x86/fault: #PF improvements, mostly related to USER bit

index 71d4b9d4d43f..91d4d2722f2e 100644
@@ -1344,13 +1344,10 @@ void do_user_addr_fault(struct pt_regs *regs,
 	 * Only do the expensive exception table search when we might be at
 	 * risk of a deadlock.  This happens if we
 	 * 1. Failed to acquire mmap_sem, and
-	 * 2. The access did not originate in userspace.  Note: either the
-	 *    hardware or earlier page fault code may set X86_PF_USER
-	 *    in sw_error_code.
+	 * 2. The access did not originate in userspace.
 	 */
 	if (unlikely(!down_read_trylock(&mm->mmap_sem))) {
-		if (!(sw_error_code & X86_PF_USER) &&
-		    !search_exception_tables(regs->ip)) {
+		if (!user_mode(regs) && !search_exception_tables(regs->ip)) {
 			/*
 			 * Fault from code in kernel from
 			 * which we do not expect faults.

index 91d4d2722f2e..eae7ee3ce89b 100644
@@ -1377,7 +1377,7 @@ void do_user_addr_fault(struct pt_regs *regs,
 		bad_area(regs, sw_error_code, address);
 		return;
 	}
-	if (sw_error_code & X86_PF_USER) {
+	if (user_mode(regs)) {
 		/*
 		 * Accessing the stack below %sp is always a bug.
 		 * The large cushion allows instructions like enter

index 33833d1909af..a5ea841cc6d2 100644
@@ -16,6 +16,12 @@
 # define DISABLE_MPX	(1<<(X86_FEATURE_MPX & 31))
 #endif
 
+#ifdef CONFIG_X86_SMAP
+# define DISABLE_SMAP	0
+#else
+# define DISABLE_SMAP	(1<<(X86_FEATURE_SMAP & 31))
+#endif
+
 #ifdef CONFIG_X86_INTEL_UMIP
 # define DISABLE_UMIP	0
 #else

@@ -68,7 +74,7 @@
 #define DISABLED_MASK6	0
 #define DISABLED_MASK7	(DISABLE_PTI)
 #define DISABLED_MASK8	0
-#define DISABLED_MASK9	(DISABLE_MPX)
+#define DISABLED_MASK9	(DISABLE_MPX|DISABLE_SMAP)
 #define DISABLED_MASK10	0
 #define DISABLED_MASK11	0
 #define DISABLED_MASK12	0

index eae7ee3ce89b..0597342d4a55 100644
@@ -1150,10 +1150,7 @@ static int fault_in_kernel_space(unsigned long address)
 
 static inline bool smap_violation(int error_code, struct pt_regs *regs)
 {
-	if (!IS_ENABLED(CONFIG_X86_SMAP))
-		return false;
-
-	if (!static_cpu_has(X86_FEATURE_SMAP))
+	if (!cpu_feature_enabled(X86_FEATURE_SMAP))
 		return false;
 
 	if (error_code & X86_PF_USER)

index 0597342d4a55..c50cd67521b6 100644
@@ -1148,20 +1148,6 @@ static int fault_in_kernel_space(unsigned long address)
 	return address >= TASK_SIZE_MAX;
 }
 
-static inline bool smap_violation(int error_code, struct pt_regs *regs)
-{
-	if (!cpu_feature_enabled(X86_FEATURE_SMAP))
-		return false;
-
-	if (error_code & X86_PF_USER)
-		return false;
-
-	if (!user_mode(regs) && (regs->flags & X86_EFLAGS_AC))
-		return false;
-
-	return true;
-}
-
 /*
  * Called for all faults where 'address' is part of the kernel address
  * space.  Might get called for faults that originate from *code* that

@@ -1249,10 +1235,13 @@ void do_user_addr_fault(struct pt_regs *regs,
 		pgtable_bad(regs, hw_error_code, address);
 
 	/*
-	 * Check for invalid kernel (supervisor) access to user
-	 * pages in the user address space.
+	 * If SMAP is on, check for invalid kernel (supervisor)
+	 * access to user pages in the user address space.
 	 */
-	if (unlikely(smap_violation(hw_error_code, regs))) {
+	if (unlikely(cpu_feature_enabled(X86_FEATURE_SMAP) &&
+		     !(hw_error_code & X86_PF_USER) &&
+		     (user_mode(regs) || !(regs->flags & X86_EFLAGS_AC))))
+	{
 		bad_area_nosemaphore(regs, hw_error_code, address);
 		return;
 	}

index c50cd67521b6..95d94d48a10d 100644
@@ -1235,12 +1235,15 @@ void do_user_addr_fault(struct pt_regs *regs,
 		pgtable_bad(regs, hw_error_code, address);
 
 	/*
-	 * If SMAP is on, check for invalid kernel (supervisor)
-	 * access to user pages in the user address space.
+	 * If SMAP is on, check for invalid kernel (supervisor) access to user
+	 * pages in the user address space.  The odd case here is WRUSS,
+	 * which, according to the preliminary documentation, does not respect
+	 * SMAP and will have the USER bit set so, in all cases, SMAP
+	 * enforcement appears to be consistent with the USER bit.
 	 */
 	if (unlikely(cpu_feature_enabled(X86_FEATURE_SMAP) &&
 		     !(hw_error_code & X86_PF_USER) &&
-		     (user_mode(regs) || !(regs->flags & X86_EFLAGS_AC))))
+		     !(regs->flags & X86_EFLAGS_AC)))
 	{
 		bad_area_nosemaphore(regs, hw_error_code, address);
 		return;

index 95d94d48a10d..bd89b6f83aa2 100644
@@ -794,7 +794,7 @@ __bad_area_nosemaphore(struct pt_regs *regs, unsigned long error_code,
 	struct task_struct *tsk = current;
 
 	/* User mode accesses just cause a SIGSEGV */
-	if (error_code & X86_PF_USER) {
+	if (user_mode(regs) && (error_code & X86_PF_USER)) {
 		/*
 		 * It's possible to have interrupts off here:
 		 */

index bd89b6f83aa2..fac931d89c23 100644
@@ -631,6 +631,24 @@ pgtable_bad(struct pt_regs *regs, unsigned long error_code,
 	oops_end(flags, regs, sig);
 }
 
+static void set_signal_archinfo(unsigned long address,
+				unsigned long error_code)
+{
+	struct task_struct *tsk = current;
+
+	/*
+	 * To avoid leaking information about the kernel page
+	 * table layout, pretend that user-mode accesses to
+	 * kernel addresses are always protection faults.
+	 */
+	if (address >= TASK_SIZE_MAX)
+		error_code |= X86_PF_PROT;
+
+	tsk->thread.trap_nr = X86_TRAP_PF;
+	tsk->thread.error_code = error_code | X86_PF_USER;
+	tsk->thread.cr2 = address;
+}
+
 static noinline void
 no_context(struct pt_regs *regs, unsigned long error_code,
 	   unsigned long address, int signal, int si_code)

@@ -656,9 +674,7 @@ no_context(struct pt_regs *regs, unsigned long error_code,
 		 * faulting through the emulate_vsyscall() logic.
 		 */
 		if (current->thread.sig_on_uaccess_err && signal) {
-			tsk->thread.trap_nr = X86_TRAP_PF;
-			tsk->thread.error_code = error_code | X86_PF_USER;
-			tsk->thread.cr2 = address;
+			set_signal_archinfo(address, error_code);
 
 			/* XXX: hwpoison faults will set the wrong code. */
 			force_sig_fault(signal, si_code, (void __user *)address,

@@ -821,9 +837,7 @@ __bad_area_nosemaphore(struct pt_regs *regs, unsigned long error_code,
 		if (likely(show_unhandled_signals))
 			show_signal_msg(regs, error_code, address, tsk);
 
-		tsk->thread.cr2		= address;
-		tsk->thread.error_code	= error_code;
-		tsk->thread.trap_nr	= X86_TRAP_PF;
+		set_signal_archinfo(address, error_code);
 
 		if (si_code == SEGV_PKUERR)
 			force_sig_pkuerr((void __user *)address, pkey);

@@ -937,9 +951,7 @@ do_sigbus(struct pt_regs *regs, unsigned long error_code, unsigned long address,
 	if (is_prefetch(regs, error_code, address))
 		return;
 
-	tsk->thread.cr2		= address;
-	tsk->thread.error_code	= error_code;
-	tsk->thread.trap_nr	= X86_TRAP_PF;
+	set_signal_archinfo(address, error_code);
 
 #ifdef CONFIG_MEMORY_FAILURE
 	if (fault & (VM_FAULT_HWPOISON|VM_FAULT_HWPOISON_LARGE)) {

index fac931d89c23..a6554110ef62 100644
@@ -621,10 +621,6 @@ pgtable_bad(struct pt_regs *regs, unsigned long error_code,
 	       tsk->comm, address);
 	dump_pagetable(address);
 
-	tsk->thread.cr2		= address;
-	tsk->thread.trap_nr	= X86_TRAP_PF;
-	tsk->thread.error_code	= error_code;
-
 	if (__die("Bad pagetable", regs, error_code))
 		sig = 0;
 

@@ -753,10 +749,6 @@ no_context(struct pt_regs *regs, unsigned long error_code,
 	if (task_stack_end_corrupted(tsk))
 		printk(KERN_EMERG "Thread overran stack, or stack corrupted\n");
 
-	tsk->thread.cr2		= address;
-	tsk->thread.trap_nr	= X86_TRAP_PF;
-	tsk->thread.error_code	= error_code;
-
 	sig = SIGKILL;
 	if (__die("Oops", regs, error_code))
 		sig = 0;

index a6554110ef62..6e18438c367f 100644
@@ -1217,7 +1217,6 @@ void do_user_addr_fault(struct pt_regs *regs,
 			unsigned long hw_error_code,
 			unsigned long address)
 {
-	unsigned long sw_error_code;
 	struct vm_area_struct *vma;
 	struct task_struct *tsk;
 	struct mm_struct *mm;

@@ -1262,13 +1261,6 @@ void do_user_addr_fault(struct pt_regs *regs,
 		return;
 	}
 
-	/*
-	 * hw_error_code is literally the "page fault error code" passed to
-	 * the kernel directly from the hardware.  But, we will shortly be
-	 * modifying it in software, so give it a new name.
-	 */
-	sw_error_code = hw_error_code;
-
 	/*
 	 * It's safe to allow irq's after cr2 has been saved and the
 	 * vmalloc fault has been handled.

@@ -1278,26 +1270,6 @@ void do_user_addr_fault(struct pt_regs *regs,
 	 */
 	if (user_mode(regs)) {
 		local_irq_enable();
-		/*
-		 * Up to this point, X86_PF_USER set in hw_error_code
-		 * indicated a user-mode access.  But, after this,
-		 * X86_PF_USER in sw_error_code will indicate either
-		 * that, *or* an implicit kernel(supervisor)-mode access
-		 * which originated from user mode.
-		 */
-		if (!(hw_error_code & X86_PF_USER)) {
-			/*
-			 * The CPU was in user mode, but the CPU says
-			 * the fault was not a user-mode access.
-			 * Must be an implicit kernel-mode access,
-			 * which we do not expect to happen in the
-			 * user address space.
-			 */
-			pr_warn_once("kernel-mode error from user-mode: %lx\n",
-					hw_error_code);
-
-			sw_error_code |= X86_PF_USER;
-		}
 		flags |= FAULT_FLAG_USER;
 	} else {
 		if (regs->flags & X86_EFLAGS_IF)

@@ -1306,9 +1278,9 @@ void do_user_addr_fault(struct pt_regs *regs,
 
 	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
 
-	if (sw_error_code & X86_PF_WRITE)
+	if (hw_error_code & X86_PF_WRITE)
 		flags |= FAULT_FLAG_WRITE;
-	if (sw_error_code & X86_PF_INSTR)
+	if (hw_error_code & X86_PF_INSTR)
 		flags |= FAULT_FLAG_INSTRUCTION;
 
 #ifdef CONFIG_X86_64

@@ -1321,7 +1293,7 @@ void do_user_addr_fault(struct pt_regs *regs,
 	 * The vsyscall page does not have a "real" VMA, so do this
 	 * emulation before we go searching for VMAs.
 	 */
-	if ((sw_error_code & X86_PF_INSTR) && is_vsyscall_vaddr(address)) {
+	if ((hw_error_code & X86_PF_INSTR) && is_vsyscall_vaddr(address)) {
 		if (emulate_vsyscall(regs, address))
 			return;
 	}

@@ -1345,7 +1317,7 @@ void do_user_addr_fault(struct pt_regs *regs,
 			 * Fault from code in kernel from
 			 * which we do not expect faults.
 			 */
-			bad_area_nosemaphore(regs, sw_error_code, address);
+			bad_area_nosemaphore(regs, hw_error_code, address);
 			return;
 		}
 retry:

@@ -1361,13 +1333,13 @@ void do_user_addr_fault(struct pt_regs *regs,
 
 	vma = find_vma(mm, address);
 	if (unlikely(!vma)) {
-		bad_area(regs, sw_error_code, address);
+		bad_area(regs, hw_error_code, address);
 		return;
 	}
 	if (likely(vma->vm_start <= address))
 		goto good_area;
 	if (unlikely(!(vma->vm_flags & VM_GROWSDOWN))) {
-		bad_area(regs, sw_error_code, address);
+		bad_area(regs, hw_error_code, address);
 		return;
 	}
 	if (user_mode(regs)) {

@@ -1378,12 +1350,12 @@ void do_user_addr_fault(struct pt_regs *regs,
 		 * 32 pointers and then decrements %sp by 65535.)
 		 */
 		if (unlikely(address + 65536 + 32 * sizeof(unsigned long) < regs->sp)) {
-			bad_area(regs, sw_error_code, address);
+			bad_area(regs, hw_error_code, address);
 			return;
 		}
 	}
 	if (unlikely(expand_stack(vma, address))) {
-		bad_area(regs, sw_error_code, address);
+		bad_area(regs, hw_error_code, address);
 		return;
 	}
 

@@ -1392,8 +1364,8 @@ void do_user_addr_fault(struct pt_regs *regs,
 	 * we can handle it..
 	 */
 good_area:
-	if (unlikely(access_error(sw_error_code, vma))) {
-		bad_area_access_error(regs, sw_error_code, address, vma);
+	if (unlikely(access_error(hw_error_code, vma))) {
+		bad_area_access_error(regs, hw_error_code, address, vma);
 		return;
 	}
 

@@ -1432,13 +1404,13 @@ void do_user_addr_fault(struct pt_regs *regs,
 			return;
 
 		/* Not returning to user mode? Handle exceptions or die: */
-		no_context(regs, sw_error_code, address, SIGBUS, BUS_ADRERR);
+		no_context(regs, hw_error_code, address, SIGBUS, BUS_ADRERR);
 		return;
 	}
 
 	up_read(&mm->mmap_sem);
 	if (unlikely(fault & VM_FAULT_ERROR)) {
-		mm_fault_error(regs, sw_error_code, address, fault);
+		mm_fault_error(regs, hw_error_code, address, fault);
 		return;
 	}
 

index 6e18438c367f..092ed6b1df8a 100644
@@ -653,6 +653,15 @@ no_context(struct pt_regs *regs, unsigned long error_code,
 	unsigned long flags;
 	int sig;
 
+	if (user_mode(regs)) {
+		/*
+		 * This is an implicit supervisor-mode access from user
+		 * mode.  Bypass all the kernel-mode recovery code and just
+		 * OOPS.
+		 */
+		goto oops;
+	}
+
 	/* Are we prepared to handle this kernel fault? */
 	if (fixup_exception(regs, X86_TRAP_PF, error_code, address)) {
 		/*

@@ -738,6 +747,7 @@ no_context(struct pt_regs *regs, unsigned long error_code,
 	if (IS_ENABLED(CONFIG_EFI))
 		efi_recover_from_page_fault(address);
 
+oops:
 	/*
 	 * Oops. The kernel tried to access some bad page. We'll have to
 	 * terminate things with extreme prejudice:

index 0e0b4288a4b2..2b8e6324fa20 100644
@@ -66,7 +66,7 @@ void __show_regs(struct pt_regs *regs, enum show_regs_mode mode)
 	unsigned long cr0 = 0L, cr2 = 0L, cr3 = 0L, cr4 = 0L, fs, gs, shadowgs;
 	unsigned long d0, d1, d2, d3, d6, d7;
 	unsigned int fsindex, gsindex;
-	unsigned int ds, cs, es;
+	unsigned int ds, es;
 
 	show_iret_regs(regs);
 

@@ -98,7 +98,6 @@ void __show_regs(struct pt_regs *regs, enum show_regs_mode mode)
 	}
 
 	asm("movl %%ds,%0" : "=r" (ds));
-	asm("movl %%cs,%0" : "=r" (cs));
 	asm("movl %%es,%0" : "=r" (es));
 	asm("movl %%fs,%0" : "=r" (fsindex));
 	asm("movl %%gs,%0" : "=r" (gsindex));

@@ -114,7 +113,7 @@ void __show_regs(struct pt_regs *regs, enum show_regs_mode mode)
 
 	printk(KERN_DEFAULT "FS:  %016lx(%04x) GS:%016lx(%04x) knlGS:%016lx\n",
 	       fs, fsindex, gs, gsindex, shadowgs);
-	printk(KERN_DEFAULT "CS:  %04x DS: %04x ES: %04x CR0: %016lx\n", cs, ds,
+	printk(KERN_DEFAULT "CS:  %04lx DS: %04x ES: %04x CR0: %016lx\n", regs->cs, ds,
 			es, cr0);
 	printk(KERN_DEFAULT "CR2: %016lx CR3: %016lx CR4: %016lx\n", cr2, cr3,
 			cr4);

index 092ed6b1df8a..f34241fcc633 100644
@@ -27,6 +27,7 @@
 #include <asm/vm86.h>			/* struct vm86			*/
 #include <asm/mmu_context.h>		/* vma_pkey()			*/
 #include <asm/efi.h>			/* efi_recover_from_page_fault()*/
+#include <asm/desc.h>			/* store_idt(), ...		*/
 
 #define CREATE_TRACE_POINTS
 #include <asm/trace/exceptions.h>

@@ -571,10 +572,53 @@ static int is_f00f_bug(struct pt_regs *regs, unsigned long address)
 	return 0;
 }
 
+static void show_ldttss(const struct desc_ptr *gdt, const char *name, u16 index)
+{
+	u32 offset = (index >> 3) * sizeof(struct desc_struct);
+	unsigned long addr;
+	struct ldttss_desc desc;
+
+	if (index == 0) {
+		pr_alert("%s: NULL\n", name);
+		return;
+	}
+
+	if (offset + sizeof(struct ldttss_desc) >= gdt->size) {
+		pr_alert("%s: 0x%hx -- out of bounds\n", name, index);
+		return;
+	}
+
+	if (probe_kernel_read(&desc, (void *)(gdt->address + offset),
+			      sizeof(struct ldttss_desc))) {
+		pr_alert("%s: 0x%hx -- GDT entry is not readable\n",
+			 name, index);
+		return;
+	}
+
+	addr = desc.base0 | (desc.base1 << 16) | (desc.base2 << 24);
+#ifdef CONFIG_X86_64
+	addr |= ((u64)desc.base3 << 32);
+#endif
+	pr_alert("%s: 0x%hx -- base=0x%lx limit=0x%x\n",
+		 name, index, addr, (desc.limit0 | (desc.limit1 << 16)));
+}
+
+static void errstr(unsigned long ec, char *buf, unsigned long mask,
+		   const char *txt)
+{
+	if (ec & mask) {
+		if (buf[0])
+			strcat(buf, " ");
+		strcat(buf, txt);
+	}
+}
+
 static void
 show_fault_oops(struct pt_regs *regs, unsigned long error_code,
 		unsigned long address)
 {
+	char errtxt[64];
+
 	if (!oops_may_print())
 		return;
 

@@ -602,6 +646,46 @@ show_fault_oops(struct pt_regs *regs, unsigned long error_code,
 		 address < PAGE_SIZE ? "NULL pointer dereference" : "paging request",
 		 (void *)address);
 
+	errtxt[0] = 0;
+	errstr(error_code, errtxt, X86_PF_PROT, "PROT");
+	errstr(error_code, errtxt, X86_PF_WRITE, "WRITE");
+	errstr(error_code, errtxt, X86_PF_USER, "USER");
+	errstr(error_code, errtxt, X86_PF_RSVD, "RSVD");
+	errstr(error_code, errtxt, X86_PF_INSTR, "INSTR");
+	errstr(error_code, errtxt, X86_PF_PK, "PK");
+	pr_alert("HW error: %s\n", error_code ? errtxt :
+		 "normal kernel read fault");
+	if (!(error_code & X86_PF_USER) && user_mode(regs)) {
+		struct desc_ptr idt, gdt;
+		u16 ldtr, tr;
+
+		pr_alert("This was a system access from user code\n");
+
+		/*
+		 * This can happen for quite a few reasons.  The more obvious
+		 * ones are faults accessing the GDT, or LDT.  Perhaps
+		 * surprisingly, if the CPU tries to deliver a benign or
+		 * contributory exception from user code and gets a page fault
+		 * during delivery, the page fault can be delivered as though
+		 * it originated directly from user code.  This could happen
+		 * due to wrong permissions on the IDT, GDT, LDT, TSS, or
+		 * kernel or IST stack.
+		 */
+		store_idt(&idt);
+
+		/* Usable even on Xen PV -- it's just slow. */
+		native_store_gdt(&gdt);
+
+		pr_alert("IDT: 0x%lx (limit=0x%hx) GDT: 0x%lx (limit=0x%hx)\n",
+			 idt.address, idt.size, gdt.address, gdt.size);
+
+		store_ldt(ldtr);
+		show_ldttss(&gdt, "LDTR", ldtr);
+
+		store_tr(tr);
+		show_ldttss(&gdt, "TR", tr);
+	}
+
 	dump_pagetable(address);
 }
 

index 85fd85d52ffd..d78bcc03e60e 100644
@@ -102,7 +102,7 @@ static bool write_ok_or_segv(unsigned long ptr, size_t size)
 	if (!access_ok(VERIFY_WRITE, (void __user *)ptr, size)) {
 		struct thread_struct *thread = &current->thread;
 
-		thread->error_code	= 6;  /* user fault, no page, write */
+		thread->error_code	= X86_PF_USER | X86_PF_WRITE;
 		thread->cr2		= ptr;
 		thread->trap_nr		= X86_TRAP_PF;