Date: Fri, 9 Jun 2017 11:38:32 +0000
From: Jayachandran C <jnair@caviumnetworks.com>
To: Robin Murphy <robin.murphy@arm.com>
Cc: Geetha sowjanya <gakula@caviumnetworks.com>, will.deacon@arm.com,
        lorenzo.pieralisi@arm.com, hanjun.guo@linaro.org, sudeep.holla@arm.com,
        iommu@lists.linux-foundation.org, robh@kernel.org,
        Charles.Garcia-Tobin@arm.com,
        Geetha Sowjanya <geethasowjanya.akula@cavium.com>,
        geethasowjanya.akula@gmail.com, jcm@redhat.com,
        linu.cherian@cavium.com, rjw@rjwysocki.net, robert.moore@intel.com,
        linux-kernel@vger.kernel.org, linux-acpi@vger.kernel.org,
        robert.richter@cavium.com, lv.zheng@intel.com, catalin.marinas@arm.com,
        sgoutham@cavium.com, linux-arm-kernel@lists.infradead.org,
        devel@acpica.org
Subject: Re: Fwd: [PATCH v7 2/3] iommu/arm-smmu-v3: Add workaround for Cavium
 ThunderX2 erratum #74
Message-ID: <20170609113832.GB106079@localhost>
References: <1496145821-3411-1-git-send-email-gakula@caviumnetworks.com>
 <1496145821-3411-3-git-send-email-gakula@caviumnetworks.com>
 <9263562d-03a3-d02e-6b7e-01613a4bd296@arm.com>
 <CA+7sy7C_44dTy0-nAE=b5BCXmc8ACQx2O6A1jCOCsemZDD5j4w@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CA+7sy7C_44dTy0-nAE=b5BCXmc8ACQx2O6A1jCOCsemZDD5j4w@mail.gmail.com>
User-Agent: Mutt/1.5.24 (2015-08-30)
SpamDiagnosticOutput: 1:99
SpamDiagnosticMetadata: NSPM
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 09 Jun 2017 11:38:38.7859 (UTC)
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-Transport-CrossTenantHeadersStamped: CY4PR07MB2999
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org
Content-Length: 12180
Lines: 298

On Fri, Jun 09, 2017 Robin Murphy wrote:
> 
> On 30/05/17 13:03, Geetha sowjanya wrote:
> > From: Linu Cherian <linu.cherian@cavium.com>
> >
> > Cavium ThunderX2 SMMU implementation doesn't support page 1 register space
> > and PAGE0_REGS_ONLY option is enabled as an errata workaround.
> > This option when turned on, replaces all page 1 offsets used for
> > EVTQ_PROD/CONS, PRIQ_PROD/CONS register access with page 0 offsets.
> >
> > SMMU resource size checks are now based on SMMU option PAGE0_REGS_ONLY,
> > since resource size can be either 64k/128k.
> > For this, arm_smmu_device_dt_probe/acpi_probe has been moved before
> > platform_get_resource call, so that SMMU options are set beforehand.
> >
> > Signed-off-by: Linu Cherian <linu.cherian@cavium.com>
> > Signed-off-by: Geetha Sowjanya <geethasowjanya.akula@cavium.com>
> > ---
> >  Documentation/arm64/silicon-errata.txt             |    1 +
> >  .../devicetree/bindings/iommu/arm,smmu-v3.txt      |    6 ++
> >  drivers/iommu/arm-smmu-v3.c                        |   64 +++++++++++++++-----
> >  3 files changed, 56 insertions(+), 15 deletions(-)
> >
> > diff --git a/Documentation/arm64/silicon-errata.txt b/Documentation/arm64/silicon-errata.txt
> > index 10f2ddd..4693a32 100644
> > --- a/Documentation/arm64/silicon-errata.txt
> > +++ b/Documentation/arm64/silicon-errata.txt
> > @@ -62,6 +62,7 @@ stable kernels.
> >  | Cavium         | ThunderX GICv3  | #23154          | CAVIUM_ERRATUM_23154        |
> >  | Cavium         | ThunderX Core   | #27456          | CAVIUM_ERRATUM_27456        |
> >  | Cavium         | ThunderX SMMUv2 | #27704          | N/A                         |
> > +| Cavium         | ThunderX2 SMMUv3| #74             | N/A                         |
> >  |                |                 |                 |                             |
> >  | Freescale/NXP  | LS2080A/LS1043A | A-008585        | FSL_ERRATUM_A008585         |
> >  |                |                 |                 |                             |
> > diff --git a/Documentation/devicetree/bindings/iommu/arm,smmu-v3.txt b/Documentation/devicetree/bindings/iommu/arm,smmu-v3.txt
> > index be57550..607e270 100644
> > --- a/Documentation/devicetree/bindings/iommu/arm,smmu-v3.txt
> > +++ b/Documentation/devicetree/bindings/iommu/arm,smmu-v3.txt
> > @@ -49,6 +49,12 @@ the PCIe specification.
> >  - hisilicon,broken-prefetch-cmd
> >                      : Avoid sending CMD_PREFETCH_* commands to the SMMU.
> >
> > +- cavium,cn9900-broken-page1-regspace
> > +                    : Replaces all page 1 offsets used for EVTQ_PROD/CONS,
> > +                                             PRIQ_PROD/CONS register access with page 0 offsets.
> > +                                             Set for Caviun ThunderX2 silicon that doesn't support
> > +                                             SMMU page1 register space.
> 
> The indentation's a bit funky here - the rest of this file is actually
> indented with spaces, but either way it's clear your editor isn't set to
> 8-space tabs ;)
> 
> > +
> >  ** Example
> >
> >          smmu@2b400000 {
> > diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c
> > index 380969a..4e80205 100644
> > --- a/drivers/iommu/arm-smmu-v3.c
> > +++ b/drivers/iommu/arm-smmu-v3.c
> > @@ -412,6 +412,9 @@
> >  #define MSI_IOVA_BASE                        0x8000000
> >  #define MSI_IOVA_LENGTH                      0x100000
> >
> > +#define ARM_SMMU_PAGE0_REGS_ONLY(smmu)               \
> > +     ((smmu)->options & ARM_SMMU_OPT_PAGE0_REGS_ONLY)
> 
> At the two places we use this macro, frankly I think it would be clearer
> to just reference smmu->options directly, as we currently do for
> SKIP_PREFETCH. The abstraction also adds more lines than it saves...
> 
> > +
> >  static bool disable_bypass;
> >  module_param_named(disable_bypass, disable_bypass, bool, S_IRUGO);
> >  MODULE_PARM_DESC(disable_bypass,
> > @@ -597,6 +600,7 @@ struct arm_smmu_device {
> >       u32                             features;
> >
> >  #define ARM_SMMU_OPT_SKIP_PREFETCH   (1 << 0)
> > +#define ARM_SMMU_OPT_PAGE0_REGS_ONLY    (1 << 1)
> 
> Whitespace again, although this time it's spaces where there should be a
> tab.
> 
> >       u32                             options;
> >
> >       struct arm_smmu_cmdq            cmdq;
> > @@ -663,9 +667,19 @@ struct arm_smmu_option_prop {
> >
> >  static struct arm_smmu_option_prop arm_smmu_options[] = {
> >       { ARM_SMMU_OPT_SKIP_PREFETCH, "hisilicon,broken-prefetch-cmd" },
> > +     { ARM_SMMU_OPT_PAGE0_REGS_ONLY, "cavium,cn9900-broken-page1-regspace"},
> >       { 0, NULL},
> >  };
> >
> > +static inline void __iomem *arm_smmu_page1_fixup(unsigned long offset,
> > +                                              struct arm_smmu_device *smmu)
> > +{
> > +     if (offset > SZ_64K && ARM_SMMU_PAGE0_REGS_ONLY(smmu))
> > +             offset -= SZ_64K;
> > +
> > +     return smmu->base + offset;
> > +}
> > +
> >  static struct arm_smmu_domain *to_smmu_domain(struct iommu_domain *dom)
> >  {
> >       return container_of(dom, struct arm_smmu_domain, domain);
> > @@ -1961,8 +1975,8 @@ static int arm_smmu_init_one_queue(struct arm_smmu_device *smmu,
> >               return -ENOMEM;
> >       }
> >
> > -     q->prod_reg     = smmu->base + prod_off;
> > -     q->cons_reg     = smmu->base + cons_off;
> > +     q->prod_reg     = arm_smmu_page1_fixup(prod_off, smmu);
> > +     q->cons_reg     = arm_smmu_page1_fixup(cons_off, smmu);
> >       q->ent_dwords   = dwords;
> >
> >       q->q_base  = Q_BASE_RWA;
> > @@ -2363,8 +2377,10 @@ static int arm_smmu_device_reset(struct arm_smmu_device *smmu, bool bypass)
> >
> >       /* Event queue */
> >       writeq_relaxed(smmu->evtq.q.q_base, smmu->base + ARM_SMMU_EVTQ_BASE);
> > -     writel_relaxed(smmu->evtq.q.prod, smmu->base + ARM_SMMU_EVTQ_PROD);
> > -     writel_relaxed(smmu->evtq.q.cons, smmu->base + ARM_SMMU_EVTQ_CONS);
> > +     writel_relaxed(smmu->evtq.q.prod,
> > +                    arm_smmu_page1_fixup(ARM_SMMU_EVTQ_PROD, smmu));
> > +     writel_relaxed(smmu->evtq.q.cons,
> > +                    arm_smmu_page1_fixup(ARM_SMMU_EVTQ_CONS, smmu));

This sequence and the arm_smmu_page1_fixup() call is repeated in quite a few
places. I think this errata code is messy because the original driver does not
make the alias page usage explicit. 

A patch like the one below (no functional changes) would clean up the original
driver and make the errata change much simpler - any comments?

-- >8 --

Date: Tue, 30 May 2017 15:43:29 +0000
Subject: [PATCH] iommu: arm-smmu-v3: make alias page usage explicit

---
 drivers/iommu/arm-smmu-v3.c | 76 +++++++++++++++++++++++++++------------------
 1 file changed, 46 insertions(+), 30 deletions(-)

diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c
index 380969a..11fdb4f 100644
--- a/drivers/iommu/arm-smmu-v3.c
+++ b/drivers/iommu/arm-smmu-v3.c
@@ -171,20 +171,19 @@
 #define STRTAB_BASE_CFG_FMT_LINEAR	(0 << STRTAB_BASE_CFG_FMT_SHIFT)
 #define STRTAB_BASE_CFG_FMT_2LVL	(1 << STRTAB_BASE_CFG_FMT_SHIFT)
 
+#define ARM_SMMU_Q_PROD(qbase)		((qbase) + 0x8)
+#define ARM_SMMU_Q_PROD_PAGE1(qbase)	((qbase) + 0x10008)
+#define ARM_SMMU_Q_CONS(qbase)		((qbase) + 0xc)
+#define ARM_SMMU_Q_CONS_PAGE1(qbase)	((qbase) + 0x1000c)
+
 #define ARM_SMMU_CMDQ_BASE		0x90
-#define ARM_SMMU_CMDQ_PROD		0x98
-#define ARM_SMMU_CMDQ_CONS		0x9c
 
 #define ARM_SMMU_EVTQ_BASE		0xa0
-#define ARM_SMMU_EVTQ_PROD		0x100a8
-#define ARM_SMMU_EVTQ_CONS		0x100ac
 #define ARM_SMMU_EVTQ_IRQ_CFG0		0xb0
 #define ARM_SMMU_EVTQ_IRQ_CFG1		0xb8
 #define ARM_SMMU_EVTQ_IRQ_CFG2		0xbc
 
 #define ARM_SMMU_PRIQ_BASE		0xc0
-#define ARM_SMMU_PRIQ_PROD		0x100c8
-#define ARM_SMMU_PRIQ_CONS		0x100cc
 #define ARM_SMMU_PRIQ_IRQ_CFG0		0xd0
 #define ARM_SMMU_PRIQ_IRQ_CFG1		0xd8
 #define ARM_SMMU_PRIQ_IRQ_CFG2		0xdc
@@ -1946,11 +1945,30 @@ static struct iommu_ops arm_smmu_ops = {
 };
 
 /* Probing and initialisation functions */
+static int arm_smmu_reset_one_queue(struct arm_smmu_device *smmu,
+				    struct arm_smmu_queue *q,
+				    unsigned long qoffset,
+				    int page_to_use)
+{
+	unsigned long prod, cons;
+
+	writeq_relaxed(q->q_base, smmu->base + qoffset);
+	if (page_to_use == 1) {
+		prod = ARM_SMMU_Q_PROD_PAGE1(qoffset);
+		cons = ARM_SMMU_Q_CONS_PAGE1(qoffset);
+	} else {
+		prod = ARM_SMMU_Q_PROD(qoffset);
+		cons = ARM_SMMU_Q_CONS(qoffset);
+	}
+	writeq_relaxed(q->prod, smmu->base + prod);
+	writeq_relaxed(q->cons, smmu->base + cons);
+}
+
 static int arm_smmu_init_one_queue(struct arm_smmu_device *smmu,
 				   struct arm_smmu_queue *q,
-				   unsigned long prod_off,
-				   unsigned long cons_off,
-				   size_t dwords)
+				   unsigned long qoffset,
+				   size_t dwords,
+				   int page_to_use)
 {
 	size_t qsz = ((1 << q->max_n_shift) * dwords) << 3;
 
@@ -1961,8 +1979,13 @@ static int arm_smmu_init_one_queue(struct arm_smmu_device *smmu,
 		return -ENOMEM;
 	}
 
-	q->prod_reg	= smmu->base + prod_off;
-	q->cons_reg	= smmu->base + cons_off;
+	if (page_to_use == 1) {
+		q->prod_reg	= smmu->base + ARM_SMMU_Q_PROD_PAGE1(qoffset);
+		q->cons_reg	= smmu->base + ARM_SMMU_Q_CONS_PAGE1(qoffset);
+	} else {
+		q->prod_reg	= smmu->base + ARM_SMMU_Q_PROD(qoffset);
+		q->cons_reg	= smmu->base + ARM_SMMU_Q_CONS(qoffset);
+	}
 	q->ent_dwords	= dwords;
 
 	q->q_base  = Q_BASE_RWA;
@@ -1980,14 +2003,14 @@ static int arm_smmu_init_queues(struct arm_smmu_device *smmu)
 
 	/* cmdq */
 	spin_lock_init(&smmu->cmdq.lock);
-	ret = arm_smmu_init_one_queue(smmu, &smmu->cmdq.q, ARM_SMMU_CMDQ_PROD,
-				      ARM_SMMU_CMDQ_CONS, CMDQ_ENT_DWORDS);
+	ret = arm_smmu_init_one_queue(smmu, &smmu->cmdq.q, ARM_SMMU_CMDQ_BASE,
+				      CMDQ_ENT_DWORDS, 0);
 	if (ret)
 		return ret;
 
 	/* evtq */
-	ret = arm_smmu_init_one_queue(smmu, &smmu->evtq.q, ARM_SMMU_EVTQ_PROD,
-				      ARM_SMMU_EVTQ_CONS, EVTQ_ENT_DWORDS);
+	ret = arm_smmu_init_one_queue(smmu, &smmu->evtq.q, ARM_SMMU_EVTQ_BASE,
+				      EVTQ_ENT_DWORDS, USE_PAGE1);
 	if (ret)
 		return ret;
 
@@ -1995,8 +2018,8 @@ static int arm_smmu_init_queues(struct arm_smmu_device *smmu)
 	if (!(smmu->features & ARM_SMMU_FEAT_PRI))
 		return 0;
 
-	return arm_smmu_init_one_queue(smmu, &smmu->priq.q, ARM_SMMU_PRIQ_PROD,
-				       ARM_SMMU_PRIQ_CONS, PRIQ_ENT_DWORDS);
+	return arm_smmu_init_one_queue(smmu, &smmu->priq.q, ARM_SMMU_PRIQ_BASE,
+				       PRIQ_ENT_DWORDS, 1);
 }
 
 static int arm_smmu_init_l1_strtab(struct arm_smmu_device *smmu)
@@ -2332,9 +2355,8 @@ static int arm_smmu_device_reset(struct arm_smmu_device *smmu, bool bypass)
 		       smmu->base + ARM_SMMU_STRTAB_BASE_CFG);
 
 	/* Command queue */
-	writeq_relaxed(smmu->cmdq.q.q_base, smmu->base + ARM_SMMU_CMDQ_BASE);
-	writel_relaxed(smmu->cmdq.q.prod, smmu->base + ARM_SMMU_CMDQ_PROD);
-	writel_relaxed(smmu->cmdq.q.cons, smmu->base + ARM_SMMU_CMDQ_CONS);
+	arm_smmu_reset_one_queue(smmu, &smmu->evtq.q,
+				 ARM_SMMU_CMDQ_BASE, 0);
 
 	enables = CR0_CMDQEN;
 	ret = arm_smmu_write_reg_sync(smmu, enables, ARM_SMMU_CR0,
@@ -2362,9 +2384,8 @@ static int arm_smmu_device_reset(struct arm_smmu_device *smmu, bool bypass)
 	arm_smmu_cmdq_issue_cmd(smmu, &cmd);
 
 	/* Event queue */
-	writeq_relaxed(smmu->evtq.q.q_base, smmu->base + ARM_SMMU_EVTQ_BASE);
-	writel_relaxed(smmu->evtq.q.prod, smmu->base + ARM_SMMU_EVTQ_PROD);
-	writel_relaxed(smmu->evtq.q.cons, smmu->base + ARM_SMMU_EVTQ_CONS);
+	arm_smmu_reset_one_queue(smmu, &smmu->evtq.q,
+				 ARM_SMMU_EVTQ_BASE, 1);
 
 	enables |= CR0_EVTQEN;
 	ret = arm_smmu_write_reg_sync(smmu, enables, ARM_SMMU_CR0,
@@ -2376,13 +2397,8 @@ static int arm_smmu_device_reset(struct arm_smmu_device *smmu, bool bypass)
 
 	/* PRI queue */
 	if (smmu->features & ARM_SMMU_FEAT_PRI) {
-		writeq_relaxed(smmu->priq.q.q_base,
-			       smmu->base + ARM_SMMU_PRIQ_BASE);
-		writel_relaxed(smmu->priq.q.prod,
-			       smmu->base + ARM_SMMU_PRIQ_PROD);
-		writel_relaxed(smmu->priq.q.cons,
-			       smmu->base + ARM_SMMU_PRIQ_CONS);
-
+		arm_smmu_reset_one_queue(smmu, &smmu->priq.q,
+					 ARM_SMMU_PRIQ_BASE, 1);
 		enables |= CR0_PRIQEN;
 		ret = arm_smmu_write_reg_sync(smmu, enables, ARM_SMMU_CR0,
 					      ARM_SMMU_CR0ACK);
-- 
2.7.4