All of lore.kernel.org
 help / color / mirror / Atom feed
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 01/42] powerpc/64s: interrupt exit improve bounding of interrupt recursion
Date: Sat, 30 Jan 2021 23:08:11 +1000	[thread overview]
Message-ID: <20210130130852.2952424-2-npiggin@gmail.com> (raw)
In-Reply-To: <20210130130852.2952424-1-npiggin@gmail.com>

When replaying pending soft-masked interrupts when an interrupt returns
to an irqs-enabled context, there is a special case required if this was
an asynchronous interrupt to avoid unbounded interrupt recursion.

This case was not tested for in the case the asynchronous interrupt hit
in user context, because a subsequent nested interrupt would by definition
hit in kernel mode, which then exits via the kernel path which does test
this case.

There is no reason to allow this for such interrupts. While recursion is
bounded at the next level, it's simpler and uses less stack to apply the
replay logic consistently.

This also expands the comment which was really pretty poor and didn't
explain the problem (I can say that because I wrote it).

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/kernel/syscall_64.c | 55 +++++++++++++++++++-------------
 1 file changed, 33 insertions(+), 22 deletions(-)

diff --git a/arch/powerpc/kernel/syscall_64.c b/arch/powerpc/kernel/syscall_64.c
index 7c85ed04a164..e0eb2a502db3 100644
--- a/arch/powerpc/kernel/syscall_64.c
+++ b/arch/powerpc/kernel/syscall_64.c
@@ -138,8 +138,12 @@ notrace long system_call_exception(long r3, long r4, long r5,
 /*
  * local irqs must be disabled. Returns false if the caller must re-enable
  * them, check for new work, and try again.
+ *
+ * This should be called with local irqs disabled, but if they were previously
+ * enabled when the interrupt handler returns (indicating a process-context /
+ * synchronous interrupt) then irqs_enabled should be true.
  */
-static notrace inline bool prep_irq_for_enabled_exit(bool clear_ri)
+static notrace inline bool prep_irq_for_enabled_exit(bool clear_ri, bool irqs_enabled)
 {
 	/* This must be done with RI=1 because tracing may touch vmaps */
 	trace_hardirqs_on();
@@ -156,6 +160,29 @@ static notrace inline bool prep_irq_for_enabled_exit(bool clear_ri)
 		trace_hardirqs_off();
 		local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
 
+		/*
+		 * Must replay pending soft-masked interrupts now. Don't just
+		 * local_irq_enabe(); local_irq_disable(); because if we are
+		 * returning from an asynchronous interrupt here, another one
+		 * might hit after irqs are enabled, and it would exit via this
+		 * same path allowing another to fire, and so on unbounded.
+		 *
+		 * If interrupts were enabled when this interrupt exited,
+		 * indicating a process context (synchronous) interrupt,
+		 * local_irq_enable/disable can be used, which will enable
+		 * interrupts rather than keeping them masked (unclear how
+		 * much benefit this is over just replaying for all cases,
+		 * because we immediately disable again, so all we're really
+		 * doing is allowing hard interrupts to execute directly for
+		 * a very small time, rather than being masked and replayed).
+		 */
+		if (irqs_enabled) {
+			local_irq_enable();
+			local_irq_disable();
+		} else {
+			replay_soft_interrupts();
+		}
+
 		return false;
 	}
 	local_paca->irq_happened = 0;
@@ -212,8 +239,9 @@ notrace unsigned long syscall_exit_prepare(unsigned long r3,
 		ret |= _TIF_RESTOREALL;
 	}
 
-again:
 	local_irq_disable();
+
+again:
 	ti_flags = READ_ONCE(*ti_flagsp);
 	while (unlikely(ti_flags & (_TIF_USER_WORK_MASK & ~_TIF_RESTORE_TM))) {
 		local_irq_enable();
@@ -258,10 +286,8 @@ notrace unsigned long syscall_exit_prepare(unsigned long r3,
 	}
 
 	/* scv need not set RI=0 because SRRs are not used */
-	if (unlikely(!prep_irq_for_enabled_exit(!scv))) {
-		local_irq_enable();
+	if (unlikely(!prep_irq_for_enabled_exit(!scv, true)))
 		goto again;
-	}
 
 #ifdef CONFIG_PPC_TRANSACTIONAL_MEM
 	local_paca->tm_scratch = regs->msr;
@@ -336,11 +362,8 @@ notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned
 		}
 	}
 
-	if (unlikely(!prep_irq_for_enabled_exit(true))) {
-		local_irq_enable();
-		local_irq_disable();
+	if (unlikely(!prep_irq_for_enabled_exit(true, !irqs_disabled_flags(flags))))
 		goto again;
-	}
 
 #ifdef CONFIG_PPC_BOOK3E
 	if (unlikely(ts->debug.dbcr0 & DBCR0_IDM)) {
@@ -403,20 +426,8 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsign
 			}
 		}
 
-		if (unlikely(!prep_irq_for_enabled_exit(true))) {
-			/*
-			 * Can't local_irq_restore to replay if we were in
-			 * interrupt context. Must replay directly.
-			 */
-			if (irqs_disabled_flags(flags)) {
-				replay_soft_interrupts();
-			} else {
-				local_irq_restore(flags);
-				local_irq_save(flags);
-			}
-			/* Took an interrupt, may have more exit work to do. */
+		if (unlikely(!prep_irq_for_enabled_exit(true, !irqs_disabled_flags(flags))))
 			goto again;
-		}
 	} else {
 		/* Returning to a kernel context with local irqs disabled. */
 		__hard_EE_RI_disable();
-- 
2.23.0


  reply	other threads:[~2021-01-30 13:18 UTC|newest]

Thread overview: 62+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-01-30 13:08 [PATCH v7 00/42] powerpc: interrupt wrappers Nicholas Piggin
2021-01-30 13:08 ` Nicholas Piggin [this message]
2021-01-30 13:08 ` [PATCH v7 02/42] KVM: PPC: Book3S HV: Context tracking exit guest context before enabling irqs Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 03/42] powerpc/32s: move DABR match out of handle_page_fault Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 04/42] powerpc/64s: " Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 05/42] powerpc/64s: move the hash fault handling logic to C Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 06/42] powerpc: remove arguments from fault handler functions Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 07/42] powerpc/fsl_booke/32: CacheLockingException remove args Nicholas Piggin
2021-02-08 10:04   ` Gautham R Shenoy
2021-02-08 11:54     ` Michael Ellerman
2021-01-30 13:08 ` [PATCH v7 08/42] powerpc: do_break get registers from regs Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 09/42] powerpc: DebugException remove args Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 10/42] powerpc/32: transfer can avoid saving r4/r5 over trace call Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 11/42] powerpc: bad_page_fault get registers from regs Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 12/42] powerpc/64s: add do_bad_page_fault_segv handler Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 13/42] powerpc: rearrange do_page_fault error case to be inside exception_enter Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 14/42] powerpc/64s: move bad_page_fault handling to C Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 15/42] powerpc/64s: split do_hash_fault Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 16/42] powerpc/mm: Remove stale do_page_fault comment referring to SLB faults Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 17/42] powerpc/64s: slb comment update Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 18/42] powerpc/traps: add NOKPROBE_SYMBOL for sreset and mce Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 19/42] powerpc/perf: move perf irq/nmi handling details into traps.c Nicholas Piggin
2021-01-31 12:30   ` Athira Rajeev
2021-01-30 13:08 ` [PATCH v7 20/42] powerpc/time: move timer_broadcast_interrupt prototype to asm/time.h Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 21/42] powerpc: add and use unknown_async_exception Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 22/42] powerpc/cell: tidy up pervasive declarations Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 23/42] powerpc: introduce die_mce Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 24/42] powerpc/mce: ensure machine check handler always tests RI Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 25/42] powerpc: improve handling of unrecoverable system reset Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 26/42] powerpc: interrupt handler wrapper functions Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 27/42] powerpc: add interrupt wrapper entry / exit stub functions Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 28/42] powerpc: convert interrupt handlers to use wrappers Nicholas Piggin
2021-02-03  1:54   ` Nicholas Piggin
2021-02-05  8:09   ` Christophe Leroy
2021-02-06  2:43     ` Nicholas Piggin
2021-02-07 12:56     ` Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 29/42] powerpc: add interrupt_cond_local_irq_enable helper Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 30/42] powerpc/64: context tracking remove _TIF_NOHZ Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 31/42] powerpc/64s/hash: improve context tracking of hash faults Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 32/42] powerpc/64: context tracking move to interrupt wrappers Nicholas Piggin
2021-02-09  5:49   ` Christophe Leroy
2021-02-09  7:45     ` Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 33/42] powerpc/64: add context tracking to asynchronous interrupts Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 34/42] powerpc: handle irq_enter/irq_exit in interrupt handler wrappers Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 35/42] powerpc/64s: move context tracking exit to interrupt exit path Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 36/42] powerpc/64s: reconcile interrupts in C Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 37/42] powerpc/64: move account_stolen_time into its own function Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 38/42] powerpc/64: entry cpu time accounting in C Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 39/42] powerpc: move NMI entry/exit code into wrapper Nicholas Piggin
2021-02-04 10:15   ` Michael Ellerman
2021-02-04 11:31     ` Nicholas Piggin
2021-02-05 23:38       ` Michael Ellerman
2021-02-06  2:46         ` Nicholas Piggin
2021-02-07 12:54           ` Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 40/42] powerpc/64s: move NMI soft-mask handling to C Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 41/42] powerpc/64s: runlatch interrupt handling in C Nicholas Piggin
2021-01-30 13:08 ` [PATCH v7 42/42] powerpc/64s: power4 nap fixup " Nicholas Piggin
2021-02-02 10:31   ` Michael Ellerman
2021-02-03  0:35     ` Nicholas Piggin
2021-02-07 12:58     ` Nicholas Piggin
2021-02-02  5:57 ` [PATCH v7 00/42] powerpc: interrupt wrappers Christophe Leroy
2021-02-10 12:57 ` Michael Ellerman

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210130130852.2952424-2-npiggin@gmail.com \
    --to=npiggin@gmail.com \
    --cc=atrajeev@linux.vnet.ibm.com \
    --cc=linuxppc-dev@lists.ozlabs.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.