| Message ID | 20220406152347.85908-8-Frank.Li@nxp.com |
|---|---|
| State | New |
| Headers | show |
| Series | Enable designware PCI EP EDMA locally | expand |
On Wed, Apr 06, 2022 at 10:23:45AM -0500, Frank Li wrote: > Add a "flags" field to the "struct dw_edma_chip" so that the controller > drivers can pass flags that are relevant to the platform. > > DW_EDMA_CHIP_LOCAL - Used by the controller drivers accessing eDMA > locally. Local eDMA access doesn't require generating MSIs to the remote. > > Signed-off-by: Frank Li <Frank.Li@nxp.com> > --- > Change from v5 to v6 > - use enum instead of define Hm, why have you decided to do that? I don't see a well justified reason to use the enumeration here, but see my next comment for details. > > Change from v4 to v5 > - split two two patch > - rework commit message > Change from v3 to v4 > none > Change from v2 to v3 > - rework commit message > - Change to DW_EDMA_CHIP_32BIT_DBI > - using DW_EDMA_CHIP_LOCAL control msi > - Apply Bjorn's comments, > if (!j) { > control |= DW_EDMA_V0_LIE; > if (!(chan->chip->flags & DW_EDMA_CHIP_LOCAL)) > control |= DW_EDMA_V0_RIE; > } > > if ((chan->chip->flags & DW_EDMA_CHIP_REG32BIT) || > !IS_ENABLED(CONFIG_64BIT)) { > SET_CH_32(...); > SET_CH_32(...); > } else { > SET_CH_64(...); > } > > > Change from v1 to v2 > - none > > drivers/dma/dw-edma/dw-edma-v0-core.c | 9 ++++++--- > include/linux/dma/edma.h | 9 +++++++++ > 2 files changed, 15 insertions(+), 3 deletions(-) > > diff --git a/drivers/dma/dw-edma/dw-edma-v0-core.c b/drivers/dma/dw-edma/dw-edma-v0-core.c > index 8ddc537d11fd6..30f8bfe6e5712 100644 > --- a/drivers/dma/dw-edma/dw-edma-v0-core.c > +++ b/drivers/dma/dw-edma/dw-edma-v0-core.c > @@ -301,6 +301,7 @@ u32 dw_edma_v0_core_status_abort_int(struct dw_edma *dw, enum dw_edma_dir dir) > static void dw_edma_v0_core_write_chunk(struct dw_edma_chunk *chunk) > { > struct dw_edma_burst *child; > + struct dw_edma_chan *chan = chunk->chan; > struct dw_edma_v0_lli __iomem *lli; > struct dw_edma_v0_llp __iomem *llp; > u32 control = 0, i = 0; > @@ -314,9 +315,11 @@ static void dw_edma_v0_core_write_chunk(struct dw_edma_chunk *chunk) > j = chunk->bursts_alloc; > list_for_each_entry(child, &chunk->burst->list, list) { > j--; > - if (!j) > - control |= (DW_EDMA_V0_LIE | DW_EDMA_V0_RIE); > - > + if (!j) { > + control |= DW_EDMA_V0_LIE; > + if (!(chan->dw->chip->flags & DW_EDMA_CHIP_LOCAL)) > + control |= DW_EDMA_V0_RIE; > + } > /* Channel control */ > SET_LL_32(&lli[i].control, control); > /* Transfer size */ > diff --git a/include/linux/dma/edma.h b/include/linux/dma/edma.h > index c2039246fc08c..bec444e2939b2 100644 > --- a/include/linux/dma/edma.h > +++ b/include/linux/dma/edma.h > @@ -33,6 +33,14 @@ enum dw_edma_map_format { > EDMA_MF_HDMA_COMPAT = 0x5 > }; > > +/** > + * enum dw_edma_chip_flags - Flags specific to an eDMA chip > + * @DW_EDMA_CHIP_LOCAL: eDMA is used locally by an endpoint > + */ > +enum dw_edma_chip_flags { > + DW_EDMA_CHIP_LOCAL = BIT(0), > +}; > + > /** > * struct dw_edma_chip - representation of DesignWare eDMA controller hardware > * @dev: struct device of the eDMA controller > @@ -53,6 +61,7 @@ struct dw_edma_chip { > int id; > int nr_irqs; > const struct dw_edma_core_ops *ops; > + enum dw_edma_chip_flags flags; There is no point in having the named enumeration here since the flags field semantics is actually a bitfield rather than a single value. If you want to stick to the enumerated flags, then please use the anonymous enum like this: +enum { + DW_EDMA_CHIP_LOCAL = BIT(0), +}; and explicit unsigned int type of the flags field. -Sergey > > void __iomem *reg_base; > > -- > 2.35.1 >
On Wed, Apr 13, 2022 at 12:18:37PM +0300, Serge Semin wrote: > On Wed, Apr 06, 2022 at 10:23:45AM -0500, Frank Li wrote: > > Add a "flags" field to the "struct dw_edma_chip" so that the controller > > drivers can pass flags that are relevant to the platform. > > > > DW_EDMA_CHIP_LOCAL - Used by the controller drivers accessing eDMA > > locally. Local eDMA access doesn't require generating MSIs to the remote. > > > > Signed-off-by: Frank Li <Frank.Li@nxp.com> > > --- > > > Change from v5 to v6 > > - use enum instead of define > > Hm, why have you decided to do that? I don't see a well justified > reason to use the enumeration here, but see my next comment for > details. It was me who suggested using the enums for flags instead of defines. Enums helps with kdoc and it also provides a neat way to group flags together. > > > > > Change from v4 to v5 > > - split two two patch > > - rework commit message > > Change from v3 to v4 > > none > > Change from v2 to v3 > > - rework commit message > > - Change to DW_EDMA_CHIP_32BIT_DBI > > - using DW_EDMA_CHIP_LOCAL control msi > > - Apply Bjorn's comments, > > if (!j) { > > control |= DW_EDMA_V0_LIE; > > if (!(chan->chip->flags & DW_EDMA_CHIP_LOCAL)) > > control |= DW_EDMA_V0_RIE; > > } > > > > if ((chan->chip->flags & DW_EDMA_CHIP_REG32BIT) || > > !IS_ENABLED(CONFIG_64BIT)) { > > SET_CH_32(...); > > SET_CH_32(...); > > } else { > > SET_CH_64(...); > > } > > > > > > Change from v1 to v2 > > - none > > > > drivers/dma/dw-edma/dw-edma-v0-core.c | 9 ++++++--- > > include/linux/dma/edma.h | 9 +++++++++ > > 2 files changed, 15 insertions(+), 3 deletions(-) > > > > diff --git a/drivers/dma/dw-edma/dw-edma-v0-core.c b/drivers/dma/dw-edma/dw-edma-v0-core.c > > index 8ddc537d11fd6..30f8bfe6e5712 100644 > > --- a/drivers/dma/dw-edma/dw-edma-v0-core.c > > +++ b/drivers/dma/dw-edma/dw-edma-v0-core.c [...] > > + enum dw_edma_chip_flags flags; > > There is no point in having the named enumeration here since the flags > field semantics is actually a bitfield rather than a single value. If > you want to stick to the enumerated flags, then please use the > anonymous enum like this: I agree with using u32 for flags field but I don't agree with anonymous enums. Enums with a name conveys information of what the enumerated types represent. If you just look at your example below, it is difficult to guess the purpose of this enum. Thanks, Mani > +enum { > + DW_EDMA_CHIP_LOCAL = BIT(0), > +}; > and explicit unsigned int type of the flags field. > > -Sergey > > > > > void __iomem *reg_base; > > > > -- > > 2.35.1 > >
On Wed, Apr 13, 2022 at 02:58:08PM +0530, Manivannan Sadhasivam wrote: > On Wed, Apr 13, 2022 at 12:18:37PM +0300, Serge Semin wrote: > > On Wed, Apr 06, 2022 at 10:23:45AM -0500, Frank Li wrote: > > > Add a "flags" field to the "struct dw_edma_chip" so that the controller > > > drivers can pass flags that are relevant to the platform. > > > > > > DW_EDMA_CHIP_LOCAL - Used by the controller drivers accessing eDMA > > > locally. Local eDMA access doesn't require generating MSIs to the remote. > > > > > > Signed-off-by: Frank Li <Frank.Li@nxp.com> > > > --- > > > > > Change from v5 to v6 > > > - use enum instead of define > > > > Hm, why have you decided to do that? I don't see a well justified > > reason to use the enumeration here, but see my next comment for > > details. > > It was me who suggested using the enums for flags instead of defines. > Enums helps with kdoc and it also provides a neat way to group flags together. > > > > > > > > > Change from v4 to v5 > > > - split two two patch > > > - rework commit message > > > Change from v3 to v4 > > > none > > > Change from v2 to v3 > > > - rework commit message > > > - Change to DW_EDMA_CHIP_32BIT_DBI > > > - using DW_EDMA_CHIP_LOCAL control msi > > > - Apply Bjorn's comments, > > > if (!j) { > > > control |= DW_EDMA_V0_LIE; > > > if (!(chan->chip->flags & DW_EDMA_CHIP_LOCAL)) > > > control |= DW_EDMA_V0_RIE; > > > } > > > > > > if ((chan->chip->flags & DW_EDMA_CHIP_REG32BIT) || > > > !IS_ENABLED(CONFIG_64BIT)) { > > > SET_CH_32(...); > > > SET_CH_32(...); > > > } else { > > > SET_CH_64(...); > > > } > > > > > > > > > Change from v1 to v2 > > > - none > > > > > > drivers/dma/dw-edma/dw-edma-v0-core.c | 9 ++++++--- > > > include/linux/dma/edma.h | 9 +++++++++ > > > 2 files changed, 15 insertions(+), 3 deletions(-) > > > > > > diff --git a/drivers/dma/dw-edma/dw-edma-v0-core.c b/drivers/dma/dw-edma/dw-edma-v0-core.c > > > index 8ddc537d11fd6..30f8bfe6e5712 100644 > > > --- a/drivers/dma/dw-edma/dw-edma-v0-core.c > > > +++ b/drivers/dma/dw-edma/dw-edma-v0-core.c > > [...] > > > > + enum dw_edma_chip_flags flags; > > > > There is no point in having the named enumeration here since the flags > > field semantics is actually a bitfield rather than a single value. If > > you want to stick to the enumerated flags, then please use the > > anonymous enum like this: > > I agree with using u32 for flags field but I don't agree with anonymous enums. > Enums with a name conveys information of what the enumerated types represent. > If you just look at your example below, it is difficult to guess the purpose of > this enum. I see your point. Ok, no anonymization then.) @Frank could you please update the field type to unsigned int or u32 then? Personally I prefer having "unsigned int" here, since that's the type used by the compiler if no negative values is enumerated. Though u32 would be ok too. -Serget > > Thanks, > Mani > > > +enum { > > + DW_EDMA_CHIP_LOCAL = BIT(0), > > +}; > > and explicit unsigned int type of the flags field. > > > > -Sergey > > > > > > > > void __iomem *reg_base; > > > > > > -- > > > 2.35.1 > > >
diff --git a/drivers/dma/dw-edma/dw-edma-v0-core.c b/drivers/dma/dw-edma/dw-edma-v0-core.c index 8ddc537d11fd6..30f8bfe6e5712 100644 --- a/drivers/dma/dw-edma/dw-edma-v0-core.c +++ b/drivers/dma/dw-edma/dw-edma-v0-core.c @@ -301,6 +301,7 @@ u32 dw_edma_v0_core_status_abort_int(struct dw_edma *dw, enum dw_edma_dir dir) static void dw_edma_v0_core_write_chunk(struct dw_edma_chunk *chunk) { struct dw_edma_burst *child; + struct dw_edma_chan *chan = chunk->chan; struct dw_edma_v0_lli __iomem *lli; struct dw_edma_v0_llp __iomem *llp; u32 control = 0, i = 0; @@ -314,9 +315,11 @@ static void dw_edma_v0_core_write_chunk(struct dw_edma_chunk *chunk) j = chunk->bursts_alloc; list_for_each_entry(child, &chunk->burst->list, list) { j--; - if (!j) - control |= (DW_EDMA_V0_LIE | DW_EDMA_V0_RIE); - + if (!j) { + control |= DW_EDMA_V0_LIE; + if (!(chan->dw->chip->flags & DW_EDMA_CHIP_LOCAL)) + control |= DW_EDMA_V0_RIE; + } /* Channel control */ SET_LL_32(&lli[i].control, control); /* Transfer size */ diff --git a/include/linux/dma/edma.h b/include/linux/dma/edma.h index c2039246fc08c..bec444e2939b2 100644 --- a/include/linux/dma/edma.h +++ b/include/linux/dma/edma.h @@ -33,6 +33,14 @@ enum dw_edma_map_format { EDMA_MF_HDMA_COMPAT = 0x5 }; +/** + * enum dw_edma_chip_flags - Flags specific to an eDMA chip + * @DW_EDMA_CHIP_LOCAL: eDMA is used locally by an endpoint + */ +enum dw_edma_chip_flags { + DW_EDMA_CHIP_LOCAL = BIT(0), +}; + /** * struct dw_edma_chip - representation of DesignWare eDMA controller hardware * @dev: struct device of the eDMA controller @@ -53,6 +61,7 @@ struct dw_edma_chip { int id; int nr_irqs; const struct dw_edma_core_ops *ops; + enum dw_edma_chip_flags flags; void __iomem *reg_base;
Add a "flags" field to the "struct dw_edma_chip" so that the controller drivers can pass flags that are relevant to the platform. DW_EDMA_CHIP_LOCAL - Used by the controller drivers accessing eDMA locally. Local eDMA access doesn't require generating MSIs to the remote. Signed-off-by: Frank Li <Frank.Li@nxp.com> --- Change from v5 to v6 - use enum instead of define Change from v4 to v5 - split two two patch - rework commit message Change from v3 to v4 none Change from v2 to v3 - rework commit message - Change to DW_EDMA_CHIP_32BIT_DBI - using DW_EDMA_CHIP_LOCAL control msi - Apply Bjorn's comments, if (!j) { control |= DW_EDMA_V0_LIE; if (!(chan->chip->flags & DW_EDMA_CHIP_LOCAL)) control |= DW_EDMA_V0_RIE; } if ((chan->chip->flags & DW_EDMA_CHIP_REG32BIT) || !IS_ENABLED(CONFIG_64BIT)) { SET_CH_32(...); SET_CH_32(...); } else { SET_CH_64(...); } Change from v1 to v2 - none drivers/dma/dw-edma/dw-edma-v0-core.c | 9 ++++++--- include/linux/dma/edma.h | 9 +++++++++ 2 files changed, 15 insertions(+), 3 deletions(-)