Message-Id: <20170819002809.111312-13-ricardo.neri-calderon@linux.intel.com>
Date:   Fri, 18 Aug 2017 17:27:53 -0700
From:   Ricardo Neri <ricardo.neri-calderon@...ux.intel.com>
To:     Ingo Molnar <mingo@...hat.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        "H. Peter Anvin" <hpa@...or.com>,
        Andy Lutomirski <luto@...nel.org>, Borislav Petkov <bp@...e.de>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Brian Gerst <brgerst@...il.com>,
        Chris Metcalf <cmetcalf@...lanox.com>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Liang Z Li <liang.z.li@...el.com>,
        Masami Hiramatsu <mhiramat@...nel.org>,
        Huang Rui <ray.huang@....com>, Jiri Slaby <jslaby@...e.cz>,
        Jonathan Corbet <corbet@....net>,
        "Michael S. Tsirkin" <mst@...hat.com>,
        Paul Gortmaker <paul.gortmaker@...driver.com>,
        Vlastimil Babka <vbabka@...e.cz>,
        Chen Yucong <slaoub@...il.com>,
        "Ravi V. Shankar" <ravi.v.shankar@...el.com>,
        Shuah Khan <shuah@...nel.org>, linux-kernel@...r.kernel.org,
        x86@...nel.org, ricardo.neri@...el.com,
        Ricardo Neri <ricardo.neri-calderon@...ux.intel.com>,
        Adam Buchbinder <adam.buchbinder@...il.com>,
        Colin Ian King <colin.king@...onical.com>,
        Lorenzo Stoakes <lstoakes@...il.com>,
        Qiaowei Ren <qiaowei.ren@...el.com>,
        Arnaldo Carvalho de Melo <acme@...hat.com>,
        Adrian Hunter <adrian.hunter@...el.com>,
        Kees Cook <keescook@...omium.org>,
        Thomas Garnier <thgarnie@...gle.com>,
        Dmitry Vyukov <dvyukov@...gle.com>
Subject: [PATCH v8 12/28] x86/insn-eval: Add utility functions to get segment selector

When computing a linear address and segmentation is used, we need to know
the base address of the segment involved in the computation. In most of
the cases, the segment base address will be zero as in USER_DS/USER32_DS.
However, it may be possible that a user space program defines its own
segments via a local descriptor table. In such a case, the segment base
address may not be zero .Thus, the segment base address is needed to
calculate correctly the linear address.

If running in protected mode, the segment selector to be used when
computing a linear address is determined by either any of segment override
prefixes in the instruction or inferred from the registers involved in the
computation of the effective address; in that order. Also, there are cases
when the segment override prefixes shall be ignored (i.e., code segments
are always selected by the CS segment register; string instructions always
use the ES segment register when using (E)DI register as operand). In long
mode, segment registers are ignored, except for FS and GS. In these two
cases, base addresses are obtained from the respective MSRs.

For clarity, this process can be split into three steps (and an equal
number of functions): parse the segment override prefixes, if any; resolve
the relevant segment register to use, and, once known, read its value to
obtain the segment selector.

The method to obtain the segment selector depends on several factors. In
32-bit builds, segment selectors are saved into a pt_regs structure
when switching to kernel mode. The same is also true for virtual-8086
mode. In 64-bit builds, segmentation is mostly ignored, except when
running a program in 32-bit legacy mode. In this case, CS and SS can be
obtained from pt_regs. DS, ES, FS and GS can be read directly from
the respective segment registers.

In order to identify the segment registers, a new set of #defines is
introduced. It also includes two special identifiers. One of them
indicates when the default segment register associated with instruction
operands shall be used. Another one indicates that the contents of the
segment register shall be ignored; this identifier is used when in long
mode.

Cc: Dave Hansen <dave.hansen@...ux.intel.com>
Cc: Adam Buchbinder <adam.buchbinder@...il.com>
Cc: Colin Ian King <colin.king@...onical.com>
Cc: Lorenzo Stoakes <lstoakes@...il.com>
Cc: Qiaowei Ren <qiaowei.ren@...el.com>
Cc: Arnaldo Carvalho de Melo <acme@...hat.com>
Cc: Masami Hiramatsu <mhiramat@...nel.org>
Cc: Adrian Hunter <adrian.hunter@...el.com>
Cc: Kees Cook <keescook@...omium.org>
Cc: Thomas Garnier <thgarnie@...gle.com>
Cc: Peter Zijlstra <peterz@...radead.org>
Cc: Borislav Petkov <bp@...e.de>
Cc: Dmitry Vyukov <dvyukov@...gle.com>
Cc: Ravi V. Shankar <ravi.v.shankar@...el.com>
Cc: x86@...nel.org
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@...ux.intel.com>
---
 arch/x86/include/asm/inat.h |  10 ++
 arch/x86/lib/insn-eval.c    | 278 ++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 288 insertions(+)

diff --git a/arch/x86/include/asm/inat.h b/arch/x86/include/asm/inat.h
index 02aff0867211..1c78580e58be 100644
--- a/arch/x86/include/asm/inat.h
+++ b/arch/x86/include/asm/inat.h
@@ -97,6 +97,16 @@
 #define INAT_MAKE_GROUP(grp)	((grp << INAT_GRP_OFFS) | INAT_MODRM)
 #define INAT_MAKE_IMM(imm)	(imm << INAT_IMM_OFFS)
 
+/* Identifiers for segment registers */
+#define INAT_SEG_REG_IGNORE	0
+#define INAT_SEG_REG_DEFAULT	1
+#define INAT_SEG_REG_CS		2
+#define INAT_SEG_REG_SS		3
+#define INAT_SEG_REG_DS		4
+#define INAT_SEG_REG_ES		5
+#define INAT_SEG_REG_FS		6
+#define INAT_SEG_REG_GS		7
+
 /* Attribute search APIs */
 extern insn_attr_t inat_get_opcode_attribute(insn_byte_t opcode);
 extern int inat_get_last_prefix_id(insn_byte_t last_pfx);
diff --git a/arch/x86/lib/insn-eval.c b/arch/x86/lib/insn-eval.c
index 25b2eb3c64c1..86f58ce6c302 100644
--- a/arch/x86/lib/insn-eval.c
+++ b/arch/x86/lib/insn-eval.c
@@ -9,6 +9,7 @@
 #include <asm/inat.h>
 #include <asm/insn.h>
 #include <asm/insn-eval.h>
+#include <asm/vm86.h>
 
 enum reg_type {
 	REG_TYPE_RM = 0,
@@ -42,6 +43,283 @@ static bool is_string_insn(struct insn *insn)
 	}
 }
 
+/**
+ * get_overridden_seg_reg() - obtain segment register to use from prefixes
+ * @insn:	Instruction structure with segment override prefixes
+ * @regs:	Structure with register values as seen when entering kernel mode
+ * @regoff:	Operand offset, in pt_regs, used to deterimine segment register
+ *
+ * The segment register to which an effective address refers depends on
+ * a) whether running in long mode (in such a case semgment override prefixes
+ * are ignored. b) Whether segment override prefixes must be ignored for certain
+ * registers: always use CS when the register is (R|E)IP; always use ES when
+ * operand register is (E)DI with a string instruction as defined in the Intel
+ * documentation. c) If segment overrides prefixes are found in the instruction
+ * prefixes. d) Use the default segment register associated with the operand
+ * register.
+ *
+ * This function returns the overridden segment register to use, if any, as per
+ * the conditions described above. Please note that this function
+ * does not return the value in the segment register (i.e., the segment
+ * selector). The segment selector needs to be obtained using
+ * get_segment_selector() and passing the segment register resolved by
+ * this function.
+ *
+ * Return: A constant identifying the segment register to use, among CS, SS, DS,
+ * ES, FS, or GS. INAT_SEG_REG_IGNORE is returned if running in long mode.
+ * INAT_SEG_REG_DEFAULT is returned if no segment override prefixes were found
+ * and the default segment register shall be used. -EINVAL in case of error.
+ */
+static int get_overridden_seg_reg(struct insn *insn, struct pt_regs *regs,
+				  int regoff)
+{
+	int i;
+	int sel_overrides = 0;
+	int seg_register = INAT_SEG_REG_DEFAULT;
+
+	/*
+	 * Segment override prefixes should not be used for (E)IP. Check this
+	 * case first as we might not have (and not needed at all) a
+	 * valid insn structure to evaluate segment override prefixes.
+	 */
+	if (regoff == offsetof(struct pt_regs, ip)) {
+		if (user_64bit_mode(regs))
+			return INAT_SEG_REG_IGNORE;
+		else
+			return INAT_SEG_REG_DEFAULT;
+	}
+
+	if (!insn)
+		return -EINVAL;
+
+	insn_get_prefixes(insn);
+
+	/* Look for any segment override prefixes. */
+	for (i = 0; i < insn->prefixes.nbytes; i++) {
+		insn_attr_t attr;
+
+		attr = inat_get_opcode_attribute(insn->prefixes.bytes[i]);
+		switch (attr) {
+		case INAT_MAKE_PREFIX(INAT_PFX_CS):
+			seg_register = INAT_SEG_REG_CS;
+			sel_overrides++;
+			break;
+		case INAT_MAKE_PREFIX(INAT_PFX_SS):
+			seg_register = INAT_SEG_REG_SS;
+			sel_overrides++;
+			break;
+		case INAT_MAKE_PREFIX(INAT_PFX_DS):
+			seg_register = INAT_SEG_REG_DS;
+			sel_overrides++;
+			break;
+		case INAT_MAKE_PREFIX(INAT_PFX_ES):
+			seg_register = INAT_SEG_REG_ES;
+			sel_overrides++;
+			break;
+		case INAT_MAKE_PREFIX(INAT_PFX_FS):
+			seg_register = INAT_SEG_REG_FS;
+			sel_overrides++;
+			break;
+		case INAT_MAKE_PREFIX(INAT_PFX_GS):
+			seg_register = INAT_SEG_REG_GS;
+			sel_overrides++;
+			break;
+		/* No default action needed. */
+		}
+	}
+
+	/*
+	 * In long mode, segment override prefixes are ignored, except for
+	 * overrides for FS and GS.
+	 */
+	if (user_64bit_mode(regs)) {
+		if (seg_register != INAT_SEG_REG_FS &&
+		    seg_register != INAT_SEG_REG_GS)
+			return INAT_SEG_REG_IGNORE;
+	/* More than one segment override prefix leads to undefined behavior. */
+	} else if (sel_overrides > 1) {
+		return -EINVAL;
+	/*
+	 * Segment override prefixes are always ignored for string instructions
+	 * that involve the use the (E)DI register.
+	 */
+	} else if ((regoff == offsetof(struct pt_regs, di)) &&
+		   is_string_insn(insn)) {
+		return INAT_SEG_REG_DEFAULT;
+	}
+
+	return seg_register;
+}
+
+/**
+ * resolve_seg_register() - obtain segment register
+ * @insn:	Instruction structure with segment override prefixes
+ * @regs:	Structure with register values as seen when entering kernel mode
+ * @regoff:	Operand offset, in pt_regs, used to deterimine segment register
+ *
+ * Determine the segment register associated with the operands and, if
+ * applicable, prefixes and the instruction pointed by insn. The function first
+ * checks if the segment register shall be ignored or has been overridden in the
+ * instruction prefixes. Otherwise, it resolves the segment register to use
+ * based on the defaults described in the Intel documentation.
+ *
+ * The operand register, regoff, is represented as the offset from the base of
+ * pt_regs. Also, regoff can be -EDOM for cases in which registers are not
+ * used as operands (e.g., displacement-only memory addressing).
+ *
+ * Return: A constant identifying the segment register to use, among CS, SS, DS,
+ * ES, FS, or GS. INAT_SEG_REG_IGNORE is returned if running in long mode.
+ * -EINVAL in case of error.
+ */
+static int resolve_seg_register(struct insn *insn, struct pt_regs *regs,
+				int regoff)
+{
+	int seg_reg;
+
+	seg_reg = get_overridden_seg_reg(insn, regs, regoff);
+
+	if (seg_reg < 0)
+		return seg_reg;
+
+	if (seg_reg == INAT_SEG_REG_IGNORE)
+		return seg_reg;
+
+	if (seg_reg != INAT_SEG_REG_DEFAULT)
+		return seg_reg;
+
+	/*
+	 * If we are here, we use the default segment register as described
+	 * in the Intel documentation:
+	 *  + DS for all references involving (E)AX, (E)CX, (E)DX, (E)BX, and
+	 * (E)SI.
+	 *  + If used in a string instruction, ES for (E)DI. Otherwise, DS.
+	 *  + AX, CX and DX are not valid register operands in 16-bit address
+	 *    encodings but are valid for 32-bit and 64-bit encodings.
+	 *  + -EDOM is reserved to identify for cases in which no register
+	 *    is used (i.e., displacement-only addressing). Use DS.
+	 *  + SS for (E)SP or (E)BP.
+	 *  + CS for (E)IP.
+	 */
+
+	switch (regoff) {
+	case offsetof(struct pt_regs, ax):
+	case offsetof(struct pt_regs, cx):
+	case offsetof(struct pt_regs, dx):
+		/* Need insn to verify address size. */
+		if (!insn || insn->addr_bytes == 2)
+			return -EINVAL;
+	case -EDOM:
+	case offsetof(struct pt_regs, bx):
+	case offsetof(struct pt_regs, si):
+		return INAT_SEG_REG_DS;
+	case offsetof(struct pt_regs, di):
+		/* Need insn to see if insn is string instruction. */
+		if (!insn)
+			return -EINVAL;
+		if (is_string_insn(insn))
+			return INAT_SEG_REG_ES;
+		return INAT_SEG_REG_DS;
+	case offsetof(struct pt_regs, bp):
+	case offsetof(struct pt_regs, sp):
+		return INAT_SEG_REG_SS;
+	case offsetof(struct pt_regs, ip):
+		return INAT_SEG_REG_CS;
+	default:
+		return -EINVAL;
+	}
+}
+
+/**
+ * get_segment_selector() - obtain segment selector
+ * @regs:	Structure with register values as seen when entering kernel mode
+ * @seg_reg:	Segment register to use
+ *
+ * Obtain the segment selector from any of the CS, SS, DS, ES, FS, GS segment
+ * registers. In CONFIG_X86_32, the segment is obtained from either pt_regs or
+ * kernel_vm86_regs as applicable. In CONFIG_X86_64, CS and SS are obtained
+ * from pt_regs. DS, ES, FS and GS are obtained by reading the actual CPU
+ * registers. This done for only for completeness as in CONFIG_X86_64 segment
+ * registers are ignored.
+ *
+ * Return: Value of the segment selector, including null when running in
+ * long mode. -1 on error.
+ */
+static short get_segment_selector(struct pt_regs *regs, int seg_reg)
+{
+#ifdef CONFIG_X86_64
+	unsigned short sel;
+
+	switch (seg_reg) {
+	case INAT_SEG_REG_IGNORE:
+		return 0;
+	case INAT_SEG_REG_CS:
+		return (unsigned short)(regs->cs & 0xffff);
+	case INAT_SEG_REG_SS:
+		return (unsigned short)(regs->ss & 0xffff);
+	case INAT_SEG_REG_DS:
+		savesegment(ds, sel);
+		return sel;
+	case INAT_SEG_REG_ES:
+		savesegment(es, sel);
+		return sel;
+	case INAT_SEG_REG_FS:
+		savesegment(fs, sel);
+		return sel;
+	case INAT_SEG_REG_GS:
+		savesegment(gs, sel);
+		return sel;
+	default:
+		return -EINVAL;
+	}
+#else /* CONFIG_X86_32 */
+	struct kernel_vm86_regs *vm86regs = (struct kernel_vm86_regs *)regs;
+
+	if (v8086_mode(regs)) {
+		switch (seg_reg) {
+		case INAT_SEG_REG_CS:
+			return (unsigned short)(regs->cs & 0xffff);
+		case INAT_SEG_REG_SS:
+			return (unsigned short)(regs->ss & 0xffff);
+		case INAT_SEG_REG_DS:
+			return vm86regs->ds;
+		case INAT_SEG_REG_ES:
+			return vm86regs->es;
+		case INAT_SEG_REG_FS:
+			return vm86regs->fs;
+		case INAT_SEG_REG_GS:
+			return vm86regs->gs;
+		case INAT_SEG_REG_IGNORE:
+			/* fall through */
+		default:
+			return -EINVAL;
+		}
+	}
+
+	switch (seg_reg) {
+	case INAT_SEG_REG_CS:
+		return (unsigned short)(regs->cs & 0xffff);
+	case INAT_SEG_REG_SS:
+		return (unsigned short)(regs->ss & 0xffff);
+	case INAT_SEG_REG_DS:
+		return (unsigned short)(regs->ds & 0xffff);
+	case INAT_SEG_REG_ES:
+		return (unsigned short)(regs->es & 0xffff);
+	case INAT_SEG_REG_FS:
+		return (unsigned short)(regs->fs & 0xffff);
+	case INAT_SEG_REG_GS:
+		/*
+		 * GS may or may not be in regs as per CONFIG_X86_32_LAZY_GS.
+		 * The macro below takes care of both cases.
+		 */
+		return get_user_gs(regs);
+	case INAT_SEG_REG_IGNORE:
+		/* fall through */
+	default:
+		return -EINVAL;
+	}
+#endif /* CONFIG_X86_64 */
+}
+
 static int get_reg_offset(struct insn *insn, struct pt_regs *regs,
 			  enum reg_type type)
 {
-- 
2.13.0

Powered by blists - more mailing lists