Merge branch 'for-next/kernel-mode-neon' into for-next/core

* for-next/kernel-mode-neon:
  arm64: neon/efi: Make EFI fpsimd save/restore variables static
  arm64: neon: Forbid when irqs are disabled
  arm64: neon: Export kernel_neon_busy to loadable modules
  arm64: neon: Temporarily add a kernel_mode_begin_partial() definition
  arm64: neon: Remove support for nested or hardirq kernel-mode NEON
  arm64: neon: Allow EFI runtime services to use FPSIMD in irq context
  arm64: fpsimd: Consistently use __this_cpu_ ops where appropriate
  arm64: neon: Add missing header guard in <asm/neon.h>
  arm64: neon: replace generic definition of may_use_simd()

arch/arm64/include/asm/Kbuild

@@ -20,7 +20,6 @@ generic-y += rwsem.h
 generic-y += segment.h
 generic-y += serial.h
 generic-y += set_memory.h
-generic-y += simd.h
 generic-y += sizes.h
 generic-y += switch_to.h
 generic-y += trace_clock.h

arch/arm64/include/asm/efi.h

@@ -3,6 +3,7 @@
 
 #include <asm/boot.h>
 #include <asm/cpufeature.h>
+#include <asm/fpsimd.h>
 #include <asm/io.h>
 #include <asm/memory.h>
 #include <asm/mmu_context.h>
@@ -21,8 +22,8 @@ int efi_set_mapping_permissions(struct mm_struct *mm, efi_memory_desc_t *md);
 
 #define arch_efi_call_virt_setup()					\
 ({									\
-	kernel_neon_begin();						\
 	efi_virtmap_load();						\
+	__efi_fpsimd_begin();						\
 })
 
 #define arch_efi_call_virt(p, f, args...)				\
@@ -34,8 +35,8 @@ int efi_set_mapping_permissions(struct mm_struct *mm, efi_memory_desc_t *md);
 
 #define arch_efi_call_virt_teardown()					\
 ({									\
+	__efi_fpsimd_end();						\
 	efi_virtmap_unload();						\
-	kernel_neon_end();						\
 })
 
 #define ARCH_EFI_IRQ_FLAGS_MASK (PSR_D_BIT | PSR_A_BIT | PSR_I_BIT | PSR_F_BIT)

arch/arm64/include/asm/fpsimd.h

@@ -41,16 +41,6 @@ struct fpsimd_state {
 	unsigned int cpu;
 };
 
-/*
- * Struct for stacking the bottom 'n' FP/SIMD registers.
- */
-struct fpsimd_partial_state {
-	u32		fpsr;
-	u32		fpcr;
-	u32		num_regs;
-	__uint128_t	vregs[32];
-};
-
 
 #if defined(__KERNEL__) && defined(CONFIG_COMPAT)
 /* Masks for extracting the FPSR and FPCR from the FPSCR */
@@ -77,9 +67,9 @@ extern void fpsimd_update_current_state(struct fpsimd_state *state);
 
 extern void fpsimd_flush_task_state(struct task_struct *target);
 
-extern void fpsimd_save_partial_state(struct fpsimd_partial_state *state,
-				      u32 num_regs);
-extern void fpsimd_load_partial_state(struct fpsimd_partial_state *state);
+/* For use by EFI runtime services calls only */
+extern void __efi_fpsimd_begin(void);
+extern void __efi_fpsimd_end(void);
 
 #endif
 

arch/arm64/include/asm/fpsimdmacros.h

@@ -75,59 +75,3 @@
 	ldr	w\tmpnr, [\state, #16 * 2 + 4]
 	fpsimd_restore_fpcr x\tmpnr, \state
 .endm
-
-.macro fpsimd_save_partial state, numnr, tmpnr1, tmpnr2
-	mrs	x\tmpnr1, fpsr
-	str	w\numnr, [\state, #8]
-	mrs	x\tmpnr2, fpcr
-	stp	w\tmpnr1, w\tmpnr2, [\state]
-	adr	x\tmpnr1, 0f
-	add	\state, \state, x\numnr, lsl #4
-	sub	x\tmpnr1, x\tmpnr1, x\numnr, lsl #1
-	br	x\tmpnr1
-	stp	q30, q31, [\state, #-16 * 30 - 16]
-	stp	q28, q29, [\state, #-16 * 28 - 16]
-	stp	q26, q27, [\state, #-16 * 26 - 16]
-	stp	q24, q25, [\state, #-16 * 24 - 16]
-	stp	q22, q23, [\state, #-16 * 22 - 16]
-	stp	q20, q21, [\state, #-16 * 20 - 16]
-	stp	q18, q19, [\state, #-16 * 18 - 16]
-	stp	q16, q17, [\state, #-16 * 16 - 16]
-	stp	q14, q15, [\state, #-16 * 14 - 16]
-	stp	q12, q13, [\state, #-16 * 12 - 16]
-	stp	q10, q11, [\state, #-16 * 10 - 16]
-	stp	q8, q9, [\state, #-16 * 8 - 16]
-	stp	q6, q7, [\state, #-16 * 6 - 16]
-	stp	q4, q5, [\state, #-16 * 4 - 16]
-	stp	q2, q3, [\state, #-16 * 2 - 16]
-	stp	q0, q1, [\state, #-16 * 0 - 16]
-0:
-.endm
-
-.macro fpsimd_restore_partial state, tmpnr1, tmpnr2
-	ldp	w\tmpnr1, w\tmpnr2, [\state]
-	msr	fpsr, x\tmpnr1
-	fpsimd_restore_fpcr x\tmpnr2, x\tmpnr1
-	adr	x\tmpnr1, 0f
-	ldr	w\tmpnr2, [\state, #8]
-	add	\state, \state, x\tmpnr2, lsl #4
-	sub	x\tmpnr1, x\tmpnr1, x\tmpnr2, lsl #1
-	br	x\tmpnr1
-	ldp	q30, q31, [\state, #-16 * 30 - 16]
-	ldp	q28, q29, [\state, #-16 * 28 - 16]
-	ldp	q26, q27, [\state, #-16 * 26 - 16]
-	ldp	q24, q25, [\state, #-16 * 24 - 16]
-	ldp	q22, q23, [\state, #-16 * 22 - 16]
-	ldp	q20, q21, [\state, #-16 * 20 - 16]
-	ldp	q18, q19, [\state, #-16 * 18 - 16]
-	ldp	q16, q17, [\state, #-16 * 16 - 16]
-	ldp	q14, q15, [\state, #-16 * 14 - 16]
-	ldp	q12, q13, [\state, #-16 * 12 - 16]
-	ldp	q10, q11, [\state, #-16 * 10 - 16]
-	ldp	q8, q9, [\state, #-16 * 8 - 16]
-	ldp	q6, q7, [\state, #-16 * 6 - 16]
-	ldp	q4, q5, [\state, #-16 * 4 - 16]
-	ldp	q2, q3, [\state, #-16 * 2 - 16]
-	ldp	q0, q1, [\state, #-16 * 0 - 16]
-0:
-.endm

arch/arm64/include/asm/neon.h

@@ -8,12 +8,22 @@
  * published by the Free Software Foundation.
  */
 
+#ifndef __ASM_NEON_H
+#define __ASM_NEON_H
+
 #include <linux/types.h>
 #include <asm/fpsimd.h>
 
 #define cpu_has_neon()		system_supports_fpsimd()
 
-#define kernel_neon_begin()	kernel_neon_begin_partial(32)
-
-void kernel_neon_begin_partial(u32 num_regs);
+void kernel_neon_begin(void);
 void kernel_neon_end(void);
+
+/*
+ * Temporary macro to allow the crypto code to compile. Note that the
+ * semantics of kernel_neon_begin_partial() are now different from the
+ * original as it does not allow being called in an interrupt context.
+ */
+#define kernel_neon_begin_partial(num_regs)	kernel_neon_begin()
+
+#endif /* ! __ASM_NEON_H */

arch/arm64/include/asm/simd.h

@@ -0,0 +1,56 @@
+/*
+ * Copyright (C) 2017 Linaro Ltd. <ard.biesheuvel@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation.
+ */
+
+#ifndef __ASM_SIMD_H
+#define __ASM_SIMD_H
+
+#include <linux/compiler.h>
+#include <linux/irqflags.h>
+#include <linux/percpu.h>
+#include <linux/preempt.h>
+#include <linux/types.h>
+
+#ifdef CONFIG_KERNEL_MODE_NEON
+
+DECLARE_PER_CPU(bool, kernel_neon_busy);
+
+/*
+ * may_use_simd - whether it is allowable at this time to issue SIMD
+ *                instructions or access the SIMD register file
+ *
+ * Callers must not assume that the result remains true beyond the next
+ * preempt_enable() or return from softirq context.
+ */
+static __must_check inline bool may_use_simd(void)
+{
+	/*
+	 * The raw_cpu_read() is racy if called with preemption enabled.
+	 * This is not a bug: kernel_neon_busy is only set when
+	 * preemption is disabled, so we cannot migrate to another CPU
+	 * while it is set, nor can we migrate to a CPU where it is set.
+	 * So, if we find it clear on some CPU then we're guaranteed to
+	 * find it clear on any CPU we could migrate to.
+	 *
+	 * If we are in between kernel_neon_begin()...kernel_neon_end(),
+	 * the flag will be set, but preemption is also disabled, so we
+	 * can't migrate to another CPU and spuriously see it become
+	 * false.
+	 */
+	return !in_irq() && !irqs_disabled() && !in_nmi() &&
+		!raw_cpu_read(kernel_neon_busy);
+}
+
+#else /* ! CONFIG_KERNEL_MODE_NEON */
+
+static __must_check inline bool may_use_simd(void) {
+	return false;
+}
+
+#endif /* ! CONFIG_KERNEL_MODE_NEON */
+
+#endif

arch/arm64/kernel/entry-fpsimd.S

@@ -41,27 +41,3 @@ ENTRY(fpsimd_load_state)
 	fpsimd_restore x0, 8
 	ret
 ENDPROC(fpsimd_load_state)
-
-#ifdef CONFIG_KERNEL_MODE_NEON
-
-/*
- * Save the bottom n FP registers.
- *
- * x0 - pointer to struct fpsimd_partial_state
- */
-ENTRY(fpsimd_save_partial_state)
-	fpsimd_save_partial x0, 1, 8, 9
-	ret
-ENDPROC(fpsimd_save_partial_state)
-
-/*
- * Load the bottom n FP registers.
- *
- * x0 - pointer to struct fpsimd_partial_state
- */
-ENTRY(fpsimd_load_partial_state)
-	fpsimd_restore_partial x0, 8, 9
-	ret
-ENDPROC(fpsimd_load_partial_state)
-
-#endif

arch/arm64/kernel/fpsimd.c

@@ -17,16 +17,19 @@
  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
+#include <linux/bottom_half.h>
 #include <linux/cpu.h>
 #include <linux/cpu_pm.h>
 #include <linux/kernel.h>
 #include <linux/init.h>
+#include <linux/percpu.h>
+#include <linux/preempt.h>
 #include <linux/sched/signal.h>
 #include <linux/signal.h>
-#include <linux/hardirq.h>
 
 #include <asm/fpsimd.h>
 #include <asm/cputype.h>
+#include <asm/simd.h>
 
 #define FPEXC_IOF	(1 << 0)
 #define FPEXC_DZF	(1 << 1)
@@ -62,6 +65,13 @@
  * CPU currently contain the most recent userland FPSIMD state of the current
  * task.
  *
+ * In order to allow softirq handlers to use FPSIMD, kernel_neon_begin() may
+ * save the task's FPSIMD context back to task_struct from softirq context.
+ * To prevent this from racing with the manipulation of the task's FPSIMD state
+ * from task context and thereby corrupting the state, it is necessary to
+ * protect any manipulation of a task's fpsimd_state or TIF_FOREIGN_FPSTATE
+ * flag with local_bh_disable() unless softirqs are already masked.
+ *
  * For a certain task, the sequence may look something like this:
  * - the task gets scheduled in; if both the task's fpsimd_state.cpu field
  *   contains the id of the current CPU, and the CPU's fpsimd_last_state per-cpu
@@ -161,9 +171,14 @@ void fpsimd_flush_thread(void)
 {
 	if (!system_supports_fpsimd())
 		return;
+
+	local_bh_disable();
+
 	memset(&current->thread.fpsimd_state, 0, sizeof(struct fpsimd_state));
 	fpsimd_flush_task_state(current);
 	set_thread_flag(TIF_FOREIGN_FPSTATE);
+
+	local_bh_enable();
 }
 
 /*
@@ -174,10 +189,13 @@ void fpsimd_preserve_current_state(void)
 {
 	if (!system_supports_fpsimd())
 		return;
-	preempt_disable();
+
+	local_bh_disable();
+
 	if (!test_thread_flag(TIF_FOREIGN_FPSTATE))
 		fpsimd_save_state(&current->thread.fpsimd_state);
-	preempt_enable();
+
+	local_bh_enable();
 }
 
 /*
@@ -189,15 +207,18 @@ void fpsimd_restore_current_state(void)
 {
 	if (!system_supports_fpsimd())
 		return;
-	preempt_disable();
+
+	local_bh_disable();
+
 	if (test_and_clear_thread_flag(TIF_FOREIGN_FPSTATE)) {
 		struct fpsimd_state *st = &current->thread.fpsimd_state;
 
 		fpsimd_load_state(st);
-		this_cpu_write(fpsimd_last_state, st);
+		__this_cpu_write(fpsimd_last_state, st);
 		st->cpu = smp_processor_id();
 	}
-	preempt_enable();
+
+	local_bh_enable();
 }
 
 /*
@@ -209,15 +230,18 @@ void fpsimd_update_current_state(struct fpsimd_state *state)
 {
 	if (!system_supports_fpsimd())
 		return;
-	preempt_disable();
+
+	local_bh_disable();
+
 	fpsimd_load_state(state);
 	if (test_and_clear_thread_flag(TIF_FOREIGN_FPSTATE)) {
 		struct fpsimd_state *st = &current->thread.fpsimd_state;
 
-		this_cpu_write(fpsimd_last_state, st);
+		__this_cpu_write(fpsimd_last_state, st);
 		st->cpu = smp_processor_id();
 	}
-	preempt_enable();
+
+	local_bh_enable();
 }
 
 /*
@@ -230,52 +254,122 @@ void fpsimd_flush_task_state(struct task_struct *t)
 
 #ifdef CONFIG_KERNEL_MODE_NEON
 
-static DEFINE_PER_CPU(struct fpsimd_partial_state, hardirq_fpsimdstate);
-static DEFINE_PER_CPU(struct fpsimd_partial_state, softirq_fpsimdstate);
+DEFINE_PER_CPU(bool, kernel_neon_busy);
+EXPORT_PER_CPU_SYMBOL(kernel_neon_busy);
 
 /*
  * Kernel-side NEON support functions
  */
-void kernel_neon_begin_partial(u32 num_regs)
+
+/*
+ * kernel_neon_begin(): obtain the CPU FPSIMD registers for use by the calling
+ * context
+ *
+ * Must not be called unless may_use_simd() returns true.
+ * Task context in the FPSIMD registers is saved back to memory as necessary.
+ *
+ * A matching call to kernel_neon_end() must be made before returning from the
+ * calling context.
+ *
+ * The caller may freely use the FPSIMD registers until kernel_neon_end() is
+ * called.
+ */
+void kernel_neon_begin(void)
 {
 	if (WARN_ON(!system_supports_fpsimd()))
 		return;
-	if (in_interrupt()) {
-		struct fpsimd_partial_state *s = this_cpu_ptr(
-			in_irq() ? &hardirq_fpsimdstate : &softirq_fpsimdstate);
 
-		BUG_ON(num_regs > 32);
-		fpsimd_save_partial_state(s, roundup(num_regs, 2));
-	} else {
-		/*
-		 * Save the userland FPSIMD state if we have one and if we
-		 * haven't done so already. Clear fpsimd_last_state to indicate
-		 * that there is no longer userland FPSIMD state in the
-		 * registers.
-		 */
-		preempt_disable();
-		if (current->mm &&
-		    !test_and_set_thread_flag(TIF_FOREIGN_FPSTATE))
-			fpsimd_save_state(&current->thread.fpsimd_state);
-		this_cpu_write(fpsimd_last_state, NULL);
-	}
+	BUG_ON(!may_use_simd());
+
+	local_bh_disable();
+
+	__this_cpu_write(kernel_neon_busy, true);
+
+	/* Save unsaved task fpsimd state, if any: */
+	if (current->mm && !test_and_set_thread_flag(TIF_FOREIGN_FPSTATE))
+		fpsimd_save_state(&current->thread.fpsimd_state);
+
+	/* Invalidate any task state remaining in the fpsimd regs: */
+	__this_cpu_write(fpsimd_last_state, NULL);
+
+	preempt_disable();
+
+	local_bh_enable();
 }
-EXPORT_SYMBOL(kernel_neon_begin_partial);
+EXPORT_SYMBOL(kernel_neon_begin);
 
+/*
+ * kernel_neon_end(): give the CPU FPSIMD registers back to the current task
+ *
+ * Must be called from a context in which kernel_neon_begin() was previously
+ * called, with no call to kernel_neon_end() in the meantime.
+ *
+ * The caller must not use the FPSIMD registers after this function is called,
+ * unless kernel_neon_begin() is called again in the meantime.
+ */
 void kernel_neon_end(void)
 {
+	bool busy;
+
 	if (!system_supports_fpsimd())
 		return;
-	if (in_interrupt()) {
-		struct fpsimd_partial_state *s = this_cpu_ptr(
-			in_irq() ? &hardirq_fpsimdstate : &softirq_fpsimdstate);
-		fpsimd_load_partial_state(s);
-	} else {
-		preempt_enable();
-	}
+
+	busy = __this_cpu_xchg(kernel_neon_busy, false);
+	WARN_ON(!busy);	/* No matching kernel_neon_begin()? */
+
+	preempt_enable();
 }
 EXPORT_SYMBOL(kernel_neon_end);
 
+static DEFINE_PER_CPU(struct fpsimd_state, efi_fpsimd_state);
+static DEFINE_PER_CPU(bool, efi_fpsimd_state_used);
+
+/*
+ * EFI runtime services support functions
+ *
+ * The ABI for EFI runtime services allows EFI to use FPSIMD during the call.
+ * This means that for EFI (and only for EFI), we have to assume that FPSIMD
+ * is always used rather than being an optional accelerator.
+ *
+ * These functions provide the necessary support for ensuring FPSIMD
+ * save/restore in the contexts from which EFI is used.
+ *
+ * Do not use them for any other purpose -- if tempted to do so, you are
+ * either doing something wrong or you need to propose some refactoring.
+ */
+
+/*
+ * __efi_fpsimd_begin(): prepare FPSIMD for making an EFI runtime services call
+ */
+void __efi_fpsimd_begin(void)
+{
+	if (!system_supports_fpsimd())
+		return;
+
+	WARN_ON(preemptible());
+
+	if (may_use_simd())
+		kernel_neon_begin();
+	else {
+		fpsimd_save_state(this_cpu_ptr(&efi_fpsimd_state));
+		__this_cpu_write(efi_fpsimd_state_used, true);
+	}
+}
+
+/*
+ * __efi_fpsimd_end(): clean up FPSIMD after an EFI runtime services call
+ */
+void __efi_fpsimd_end(void)
+{
+	if (!system_supports_fpsimd())
+		return;
+
+	if (__this_cpu_xchg(efi_fpsimd_state_used, false))
+		fpsimd_load_state(this_cpu_ptr(&efi_fpsimd_state));
+	else
+		kernel_neon_end();
+}
+
 #endif /* CONFIG_KERNEL_MODE_NEON */
 
 #ifdef CONFIG_CPU_PM