linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Paul Burton <paul.burton@imgtec.com>
To: <linux-mips@linux-mips.org>, Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Burton <paul.burton@imgtec.com>,
	<linux-kernel@vger.kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	James Hogan <james.hogan@imgtec.com>,
	Qais Yousef <qsyousef@gmail.com>
Subject: [PATCH v2 19/26] MIPS: Stengthen IPI IRQ domain sanity check
Date: Tue, 30 Aug 2016 18:29:22 +0100	[thread overview]
Message-ID: <20160830172929.16948-20-paul.burton@imgtec.com> (raw)
In-Reply-To: <20160830172929.16948-1-paul.burton@imgtec.com>

Commit fbde2d7d8290 ("MIPS: Add generic SMP IPI support") introduced a
sanity check that an IPI IRQ domain can be found during boot, in order
to ensure that IPIs are able to be set up in systems using such domains.
However it was added at a point where systems may have used an IPI IRQ
domain in some situations but not others, and we could not know which
were the case until runtime, so commit 578bffc82ec5 ("MIPS: Don't BUG_ON
when no IPI domain is found") made that check simply skip IPI init if no
domain were found in order to fix the boot for systems such as QEMU
Malta.

We now use IPI IRQ domains for the MIPS CPU interrupt controller, which
means systems which make use of IPI IRQ domains will always do so when
running on multiple CPUs. As a result we now strengthen the sanity check
to ensure that an IPI IRQ domain is found when multiple CPUs are present
in the system.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
---

Changes in v2: None

 arch/mips/kernel/smp.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

diff --git a/arch/mips/kernel/smp.c b/arch/mips/kernel/smp.c
index f95f094..0155e85 100644
--- a/arch/mips/kernel/smp.c
+++ b/arch/mips/kernel/smp.c
@@ -257,16 +257,20 @@ static int __init mips_smp_ipi_init(void)
 		ipidomain = irq_find_matching_host(NULL, DOMAIN_BUS_IPI);
 
 	/*
-	 * There are systems which only use IPI domains some of the time,
-	 * depending upon configuration we don't know until runtime. An
-	 * example is Malta where we may compile in support for GIC & the
-	 * MT ASE, but run on a system which has multiple VPEs in a single
-	 * core and doesn't include a GIC. Until all IPI implementations
-	 * have been converted to use IPI domains the best we can do here
-	 * is to return & hope some other code sets up the IPIs.
+	 * There are systems which use IPI IRQ domains, but only have one
+	 * registered when some runtime condition is met. For example a Malta
+	 * kernel may include support for GIC & CPU interrupt controller IPI
+	 * IRQ domains, but if run on a system with no GIC & no MT ASE then
+	 * neither will be supported or registered.
+	 *
+	 * We only have a problem if we're actually using multiple CPUs so fail
+	 * loudly if that is the case. Otherwise simply return, skipping IPI
+	 * setup, if we're running with only a single CPU.
 	 */
-	if (!ipidomain)
+	if (!ipidomain) {
+		BUG_ON(num_present_cpus() > 1);
 		return 0;
+	}
 
 	call_virq = irq_reserve_ipi(ipidomain, cpu_possible_mask);
 	BUG_ON(!call_virq);
-- 
2.9.3

  parent reply	other threads:[~2016-08-30 17:35 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-08-30 17:29 [PATCH v2 00/26] MIPS generic kernels, SEAD-3 & Boston support Paul Burton
2016-08-30 17:29 ` [PATCH v2 01/26] MIPS: PCI: Support for CONFIG_PCI_DOMAINS_GENERIC Paul Burton
2016-08-30 17:29 ` [PATCH v2 02/26] MIPS: PCI: Make pcibios_set_cache_line_size an initcall Paul Burton
2016-08-30 17:29 ` [PATCH v2 03/26] MIPS: PCI: Inline pcibios_assign_all_busses Paul Burton
2016-08-30 17:29 ` [PATCH v2 04/26] MIPS: PCI: Split pci.c into pci.c & pci-legacy.c Paul Burton
2016-08-30 17:29 ` [PATCH v2 05/26] MIPS: PCI: Introduce CONFIG_PCI_DRIVERS_LEGACY Paul Burton
2016-08-30 17:29 ` [PATCH v2 06/26] MIPS: PCI: Support generic drivers Paul Burton
2016-08-30 17:29 ` [PATCH v2 07/26] MIPS: Sanitise coherentio semantics Paul Burton
2016-08-30 17:29 ` [PATCH v2 08/26] MIPS: dma-default: Don't check hw_coherentio if device is non-coherent Paul Burton
2016-08-30 17:29 ` [PATCH v2 09/26] MIPS: Support per-device DMA coherence Paul Burton
2016-08-30 17:29 ` [PATCH v2 10/26] MIPS: Print CM error reports upon bus errors Paul Burton
2016-08-30 17:29 ` [PATCH v2 11/26] dt-bindings: Document mti,mips-cpc binding Paul Burton
2016-08-30 17:29 ` [PATCH v2 12/26] MIPS: CPC: Provide a default mips_cpc_default_phys_base Paul Burton
2016-08-31 10:21   ` Sergei Shtylyov
2016-08-30 17:29 ` [PATCH v2 13/26] dt-bindings: Document mti,mips-cdmm binding Paul Burton
2016-08-30 17:29 ` [PATCH v2 14/26] MIPS: CDMM: Allow CDMM base address to be specified via DT Paul Burton
2016-08-30 17:29 ` [PATCH v2 15/26] irqchip: mips-cpu: Replace magic 0x100 with IE_SW0 Paul Burton
2016-08-30 17:29 ` [PATCH v2 16/26] irqchip: mips-cpu: Prepare for non-legacy IRQ domains Paul Burton
2016-08-30 17:29 ` [PATCH v2 17/26] irqchip: mips-cpu: Introduce IPI IRQ domain support Paul Burton
2016-09-01 19:46   ` Jason Cooper
2016-08-30 17:29 ` [PATCH v2 18/26] MIPS: smp-mt: Use CPU interrupt controller " Paul Burton
2016-08-30 17:29 ` Paul Burton [this message]
2016-08-30 17:29 ` [PATCH v2 20/26] MIPS: Adjust MIPS64 CAC_BASE to reflect Config.K0 Paul Burton
2016-08-30 17:29 ` [PATCH v2 21/26] MIPS: Support generating Flattened Image Trees (.itb) Paul Burton
2016-08-30 17:29 ` [PATCH v2 22/26] MIPS: generic: Introduce generic DT-based board support Paul Burton
2016-08-30 17:29 ` [PATCH v2 23/26] MIPS: generic: Convert SEAD-3 to a generic board Paul Burton
2016-08-30 17:29 ` [PATCH v2 24/26] dt-bindings: Document img,boston-clock binding Paul Burton
2016-08-30 17:29 ` [PATCH v2 25/26] clk: boston: Add a driver for MIPS Boston board clocks Paul Burton
2016-08-30 17:29 ` [PATCH v2 26/26] MIPS: generic: Support MIPS Boston development boards Paul Burton

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=20160830172929.16948-20-paul.burton@imgtec.com \
    --to=paul.burton@imgtec.com \
    --cc=james.hogan@imgtec.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mips@linux-mips.org \
    --cc=qsyousef@gmail.com \
    --cc=ralf@linux-mips.org \
    --cc=tglx@linutronix.de \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).