msm: iommu: Further improvements to the MSM IOMMU driver
Subsystems (3)
cover msm: iommu: Further improvements to the MSM IOMMU driver
This depends on the clock control driver. Please hold off on this until the clock driver is in.
Simplify the IOMMU clock control, probe functions, and initialization, since we no longer have to work around having a dummy clock driver. As a result, all the IOMMU devices can now be safely be probed and initialized. Also add clock control for the IOMMU bus interconnect, which is only needed for access to IOMMU registers and now can be disabled for most of the time.
Stepan Moskovchenko (3):
msm: iommu: Add bus clocks to platform data
msm: iommu: Clock control for the IOMMU driver
msm: iommu: Simplify the platform clock code
arch/arm/mach-msm/devices-msm8x60-iommu.c | 5 - arch/arm/mach-msm/include/mach/iommu.h | 12 +- arch/arm/mach-msm/iommu.c | 82 ++++++++++-- arch/arm/mach-msm/iommu_dev.c | 204 +++++++++++++++++------------ 4 files changed, 196 insertions(+), 107 deletions(-)
Sent by an employee of the Qualcomm Innovation Center, Inc. The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.
PATCH 1/3 msm: iommu: Add bus clocks to platform data
Add the IOMMU bus clock and the IOMMU bus interconnect clocks to the platform data for an IOMMU device. These clocks are needed to access the IOMMU registers.
---
Please hold off on this until the clock driver is in.
arch/arm/mach-msm/include/mach/iommu.h | 7 ++++++- 1 files changed, 6 insertions(+), 1 deletions(-)
arch/arm/mach-msm/include/mach/iommu.h +6 -1
@@ -19,6 +19,7 @@#define MSM_IOMMU_H #include <linux/interrupt.h> +#include <linux/clk.h> /* Sharability attributes of MSM IOMMU mappings */ #define MSM_IOMMU_ATTR_NON_SH 0x0
@@ -74,13 +75,17 @@ struct msm_iommu_ctx_dev {* struct msm_iommu_drvdata - A single IOMMU hardware instance * @base: IOMMU config port base address (VA) * @irq: Interrupt number -* +* @clk: The bus clock for this IOMMU hardware instance +* @pclk: The clock for the IOMMU bus interconnect +* * A msm_iommu_drvdata holds the global driver data about a single piece * of an IOMMU hardware instance. */ struct msm_iommu_drvdata {void __iomem *base;int irq; +struct clk *clk; +struct clk *pclk; };/**
--
1.7.0.2
Sent by an employee of the Qualcomm Innovation Center, Inc. The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.
PATCH 2/3 msm: iommu: Clock control for the IOMMU driver
In reply to Stepan MoskovchenkoAdd clock control to the IOMMU driver. The IOMMU bus clock (and potentially an AXI clock) need to be on to gain access to IOMMU registers. Actively control these clocks when needed instead of leaving them on.
---
Please hold off on this until the clock driver is in.
arch/arm/mach-msm/iommu.c | 82 ++++++++++++++++++++++++++++++++++++++------ 1 files changed, 70 insertions(+), 12 deletions(-)
arch/arm/mach-msm/iommu.c +70 -12
@@ -26,6 +26,7 @@#include <linux/spinlock.h> #include <linux/slab.h> #include <linux/iommu.h> +#include <linux/clk.h> #include <asm/cacheflush.h> #include <asm/sizes.h>
@@ -50,12 +51,36 @@ struct msm_priv {struct list_head list_attached; }; -static void __flush_iotlb (struct iommu_domain *domain) +static int __enable_clocks (struct msm_iommu_drvdata *drvdata) +{ +int ret; + + ret =clk_enable (drvdata->pclk); +if (ret) + goto fail; + +if (drvdata->clk) { + ret =clk_enable (drvdata->clk); +if (ret) +clk_disable (drvdata->pclk); + } +fail: +return ret; +} + +static void __disable_clocks (struct msm_iommu_drvdata *drvdata) +{ +if (drvdata->clk) +clk_disable (drvdata->clk); +clk_disable (drvdata->pclk); +} + +static int __flush_iotlb (struct iommu_domain *domain) {struct msm_priv *priv = domain->priv;struct msm_iommu_drvdata *iommu_drvdata;struct msm_iommu_ctx_drvdata *ctx_drvdata; - +int ret =0 ;#ifndef CONFIG_IOMMU_PGTABLES_L2 unsigned long *fl_table = priv->pgtable;int i;
@@ -77,8 +102,18 @@ static void __flush_iotlb(struct iommu_domain *domain)BUG (); iommu_drvdata =dev_get_drvdata (ctx_drvdata->pdev->dev.parent); +if (!iommu_drvdata) +BUG (); + + ret =__enable_clocks (iommu_drvdata); +if (ret) + goto fail; +SET_CTX_TLBIALL (iommu_drvdata->base, ctx_drvdata->num,0 ); +__disable_clocks (iommu_drvdata); } +fail: +return ret; }static void __reset_context (void __iomem *base,int ctx)
@@ -263,11 +298,16 @@ static int msm_iommu_attach_dev(struct iommu_domain *domain, struct device *dev) goto fail; } + ret =__enable_clocks (iommu_drvdata); +if (ret) + goto fail; +__program_context (iommu_drvdata->base, ctx_dev->num,__pa (priv->pgtable)); +__disable_clocks (iommu_drvdata);list_add (&(ctx_drvdata->attached_elm), &priv->list_attached); -__flush_iotlb (domain); + ret =__flush_iotlb (domain); fail:spin_unlock_irqrestore (&msm_iommu_lock, flags);
@@ -282,6 +322,7 @@ static void msm_iommu_detach_dev(struct iommu_domain *domain,struct msm_iommu_drvdata *iommu_drvdata;struct msm_iommu_ctx_drvdata *ctx_drvdata;unsigned long flags; +int ret;spin_lock_irqsave (&msm_iommu_lock, flags); priv = domain->priv;
@@ -296,8 +337,16 @@ static void msm_iommu_detach_dev(struct iommu_domain *domain,if (!iommu_drvdata || !ctx_drvdata || !ctx_dev) goto fail; -__flush_iotlb (domain); + ret =__flush_iotlb (domain); +if (ret) + goto fail; + + ret =__enable_clocks (iommu_drvdata); +if (ret) + goto fail; +__reset_context (iommu_drvdata->base, ctx_dev->num); +__disable_clocks (iommu_drvdata);list_del_init (&ctx_drvdata->attached_elm); fail:
@@ -410,7 +459,7 @@ static int msm_iommu_map(struct iommu_domain *domain, unsigned long va,SL_AP1 |SL_SHARED |SL_TYPE_LARGE | pgprot; } -__flush_iotlb (domain); + ret =__flush_iotlb (domain); fail:spin_unlock_irqrestore (&msm_iommu_lock, flags);return ret;
@@ -495,7 +544,7 @@ static int msm_iommu_unmap(struct iommu_domain *domain, unsigned long va, } } -__flush_iotlb (domain); + ret =__flush_iotlb (domain); fail:spin_unlock_irqrestore (&msm_iommu_lock, flags);return ret;
@@ -526,13 +575,14 @@ static phys_addr_t msm_iommu_iova_to_phys(struct iommu_domain *domain, base = iommu_drvdata->base; ctx = ctx_drvdata->num; + ret =__enable_clocks (iommu_drvdata); +if (ret) + goto fail; +/* Invalidate context TLB */ SET_CTX_TLBIALL (base, ctx,0 );SET_V2PPR_VA (base, ctx, va >> V2Pxx_VA_SHIFT); -if (GET_FAULT (base, ctx)) - goto fail; - par =GET_PAR (base, ctx);/* We are dealing with a supersection */
@@ -541,6 +591,10 @@ static phys_addr_t msm_iommu_iova_to_phys(struct iommu_domain *domain, else /* Upper 20 bits from PAR, lower 12 from VA */ ret = (par &0xFFFFF000 ) | (va &0x00000FFF ); +if (GET_FAULT (base, ctx)) + ret =0 ; + +__disable_clocks (iommu_drvdata); fail:spin_unlock_irqrestore (&msm_iommu_lock, flags);return ret;
@@ -583,8 +637,8 @@ irqreturn_t msm_iommu_fault_handler(int irq, void *dev_id) {struct msm_iommu_drvdata *drvdata = dev_id;void __iomem *base; -unsigned int fsr =0 ; -int ncb =0 , i =0 ; +unsigned int fsr; +int ncb, i, ret;spin_lock (&msm_iommu_lock);
@@ -595,10 +649,13 @@ irqreturn_t msm_iommu_fault_handler(int irq, void *dev_id) base = drvdata->base; -pr_err ("===== WOAH! =====\n" );pr_err ("Unexpected IOMMU page fault!\n" );pr_err ("base = %08x\n" , (unsigned int ) base); + ret =__enable_clocks (drvdata); +if (ret) + goto fail; + ncb =GET_NCB (base)+1 ;for (i =0 ; i < ncb; i++) { fsr =GET_FSR (base, i);
@@ -609,6 +666,7 @@ irqreturn_t msm_iommu_fault_handler(int irq, void *dev_id)SET_FSR (base, i,0x4000000F ); } } +__disable_clocks (drvdata); fail:spin_unlock (&msm_iommu_lock);return 0 ;
--
1.7.0.2
Sent by an employee of the Qualcomm Innovation Center, Inc. The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.
PATCH 3/3 msm: iommu: Rework clock logic and add IOMMU bus clock control
In reply to Stepan MoskovchenkoClean up the clock control code in the probe calls, and add support for controlling the clock for the IOMMU bus interconnect. With the (proper) clock driver in place, the clock control logic in the probe function can be made much cleaner since it does not have to deal with the placeholder driver anymore.
---
Please hold off on this until the clock driver is in. The patch seems like it is a lot of changes, but a lot of it comes from removing one condition, which causes a bunch of code to be unindented by one level. This implementation is a lot cleaner IMO.
arch/arm/mach-msm/devices-msm8x60-iommu.c | 5 - arch/arm/mach-msm/include/mach/iommu.h | 5 - arch/arm/mach-msm/iommu_dev.c | 204 +++++++++++++++++------------ 3 files changed, 120 insertions(+), 94 deletions(-)
arch/arm/mach-msm/devices-msm8x60-iommu.c +0 -5
@@ -282,7 +282,6 @@ static struct platform_device msm_root_iommu_dev = {static struct msm_iommu_dev jpegd_iommu = { .name ="jpegd" , - .clk_rate =-1 };static struct msm_iommu_dev vpe_iommu = {
@@ -307,7 +306,6 @@ static struct msm_iommu_dev ijpeg_iommu = {static struct msm_iommu_dev vfe_iommu = { .name ="vfe" , - .clk_rate =-1 };static struct msm_iommu_dev vcodec_a_iommu = {
@@ -320,17 +318,14 @@ static struct msm_iommu_dev vcodec_b_iommu = {static struct msm_iommu_dev gfx3d_iommu = { .name ="gfx3d" , - .clk_rate =27000000 };static struct msm_iommu_dev gfx2d0_iommu = { .name ="gfx2d0" , - .clk_rate =27000000 };static struct msm_iommu_dev gfx2d1_iommu = { .name ="gfx2d1" , - .clk_rate =27000000 };static struct platform_device msm_device_iommu_jpegd = {
arch/arm/mach-msm/include/mach/iommu.h +0 -5
@@ -45,14 +45,9 @@/** * struct msm_iommu_dev - a single IOMMU hardware instance * name Human-readable name given to this IOMMU HW instance -* clk_rate Rate to set for this IOMMU's clock, if applicable to this -* particular IOMMU. 0 means don't set a rate. -* -1 means it is an AXI clock with no valid rate -* */ struct msm_iommu_dev {const char *name; -int clk_rate; };/**
arch/arm/mach-msm/iommu_dev.c +120 -84
271 lines not on this page · raw message
--
1.7.0.2
Sent by an employee of the Qualcomm Innovation Center, Inc. The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.
Re: [PATCH 2/3] msm: iommu: Clock control for the IOMMU driver
In reply to Stepan MoskovchenkoQuoted 24 lines · Stepan Moskovchenko
Add clock control to the IOMMU driver. The IOMMU bus clock (and potentially an AXI clock) need to be on to gain access to IOMMU registers. Actively control these clocks when needed instead of leaving them on.
Signed-off-by Stepan Moskovchenko <[email protected]>---
Please hold off on this until the clock driver is in.arch/arm/mach-msm/iommu.c | 82 ++++++++++++++++++++++++++++++++++++++------ 1 files changed, 70 insertions(+), 12 deletions(-)arch/arm/mach-msm/iommu.c +1
-0@@ -26,6 +26,7 @@#include <linux/spinlock.h> #include <linux/slab.h> #include <linux/iommu.h> +#include <linux/clk.h> #include <asm/cacheflush.h> #include <asm/sizes.h>
arch/arm/mach-msm/iommu.c struct msm_priv { · line 58
Quoted 9 lines
arch/arm/mach-msm/iommu.c +5
-1@@ -50,12 +51,36 @@ struct msm_priv {struct list_head list_attached; }; -static void __flush_iotlb (struct iommu_domain *domain) +static int __enable_clocks (struct msm_iommu_drvdata *drvdata) +{ +int ret; + + ret =clk_enable (drvdata->pclk);
You don't need to check if pclk is null ?
Quoted 30 lines
arch/arm/mach-msm/iommu.c +21
-1+if (ret) + goto fail; + +if (drvdata->clk) { + ret =clk_enable (drvdata->clk); +if (ret) +clk_disable (drvdata->pclk); + } +fail: +return ret; +} + +static void __disable_clocks (struct msm_iommu_drvdata *drvdata) +{ +if (drvdata->clk) +clk_disable (drvdata->clk); +clk_disable (drvdata->pclk); +} + +static int __flush_iotlb (struct iommu_domain *domain) {struct msm_priv *priv = domain->priv;struct msm_iommu_drvdata *iommu_drvdata;struct msm_iommu_ctx_drvdata *ctx_drvdata; - +int ret =0 ;#ifndef CONFIG_IOMMU_PGTABLES_L2 unsigned long *fl_table = priv->pgtable;int i;
arch/arm/mach-msm/iommu.c static void __flush_iotlb(struct iommu_domain *domain) · line 106
Quoted 5 lines
arch/arm/mach-msm/iommu.c +2
-0@@ -77,8 +102,18 @@ static void __flush_iotlb(struct iommu_domain *domain)BUG (); iommu_drvdata =dev_get_drvdata (ctx_drvdata->pdev->dev.parent); +if (!iommu_drvdata) +BUG ();
Just do,
BUG_ON(!iommu_drvdata);
Quoted 39 lines
arch/arm/mach-msm/iommu.c +14
-1+ ret =__enable_clocks (iommu_drvdata); +if (ret) + goto fail; +SET_CTX_TLBIALL (iommu_drvdata->base, ctx_drvdata->num,0 ); +__disable_clocks (iommu_drvdata); } +fail: +return ret; }static void __reset_context (void __iomem *base,int ctx)@@ -263,11 +298,16 @@ static int msm_iommu_attach_dev(struct iommu_domain *domain, struct device *dev) goto fail; } + ret =__enable_clocks (iommu_drvdata); +if (ret) + goto fail; +__program_context (iommu_drvdata->base, ctx_dev->num,__pa (priv->pgtable)); +__disable_clocks (iommu_drvdata);list_add (&(ctx_drvdata->attached_elm), &priv->list_attached); -__flush_iotlb (domain); + ret =__flush_iotlb (domain); fail:spin_unlock_irqrestore (&msm_iommu_lock, flags);@@ -282,6 +322,7 @@ static void msm_iommu_detach_dev(struct iommu_domain *domain,struct msm_iommu_drvdata *iommu_drvdata;struct msm_iommu_ctx_drvdata *ctx_drvdata;unsigned long flags; +int ret;spin_lock_irqsave (&msm_iommu_lock, flags); priv = domain->priv;
arch/arm/mach-msm/iommu.c static void msm_iommu_detach_dev(struct iommu_domain *domain, · line 340
Quoted 5 lines
arch/arm/mach-msm/iommu.c +1
-1@@ -296,8 +337,16 @@ static void msm_iommu_detach_dev(struct iommu_domain *domain,if (!iommu_drvdata || !ctx_drvdata || !ctx_dev) goto fail; -__flush_iotlb (domain); + ret =__flush_iotlb (domain);
What the relationship between this __flush_iotlb() and turning the clocks on/off.
Quoted 71 lines
arch/arm/mach-msm/iommu.c +20
-7+if (ret) + goto fail; + + ret =__enable_clocks (iommu_drvdata); +if (ret) + goto fail; +__reset_context (iommu_drvdata->base, ctx_dev->num); +__disable_clocks (iommu_drvdata);list_del_init (&ctx_drvdata->attached_elm); fail:@@ -410,7 +459,7 @@ static int msm_iommu_map(struct iommu_domain *domain, unsigned long va,SL_AP1 |SL_SHARED |SL_TYPE_LARGE | pgprot; } -__flush_iotlb (domain); + ret =__flush_iotlb (domain); fail:spin_unlock_irqrestore (&msm_iommu_lock, flags);return ret;@@ -495,7 +544,7 @@ static int msm_iommu_unmap(struct iommu_domain *domain, unsigned long va, } } -__flush_iotlb (domain); + ret =__flush_iotlb (domain); fail:spin_unlock_irqrestore (&msm_iommu_lock, flags);return ret;@@ -526,13 +575,14 @@ static phys_addr_t msm_iommu_iova_to_phys(struct iommu_domain *domain, base = iommu_drvdata->base; ctx = ctx_drvdata->num; + ret =__enable_clocks (iommu_drvdata); +if (ret) + goto fail; +/* Invalidate context TLB */ SET_CTX_TLBIALL (base, ctx,0 );SET_V2PPR_VA (base, ctx, va >> V2Pxx_VA_SHIFT); -if (GET_FAULT (base, ctx)) - goto fail; - par =GET_PAR (base, ctx);/* We are dealing with a supersection */ @@ -541,6 +591,10 @@ static phys_addr_t msm_iommu_iova_to_phys(struct iommu_domain *domain, else /* Upper 20 bits from PAR, lower 12 from VA */ ret = (par &0xFFFFF000 ) | (va &0x00000FFF ); +if (GET_FAULT (base, ctx)) + ret =0 ; + +__disable_clocks (iommu_drvdata); fail:spin_unlock_irqrestore (&msm_iommu_lock, flags);return ret;@@ -583,8 +637,8 @@ irqreturn_t msm_iommu_fault_handler(int irq, void *dev_id) {struct msm_iommu_drvdata *drvdata = dev_id;void __iomem *base; -unsigned int fsr =0 ; -int ncb =0 , i =0 ; +unsigned int fsr; +int ncb, i, ret;spin_lock (&msm_iommu_lock);
arch/arm/mach-msm/iommu.c irqreturn_t msm_iommu_fault_handler(int irq, void *dev_id) · line 598
Quoted 4 lines
arch/arm/mach-msm/iommu.c +0
-1@@ -595,10 +649,13 @@ irqreturn_t msm_iommu_fault_handler(int irq, void *dev_id) base = drvdata->base; -pr_err ("===== WOAH! =====\n" );
Cleanup right? It doesn't need it's own patch, but you could mention in the description that you've done "minor cleanups" or something to that effect.
Daniel
Signature · 3 lines
Sent by an consultant of the Qualcomm Innovation Center, Inc. The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.
Re: [PATCH 3/3] msm: iommu: Rework clock logic and add IOMMU bus clock control
In reply to Stepan MoskovchenkoQuoted 8 lines · Stepan Moskovchenko
Clean up the clock control code in the probe calls, and add support for controlling the clock for the IOMMU bus interconnect. With the (proper) clock driver in place, the clock control logic in the probe function can be made much cleaner since it does not have to deal with the placeholder driver anymore.
Change-Id: I1040bc4e18f4ab4b7cc0dd5fe667f9df83b9f1f5
You need to remove this Change-Id ..
Quoted 184 lines
Signed-off-by Stepan Moskovchenko <[email protected]>---
Please hold off on this until the clock driver is in. The patch seems like it is a lot of changes, but a lot of it comes from removing one condition, which causes a bunch of code to be unindented by one level. This implementation is a lot cleaner IMO.arch/arm/mach-msm/devices-msm8x60-iommu.c | 5 - arch/arm/mach-msm/include/mach/iommu.h | 5 - arch/arm/mach-msm/iommu_dev.c | 204 +++++++++++++++++------------ 3 files changed, 120 insertions(+), 94 deletions(-)arch/arm/mach-msm/devices-msm8x60-iommu.c +0
-5@@ -282,7 +282,6 @@ static struct platform_device msm_root_iommu_dev = {static struct msm_iommu_dev jpegd_iommu = { .name ="jpegd" , - .clk_rate =-1 };static struct msm_iommu_dev vpe_iommu = {@@ -307,7 +306,6 @@ static struct msm_iommu_dev ijpeg_iommu = {static struct msm_iommu_dev vfe_iommu = { .name ="vfe" , - .clk_rate =-1 };static struct msm_iommu_dev vcodec_a_iommu = {@@ -320,17 +318,14 @@ static struct msm_iommu_dev vcodec_b_iommu = {static struct msm_iommu_dev gfx3d_iommu = { .name ="gfx3d" , - .clk_rate =27000000 };static struct msm_iommu_dev gfx2d0_iommu = { .name ="gfx2d0" , - .clk_rate =27000000 };static struct msm_iommu_dev gfx2d1_iommu = { .name ="gfx2d1" , - .clk_rate =27000000 };static struct platform_device msm_device_iommu_jpegd = {arch/arm/mach-msm/include/mach/iommu.h +0
-5@@ -45,14 +45,9 @@/** * struct msm_iommu_dev - a single IOMMU hardware instance * name Human-readable name given to this IOMMU HW instance -* clk_rate Rate to set for this IOMMU's clock, if applicable to this -* particular IOMMU. 0 means don't set a rate. -* -1 means it is an AXI clock with no valid rate -* */ struct msm_iommu_dev {const char *name; -int clk_rate; };/** arch/arm/mach-msm/iommu_dev.c +43
-44@@ -29,6 +29,7 @@#include <mach/iommu_hw-8xxx.h> #include <mach/iommu.h> +#include <mach/clk.h> struct iommu_ctx_iter_data {/* input */ @@ -130,117 +131,134 @@ static int msm_iommu_probe(struct platform_device *pdev) {struct resource *r, *r2;struct clk *iommu_clk; +struct clk *iommu_pclk;struct msm_iommu_drvdata *drvdata;struct msm_iommu_dev *iommu_dev = pdev->dev.platform_data;void __iomem *regs_base;resource_size_t len; -int ret =0 , ncb, nm2v, irq; +int ret, ncb, nm2v, irq; -if (pdev->id !=-1 ) { - drvdata =kzalloc (sizeof (*drvdata),GFP_KERNEL ); +if (pdev->id ==-1 ) { + msm_iommu_root_dev = pdev; +return 0 ; + } -if (!drvdata) { - ret = -ENOMEM ; - goto fail; - } + drvdata =kzalloc (sizeof (*drvdata),GFP_KERNEL ); -if (!iommu_dev) { - ret = -ENODEV ; - goto fail; - } +if (!drvdata) { + ret = -ENOMEM ; + goto fail; + } + +if (!iommu_dev) { + ret = -ENODEV ; + goto fail; + } -if (iommu_dev->clk_rate !=0 ) { - iommu_clk =clk_get (&pdev->dev,"iommu_clk" ); - -if (IS_ERR (iommu_clk)) { - ret = -ENODEV ; - goto fail; - } - -if (iommu_dev->clk_rate >0 ) { - ret =clk_set_rate (iommu_clk, - iommu_dev->clk_rate); -if (ret) { -clk_put (iommu_clk); - goto fail; - } - } - - ret =clk_enable (iommu_clk); -if (ret) { -clk_put (iommu_clk); - goto fail; - } + iommu_pclk =clk_get (NULL ,"smmu_pclk" ); +if (IS_ERR (iommu_pclk)) { + ret = -ENODEV ; + goto fail; + } + + ret =clk_enable (iommu_pclk); +if (ret) + goto fail_enable; + + iommu_clk =clk_get (&pdev->dev,"iommu_clk" ); + +if (!IS_ERR (iommu_clk)) { +if (clk_get_rate (iommu_clk) ==0 ) +clk_set_min_rate (iommu_clk,1 ); + + ret =clk_enable (iommu_clk); +if (ret) {clk_put (iommu_clk); + goto fail_pclk; } + }else + iommu_clk =NULL ; - r =platform_get_resource_byname (pdev,IORESOURCE_MEM , -"physbase" ); -if (!r) { - ret = -ENODEV ; - goto fail; - } + r =platform_get_resource_byname (pdev,IORESOURCE_MEM ,"physbase" ); - len = r->end - r->start +1 ; +if (!r) { + ret = -ENODEV ; + goto fail_clk; + } - r2 =request_mem_region (r->start, len, r->name); -if (!r2) { -pr_err ("Could not request memory region: " -"start=%p, len=%d\n" , (void *) r->start, len);
arch/arm/mach-msm/iommu_dev.c static int msm_iommu_probe(struct platform_device *pdev) · line 194
Quoted 10 lines
arch/arm/mach-msm/iommu_dev.c +5
-4- ret = -EBUSY ; - goto fail; - } + len = r->end - r->start +1 ; - regs_base =ioremap (r2->start, len); + r2 =request_mem_region (r->start, len, r->name); +if (!r2) { +pr_err ("Could not request memory region: start=%p, len=%d\n" , + (void *) r->start, len);(void *) r->start, len);
You usually just tab over till you get to the " like this,
pr_err("Could not request memory region: start=%p, len=%d\n",
(void *) r->start, len);
Quoted 104 lines
arch/arm/mach-msm/iommu_dev.c +51
-33+ ret = -EBUSY ; + goto fail_clk; + } -if (!regs_base) { -pr_err ("Could not ioremap: start=%p, len=%d\n" , - (void *) r2->start, len); - ret = -EBUSY ; - goto fail_mem; - } + regs_base =ioremap (r2->start, len); - irq =platform_get_irq_byname (pdev,"secure_irq" ); -if (irq <0 ) { - ret = -ENODEV ; - goto fail_io; - } +if (!regs_base) { +pr_err ("Could not ioremap: start=%p, len=%d\n" , + (void *) r2->start, len); + ret = -EBUSY ; + goto fail_mem; + } + + irq =platform_get_irq_byname (pdev,"secure_irq" ); +if (irq <0 ) { + ret = -ENODEV ; + goto fail_io; + } -mb (); +mb (); -if (GET_IDR (regs_base) ==0 ) { -pr_err ("Invalid IDR value detected\n" ); - ret = -ENODEV ; - goto fail_io; - } +if (GET_IDR (regs_base) ==0 ) { +pr_err ("Invalid IDR value detected\n" ); + ret = -ENODEV ; + goto fail_io; + } - ret =request_irq (irq, msm_iommu_fault_handler,0 , -"msm_iommu_secure_irpt_handler" , drvdata); -if (ret) { -pr_err ("Request IRQ %d failed with ret=%d\n" , irq, ret); - goto fail_io; - } + ret =request_irq (irq, msm_iommu_fault_handler,0 , +"msm_iommu_secure_irpt_handler" , drvdata); +if (ret) { +pr_err ("Request IRQ %d failed with ret=%d\n" , irq, ret); + goto fail_io; + } -msm_iommu_reset (regs_base); - drvdata->base = regs_base; - drvdata->irq = irq; +msm_iommu_reset (regs_base); + drvdata->pclk = iommu_pclk; + drvdata->clk = iommu_clk; + drvdata->base = regs_base; + drvdata->irq = irq; - nm2v =GET_NM2VCBMT ((unsigned long ) regs_base); - ncb =GET_NCB ((unsigned long ) regs_base); + nm2v =GET_NM2VCBMT ((unsigned long ) regs_base); + ncb =GET_NCB ((unsigned long ) regs_base); -pr_info ("device %s mapped at %p, irq %d with %d ctx banks\n" , +pr_info ("device %s mapped at %p, irq %d with %d ctx banks\n" , iommu_dev->name, regs_base, irq, ncb+1 ); -platform_set_drvdata (pdev, drvdata); - }else - msm_iommu_root_dev = pdev; +platform_set_drvdata (pdev, drvdata); -return 0 ; +if (iommu_clk) +clk_disable (iommu_clk); + +clk_disable (iommu_pclk); +return 0 ; fail_io:iounmap (regs_base); fail_mem:release_mem_region (r->start, len); +fail_clk: +if (iommu_clk) { +clk_disable (iommu_clk); +clk_put (iommu_clk); + } +fail_pclk: +clk_disable (iommu_pclk); +fail_enable: +clk_put (iommu_pclk); fail:kfree (drvdata);return ret;
arch/arm/mach-msm/iommu_dev.c static int msm_iommu_remove(struct platform_device *pdev) · line 276
Quoted 8 lines
arch/arm/mach-msm/iommu_dev.c +4
-1@@ -252,7 +270,10 @@ static int msm_iommu_remove(struct platform_device *pdev) drv =platform_get_drvdata (pdev);if (drv) { -memset (drv,0 ,sizeof (struct msm_iommu_drvdata )); +if (drv->clk) +clk_put (drv->clk); +clk_put (drv->pclk); +memset (drv,0 ,sizeof (*drv));
Do you really need the memset ?
Quoted 17 lines
arch/arm/mach-msm/iommu_dev.c +2
-1kfree (drv);platform_set_drvdata (pdev,NULL ); }@@ -264,7 +285,7 @@ static int msm_iommu_ctx_probe(struct platform_device *pdev)struct msm_iommu_ctx_dev *c = pdev->dev.platform_data;struct msm_iommu_drvdata *drvdata;struct msm_iommu_ctx_drvdata *ctx_drvdata =NULL ; -int i, ret =0 ; +int i, ret;if (!c || !pdev->dev.parent) { ret = -EINVAL ; goto fail;@@ -288,6 +309,18 @@ static int msm_iommu_ctx_probe(struct platform_device *pdev)INIT_LIST_HEAD (&ctx_drvdata->attached_elm);platform_set_drvdata (pdev, ctx_drvdata); + ret =clk_enable (drvdata->pclk);
arch/arm/mach-msm/iommu_dev.c static int msm_iommu_ctx_probe(struct platform_device *pdev) · line 322
Quoted 10 lines
arch/arm/mach-msm/iommu_dev.c +10
-0+if (ret) + goto fail; + +if (drvdata->clk) { + ret =clk_enable (drvdata->clk); +if (ret) { +clk_disable (drvdata->pclk); + goto fail; + } + }
You did this in a prior patch also, you could combine them into a single helper function. Maybe do the same for the disable side too.
Daniel
Signature · 3 lines
Sent by an consultant of the Qualcomm Innovation Center, Inc. The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.
Re: [PATCH 2/3] msm: iommu: Clock control for the IOMMU driver
In reply to Daniel WalkerQuoted 6 lines · Daniel Walker
arch/arm/mach-msm/iommu.c struct msm_priv { · line 58
Quoted 3 lines · Stepan Moskovchenko
arch/arm/mach-msm/iommu.c +3
-0+int ret; + + ret =clk_enable (drvdata->pclk);You don't need to check if pclk is null ?
Nope. If we are here, the pclk will always be non-null, which is something that may not necessarily be said for for AXI clock.
Quoted 6 lines
arch/arm/mach-msm/iommu.c static void __flush_iotlb(struct iommu_domain *domain) · line 106
Quoted 3 lines
iommu_drvdata = dev_get_drvdata(ctx_drvdata->pdev->dev.parent);arch/arm/mach-msm/iommu.c +2
-0+if (!iommu_drvdata) +BUG ();Just do,
BUG_ON(!iommu_drvdata);
Will fix in v2.
Quoted 5 lines
Review
Quoted 2 lines
(quoted diff) +1
-1- __flush_iotlb(domain); + ret = __flush_iotlb(domain);What the relationship between this
__flush_iotlb()and turning the clocks on/off.
The flush_iotlb function needs to handle clock control as well, which means it can now fail. As a result, we now need to check its return value, instead of assuming it succeeded.
Quoted 5 lines
arch/arm/mach-msm/iommu.c irqreturn_t msm_iommu_fault_handler(int irq, void *dev_id) · line 598
Quoted 1 line
arch/arm/mach-msm/iommu.c +0
-1-pr_err ("===== WOAH! =====\n" );Cleanup right? It doesn't need it's own patch, but you could mention in the description that you've done "minor cleanups" or something to that effect.
Will fix in v2.
Re: [PATCH 3/3] msm: iommu: Rework clock logic and add IOMMU bus clock control
In reply to Daniel WalkerQuoted 10 lines · Daniel Walker
Quoted 8 lines · Stepan Moskovchenko
Clean up the clock control code in the probe calls, and add support for controlling the clock for the IOMMU bus interconnect. With the (proper) clock driver in place, the clock control logic in the probe function can be made much cleaner since it does not have to deal with the placeholder driver anymore.
Change-Id: I1040bc4e18f4ab4b7cc0dd5fe667f9df83b9f1f5
You need to remove this Change-Id ..
Fixed in v2.
Quoted 7 lines
Review
Quoted 3 lines
(quoted diff) +2
-0+ pr_err("Could not request memory region: start=%p, len=%d\n", + (void *) r->start, len);(void *) r->start, len);You usually just tab over till you get to the " like this,
pr_err("Could not request memory region: start=%p, len=%d\n",
(void *) r->start, len);
Fixed in v2.
Quoted 8 lines
arch/arm/mach-msm/iommu_dev.c static int msm_iommu_remove(struct platform_device *pdev) · line 276
Quoted 7 lines
drv = platform_get_drvdata(pdev); if (drv) {arch/arm/mach-msm/iommu_dev.c +4
-1-memset (drv,0 ,sizeof (struct msm_iommu_drvdata )); +if (drv->clk) +clk_put (drv->clk); +clk_put (drv->pclk); +memset (drv,0 ,sizeof (*drv));Do you really need the memset ?
I guess not.. it seemed like good practice to poison the memory in case someone else had a stale reference to it. I guess I can remove them.
Quoted 4 lines
Quoted 4 lines
arch/arm/mach-msm/iommu_dev.c +4
-0+if (ret) + goto fail; + +if (drvdata->clk) {
Quoted 8 lines
arch/arm/mach-msm/iommu_dev.c static int msm_iommu_ctx_probe(struct platform_device *pdev) · line 322
Quoted 6 lines
arch/arm/mach-msm/iommu_dev.c +6
-0+ ret =clk_enable (drvdata->clk); +if (ret) { +clk_disable (drvdata->pclk); + goto fail; + } + }You did this in a prior patch also, you could combine them into a single helper function. Maybe do the same for the disable side too.
That was in a different file (where it gets used much more extensively than just in the one case here). I would rather not make a bunch of my internal stuff visible to the global namespace if I can help it, and it's a trivial enough operation to enable two clocks.
I will send out v2 in a few minutes.
Steve
---
Sent by an employee of the Qualcomm Innovation Center, Inc. The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.
PATCH v2 2/3 msm: iommu: Clock control for the IOMMU driver
In reply to Stepan MoskovchenkoAdd clock control to the IOMMU driver. The IOMMU bus clock (and potentially an AXI clock) need to be on to gain access to IOMMU registers. Actively control these clocks when needed instead of leaving them on. Additionally, perform minor code cleanups.
---
Please hold off on this until the clock driver is in. Changes in v2:
* Two minor cleanups of the form of changing if (..) BUG() to BUG_ON()
* Amend the commit text to that effect
arch/arm/mach-msm/iommu.c | 84 +++++++++++++++++++++++++++++++++++++------- 1 files changed, 70 insertions(+), 14 deletions(-)
arch/arm/mach-msm/iommu.c +70 -14
@@ -26,6 +26,7 @@#include <linux/spinlock.h> #include <linux/slab.h> #include <linux/iommu.h> +#include <linux/clk.h> #include <asm/cacheflush.h> #include <asm/sizes.h>
@@ -50,12 +51,36 @@ struct msm_priv {struct list_head list_attached; }; -static void __flush_iotlb (struct iommu_domain *domain) +static int __enable_clocks (struct msm_iommu_drvdata *drvdata) +{ +int ret; + + ret =clk_enable (drvdata->pclk); +if (ret) + goto fail; + +if (drvdata->clk) { + ret =clk_enable (drvdata->clk); +if (ret) +clk_disable (drvdata->pclk); + } +fail: +return ret; +} + +static void __disable_clocks (struct msm_iommu_drvdata *drvdata) +{ +if (drvdata->clk) +clk_disable (drvdata->clk); +clk_disable (drvdata->pclk); +} + +static int __flush_iotlb (struct iommu_domain *domain) {struct msm_priv *priv = domain->priv;struct msm_iommu_drvdata *iommu_drvdata;struct msm_iommu_ctx_drvdata *ctx_drvdata; - +int ret =0 ;#ifndef CONFIG_IOMMU_PGTABLES_L2 unsigned long *fl_table = priv->pgtable;int i;
@@ -73,12 +98,20 @@ static void __flush_iotlb(struct iommu_domain *domain)#endif list_for_each_entry (ctx_drvdata, &priv->list_attached, attached_elm) { -if (!ctx_drvdata->pdev || !ctx_drvdata->pdev->dev.parent) -BUG (); +BUG_ON (!ctx_drvdata->pdev || !ctx_drvdata->pdev->dev.parent); iommu_drvdata =dev_get_drvdata (ctx_drvdata->pdev->dev.parent); +BUG_ON (!iommu_drvdata); + + ret =__enable_clocks (iommu_drvdata); +if (ret) + goto fail; +SET_CTX_TLBIALL (iommu_drvdata->base, ctx_drvdata->num,0 ); +__disable_clocks (iommu_drvdata); } +fail: +return ret; }static void __reset_context (void __iomem *base,int ctx)
@@ -263,11 +296,16 @@ static int msm_iommu_attach_dev(struct iommu_domain *domain, struct device *dev) goto fail; } + ret =__enable_clocks (iommu_drvdata); +if (ret) + goto fail; +__program_context (iommu_drvdata->base, ctx_dev->num,__pa (priv->pgtable)); +__disable_clocks (iommu_drvdata);list_add (&(ctx_drvdata->attached_elm), &priv->list_attached); -__flush_iotlb (domain); + ret =__flush_iotlb (domain); fail:spin_unlock_irqrestore (&msm_iommu_lock, flags);
@@ -282,6 +320,7 @@ static void msm_iommu_detach_dev(struct iommu_domain *domain,struct msm_iommu_drvdata *iommu_drvdata;struct msm_iommu_ctx_drvdata *ctx_drvdata;unsigned long flags; +int ret;spin_lock_irqsave (&msm_iommu_lock, flags); priv = domain->priv;
@@ -296,8 +335,16 @@ static void msm_iommu_detach_dev(struct iommu_domain *domain,if (!iommu_drvdata || !ctx_drvdata || !ctx_dev) goto fail; -__flush_iotlb (domain); + ret =__flush_iotlb (domain); +if (ret) + goto fail; + + ret =__enable_clocks (iommu_drvdata); +if (ret) + goto fail; +__reset_context (iommu_drvdata->base, ctx_dev->num); +__disable_clocks (iommu_drvdata);list_del_init (&ctx_drvdata->attached_elm); fail:
@@ -410,7 +457,7 @@ static int msm_iommu_map(struct iommu_domain *domain, unsigned long va,SL_AP1 |SL_SHARED |SL_TYPE_LARGE | pgprot; } -__flush_iotlb (domain); + ret =__flush_iotlb (domain); fail:spin_unlock_irqrestore (&msm_iommu_lock, flags);return ret;
@@ -495,7 +542,7 @@ static int msm_iommu_unmap(struct iommu_domain *domain, unsigned long va, } } -__flush_iotlb (domain); + ret =__flush_iotlb (domain); fail:spin_unlock_irqrestore (&msm_iommu_lock, flags);return ret;
@@ -526,13 +573,14 @@ static phys_addr_t msm_iommu_iova_to_phys(struct iommu_domain *domain, base = iommu_drvdata->base; ctx = ctx_drvdata->num; + ret =__enable_clocks (iommu_drvdata); +if (ret) + goto fail; +/* Invalidate context TLB */ SET_CTX_TLBIALL (base, ctx,0 );SET_V2PPR_VA (base, ctx, va >> V2Pxx_VA_SHIFT); -if (GET_FAULT (base, ctx)) - goto fail; - par =GET_PAR (base, ctx);/* We are dealing with a supersection */
@@ -541,6 +589,10 @@ static phys_addr_t msm_iommu_iova_to_phys(struct iommu_domain *domain, else /* Upper 20 bits from PAR, lower 12 from VA */ ret = (par &0xFFFFF000 ) | (va &0x00000FFF ); +if (GET_FAULT (base, ctx)) + ret =0 ; + +__disable_clocks (iommu_drvdata); fail:spin_unlock_irqrestore (&msm_iommu_lock, flags);return ret;
@@ -583,8 +635,8 @@ irqreturn_t msm_iommu_fault_handler(int irq, void *dev_id) {struct msm_iommu_drvdata *drvdata = dev_id;void __iomem *base; -unsigned int fsr =0 ; -int ncb =0 , i =0 ; +unsigned int fsr; +int ncb, i, ret;spin_lock (&msm_iommu_lock);
@@ -595,10 +647,13 @@ irqreturn_t msm_iommu_fault_handler(int irq, void *dev_id) base = drvdata->base; -pr_err ("===== WOAH! =====\n" );pr_err ("Unexpected IOMMU page fault!\n" );pr_err ("base = %08x\n" , (unsigned int ) base); + ret =__enable_clocks (drvdata); +if (ret) + goto fail; + ncb =GET_NCB (base)+1 ;for (i =0 ; i < ncb; i++) { fsr =GET_FSR (base, i);
@@ -609,6 +664,7 @@ irqreturn_t msm_iommu_fault_handler(int irq, void *dev_id)SET_FSR (base, i,0x4000000F ); } } +__disable_clocks (drvdata); fail:spin_unlock (&msm_iommu_lock);return 0 ;
--
1.7.0.2
Sent by an employee of the Qualcomm Innovation Center, Inc. The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.
PATCH v2 3/3 msm: iommu: Rework clock logic and add IOMMU bus clock control
In reply to Stepan MoskovchenkoClean up the clock control code in the probe calls, and add support for controlling the clock for the IOMMU bus interconnect. With the (proper) clock driver in place, the clock control logic in the probe function can be made much cleaner since it does not have to deal with the placeholder driver anymore.
---
Please hold off on this until the clock driver is in. The patch seems like it is a lot of changes, but a lot of it comes from removing one condition, which causes a bunch of code to be unindented by one level. This implementation is a lot cleaner IMO.
Changes in v2:
* Remove two unnecessart memset calls
* Fix alignment / whitespace
arch/arm/mach-msm/devices-msm8x60-iommu.c | 5 - arch/arm/mach-msm/include/mach/iommu.h | 5 - arch/arm/mach-msm/iommu_dev.c | 204 +++++++++++++++++------------ 3 files changed, 119 insertions(+), 95 deletions(-)
arch/arm/mach-msm/devices-msm8x60-iommu.c +0 -5
@@ -282,7 +282,6 @@ static struct platform_device msm_root_iommu_dev = {static struct msm_iommu_dev jpegd_iommu = { .name ="jpegd" , - .clk_rate =-1 };static struct msm_iommu_dev vpe_iommu = {
@@ -307,7 +306,6 @@ static struct msm_iommu_dev ijpeg_iommu = {static struct msm_iommu_dev vfe_iommu = { .name ="vfe" , - .clk_rate =-1 };static struct msm_iommu_dev vcodec_a_iommu = {
@@ -320,17 +318,14 @@ static struct msm_iommu_dev vcodec_b_iommu = {static struct msm_iommu_dev gfx3d_iommu = { .name ="gfx3d" , - .clk_rate =27000000 };static struct msm_iommu_dev gfx2d0_iommu = { .name ="gfx2d0" , - .clk_rate =27000000 };static struct msm_iommu_dev gfx2d1_iommu = { .name ="gfx2d1" , - .clk_rate =27000000 };static struct platform_device msm_device_iommu_jpegd = {
arch/arm/mach-msm/include/mach/iommu.h +0 -5
@@ -45,14 +45,9 @@/** * struct msm_iommu_dev - a single IOMMU hardware instance * name Human-readable name given to this IOMMU HW instance -* clk_rate Rate to set for this IOMMU's clock, if applicable to this -* particular IOMMU. 0 means don't set a rate. -* -1 means it is an AXI clock with no valid rate -* */ struct msm_iommu_dev {const char *name; -int clk_rate; };/**
arch/arm/mach-msm/iommu_dev.c +119 -85
277 lines not on this page · raw message
--
1.7.0.2
Sent by an employee of the Qualcomm Innovation Center, Inc. The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.
Re: [PATCH 3/3] msm: iommu: Rework clock logic and add IOMMU bus clock control
In reply to Stepan MoskovchenkoQuoted 30 lines · Stepan Moskovchenko
Quoted 10 lines · Daniel Walker
Quoted 8 lines · Stepan Moskovchenko
Clean up the clock control code in the probe calls, and add support for controlling the clock for the IOMMU bus interconnect. With the (proper) clock driver in place, the clock control logic in the probe function can be made much cleaner since it does not have to deal with the placeholder driver anymore.
Change-Id: I1040bc4e18f4ab4b7cc0dd5fe667f9df83b9f1f5
You need to remove this Change-Id ..
Fixed in v2.
Quoted 7 lines
Review
Quoted 3 lines
(quoted diff) +2
-0+ pr_err("Could not request memory region: start=%p, len=%d\n", + (void *) r->start, len);(void *) r->start, len);You usually just tab over till you get to the " like this,
pr_err("Could not request memory region: start=%p, len=%d\n",
(void *) r->start, len);Fixed in v2.
Quoted 8 lines
arch/arm/mach-msm/iommu_dev.c static int msm_iommu_remove(struct platform_device *pdev) · line 276
Quoted 7 lines
drv = platform_get_drvdata(pdev); if (drv) {arch/arm/mach-msm/iommu_dev.c +4
-1-memset (drv,0 ,sizeof (struct msm_iommu_drvdata )); +if (drv->clk) +clk_put (drv->clk); +clk_put (drv->pclk); +memset (drv,0 ,sizeof (*drv));Do you really need the memset ?
Quoted 2 lines
I guess not.. it seemed like good practice to poison the memory in case someone else had a stale reference to it. I guess I can remove them.
The memory allocator already has a poison debugging option. If you suspect memory issues you can always turn that on.
Quoted 12 lines
Quoted 4 lines
Quoted 4 lines
arch/arm/mach-msm/iommu_dev.c +4
-0+if (ret) + goto fail; + +if (drvdata->clk) {Quoted 8 lines
arch/arm/mach-msm/iommu_dev.c static int msm_iommu_ctx_probe(struct platform_device *pdev) · line 322
Quoted 6 lines
arch/arm/mach-msm/iommu_dev.c +6
-0+ ret =clk_enable (drvdata->clk); +if (ret) { +clk_disable (drvdata->pclk); + goto fail; + } + }You did this in a prior patch also, you could combine them into a single helper function. Maybe do the same for the disable side too.
Quoted 4 lines
That was in a different file (where it gets used much more extensively than just in the one case here). I would rather not make a bunch of my internal stuff visible to the global namespace if I can help it, and it's a trivial enough operation to enable two clocks.
I was meaning making it a static inline inside a header.. You wouldn't want this to be a global function. If you can it's usually a good idea to combine code that's similar, that way there's only one place to make modifications.
Daniel
Signature · 3 lines
Sent by an consultant of the Qualcomm Innovation Center, Inc. The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.