sparc: Fix typos

Fix typos, most reported by "codespell arch/sparc". Only touches
comments, no code changes.

Signed-off-by: Bjorn Helgaas <[e-mail]>
Cc: [e-mail]
Reviewed-by: Randy Dunlap <[e-mail]>
Signed-off-by: Andreas Larsson <[e-mail]>
Link: https://lore.kernel.org/r/[e-mail]

arch/sparc/include/asm/hypervisor.h

···
  * ERRORS:	No errors defined.
  *
  * Return the hypervisor ID handle for the current CPU.  Use by a
- * virtual CPU to discover it's own identity.
+ * virtual CPU to discover its own identity.
  */
 #define HV_FAST_CPU_MYID		0x16
 
···
  *		EBADALIGNED	software state description is not correctly
  *				aligned
  *
- * This allows the guest to report it's soft state to the hypervisor.  There
+ * This allows the guest to report its soft state to the hypervisor.  There
  * are two primary components to this state.  The first part states whether
  * the guest software is running or not.  The second containts optional
  * details specific to the software.
···
  * configuration error of some sort.
  *
  * The dump services provide an opaque buffer into which the
- * hypervisor can place it's internal state in order to assist in
+ * hypervisor can place its internal state in order to assist in
  * debugging such situations.  The contents are opaque and extremely
  * platform and hypervisor implementation specific.  The guest, during
  * a core dump, requests that the hypervisor update any information in

arch/sparc/include/asm/ldc.h

···
  * or data becomes available on the receive side.
  *
  * For non-RAW links, if the LDC_EVENT_RESET event arrives the
- * driver should reset all of it's internal state and reinvoke
+ * driver should reset all of its internal state and reinvoke
  * ldc_connect() to try and bring the link up again.
  *
  * For RAW links, ldc_connect() is not used.  Instead the driver

arch/sparc/include/asm/mmu_context_64.h

···
 
 	/* We have to be extremely careful here or else we will miss
 	 * a TSB grow if we switch back and forth between a kernel
-	 * thread and an address space which has it's TSB size increased
+	 * thread and an address space which has its TSB size increased
 	 * on another processor.
 	 *
 	 * It is possible to play some games in order to optimize the
···
 	 *
 	 * At that point cpu0 continues to use a stale TSB, the one from
 	 * before the TSB grow performed on cpu1.  cpu1 did not cross-call
-	 * cpu0 to update it's TSB because at that point the cpu_vm_mask
+	 * cpu0 to update its TSB because at that point the cpu_vm_mask
 	 * only had cpu1 set in it.
 	 */
 	tsb_context_switch_ctx(mm, CTX_HWBITS(mm->context));

arch/sparc/include/asm/switch_to_64.h

···
 	 * for l0/l1.  It will use one for 'next' and the other to hold
 	 * the output value of 'last'.  'next' is not referenced again
 	 * past the invocation of switch_to in the scheduler, so we need
-	 * not preserve it's value.  Hairy, but it lets us remove 2 loads
+	 * not preserve its value.  Hairy, but it lets us remove 2 loads
 	 * and 2 stores in this critical code path.  -DaveM
 	 */
 #define switch_to(prev, next, last)					\

arch/sparc/kernel/irq_64.c

···
  *
  * On SMP this gets invoked from the CPU trampoline before
  * the cpu has fully taken over the trap table from OBP,
- * and it's kernel stack + %g6 thread register state is
+ * and its kernel stack + %g6 thread register state is
  * not fully cooked yet.
  *
  * Therefore you cannot make any OBP calls, not even prom_printf,

arch/sparc/kernel/kprobes.c

···
 	return regs->tnpc;
 }
 
-/* If INSN is an instruction which writes it's PC location
+/* If INSN is an instruction which writes its PC location
  * into a destination register, fix that up.
  */
 static void __kprobes retpc_fixup(struct pt_regs *regs, u32 insn,

arch/sparc/kernel/ldc.c

···
 			 * This seems the best behavior because this allows
 			 * a user of the LDC layer to start with a small
 			 * RX buffer for ldc_read() calls and use -EMSGSIZE
-			 * as a cue to enlarge it's read buffer.
+			 * as a cue to enlarge its read buffer.
 			 */
 			err = -EMSGSIZE;
 			break;

arch/sparc/kernel/leon_pci_grpci2.c

···
 	REGSTORE(regs->io_map, REGLOAD(regs->io_map) & 0x0000ffff);
 
 	/* set 1:1 mapping between AHB -> PCI memory space, for all Masters
-	 * Each AHB master has it's own mapping registers. Max 16 AHB masters.
+	 * Each AHB master has its own mapping registers. Max 16 AHB masters.
 	 */
 	for (i = 0; i < 16; i++)
 		REGSTORE(regs->ahbmst_map[i], priv->pci_area);

arch/sparc/kernel/of_device_64.c

···
 	 *
 	 * If we hit a bus type or situation we cannot handle, we
 	 * stop and assume that the original IRQ number was in a
-	 * format which has special meaning to it's immediate parent.
+	 * format which has special meaning to its immediate parent.
 	 */
 	pp = dp->parent;
 	ip = NULL;

arch/sparc/kernel/pci.c

···
 	/* We can't actually use the firmware value, we have
 	 * to read what is in the register right now.  One
 	 * reason is that in the case of IDE interfaces the
-	 * firmware can sample the value before the the IDE
+	 * firmware can sample the value before the IDE
 	 * interface is programmed into native mode.
 	 */
 	pci_read_config_dword(dev, PCI_CLASS_REVISION, &class);

arch/sparc/kernel/pci_impl.h

···
  * each with one (Sabre) or two (PSYCHO/SCHIZO) PCI bus modules
  * underneath.  Each PCI bus module uses an IOMMU (shared by both
  * PBMs of a controller, or per-PBM), and if a streaming buffer
- * is present, each PCI bus module has it's own. (ie. the IOMMU
+ * is present, each PCI bus module has its own. (ie. the IOMMU
  * might be shared between PBMs, the STC is never shared)
- * Furthermore, each PCI bus module controls it's own autonomous
+ * Furthermore, each PCI bus module controls its own autonomous
  * PCI bus.
  */
 

arch/sparc/kernel/pci_schizo.c

···
 
 	/* This is __REALLY__ dangerous.  When we put the
 	 * streaming buffer into diagnostic mode to probe
-	 * it's tags and error status, we _must_ clear all
+	 * its tags and error status, we _must_ clear all
 	 * of the line tag valid bits before re-enabling
 	 * the streaming buffer.  If any dirty data lives
 	 * in the STC when we do this, we will end up
···
 		       pbm->name, type_string);
 
 		/* Put the IOMMU into diagnostic mode and probe
-		 * it's TLB for entries with error status.
+		 * its TLB for entries with error status.
 		 *
 		 * It is very possible for another DVMA to occur
 		 * while we do this probe, and corrupt the system

arch/sparc/kernel/perf_event.c

···
 
 static void sparc_pmu_start(struct perf_event *event, int flags);
 
-/* On this PMU each PIC has it's own PCR control register.  */
+/* On this PMU each PIC has its own PCR control register.  */
 static void calculate_multiple_pcrs(struct cpu_hw_events *cpuc)
 {
 	int i;

arch/sparc/kernel/prom_irqtrans.c

···
 	iclr = schizo_ino_to_iclr(pbm_regs, ino);
 
 	/* On Schizo, no inofixup occurs.  This is because each
-	 * INO has it's own IMAP register.  On Psycho and Sabre
+	 * INO has its own IMAP register.  On Psycho and Sabre
 	 * there is only one IMAP register for each PCI slot even
 	 * though four different INOs can be generated by each
 	 * PCI slot.

arch/sparc/kernel/psycho_common.c

···
 	spin_lock(&stc_buf_lock);
 
 	/* This is __REALLY__ dangerous.  When we put the streaming
-	 * buffer into diagnostic mode to probe it's tags and error
+	 * buffer into diagnostic mode to probe its tags and error
 	 * status, we _must_ clear all of the line tag valid bits
 	 * before re-enabling the streaming buffer.  If any dirty data
 	 * lives in the STC when we do this, we will end up

arch/sparc/kernel/signal_32.c

···
 	 *
 	 * %g7 is used as the "thread register".   %g6 is not used in
 	 * any fixed manner.  %g6 is used as a scratch register and
-	 * a compiler temporary, but it's value is never used across
+	 * a compiler temporary, but its value is never used across
 	 * a system call.  Therefore %g6 is usable for orig_i0 storage.
 	 */
 	if (pt_regs_is_syscall(regs) && (regs->psr & PSR_C))

arch/sparc/kernel/signal_64.c

···
 	 *
 	 * %g7 is used as the "thread register".   %g6 is not used in
 	 * any fixed manner.  %g6 is used as a scratch register and
-	 * a compiler temporary, but it's value is never used across
+	 * a compiler temporary, but its value is never used across
 	 * a system call.  Therefore %g6 is usable for orig_i0 storage.
 	 */
 	if (pt_regs_is_syscall(regs) &&

arch/sparc/mm/srmmu.c

···
 
 		/*
 		 * We need this to make sure old viking takes no hits
-		 * on it's cache for dma snoops to workaround the
+		 * on its cache for dma snoops to workaround the
 		 * "load from non-cacheable memory" interrupt bug.
 		 * This is only necessary because of the new way in
 		 * which we use the IOMMU.

arch/sparc/mm/tsb.c

···
  * will not trigger any longer.
  *
  * The TSB can be anywhere from 8K to 1MB in size, in increasing powers
- * of two.  The TSB must be aligned to it's size, so f.e. a 512K TSB
+ * of two.  The TSB must be aligned to its size, so f.e. a 512K TSB
  * must be 512K aligned.  It also must be physically contiguous, so we
  * cannot use vmalloc().
  *

arch/sparc/net/bpf_jit_comp_32.c

···
  *
  * The most common case is to emit a branch at the end of such
  * a code sequence.  So this would be two instructions, the
- * branch and it's delay slot.
+ * branch and its delay slot.
  *
  * Therefore by default the branch emitters calculate the branch
  * offset field as:
···
  *
  * This "addrs[i] - 8" is the address of the branch itself or
  * what "." would be in assembler notation.  The "8" part is
- * how we take into consideration the branch and it's delay
+ * how we take into consideration the branch and its delay
  * slot mentioned above.
  *
  * Sometimes we need to emit a branch earlier in the code
  * sequence.  And in these situations we adjust "destination"
  * to accommodate this difference.  For example, if we needed
- * to emit a branch (and it's delay slot) right before the
+ * to emit a branch (and its delay slot) right before the
  * final instruction emitted for a BPF opcode, we'd use
  * "destination + 4" instead of just plain "destination" above.
  *