__kernel_pa(vaddress);
+#else
+	return (struct insn_ril *)vaddress;
+#endif
+}
+
+static unsigned long insn_to_vaddress(struct insn_ril *insn)
+{
+#ifdef __DECOMPRESSOR
+	return (unsigned long)__kernel_va(insn);
+#else
+	return (unsigned long)insn;
+#endif
+}
+
+#define INSN_RIL_STRING_SIZE (sizeof(struct insn_ril) * 2 + 1)
+
+static void insn_ril_to_string(char *str, struct insn_ril *insn)
+{
+	u8 *ptr = (u8 *)insn;
+	int i;
+
+	for (i = 0; i < sizeof(*insn); i++)
+		hex_byte_pack(&str[2 * i], ptr[i]);
+	str[2 * i] = 0;
+}
+
+static void stack_protector_dump(struct insn_ril *old, struct insn_ril *new)
+{
+	char ostr[INSN_RIL_STRING_SIZE];
+	char nstr[INSN_RIL_STRING_SIZE];
+
+	insn_ril_to_string(ostr, old);
+	insn_ril_to_string(nstr, new);
+	DEBUGP("%016lx: %s -> %s\n", insn_to_vaddress(old), ostr, nstr);
+}
+
+static int stack_protector_verify(struct insn_ril *insn, unsigned long kernel_start)
+{
+	char istr[INSN_RIL_STRING_SIZE];
+	unsigned long vaddress, offset;
+
+	/* larl */
+	if (insn->opc1 == 0xc0 && insn->opc2 == 0x0)
+		return 0;
+	/* lgrl */
+	if (insn->opc1 == 0xc4 && insn->opc2 == 0x8)
+		return 0;
+	insn_ril_to_string(istr, insn);
+	vaddress = insn_to_vaddress(insn);
+	if (__is_defined(__DECOMPRESSOR)) {
+		offset = (unsigned long)insn - kernel_start + TEXT_OFFSET;
+		EMERGP("Unexpected instruction at %016lx/%016lx: %s\n", vaddress, offset, istr);
+		PANIC("Stackprotector error\n");
+	} else {
+		EMERGP("Unexpected instruction at %016lx: %s\n", vaddress, istr);
+	}
+	return -EINVAL;
+}
+
+int __stack_protector_apply(unsigned long *start, unsigned long *end, unsigned long kernel_start)
+{
+	unsigned long canary, *loc;
+	struct insn_ril *insn, new;
+	int rc;
+
+	/*
+	 * Convert LARL/LGRL instructions to LLILF so register R1 contains the
+	 * address of the per-cpu / per-process stack canary:
+	 *
+	 * LARL/LGRL R1,__stack_chk_guard => LLILF R1,__lc_stack_canary
+	 */
+	canary = __LC_STACK_CANARY;
+	if (machine_has_relocated_lowcore())
+		canary += LOWCORE_ALT_ADDRESS;
+	for (loc = start; loc < end; loc++) {
+		insn = vaddress_to_insn(*loc);
+		rc = stack_protector_verify(insn, kernel_start);
+		if (rc)
+			return rc;
+		new = *insn;
+		new.opc1 = 0xc0;
+		new.opc2 = 0xf;
+		new.imm = canary;
+		if (stack_protector_debug)
+			stack_protector_dump(insn, &new);
+		s390_kernel_write(insn, &new, sizeof(*insn));
+	}
+	return 0;
+}
+
+#ifdef __DECOMPRESSOR
+void __stack_protector_apply_early(unsigned long kernel_start)
+{
+	unsigned long *start, *end;
+
+	start = (unsigned long *)vmlinux.stack_prot_start;
+	end = (unsigned long *)vmlinux.stack_prot_end;
+	__stack_protector_apply(start, end, kernel_start);
+}
+#endif