Message ID | 1457451357-14000-5-git-send-email-stuart.yoder@nxp.com |
---|---|
State | Superseded |
Delegated to: | York Sun |
Headers | show |
On 03/08/2016 07:56 AM, Stuart Yoder wrote: > From: Stuart Yoder <stuart.yoder@nxp.com> > > The per-PCI controller LUT (Look-Up-Table) is a 32-entry table > that maps PCI requester IDs (bus/dev/fun) to a stream ID. > > This patch implements infrastructure to enable LUT initialization: > -define registers offsets > -add an index to 'struct ls_pcie' to track next available slot in LUT > -add function to allocate the next available entry index > -add function to program a LUT entry > > Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com> > --- > -v4: put new support under LS2 #ifdef > > .../include/asm/arch-fsl-layerscape/immap_lsch3.h | 4 +++ > drivers/pci/pcie_layerscape.c | 30 ++++++++++++++++++++ > 2 files changed, 34 insertions(+) > > diff --git a/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h b/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h > index 91f3ce8..d04e336 100644 > --- a/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h > +++ b/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h > @@ -86,6 +86,10 @@ > #define PCIE_LUT_BASE 0x80000 > #define PCIE_LUT_LCTRL0 0x7F8 > #define PCIE_LUT_DBG 0x7FC > +#define PCIE_LUT_UDR(n) (0x800 + (n) * 8) > +#define PCIE_LUT_LDR(n) (0x804 + (n) * 8) > +#define PCIE_LUT_ENABLE (1 << 31) > +#define PCIE_LUT_ENTRY_COUNT 32 > > /* Device Configuration */ > #define DCFG_BASE 0x01e00000 > diff --git a/drivers/pci/pcie_layerscape.c b/drivers/pci/pcie_layerscape.c > index bb29222..8435446 100644 > --- a/drivers/pci/pcie_layerscape.c > +++ b/drivers/pci/pcie_layerscape.c > @@ -93,6 +93,7 @@ struct ls_pcie { > void __iomem *dbi; > void __iomem *va_cfg0; > void __iomem *va_cfg1; > + int next_lut_index; > struct pci_controller hose; > }; > > @@ -482,6 +483,34 @@ static void ls_pcie_setup_ep(struct ls_pcie *pcie, struct ls_pcie_info *info) > } > } > > +#if defined(CONFIG_LS2080A) || defined(CONFIG_LS2085A) Would CONFIG_FSL_LSCH3 serve this purpose better? We will have other SoC in the same family. York
> -----Original Message----- > From: york sun > Sent: Tuesday, March 08, 2016 11:30 AM > To: Stuart Yoder <stuart.yoder@nxp.com>; u-boot@lists.denx.de > Cc: Prabhakar Kushwaha <prabhakar.kushwaha@nxp.com>; Yang-Leo Li <leoyang.li@nxp.com> > Subject: Re: [PATCH 4/7 v4] pci/layerscape: add support for LUT > > On 03/08/2016 07:56 AM, Stuart Yoder wrote: > > From: Stuart Yoder <stuart.yoder@nxp.com> > > > > The per-PCI controller LUT (Look-Up-Table) is a 32-entry table > > that maps PCI requester IDs (bus/dev/fun) to a stream ID. > > > > This patch implements infrastructure to enable LUT initialization: > > -define registers offsets > > -add an index to 'struct ls_pcie' to track next available slot in LUT > > -add function to allocate the next available entry index > > -add function to program a LUT entry > > > > Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com> > > --- > > -v4: put new support under LS2 #ifdef > > > > .../include/asm/arch-fsl-layerscape/immap_lsch3.h | 4 +++ > > drivers/pci/pcie_layerscape.c | 30 ++++++++++++++++++++ > > 2 files changed, 34 insertions(+) > > > > diff --git a/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h > b/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h > > index 91f3ce8..d04e336 100644 > > --- a/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h > > +++ b/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h > > @@ -86,6 +86,10 @@ > > #define PCIE_LUT_BASE 0x80000 > > #define PCIE_LUT_LCTRL0 0x7F8 > > #define PCIE_LUT_DBG 0x7FC > > +#define PCIE_LUT_UDR(n) (0x800 + (n) * 8) > > +#define PCIE_LUT_LDR(n) (0x804 + (n) * 8) > > +#define PCIE_LUT_ENABLE (1 << 31) > > +#define PCIE_LUT_ENTRY_COUNT 32 > > > > /* Device Configuration */ > > #define DCFG_BASE 0x01e00000 > > diff --git a/drivers/pci/pcie_layerscape.c b/drivers/pci/pcie_layerscape.c > > index bb29222..8435446 100644 > > --- a/drivers/pci/pcie_layerscape.c > > +++ b/drivers/pci/pcie_layerscape.c > > @@ -93,6 +93,7 @@ struct ls_pcie { > > void __iomem *dbi; > > void __iomem *va_cfg0; > > void __iomem *va_cfg1; > > + int next_lut_index; > > struct pci_controller hose; > > }; > > > > @@ -482,6 +483,34 @@ static void ls_pcie_setup_ep(struct ls_pcie *pcie, struct > ls_pcie_info *info) > > } > > } > > > > +#if defined(CONFIG_LS2080A) || defined(CONFIG_LS2085A) > > Would CONFIG_FSL_LSCH3 serve this purpose better? We will have other SoC in the > same family. Maybe, I'm open to a common define if we can easily get agreement. I see at least 7 other references to: defined(CONFIG_LS2080A) || defined(CONFIG_LS2085A) I'd rather not see agreement on a common #define get mixed up in this patch series which is focused on enabling MSIs. But, perhaps we need to agree on a name and then clean all the common platform references up. I'm not sure I like FSL_LSCH3. A while back we changed the name of a u-boot directory that had "ch3" in it and changed to: "fsl-layerscape". So, I thought we were moving in the direction of removing references to "ch3". In Linux we use CONFIG_ARCH_LAYERSCAPE to refer to the ARMv8 family of SoCs. While LAYERSCAPE is not ideal, it may be better than CH3 which is a reference to an internal name that no one outside of NXP will ever understand. My vote would be CONFIG_LAYERSCAPE. Stuart
On 03/08/2016 11:36 AM, Stuart Yoder wrote: > > >> -----Original Message----- >> From: york sun >> Sent: Tuesday, March 08, 2016 11:30 AM >> To: Stuart Yoder <stuart.yoder@nxp.com>; u-boot@lists.denx.de >> Cc: Prabhakar Kushwaha <prabhakar.kushwaha@nxp.com>; Yang-Leo Li <leoyang.li@nxp.com> >> Subject: Re: [PATCH 4/7 v4] pci/layerscape: add support for LUT >> >> On 03/08/2016 07:56 AM, Stuart Yoder wrote: >>> From: Stuart Yoder <stuart.yoder@nxp.com> >>> >>> The per-PCI controller LUT (Look-Up-Table) is a 32-entry table >>> that maps PCI requester IDs (bus/dev/fun) to a stream ID. >>> >>> This patch implements infrastructure to enable LUT initialization: >>> -define registers offsets >>> -add an index to 'struct ls_pcie' to track next available slot in LUT >>> -add function to allocate the next available entry index >>> -add function to program a LUT entry >>> >>> Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com> >>> --- >>> -v4: put new support under LS2 #ifdef >>> >>> .../include/asm/arch-fsl-layerscape/immap_lsch3.h | 4 +++ >>> drivers/pci/pcie_layerscape.c | 30 ++++++++++++++++++++ >>> 2 files changed, 34 insertions(+) >>> >>> diff --git a/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h >> b/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h >>> index 91f3ce8..d04e336 100644 >>> --- a/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h >>> +++ b/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h >>> @@ -86,6 +86,10 @@ >>> #define PCIE_LUT_BASE 0x80000 >>> #define PCIE_LUT_LCTRL0 0x7F8 >>> #define PCIE_LUT_DBG 0x7FC >>> +#define PCIE_LUT_UDR(n) (0x800 + (n) * 8) >>> +#define PCIE_LUT_LDR(n) (0x804 + (n) * 8) >>> +#define PCIE_LUT_ENABLE (1 << 31) >>> +#define PCIE_LUT_ENTRY_COUNT 32 >>> >>> /* Device Configuration */ >>> #define DCFG_BASE 0x01e00000 >>> diff --git a/drivers/pci/pcie_layerscape.c b/drivers/pci/pcie_layerscape.c >>> index bb29222..8435446 100644 >>> --- a/drivers/pci/pcie_layerscape.c >>> +++ b/drivers/pci/pcie_layerscape.c >>> @@ -93,6 +93,7 @@ struct ls_pcie { >>> void __iomem *dbi; >>> void __iomem *va_cfg0; >>> void __iomem *va_cfg1; >>> + int next_lut_index; >>> struct pci_controller hose; >>> }; >>> >>> @@ -482,6 +483,34 @@ static void ls_pcie_setup_ep(struct ls_pcie *pcie, struct >> ls_pcie_info *info) >>> } >>> } >>> >>> +#if defined(CONFIG_LS2080A) || defined(CONFIG_LS2085A) >> >> Would CONFIG_FSL_LSCH3 serve this purpose better? We will have other SoC in the >> same family. > > Maybe, I'm open to a common define if we can easily get agreement. > > I see at least 7 other references to: > defined(CONFIG_LS2080A) || defined(CONFIG_LS2085A) > > I'd rather not see agreement on a common #define get mixed > up in this patch series which is focused on enabling > MSIs. But, perhaps we need to agree on a name and > then clean all the common platform references up. > > I'm not sure I like FSL_LSCH3. A while back we changed the > name of a u-boot directory that had "ch3" in it and changed to: > "fsl-layerscape". So, I thought we were moving in the direction > of removing references to "ch3". > > In Linux we use CONFIG_ARCH_LAYERSCAPE to refer to the ARMv8 > family of SoCs. While LAYERSCAPE is not ideal, it may be > better than CH3 which is a reference to an internal name > that no one outside of NXP will ever understand. > > My vote would be CONFIG_LAYERSCAPE. > We already have CONFIG_FSL_LSCH3 and CONFIG_FSL_LSCH2 in U-Boot. My concern is not the name. If you see your change fits CONFIG_FSL_LSCH3, you can go ahead to use it. Or if it only applies to CONFIG_LS2080A and CONFIG_LS2085A, you can keep your change. With aother LSCH3 SoC around the corner, I am trying to save you the trouble to fix it when that comes. York
> -----Original Message----- > From: york sun > Sent: Tuesday, March 08, 2016 1:46 PM > To: Stuart Yoder <stuart.yoder@nxp.com>; u-boot@lists.denx.de > Cc: Prabhakar Kushwaha <prabhakar.kushwaha@nxp.com>; Yang-Leo Li <leoyang.li@nxp.com> > Subject: Re: [PATCH 4/7 v4] pci/layerscape: add support for LUT > > On 03/08/2016 11:36 AM, Stuart Yoder wrote: > > > > > >> -----Original Message----- > >> From: york sun > >> Sent: Tuesday, March 08, 2016 11:30 AM > >> To: Stuart Yoder <stuart.yoder@nxp.com>; u-boot@lists.denx.de > >> Cc: Prabhakar Kushwaha <prabhakar.kushwaha@nxp.com>; Yang-Leo Li <leoyang.li@nxp.com> > >> Subject: Re: [PATCH 4/7 v4] pci/layerscape: add support for LUT > >> > >> On 03/08/2016 07:56 AM, Stuart Yoder wrote: > >>> From: Stuart Yoder <stuart.yoder@nxp.com> > >>> > >>> The per-PCI controller LUT (Look-Up-Table) is a 32-entry table > >>> that maps PCI requester IDs (bus/dev/fun) to a stream ID. > >>> > >>> This patch implements infrastructure to enable LUT initialization: > >>> -define registers offsets > >>> -add an index to 'struct ls_pcie' to track next available slot in LUT > >>> -add function to allocate the next available entry index > >>> -add function to program a LUT entry > >>> > >>> Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com> > >>> --- > >>> -v4: put new support under LS2 #ifdef > >>> > >>> .../include/asm/arch-fsl-layerscape/immap_lsch3.h | 4 +++ > >>> drivers/pci/pcie_layerscape.c | 30 ++++++++++++++++++++ > >>> 2 files changed, 34 insertions(+) > >>> > >>> diff --git a/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h > >> b/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h > >>> index 91f3ce8..d04e336 100644 > >>> --- a/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h > >>> +++ b/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h > >>> @@ -86,6 +86,10 @@ > >>> #define PCIE_LUT_BASE 0x80000 > >>> #define PCIE_LUT_LCTRL0 0x7F8 > >>> #define PCIE_LUT_DBG 0x7FC > >>> +#define PCIE_LUT_UDR(n) (0x800 + (n) * 8) > >>> +#define PCIE_LUT_LDR(n) (0x804 + (n) * 8) > >>> +#define PCIE_LUT_ENABLE (1 << 31) > >>> +#define PCIE_LUT_ENTRY_COUNT 32 > >>> > >>> /* Device Configuration */ > >>> #define DCFG_BASE 0x01e00000 > >>> diff --git a/drivers/pci/pcie_layerscape.c b/drivers/pci/pcie_layerscape.c > >>> index bb29222..8435446 100644 > >>> --- a/drivers/pci/pcie_layerscape.c > >>> +++ b/drivers/pci/pcie_layerscape.c > >>> @@ -93,6 +93,7 @@ struct ls_pcie { > >>> void __iomem *dbi; > >>> void __iomem *va_cfg0; > >>> void __iomem *va_cfg1; > >>> + int next_lut_index; > >>> struct pci_controller hose; > >>> }; > >>> > >>> @@ -482,6 +483,34 @@ static void ls_pcie_setup_ep(struct ls_pcie *pcie, struct > >> ls_pcie_info *info) > >>> } > >>> } > >>> > >>> +#if defined(CONFIG_LS2080A) || defined(CONFIG_LS2085A) > >> > >> Would CONFIG_FSL_LSCH3 serve this purpose better? We will have other SoC in the > >> same family. > > > > Maybe, I'm open to a common define if we can easily get agreement. > > > > I see at least 7 other references to: > > defined(CONFIG_LS2080A) || defined(CONFIG_LS2085A) > > > > I'd rather not see agreement on a common #define get mixed > > up in this patch series which is focused on enabling > > MSIs. But, perhaps we need to agree on a name and > > then clean all the common platform references up. > > > > I'm not sure I like FSL_LSCH3. A while back we changed the > > name of a u-boot directory that had "ch3" in it and changed to: > > "fsl-layerscape". So, I thought we were moving in the direction > > of removing references to "ch3". > > > > In Linux we use CONFIG_ARCH_LAYERSCAPE to refer to the ARMv8 > > family of SoCs. While LAYERSCAPE is not ideal, it may be > > better than CH3 which is a reference to an internal name > > that no one outside of NXP will ever understand. > > > > My vote would be CONFIG_LAYERSCAPE. > > > > We already have CONFIG_FSL_LSCH3 and CONFIG_FSL_LSCH2 in U-Boot. My concern is > not the name. If you see your change fits CONFIG_FSL_LSCH3, you can go ahead to > use it. Or if it only applies to CONFIG_LS2080A and CONFIG_LS2085A, you can keep > your change. With aother LSCH3 SoC around the corner, I am trying to save you > the trouble to fix it when that comes. Sure, makes sense...I'll change to use the existing CH3 define. I suspect that the existing references to defined(CONFIG_LS2080A) || defined(CONFIG_LS2085A) should probably change at some point. Stuart
diff --git a/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h b/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h index 91f3ce8..d04e336 100644 --- a/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h +++ b/arch/arm/include/asm/arch-fsl-layerscape/immap_lsch3.h @@ -86,6 +86,10 @@ #define PCIE_LUT_BASE 0x80000 #define PCIE_LUT_LCTRL0 0x7F8 #define PCIE_LUT_DBG 0x7FC +#define PCIE_LUT_UDR(n) (0x800 + (n) * 8) +#define PCIE_LUT_LDR(n) (0x804 + (n) * 8) +#define PCIE_LUT_ENABLE (1 << 31) +#define PCIE_LUT_ENTRY_COUNT 32 /* Device Configuration */ #define DCFG_BASE 0x01e00000 diff --git a/drivers/pci/pcie_layerscape.c b/drivers/pci/pcie_layerscape.c index bb29222..8435446 100644 --- a/drivers/pci/pcie_layerscape.c +++ b/drivers/pci/pcie_layerscape.c @@ -93,6 +93,7 @@ struct ls_pcie { void __iomem *dbi; void __iomem *va_cfg0; void __iomem *va_cfg1; + int next_lut_index; struct pci_controller hose; }; @@ -482,6 +483,34 @@ static void ls_pcie_setup_ep(struct ls_pcie *pcie, struct ls_pcie_info *info) } } +#if defined(CONFIG_LS2080A) || defined(CONFIG_LS2085A) +/* + * Return next available LUT index. + */ +static int ls_pcie_next_lut_index(struct ls_pcie *pcie) +{ + if (pcie->next_lut_index < PCIE_LUT_ENTRY_COUNT) + return pcie->next_lut_index++; + else + return -1; /* LUT is full */ +} + +/* + * Program a single LUT entry + */ +static void ls_pcie_lut_set_mapping(struct ls_pcie *pcie, int index, u32 devid, + u32 streamid) +{ + void __iomem *lut; + + lut = pcie->dbi + PCIE_LUT_BASE; + + /* leave mask as all zeroes, want to match all bits */ + writel((devid << 16), lut + PCIE_LUT_UDR(index)); + writel(streamid | PCIE_LUT_ENABLE, lut + PCIE_LUT_LDR(index)); +} +#endif + int ls_pcie_init_ctrl(int busno, enum srds_prtcl dev, struct ls_pcie_info *info) { struct ls_pcie *pcie; @@ -513,6 +542,7 @@ int ls_pcie_init_ctrl(int busno, enum srds_prtcl dev, struct ls_pcie_info *info) pcie->va_cfg1 = map_physmem(info->cfg1_phys, info->cfg1_size, MAP_NOCACHE); + pcie->next_lut_index = 0; /* outbound memory */ pci_set_region(&hose->regions[0],