[PATCH v7 17/42] powerpc/64s: slb comment update

From: Nicholas Piggin <npiggin@gmail.com>
To: linuxppc-dev@lists.ozlabs.org
Cc: Athira Rajeev <atrajeev@linux.vnet.ibm.com>,
	Nicholas Piggin <npiggin@gmail.com>
Subject: [PATCH v7 17/42] powerpc/64s: slb comment update
Date: Sat, 30 Jan 2021 23:08:27 +1000	[thread overview]
Message-ID: <20210130130852.2952424-18-npiggin@gmail.com> (raw)
In-Reply-To: <20210130130852.2952424-1-npiggin@gmail.com>

This makes a small improvement to the description of the SLB interrupt
environment. Move the memory access restrictions into one paragraph,
and the interrupt restrictions into the next rather than mix them.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/mm/book3s64/slb.c | 28 +++++++++++++++-------------
 1 file changed, 15 insertions(+), 13 deletions(-)

diff --git a/arch/powerpc/mm/book3s64/slb.c b/arch/powerpc/mm/book3s64/slb.c
index c581548b533f..14c62b685f0c 100644
--- a/arch/powerpc/mm/book3s64/slb.c
+++ b/arch/powerpc/mm/book3s64/slb.c
@@ -825,19 +825,21 @@ long do_slb_fault(struct pt_regs *regs)
 		return -EINVAL;
 
 	/*
-	 * SLB kernel faults must be very careful not to touch anything
-	 * that is not bolted. E.g., PACA and global variables are okay,
-	 * mm->context stuff is not.
-	 *
-	 * SLB user faults can access all of kernel memory, but must be
-	 * careful not to touch things like IRQ state because it is not
-	 * "reconciled" here. The difficulty is that we must use
-	 * fast_exception_return to return from kernel SLB faults without
-	 * looking at possible non-bolted memory. We could test user vs
-	 * kernel faults in the interrupt handler asm and do a full fault,
-	 * reconcile, ret_from_except for user faults which would make them
-	 * first class kernel code. But for performance it's probably nicer
-	 * if they go via fast_exception_return too.
+	 * SLB kernel faults must be very careful not to touch anything that is
+	 * not bolted. E.g., PACA and global variables are okay, mm->context
+	 * stuff is not. SLB user faults may access all of memory (and induce
+	 * one recursive SLB kernel fault), so the kernel fault must not
+	 * trample on the user fault state at those points.
+	 */
+
+	/*
+	 * The interrupt state is not reconciled, for performance, so that
+	 * fast_interrupt_return can be used. The handler must not touch local
+	 * irq state, or schedule. We could test for usermode and upgrade to a
+	 * normal process context (synchronous) interrupt for those, which
+	 * would make them first-class kernel code and able to be traced and
+	 * instrumented, although performance would suffer a bit, it would
+	 * probably be a good tradeoff.
 	 */
 	if (id >= LINEAR_MAP_REGION_ID) {
 		long err;
-- 
2.23.0

