Merge tag 'parisc-for-6.4-3' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux

Pull parisc architecture fixes from Helge Deller:
 "Quite a bunch of real bugfixes in here and most of them are tagged for
  backporting: A fix for cache flushing from irq context, a kprobes &
  kgdb breakpoint handling fix, and a fix in the alternative code
  patching function to take care of CPU hotplugging.

  parisc now provides LOCKDEP support and comes with a lightweight
  spinlock check. Both features helped me to find the cache flush bug.

  Additionally writing the AGP gatt has been fixed, the machine allows
  the user to reboot after a system halt and arch_sync_dma_for_cpu() has
  been optimized for PCXL PCUs.

  Summary:

   - Fix flush_dcache_page() for usage from irq context

   - Handle kprobes breakpoints only in kernel context

   - Handle kgdb breakpoints only in kernel context

   - Use num_present_cpus() in alternative patching code

   - Enable LOCKDEP support

   - Add lightweight spinlock checks

   - Flush AGP gatt writes and adjust gatt mask in parisc_agp_mask_memory()

   - Allow to reboot machine after system halt

   - Improve cache flushing for PCXL in arch_sync_dma_for_cpu()"

* tag 'parisc-for-6.4-3' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: Fix flush_dcache_page() for usage from irq context
  parisc: Handle kgdb breakpoints only in kernel context
  parisc: Handle kprobes breakpoints only in kernel context
  parisc: Allow to reboot machine after system halt
  parisc: Enable LOCKDEP support
  parisc: Add lightweight spinlock checks
  parisc: Use num_present_cpus() in alternative patching code
  parisc: Flush gatt writes and adjust gatt mask in parisc_agp_mask_memory()
  parisc: Improve cache flushing for PCXL in arch_sync_dma_for_cpu()

11 files changed, 116 insertions, 19 deletions

diff --git a/arch/parisc/Kconfig b/arch/parisc/Kconfig
index 466a25525364d2..967bde65dd0ec8 100644
--- a/arch/parisc/Kconfig
+++ b/arch/parisc/Kconfig
@@ -130,6 +130,10 @@ config PM
 config STACKTRACE_SUPPORT
 	def_bool y
 
+config LOCKDEP_SUPPORT
+	bool
+	default y
+
 config ISA_DMA_API
 	bool
 
diff --git a/arch/parisc/Kconfig.debug b/arch/parisc/Kconfig.debug
index f66554cd5c4518..3a059cb5e112fb 100644
--- a/arch/parisc/Kconfig.debug
+++ b/arch/parisc/Kconfig.debug
@@ -1 +1,12 @@
 # SPDX-License-Identifier: GPL-2.0
+#
+config LIGHTWEIGHT_SPINLOCK_CHECK
+	bool "Enable lightweight spinlock checks"
+	depends on SMP && !DEBUG_SPINLOCK
+	default y
+	help
+	  Add checks with low performance impact to the spinlock functions
+	  to catch memory overwrites at runtime. For more advanced
+	  spinlock debugging you should choose the DEBUG_SPINLOCK option
+	  which will detect unitialized spinlocks too.
+	  If unsure say Y here.
diff --git a/arch/parisc/include/asm/cacheflush.h b/arch/parisc/include/asm/cacheflush.h
index 0bdee672413206..c8b6928cee1ee4 100644
--- a/arch/parisc/include/asm/cacheflush.h
+++ b/arch/parisc/include/asm/cacheflush.h
@@ -48,6 +48,10 @@ void flush_dcache_page(struct page *page);
 
 #define flush_dcache_mmap_lock(mapping)		xa_lock_irq(&mapping->i_pages)
 #define flush_dcache_mmap_unlock(mapping)	xa_unlock_irq(&mapping->i_pages)
+#define flush_dcache_mmap_lock_irqsave(mapping, flags)		\
+		xa_lock_irqsave(&mapping->i_pages, flags)
+#define flush_dcache_mmap_unlock_irqrestore(mapping, flags)	\
+		xa_unlock_irqrestore(&mapping->i_pages, flags)
 
 #define flush_icache_page(vma,page)	do { 		\
 	flush_kernel_dcache_page_addr(page_address(page)); \
diff --git a/arch/parisc/include/asm/spinlock.h b/arch/parisc/include/asm/spinlock.h
index a6e5d66a76562a..edfcb9858bcb7c 100644
--- a/arch/parisc/include/asm/spinlock.h
+++ b/arch/parisc/include/asm/spinlock.h
@@ -7,10 +7,26 @@
 #include <asm/processor.h>
 #include <asm/spinlock_types.h>
 
+#define SPINLOCK_BREAK_INSN	0x0000c006	/* break 6,6 */
+
+static inline void arch_spin_val_check(int lock_val)
+{
+	if (IS_ENABLED(CONFIG_LIGHTWEIGHT_SPINLOCK_CHECK))
+		asm volatile(	"andcm,= %0,%1,%%r0\n"
+				".word %2\n"
+		: : "r" (lock_val), "r" (__ARCH_SPIN_LOCK_UNLOCKED_VAL),
+			"i" (SPINLOCK_BREAK_INSN));
+}
+
 static inline int arch_spin_is_locked(arch_spinlock_t *x)
 {
-	volatile unsigned int *a = __ldcw_align(x);
-	return READ_ONCE(*a) == 0;
+	volatile unsigned int *a;
+	int lock_val;
+
+	a = __ldcw_align(x);
+	lock_val = READ_ONCE(*a);
+	arch_spin_val_check(lock_val);
+	return (lock_val == 0);
 }
 
 static inline void arch_spin_lock(arch_spinlock_t *x)
@@ -18,9 +34,18 @@ static inline void arch_spin_lock(arch_spinlock_t *x)
 	volatile unsigned int *a;
 
 	a = __ldcw_align(x);
-	while (__ldcw(a) == 0)
+	do {
+		int lock_val_old;
+
+		lock_val_old = __ldcw(a);
+		arch_spin_val_check(lock_val_old);
+		if (lock_val_old)
+			return;	/* got lock */
+
+		/* wait until we should try to get lock again */
 		while (*a == 0)
 			continue;
+	} while (1);
 }
 
 static inline void arch_spin_unlock(arch_spinlock_t *x)
@@ -29,15 +54,19 @@ static inline void arch_spin_unlock(arch_spinlock_t *x)
 
 	a = __ldcw_align(x);
 	/* Release with ordered store. */
-	__asm__ __volatile__("stw,ma %0,0(%1)" : : "r"(1), "r"(a) : "memory");
+	__asm__ __volatile__("stw,ma %0,0(%1)"
+		: : "r"(__ARCH_SPIN_LOCK_UNLOCKED_VAL), "r"(a) : "memory");
 }
 
 static inline int arch_spin_trylock(arch_spinlock_t *x)
 {
 	volatile unsigned int *a;
+	int lock_val;
 
 	a = __ldcw_align(x);
-	return __ldcw(a) != 0;
+	lock_val = __ldcw(a);
+	arch_spin_val_check(lock_val);
+	return lock_val != 0;
 }
 
 /*
diff --git a/arch/parisc/include/asm/spinlock_types.h b/arch/parisc/include/asm/spinlock_types.h
index ca39ee350c3f42..d65934079ebdb0 100644
--- a/arch/parisc/include/asm/spinlock_types.h
+++ b/arch/parisc/include/asm/spinlock_types.h
@@ -2,13 +2,17 @@
 #ifndef __ASM_SPINLOCK_TYPES_H
 #define __ASM_SPINLOCK_TYPES_H
 
+#define __ARCH_SPIN_LOCK_UNLOCKED_VAL	0x1a46
+
 typedef struct {
 #ifdef CONFIG_PA20
 	volatile unsigned int slock;
-# define __ARCH_SPIN_LOCK_UNLOCKED { 1 }
+# define __ARCH_SPIN_LOCK_UNLOCKED { __ARCH_SPIN_LOCK_UNLOCKED_VAL }
 #else
 	volatile unsigned int lock[4];
-# define __ARCH_SPIN_LOCK_UNLOCKED	{ { 1, 1, 1, 1 } }
+# define __ARCH_SPIN_LOCK_UNLOCKED	\
+	{ { __ARCH_SPIN_LOCK_UNLOCKED_VAL, __ARCH_SPIN_LOCK_UNLOCKED_VAL, \
+	    __ARCH_SPIN_LOCK_UNLOCKED_VAL, __ARCH_SPIN_LOCK_UNLOCKED_VAL } }
 #endif
 } arch_spinlock_t;
 
diff --git a/arch/parisc/kernel/alternative.c b/arch/parisc/kernel/alternative.c
index 66f5672c70bd46..25c4d6c3375db3 100644
--- a/arch/parisc/kernel/alternative.c
+++ b/arch/parisc/kernel/alternative.c
@@ -25,7 +25,7 @@ void __init_or_module apply_alternatives(struct alt_instr *start,
 {
 	struct alt_instr *entry;
 	int index = 0, applied = 0;
-	int num_cpus = num_online_cpus();
+	int num_cpus = num_present_cpus();
 	u16 cond_check;
 
 	cond_check = ALT_COND_ALWAYS |
diff --git a/arch/parisc/kernel/cache.c b/arch/parisc/kernel/cache.c
index 1d3b8bc8a62336..ca4a302d4365fe 100644
--- a/arch/parisc/kernel/cache.c
+++ b/arch/parisc/kernel/cache.c
@@ -399,6 +399,7 @@ void flush_dcache_page(struct page *page)
 	unsigned long offset;
 	unsigned long addr, old_addr = 0;
 	unsigned long count = 0;
+	unsigned long flags;
 	pgoff_t pgoff;
 
 	if (mapping && !mapping_mapped(mapping)) {
@@ -420,7 +421,7 @@ void flush_dcache_page(struct page *page)
 	 * to flush one address here for them all to become coherent
 	 * on machines that support equivalent aliasing
 	 */
-	flush_dcache_mmap_lock(mapping);
+	flush_dcache_mmap_lock_irqsave(mapping, flags);
 	vma_interval_tree_foreach(mpnt, &mapping->i_mmap, pgoff, pgoff) {
 		offset = (pgoff - mpnt->vm_pgoff) << PAGE_SHIFT;
 		addr = mpnt->vm_start + offset;
@@ -460,7 +461,7 @@ void flush_dcache_page(struct page *page)
 		}
 		WARN_ON(++count == 4096);
 	}
-	flush_dcache_mmap_unlock(mapping);
+	flush_dcache_mmap_unlock_irqrestore(mapping, flags);
 }
 EXPORT_SYMBOL(flush_dcache_page);
 
diff --git a/arch/parisc/kernel/pci-dma.c b/arch/parisc/kernel/pci-dma.c
index ba87f791323be0..71ed5391f29d6f 100644
--- a/arch/parisc/kernel/pci-dma.c
+++ b/arch/parisc/kernel/pci-dma.c
@@ -446,11 +446,27 @@ void arch_dma_free(struct device *dev, size_t size, void *vaddr,
 void arch_sync_dma_for_device(phys_addr_t paddr, size_t size,
 		enum dma_data_direction dir)
 {
+	/*
+	 * fdc: The data cache line is written back to memory, if and only if
+	 * it is dirty, and then invalidated from the data cache.
+	 */
 	flush_kernel_dcache_range((unsigned long)phys_to_virt(paddr), size);
 }
 
 void arch_sync_dma_for_cpu(phys_addr_t paddr, size_t size,
 		enum dma_data_direction dir)
 {
-	flush_kernel_dcache_range((unsigned long)phys_to_virt(paddr), size);
+	unsigned long addr = (unsigned long) phys_to_virt(paddr);
+
+	switch (dir) {
+	case DMA_TO_DEVICE:
+	case DMA_BIDIRECTIONAL:
+		flush_kernel_dcache_range(addr, size);
+		return;
+	case DMA_FROM_DEVICE:
+		purge_kernel_dcache_range_asm(addr, addr + size);
+		return;
+	default:
+		BUG();
+	}
 }
diff --git a/arch/parisc/kernel/process.c b/arch/parisc/kernel/process.c
index 97c6f875bd0e12..24411ab79c307d 100644
--- a/arch/parisc/kernel/process.c
+++ b/arch/parisc/kernel/process.c
@@ -122,13 +122,18 @@ void machine_power_off(void)
 	/* It seems we have no way to power the system off via
 	 * software. The user has to press the button himself. */
 
-	printk(KERN_EMERG "System shut down completed.\n"
-	       "Please power this system off now.");
+	printk("Power off or press RETURN to reboot.\n");
 
 	/* prevent soft lockup/stalled CPU messages for endless loop. */
 	rcu_sysrq_start();
 	lockup_detector_soft_poweroff();
-	for (;;);
+	while (1) {
+		/* reboot if user presses RETURN key */
+		if (pdc_iodc_getc() == 13) {
+			printk("Rebooting...\n");
+			machine_restart(NULL);
+		}
+	}
 }
 
 void (*pm_power_off)(void);
diff --git a/arch/parisc/kernel/traps.c b/arch/parisc/kernel/traps.c
index f9696fbf646c47..304eebd1c83e7d 100644
--- a/arch/parisc/kernel/traps.c
+++ b/arch/parisc/kernel/traps.c
@@ -47,6 +47,10 @@
 #include <linux/kgdb.h>
 #include <linux/kprobes.h>
 
+#if defined(CONFIG_LIGHTWEIGHT_SPINLOCK_CHECK)
+#include <asm/spinlock.h>
+#endif
+
 #include "../math-emu/math-emu.h"	/* for handle_fpe() */
 
 static void parisc_show_stack(struct task_struct *task,
@@ -291,24 +295,30 @@ static void handle_break(struct pt_regs *regs)
 	}
 
 #ifdef CONFIG_KPROBES
-	if (unlikely(iir == PARISC_KPROBES_BREAK_INSN)) {
+	if (unlikely(iir == PARISC_KPROBES_BREAK_INSN && !user_mode(regs))) {
 		parisc_kprobe_break_handler(regs);
 		return;
 	}
-	if (unlikely(iir == PARISC_KPROBES_BREAK_INSN2)) {
+	if (unlikely(iir == PARISC_KPROBES_BREAK_INSN2 && !user_mode(regs))) {
 		parisc_kprobe_ss_handler(regs);
 		return;
 	}
 #endif
 
 #ifdef CONFIG_KGDB
-	if (unlikely(iir == PARISC_KGDB_COMPILED_BREAK_INSN ||
-		iir == PARISC_KGDB_BREAK_INSN)) {
+	if (unlikely((iir == PARISC_KGDB_COMPILED_BREAK_INSN ||
+		iir == PARISC_KGDB_BREAK_INSN)) && !user_mode(regs)) {
 		kgdb_handle_exception(9, SIGTRAP, 0, regs);
 		return;
 	}
 #endif
 
+#ifdef CONFIG_LIGHTWEIGHT_SPINLOCK_CHECK
+        if ((iir == SPINLOCK_BREAK_INSN) && !user_mode(regs)) {
+		die_if_kernel("Spinlock was trashed", regs, 1);
+	}
+#endif
+
 	if (unlikely(iir != GDB_BREAK_INSN))
 		parisc_printk_ratelimited(0, regs,
 			KERN_DEBUG "break %d,%d: pid=%d command='%s'\n",
diff --git a/drivers/char/agp/parisc-agp.c b/drivers/char/agp/parisc-agp.c
index d68d05d5d38388..514f9f287a781d 100644
--- a/drivers/char/agp/parisc-agp.c
+++ b/drivers/char/agp/parisc-agp.c
@@ -90,6 +90,9 @@ parisc_agp_tlbflush(struct agp_memory *mem)
 {
 	struct _parisc_agp_info *info = &parisc_agp_info;
 
+	/* force fdc ops to be visible to IOMMU */
+	asm_io_sync();
+
 	writeq(info->gart_base | ilog2(info->gart_size), info->ioc_regs+IOC_PCOM);
 	readq(info->ioc_regs+IOC_PCOM);	/* flush */
 }
@@ -158,6 +161,7 @@ parisc_agp_insert_memory(struct agp_memory *mem, off_t pg_start, int type)
 			info->gatt[j] =
 				parisc_agp_mask_memory(agp_bridge,
 					paddr, type);
+			asm_io_fdc(&info->gatt[j]);
 		}
 	}
 
@@ -191,7 +195,16 @@ static unsigned long
 parisc_agp_mask_memory(struct agp_bridge_data *bridge, dma_addr_t addr,
 		       int type)
 {
-	return SBA_PDIR_VALID_BIT | addr;
+	unsigned ci;			/* coherent index */
+	dma_addr_t pa;
+
+	pa = addr & IOVP_MASK;
+	asm("lci 0(%1), %0" : "=r" (ci) : "r" (phys_to_virt(pa)));
+
+	pa |= (ci >> PAGE_SHIFT) & 0xff;/* move CI (8 bits) into lowest byte */
+	pa |= SBA_PDIR_VALID_BIT;	/* set "valid" bit */
+
+	return cpu_to_le64(pa);
 }
 
 static void