riscv64: Move pcb out of kstack into a new UMA zone

Similar to arm64, riscv's pcb embeds a copy of the floating point
registers and is too large to store directly in struct mdthread as is
done on amd64.  Instead, use a separate UMA zone for pcbs.  riscv's
floating point state is not as large as arm64's, so its pcb is also
somewhat smaller and a single 4k page can hold 6 pcbs.

Reviewed by:	kib, jrtc27
Sponsored by:	AFRL, DARPA
Pull Request:	https://ron-dev.freebsd.org/FreeBSD/src/pulls/23

sys/riscv/include/stack.h

@@ -48,12 +48,10 @@ bool unwind_frame(struct thread *, struct unwind_state *);
 
 #ifdef _SYS_PROC_H_
 
-#include <machine/pcb.h>
-
 /* Get the current kernel thread stack usage. */
 #define	GET_STACK_USAGE(total, used) do {				\
 	struct thread *td = curthread;					\
-	(total) = ptoa(td->td_kstack_pages) - sizeof(struct pcb);	\
+	(total) = ptoa(td->td_kstack_pages);				\
 	(used) = td->td_kstack + (total) - (char *)&td;			\
 } while (0)
 
@@ -61,7 +59,7 @@ static __inline bool
 kstack_contains(struct thread *td, vm_offset_t va, size_t len)
 {
 	return (va >= (vm_offset_t)td->td_kstack && va + len >= va &&
-	    va + len <= (vm_offset_t)td_kstack_top(td) - sizeof(struct pcb));
+	    va + len <= (vm_offset_t)td_kstack_top(td));
 }
 #endif	/* _SYS_PROC_H_ */
 

sys/riscv/riscv/genassym.c

@@ -63,7 +63,6 @@ ASSYM(PMAP_MAPDEV_EARLY_SIZE, PMAP_MAPDEV_EARLY_SIZE);
 ASSYM(PM_SATP, offsetof(struct pmap, pm_satp));
 
 ASSYM(PCB_ONFAULT, offsetof(struct pcb, pcb_onfault));
-ASSYM(PCB_SIZE, sizeof(struct pcb));
 ASSYM(PCB_RA, offsetof(struct pcb, pcb_ra));
 ASSYM(PCB_SP, offsetof(struct pcb, pcb_sp));
 ASSYM(PCB_GP, offsetof(struct pcb, pcb_gp));

sys/riscv/riscv/locore.S

@@ -241,8 +241,8 @@ va:
 	/* Clear frame pointer */
 	mv	s0, zero
 
-	/* Allocate space for thread0 PCB and riscv_bootparams */
-	addi	sp, sp, -(PCB_SIZE + RISCV_BOOTPARAMS_SIZE) & ~STACKALIGNBYTES
+	/* Allocate space for riscv_bootparams */
+	addi	sp, sp, -RISCV_BOOTPARAMS_SIZE & ~STACKALIGNBYTES
 
 	/* Clear BSS */
 	la	t0, _C_LABEL(__bss_start)

sys/riscv/riscv/machdep.c

@@ -105,6 +105,7 @@
 struct pcpu __pcpu[MAXCPU];
 
 static struct trapframe proc0_tf;
+static struct pcb pcb0;
 
 int early_boot = 1;
 int cold = 1;
@@ -296,7 +297,7 @@ init_proc0(void *kstack)
 	proc_linkup0(&proc0, &thread0);
 	thread0.td_kstack = kstack;
 	thread0.td_kstack_pages = KSTACK_PAGES;
-	thread0.td_pcb = (struct pcb *)td_kstack_top(&thread0) - 1;
+	thread0.td_pcb = &pcb0;
 	thread0.td_pcb->pcb_fpflags = 0;
 	thread0.td_frame = &proc0_tf;
 	pcpup->pc_curpcb = thread0.td_pcb;

sys/riscv/riscv/vm_machdep.c

@@ -32,8 +32,8 @@
  * SUCH DAMAGE.
  */
 
-#include <sys/param.h>
 #include <sys/systm.h>
+#include <sys/kernel.h>
 #include <sys/limits.h>
 #include <sys/proc.h>
 #include <sys/sf_buf.h>
@@ -58,23 +58,23 @@
 #define	TP_OFFSET	16	/* sizeof(struct tcb) */
 #endif
 
+static uma_zone_t pcb_zone;
+
 void
 cpu_thread_new_kstack(struct thread *td)
 {
-	td->td_pcb = (struct pcb *)td_kstack_top(td) - 1;
-
 	/*
 	 * td->td_frame + TF_SIZE will be the saved kernel stack pointer whilst
 	 * in userspace, so keep it aligned so it's also aligned when we
 	 * subtract TF_SIZE in the trap handler (and here for the initial stack
 	 * pointer). This also keeps the struct kernframe just afterwards
-	 * aligned no matter what's in it or struct pcb.
+	 * aligned no matter what's in it.
 	 *
 	 * NB: TF_SIZE not sizeof(struct trapframe) as we need the rounded
 	 * value to match the trap handler.
 	 */
 	td->td_frame = (struct trapframe *)(STACKALIGN(
-	    (char *)td->td_pcb - sizeof(struct kernframe)) - TF_SIZE);
+	    td_kstack_top(td) - sizeof(struct kernframe)) - TF_SIZE);
 }
 
 /*
@@ -227,11 +227,13 @@ cpu_thread_exit(struct thread *td)
 void
 cpu_thread_alloc(struct thread *td)
 {
+	td->td_pcb = uma_zalloc(pcb_zone, M_WAITOK);
 }
 
 void
 cpu_thread_free(struct thread *td)
 {
+	uma_zfree(pcb_zone, td->td_pcb);
 }
 
 void
@@ -285,3 +287,11 @@ cpu_sync_core(void)
 {
 	fence_i();
 }
+
+static void
+pcbinit(void *dummy __unused)
+{
+	pcb_zone = uma_zcreate("pcb", sizeof(struct pcb), NULL, NULL, NULL,
+	    NULL, UMA_ALIGNOF(struct pcb), 0);
+}
+SYSINIT(pcbinit, SI_SUB_INTRINSIC, SI_ORDER_ANY, pcbinit, NULL);