PCI: designware: Move LTSSM state definitions to pcie-designware.h
Subsystems (2)
Fabio Estevam
reviewerPATCH v2 1/2 PCI: designware: Move LTSSM state definitions to pcie-designware.h
From: Fabio Estevam <[email protected]>
Move LTSSM state definitions to common pcie-designware.h so that other drivers can make use of them.
Also, in order to avoid name collision define LTSSM_STATE_MASK_KS, as keystone uses a different LTSSM mask.
---
Changes since v1:
- Fix build warning noticed by kbuild robot
drivers/pci/host/pci-keystone-dw.c | 4 ++-- drivers/pci/host/pci-layerscape.c | 1 - drivers/pci/host/pcie-designware.h | 34 ++++++++++++++++++++++++++++++++++ drivers/pci/host/pcie-spear13xx.c | 33 --------------------------------- 4 files changed, 36 insertions(+), 36 deletions(-)
drivers/pci/host/pci-keystone-dw.c +2 -2
@@ -25,7 +25,7 @@/* Application register defines */ #define LTSSM_EN_VAL 1 -#define LTSSM_STATE_MASK 0x1f +#define LTSSM_STATE_MASK_KS 0x1f#define LTSSM_STATE_L0 0x11#define DBI_CS2_EN_VAL 0x20#define OB_XLAT_EN_VAL 2
@@ -445,7 +445,7 @@ int ks_dw_pcie_link_up(struct pcie_port *pp) {u32 val =readl (pp->dbi_base +DEBUG0 ); -return (val <SSM_STATE_MASK ) ==LTSSM_STATE_L0 ; +return (val <SSM_STATE_MASK_KS ) ==LTSSM_STATE_L0 ; }void ks_dw_pcie_initiate_link_train (struct keystone_pcie *ks_pcie)
drivers/pci/host/pci-layerscape.c +0 -1
@@ -29,7 +29,6 @@/* PEX1/2 Misc Ports Status Register */ #define SCFG_PEXMSCPORTSR (pex_idx) (0x94 + (pex_idx) * 4)#define LTSSM_STATE_SHIFT 20 -#define LTSSM_STATE_MASK 0x3f#define LTSSM_PCIE_L0 0x11/* L0 state */ /* Symbol Timer Register and Filter Mask Register 1 */
drivers/pci/host/pcie-designware.h +34 -0
@@ -22,6 +22,40 @@#define MAX_MSI_IRQS 32#define MAX_MSI_CTRLS (MAX_MSI_IRQS / 32) +#define LTSSM_STATE_DETECT_QUIET 0x00 +#define LTSSM_STATE_DETECT_ACT 0x01 +#define LTSSM_STATE_POLL_ACTIVE 0x02 +#define LTSSM_STATE_POLL_COMPLIANCE 0x03 +#define LTSSM_STATE_POLL_CONFIG 0x04 +#define LTSSM_STATE_PRE_DETECT_QUIET 0x05 +#define LTSSM_STATE_DETECT_WAIT 0x06 +#define LTSSM_STATE_CFG_LINKWD_START 0x07 +#define LTSSM_STATE_CFG_LINKWD_ACEPT 0x08 +#define LTSSM_STATE_CFG_LANENUM_WAIT 0x09 +#define LTSSM_STATE_CFG_LANENUM_ACEPT 0x0A +#define LTSSM_STATE_CFG_COMPLETE 0x0B +#define LTSSM_STATE_CFG_IDLE 0x0C +#define LTSSM_STATE_RCVRY_LOCK 0x0D +#define LTSSM_STATE_RCVRY_SPEED 0x0E +#define LTSSM_STATE_RCVRY_RCVRCFG 0x0F +#define LTSSM_STATE_RCVRY_IDLE 0x10 +#define LTSSM_STATE_L0 0x11 +#define LTSSM_STATE_L0S 0x12 +#define LTSSM_STATE_L123_SEND_EIDLE 0x13 +#define LTSSM_STATE_L1_IDLE 0x14 +#define LTSSM_STATE_L2_IDLE 0x15 +#define LTSSM_STATE_L2_WAKE 0x16 +#define LTSSM_STATE_DISABLED_ENTRY 0x17 +#define LTSSM_STATE_DISABLED_IDLE 0x18 +#define LTSSM_STATE_DISABLED 0x19 +#define LTSSM_STATE_LPBK_ENTRY 0x1A +#define LTSSM_STATE_LPBK_ACTIVE 0x1B +#define LTSSM_STATE_LPBK_EXIT 0x1C +#define LTSSM_STATE_LPBK_EXIT_TIMEOUT 0x1D +#define LTSSM_STATE_HOT_RESET_ENTRY 0x1E +#define LTSSM_STATE_HOT_RESET 0x1F +#define LTSSM_STATE_MASK 0x3F +struct pcie_port {struct device *dev;u8 root_bus_nr;
drivers/pci/host/pcie-spear13xx.c +0 -33
@@ -84,39 +84,6 @@ struct pcie_app_reg {#define APPS_PM_XMT_PME_ID 5/* CR3 ID */ -#define XMLH_LTSSM_STATE_DETECT_QUIET 0x00 -#define XMLH_LTSSM_STATE_DETECT_ACT 0x01 -#define XMLH_LTSSM_STATE_POLL_ACTIVE 0x02 -#define XMLH_LTSSM_STATE_POLL_COMPLIANCE 0x03 -#define XMLH_LTSSM_STATE_POLL_CONFIG 0x04 -#define XMLH_LTSSM_STATE_PRE_DETECT_QUIET 0x05 -#define XMLH_LTSSM_STATE_DETECT_WAIT 0x06 -#define XMLH_LTSSM_STATE_CFG_LINKWD_START 0x07 -#define XMLH_LTSSM_STATE_CFG_LINKWD_ACEPT 0x08 -#define XMLH_LTSSM_STATE_CFG_LANENUM_WAIT 0x09 -#define XMLH_LTSSM_STATE_CFG_LANENUM_ACEPT 0x0A -#define XMLH_LTSSM_STATE_CFG_COMPLETE 0x0B -#define XMLH_LTSSM_STATE_CFG_IDLE 0x0C -#define XMLH_LTSSM_STATE_RCVRY_LOCK 0x0D -#define XMLH_LTSSM_STATE_RCVRY_SPEED 0x0E -#define XMLH_LTSSM_STATE_RCVRY_RCVRCFG 0x0F -#define XMLH_LTSSM_STATE_RCVRY_IDLE 0x10 -#define XMLH_LTSSM_STATE_L0 0x11 -#define XMLH_LTSSM_STATE_L0S 0x12 -#define XMLH_LTSSM_STATE_L123_SEND_EIDLE 0x13 -#define XMLH_LTSSM_STATE_L1_IDLE 0x14 -#define XMLH_LTSSM_STATE_L2_IDLE 0x15 -#define XMLH_LTSSM_STATE_L2_WAKE 0x16 -#define XMLH_LTSSM_STATE_DISABLED_ENTRY 0x17 -#define XMLH_LTSSM_STATE_DISABLED_IDLE 0x18 -#define XMLH_LTSSM_STATE_DISABLED 0x19 -#define XMLH_LTSSM_STATE_LPBK_ENTRY 0x1A -#define XMLH_LTSSM_STATE_LPBK_ACTIVE 0x1B -#define XMLH_LTSSM_STATE_LPBK_EXIT 0x1C -#define XMLH_LTSSM_STATE_LPBK_EXIT_TIMEOUT 0x1D -#define XMLH_LTSSM_STATE_HOT_RESET_ENTRY 0x1E -#define XMLH_LTSSM_STATE_HOT_RESET 0x1F -#define XMLH_LTSSM_STATE_MASK 0x3F#define XMLH_LINK_UP (1 << 6)/* CR4 ID */
Signature · 1 line
1.9.1
Fabio Estevam
reviewerPATCH v2 2/2 PCI: imx6: Use define instead of hard coded value
From: Fabio Estevam <[email protected]>
Use LTSSM_STATE_RCVRY_LOCK define instead of hard coded value in order to improve the code readability;
---
Changes since v1:
- None
drivers/pci/host/pci-imx6.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivePCIE_PHY_DEBUG_R0_LTSSM_MASKrs/pci/host/pci-imx6.c renamed from drivers/pci/host/pci-imx6.c +1 -1
@@ -508,7 +508,7 @@ static int imx6_pcie_link_up(struct pcie_port *pp)if (rx_valid &PCIE_PHY_RX_ASIC_OUT_VALID )return 0 ; -if ((debug_r0 &PCIE_PHY_DEBUG_R0_LTSSM_MASK ) !=0x0d ) +if ((debug_r0 &PCIE_PHY_DEBUG_R0_LTSSM_MASK ) !=LTSSM_STATE_RCVRY_LOCK )return 0 ;dev_err (pp->dev,"transition to gen2 is stuck, reset PHY!\n" );
Signature · 1 line
1.9.1
Re: [PATCH v2 2/2] PCI: imx6: Use define instead of hard coded value
Am Donnerstag, den 08.10.2015, 22:48 -0300 schrieb Fabio Estevam:
Quoted 5 lines
From: Fabio Estevam <[email protected]>
Use
LTSSM_STATE_RCVRY_LOCKdefine instead of hard coded value in order to improve the code readability;
Normally we end sentences to be parsed by humans with a full stop instead of a semicolon. ;)
Quoted 1 line
Signed-off-by Fabio Estevam <[email protected]>
Quoted 20 lines
---
Changes since v1:
- Nonedrivers/pci/host/pci-imx6.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivePCIE_PHY_DEBUG_R0_LTSSM_MASKrs/pci/host/pci-imx6.c renamed from drivers/pci/host/pci-imx6.c +1
-1@@ -508,7 +508,7 @@ static int imx6_pcie_link_up(struct pcie_port *pp)if (rx_valid &PCIE_PHY_RX_ASIC_OUT_VALID )return 0 ; -if ((debug_r0 &PCIE_PHY_DEBUG_R0_LTSSM_MASK ) !=0x0d ) +if ((debug_r0 &PCIE_PHY_DEBUG_R0_LTSSM_MASK ) !=LTSSM_STATE_RCVRY_LOCK )return 0 ;dev_err (pp->dev,"transition to gen2 is stuck, reset PHY!\n" );
Signature · 2 lines
Pengutronix e.K. | Lucas Stach | Industrial Linux Solutions | http://www.pengutronix.de/ |
Am Donnerstag, den 08.10.2015, 22:48 -0300 schrieb Fabio Estevam:
Quoted 8 lines
From: Fabio Estevam <[email protected]>
Move LTSSM state definitions to common
pcie-designware.hso that other drivers can make use of them.Also, in order to avoid name collision define
LTSSM_STATE_MASK_KS, as keystone uses a different LTSSM mask.
Why? Is this some kind of difference between the DW PCIe core revisions?
I don't see the other drivers using any LTSSM state value that would not fit inside the KS mask. Can we just use the narrower KS mask, even if the others have one bit more allocated in the register? This would spare us this needless differentiation in the software.
Quoted 78 lines
Signed-off-by Fabio Estevam <[email protected]>Acked-by Pratyush Anand <[email protected]>---
Changes since v1:
- Fix build warning noticed by kbuild robotdrivers/pci/host/pci-keystone-dw.c | 4 ++-- drivers/pci/host/pci-layerscape.c | 1 - drivers/pci/host/pcie-designware.h | 34 ++++++++++++++++++++++++++++++++++ drivers/pci/host/pcie-spear13xx.c | 33 --------------------------------- 4 files changed, 36 insertions(+), 36 deletions(-)drivers/pci/host/pci-keystone-dw.c +2
-2@@ -25,7 +25,7 @@/* Application register defines */ #define LTSSM_EN_VAL 1 -#define LTSSM_STATE_MASK 0x1f +#define LTSSM_STATE_MASK_KS 0x1f#define LTSSM_STATE_L0 0x11#define DBI_CS2_EN_VAL 0x20#define OB_XLAT_EN_VAL 2@@ -445,7 +445,7 @@ int ks_dw_pcie_link_up(struct pcie_port *pp) {u32 val =readl (pp->dbi_base +DEBUG0 ); -return (val <SSM_STATE_MASK ) ==LTSSM_STATE_L0 ; +return (val <SSM_STATE_MASK_KS ) ==LTSSM_STATE_L0 ; }void ks_dw_pcie_initiate_link_train (struct keystone_pcie *ks_pcie)drivers/pci/host/pci-layerscape.c +0
-1@@ -29,7 +29,6 @@/* PEX1/2 Misc Ports Status Register */ #define SCFG_PEXMSCPORTSR (pex_idx) (0x94 + (pex_idx) * 4)#define LTSSM_STATE_SHIFT 20 -#define LTSSM_STATE_MASK 0x3f#define LTSSM_PCIE_L0 0x11/* L0 state */ /* Symbol Timer Register and Filter Mask Register 1 */ drivers/pci/host/pcie-designware.h +24
-0@@ -22,6 +22,40 @@#define MAX_MSI_IRQS 32#define MAX_MSI_CTRLS (MAX_MSI_IRQS / 32) +#define LTSSM_STATE_DETECT_QUIET 0x00 +#define LTSSM_STATE_DETECT_ACT 0x01 +#define LTSSM_STATE_POLL_ACTIVE 0x02 +#define LTSSM_STATE_POLL_COMPLIANCE 0x03 +#define LTSSM_STATE_POLL_CONFIG 0x04 +#define LTSSM_STATE_PRE_DETECT_QUIET 0x05 +#define LTSSM_STATE_DETECT_WAIT 0x06 +#define LTSSM_STATE_CFG_LINKWD_START 0x07 +#define LTSSM_STATE_CFG_LINKWD_ACEPT 0x08 +#define LTSSM_STATE_CFG_LANENUM_WAIT 0x09 +#define LTSSM_STATE_CFG_LANENUM_ACEPT 0x0A +#define LTSSM_STATE_CFG_COMPLETE 0x0B +#define LTSSM_STATE_CFG_IDLE 0x0C +#define LTSSM_STATE_RCVRY_LOCK 0x0D +#define LTSSM_STATE_RCVRY_SPEED 0x0E +#define LTSSM_STATE_RCVRY_RCVRCFG 0x0F +#define LTSSM_STATE_RCVRY_IDLE 0x10 +#define LTSSM_STATE_L0 0x11 +#define LTSSM_STATE_L0S 0x12 +#define LTSSM_STATE_L123_SEND_EIDLE 0x13 +#define LTSSM_STATE_L1_IDLE 0x14 +#define LTSSM_STATE_L2_IDLE 0x15 +#define LTSSM_STATE_L2_WAKE 0x16 +#define LTSSM_STATE_DISABLED_ENTRY 0x17
drivers/pci/host/pcie-designware.h line 58
Quoted 10 lines
drivers/pci/host/pcie-designware.h +10
-0+#define LTSSM_STATE_DISABLED_IDLE 0x18 +#define LTSSM_STATE_DISABLED 0x19 +#define LTSSM_STATE_LPBK_ENTRY 0x1A +#define LTSSM_STATE_LPBK_ACTIVE 0x1B +#define LTSSM_STATE_LPBK_EXIT 0x1C +#define LTSSM_STATE_LPBK_EXIT_TIMEOUT 0x1D +#define LTSSM_STATE_HOT_RESET_ENTRY 0x1E +#define LTSSM_STATE_HOT_RESET 0x1F +#define LTSSM_STATE_MASK 0x3F +
I know this is just copy-and-paste, but can we use lowercase letters for the hex values please? This would make this fit in better with the rest of the header and is much more like any other part of the kernel.
Regards,
Lucas
Quoted 47 lines
drivers/pci/host/pcie-designware.h +0
-0struct pcie_port {struct device *dev;u8 root_bus_nr;drivers/pci/host/pcie-spear13xx.c +0
-33@@ -84,39 +84,6 @@ struct pcie_app_reg {#define APPS_PM_XMT_PME_ID 5/* CR3 ID */ -#define XMLH_LTSSM_STATE_DETECT_QUIET 0x00 -#define XMLH_LTSSM_STATE_DETECT_ACT 0x01 -#define XMLH_LTSSM_STATE_POLL_ACTIVE 0x02 -#define XMLH_LTSSM_STATE_POLL_COMPLIANCE 0x03 -#define XMLH_LTSSM_STATE_POLL_CONFIG 0x04 -#define XMLH_LTSSM_STATE_PRE_DETECT_QUIET 0x05 -#define XMLH_LTSSM_STATE_DETECT_WAIT 0x06 -#define XMLH_LTSSM_STATE_CFG_LINKWD_START 0x07 -#define XMLH_LTSSM_STATE_CFG_LINKWD_ACEPT 0x08 -#define XMLH_LTSSM_STATE_CFG_LANENUM_WAIT 0x09 -#define XMLH_LTSSM_STATE_CFG_LANENUM_ACEPT 0x0A -#define XMLH_LTSSM_STATE_CFG_COMPLETE 0x0B -#define XMLH_LTSSM_STATE_CFG_IDLE 0x0C -#define XMLH_LTSSM_STATE_RCVRY_LOCK 0x0D -#define XMLH_LTSSM_STATE_RCVRY_SPEED 0x0E -#define XMLH_LTSSM_STATE_RCVRY_RCVRCFG 0x0F -#define XMLH_LTSSM_STATE_RCVRY_IDLE 0x10 -#define XMLH_LTSSM_STATE_L0 0x11 -#define XMLH_LTSSM_STATE_L0S 0x12 -#define XMLH_LTSSM_STATE_L123_SEND_EIDLE 0x13 -#define XMLH_LTSSM_STATE_L1_IDLE 0x14 -#define XMLH_LTSSM_STATE_L2_IDLE 0x15 -#define XMLH_LTSSM_STATE_L2_WAKE 0x16 -#define XMLH_LTSSM_STATE_DISABLED_ENTRY 0x17 -#define XMLH_LTSSM_STATE_DISABLED_IDLE 0x18 -#define XMLH_LTSSM_STATE_DISABLED 0x19 -#define XMLH_LTSSM_STATE_LPBK_ENTRY 0x1A -#define XMLH_LTSSM_STATE_LPBK_ACTIVE 0x1B -#define XMLH_LTSSM_STATE_LPBK_EXIT 0x1C -#define XMLH_LTSSM_STATE_LPBK_EXIT_TIMEOUT 0x1D -#define XMLH_LTSSM_STATE_HOT_RESET_ENTRY 0x1E -#define XMLH_LTSSM_STATE_HOT_RESET 0x1F -#define XMLH_LTSSM_STATE_MASK 0x3F#define XMLH_LINK_UP (1 << 6)/* CR4 ID */
Signature · 2 lines
Pengutronix e.K. | Lucas Stach | Industrial Linux Solutions | http://www.pengutronix.de/ |
Am Montag, den 12.10.2015, 12:19 +0200 schrieb Lucas Stach:
Quoted 9 lines
Am Donnerstag, den 08.10.2015, 22:48 -0300 schrieb Fabio Estevam:
Quoted 8 lines
From: Fabio Estevam <[email protected]>
Move LTSSM state definitions to common
pcie-designware.hso that other drivers can make use of them.Also, in order to avoid name collision define
LTSSM_STATE_MASK_KS, as keystone uses a different LTSSM mask.
Quoted 7 lines
Why? Is this some kind of difference between the DW PCIe core revisions?
I don't see the other drivers using any LTSSM state value that would not fit inside the KS mask. Can we just use the narrower KS mask, even if the others have one bit more allocated in the register? This would spare us this needless differentiation in the software.
Actually the TRM states that i.MX6 is using the same narrower mask of 0x1f for the LTSSM state, so another reason to just use that.
Quoted 143 lines
Quoted 78 lines
Signed-off-by Fabio Estevam <[email protected]>Acked-by Pratyush Anand <[email protected]>---
Changes since v1:
- Fix build warning noticed by kbuild robotdrivers/pci/host/pci-keystone-dw.c | 4 ++-- drivers/pci/host/pci-layerscape.c | 1 - drivers/pci/host/pcie-designware.h | 34 ++++++++++++++++++++++++++++++++++ drivers/pci/host/pcie-spear13xx.c | 33 --------------------------------- 4 files changed, 36 insertions(+), 36 deletions(-)drivers/pci/host/pci-keystone-dw.c +2
-2@@ -25,7 +25,7 @@/* Application register defines */ #define LTSSM_EN_VAL 1 -#define LTSSM_STATE_MASK 0x1f +#define LTSSM_STATE_MASK_KS 0x1f#define LTSSM_STATE_L0 0x11#define DBI_CS2_EN_VAL 0x20#define OB_XLAT_EN_VAL 2@@ -445,7 +445,7 @@ int ks_dw_pcie_link_up(struct pcie_port *pp) {u32 val =readl (pp->dbi_base +DEBUG0 ); -return (val <SSM_STATE_MASK ) ==LTSSM_STATE_L0 ; +return (val <SSM_STATE_MASK_KS ) ==LTSSM_STATE_L0 ; }void ks_dw_pcie_initiate_link_train (struct keystone_pcie *ks_pcie)drivers/pci/host/pci-layerscape.c +0
-1@@ -29,7 +29,6 @@/* PEX1/2 Misc Ports Status Register */ #define SCFG_PEXMSCPORTSR (pex_idx) (0x94 + (pex_idx) * 4)#define LTSSM_STATE_SHIFT 20 -#define LTSSM_STATE_MASK 0x3f#define LTSSM_PCIE_L0 0x11/* L0 state */ /* Symbol Timer Register and Filter Mask Register 1 */ drivers/pci/host/pcie-designware.h +24
-0@@ -22,6 +22,40 @@#define MAX_MSI_IRQS 32#define MAX_MSI_CTRLS (MAX_MSI_IRQS / 32) +#define LTSSM_STATE_DETECT_QUIET 0x00 +#define LTSSM_STATE_DETECT_ACT 0x01 +#define LTSSM_STATE_POLL_ACTIVE 0x02 +#define LTSSM_STATE_POLL_COMPLIANCE 0x03 +#define LTSSM_STATE_POLL_CONFIG 0x04 +#define LTSSM_STATE_PRE_DETECT_QUIET 0x05 +#define LTSSM_STATE_DETECT_WAIT 0x06 +#define LTSSM_STATE_CFG_LINKWD_START 0x07 +#define LTSSM_STATE_CFG_LINKWD_ACEPT 0x08 +#define LTSSM_STATE_CFG_LANENUM_WAIT 0x09 +#define LTSSM_STATE_CFG_LANENUM_ACEPT 0x0A +#define LTSSM_STATE_CFG_COMPLETE 0x0B +#define LTSSM_STATE_CFG_IDLE 0x0C +#define LTSSM_STATE_RCVRY_LOCK 0x0D +#define LTSSM_STATE_RCVRY_SPEED 0x0E +#define LTSSM_STATE_RCVRY_RCVRCFG 0x0F +#define LTSSM_STATE_RCVRY_IDLE 0x10 +#define LTSSM_STATE_L0 0x11 +#define LTSSM_STATE_L0S 0x12 +#define LTSSM_STATE_L123_SEND_EIDLE 0x13 +#define LTSSM_STATE_L1_IDLE 0x14 +#define LTSSM_STATE_L2_IDLE 0x15 +#define LTSSM_STATE_L2_WAKE 0x16 +#define LTSSM_STATE_DISABLED_ENTRY 0x17drivers/pci/host/pcie-designware.h line 58
Quoted 10 lines
drivers/pci/host/pcie-designware.h +10
-0+#define LTSSM_STATE_DISABLED_IDLE 0x18 +#define LTSSM_STATE_DISABLED 0x19 +#define LTSSM_STATE_LPBK_ENTRY 0x1A +#define LTSSM_STATE_LPBK_ACTIVE 0x1B +#define LTSSM_STATE_LPBK_EXIT 0x1C +#define LTSSM_STATE_LPBK_EXIT_TIMEOUT 0x1D +#define LTSSM_STATE_HOT_RESET_ENTRY 0x1E +#define LTSSM_STATE_HOT_RESET 0x1F +#define LTSSM_STATE_MASK 0x3F +I know this is just copy-and-paste, but can we use lowercase letters for the hex values please? This would make this fit in better with the rest of the header and is much more like any other part of the kernel.
Regards,
LucasQuoted 47 lines
drivers/pci/host/pcie-designware.h +0
-0struct pcie_port {struct device *dev;u8 root_bus_nr;drivers/pci/host/pcie-spear13xx.c +0
-33@@ -84,39 +84,6 @@ struct pcie_app_reg {#define APPS_PM_XMT_PME_ID 5/* CR3 ID */ -#define XMLH_LTSSM_STATE_DETECT_QUIET 0x00 -#define XMLH_LTSSM_STATE_DETECT_ACT 0x01 -#define XMLH_LTSSM_STATE_POLL_ACTIVE 0x02 -#define XMLH_LTSSM_STATE_POLL_COMPLIANCE 0x03 -#define XMLH_LTSSM_STATE_POLL_CONFIG 0x04 -#define XMLH_LTSSM_STATE_PRE_DETECT_QUIET 0x05 -#define XMLH_LTSSM_STATE_DETECT_WAIT 0x06 -#define XMLH_LTSSM_STATE_CFG_LINKWD_START 0x07 -#define XMLH_LTSSM_STATE_CFG_LINKWD_ACEPT 0x08 -#define XMLH_LTSSM_STATE_CFG_LANENUM_WAIT 0x09 -#define XMLH_LTSSM_STATE_CFG_LANENUM_ACEPT 0x0A -#define XMLH_LTSSM_STATE_CFG_COMPLETE 0x0B -#define XMLH_LTSSM_STATE_CFG_IDLE 0x0C -#define XMLH_LTSSM_STATE_RCVRY_LOCK 0x0D -#define XMLH_LTSSM_STATE_RCVRY_SPEED 0x0E -#define XMLH_LTSSM_STATE_RCVRY_RCVRCFG 0x0F -#define XMLH_LTSSM_STATE_RCVRY_IDLE 0x10 -#define XMLH_LTSSM_STATE_L0 0x11 -#define XMLH_LTSSM_STATE_L0S 0x12 -#define XMLH_LTSSM_STATE_L123_SEND_EIDLE 0x13 -#define XMLH_LTSSM_STATE_L1_IDLE 0x14 -#define XMLH_LTSSM_STATE_L2_IDLE 0x15 -#define XMLH_LTSSM_STATE_L2_WAKE 0x16 -#define XMLH_LTSSM_STATE_DISABLED_ENTRY 0x17 -#define XMLH_LTSSM_STATE_DISABLED_IDLE 0x18 -#define XMLH_LTSSM_STATE_DISABLED 0x19 -#define XMLH_LTSSM_STATE_LPBK_ENTRY 0x1A -#define XMLH_LTSSM_STATE_LPBK_ACTIVE 0x1B -#define XMLH_LTSSM_STATE_LPBK_EXIT 0x1C -#define XMLH_LTSSM_STATE_LPBK_EXIT_TIMEOUT 0x1D -#define XMLH_LTSSM_STATE_HOT_RESET_ENTRY 0x1E -#define XMLH_LTSSM_STATE_HOT_RESET 0x1F -#define XMLH_LTSSM_STATE_MASK 0x3F#define XMLH_LINK_UP (1 << 6)/* CR4 ID */
Signature · 2 lines
Pengutronix e.K. | Lucas Stach | Industrial Linux Solutions | http://www.pengutronix.de/ |
Fabio Estevam
reviewerQuoted 2 lines · Lucas Stach
Actually the TRM states that i.MX6 is using the same narrower mask of
0x1ffor the LTSSM state, so another reason to just use that.
The manual I have seems to tell a different value :-)
Please check: http://cache.freescale.com/files/32bit/doc/ref_manual/IMX6DQRM.pdf
"48.12.11
Debug Register 0 (PCIE_PL_DEBUG0)"
says "[5:0]: xmlh_ltssm_state LTSSM current state. See source for encodings"
Regards,
Fabio Estevam
Am Montag, den 12.10.2015, 09:31 -0300 schrieb Fabio Estevam:
Quoted 5 lines · Lucas Stach
Quoted 3 lines
Actually the TRM states that i.MX6 is using the same narrower mask of
0x1ffor the LTSSM state, so another reason to just use that.
Quoted 9 lines
The manual I have seems to tell a different value :-)
Please check: http://cache.freescale.com/files/32bit/doc/ref_manual/IMX6DQRM.pdf
"48.12.11
Debug Register 0 (PCIE_PL_DEBUG0)"says "[5:0]:
xmlh_ltssm_stateLTSSM current state. See source for encodings"
Urgh. Seems we got confused by all the different defines.
In fact [5:0] is 0x1f, so i.MX6 uses the same mask as Keystone, yet the common and shared header now states a mask of 0x3f with Keystone being the exception. That doesn't reflect reality which isn't nice IMHO. I think it would be better to just use 0x1f as the LTSSM mask in the shared header and use this for all drivers until a need arises for using the extra bit on SPEAR.
Regards,
Lucas
Signature · 2 lines
Pengutronix e.K. | Lucas Stach | Industrial Linux Solutions | http://www.pengutronix.de/ |
Fabio Estevam
reviewerQuoted 10 lines · Lucas Stach
Am Montag, den 12.10.2015, 09:31 -0300 schrieb Fabio Estevam:
Quoted 9 lines · Lucas Stach
Quoted 3 lines
Actually the TRM states that i.MX6 is using the same narrower mask of
0x1ffor the LTSSM state, so another reason to just use that.The manual I have seems to tell a different value :-)
Please check: http://cache.freescale.com/files/32bit/doc/ref_manual/IMX6DQRM.pdf
Quoted 8 lines
Quoted 5 lines
"48.12.11
Debug Register 0 (PCIE_PL_DEBUG0)"says "[5:0]:
xmlh_ltssm_stateLTSSM current state. See source for encodings"Urgh. Seems we got confused by all the different defines.
In fact [5:0] is
0x1f, so i.MX6 uses the same mask as Keystone, yet the
No, bits [5..0] means 0x3f. What am I missing here?
Am Montag, den 12.10.2015, 11:04 -0300 schrieb Fabio Estevam:
Quoted 14 lines · Lucas Stach
Quoted 13 lines
Am Montag, den 12.10.2015, 09:31 -0300 schrieb Fabio Estevam:
Quoted 12 lines · Lucas Stach
Quoted 3 lines
Actually the TRM states that i.MX6 is using the same narrower mask of
0x1ffor the LTSSM state, so another reason to just use that.The manual I have seems to tell a different value :-)
Please check: http://cache.freescale.com/files/32bit/doc/ref_manual/IMX6DQRM.pdf
"48.12.11
Debug Register 0 (PCIE_PL_DEBUG0)"
Quoted 7 lines
Quoted 5 lines
Quoted 2 lines
says "[5:0]:
xmlh_ltssm_stateLTSSM current state. See source for encodings"Urgh. Seems we got confused by all the different defines.
In fact [5:0] is
0x1f, so i.MX6 uses the same mask as Keystone, yet theNo, bits [5..0] means
0x3f. What am I missing here?
That you have to deal with people who are apparently unable to count to
5. *me puts on the brown paper bag*
Still I would love to have a definition in the common header that's usable for all drivers. So unless there is a reason to use 0x3f (and we don't have any LTSSM state defines which would use the extra bit) keep it at 0x1f. Simpler, cleaner... IMHO.
Regards,
Lucas
Signature · 2 lines
Pengutronix e.K. | Lucas Stach | Industrial Linux Solutions | http://www.pengutronix.de/ |
Fabio Estevam
reviewerQuoted 4 lines · Lucas Stach
Still I would love to have a definition in the common header that's usable for all drivers. So unless there is a reason to use
0x3f(and we don't have any LTSSM state defines which would use the extra bit) keep it at0x1f. Simpler, cleaner... IMHO.
Sure, just did as suggested in v3. Thanks
Fabio Estevam
reviewerHi Lucas,
Quoted 10 lines · Lucas Stach
Am Donnerstag, den 08.10.2015, 22:48 -0300 schrieb Fabio Estevam:
Quoted 8 lines
From: Fabio Estevam <[email protected]>
Move LTSSM state definitions to common
pcie-designware.hso that other drivers can make use of them.Also, in order to avoid name collision define
LTSSM_STATE_MASK_KS, as keystone uses a different LTSSM mask.Why? Is this some kind of difference between the DW PCIe core revisions?
This patch does not alter the LTSSM masks that are used currently and they match the ones stated in the SoC manuals.
For keystone:
http://www.ti.com/lit/ug/sprugs6d/sprugs6d.pdf
Section "48.9.11 Debug Register 0 (PCIE_PL_DEBUG0)" shows that LTSSM current state has 6 bits.
For i.MX6:
Section "3.9.11 Debug 0 Register (DEBUG0)" shows that LTSSM_STATE field has 5 bits.
Regards,
Fabio Estevam