add PCI_DMA_{64,32}BIT constants
Subsystems (4)
PATCH 1/1 add PCI_DMA_{64,32}BIT constants
Been meaning to do this for ages...
Another one for the janitors.
Please do a
bk pull bk://kernel.bkbits.net/jgarzik/pci-dma-mask-2.6
This will update the following files:
Documentation/DMA-mapping.txt | 16 ++++++++-------- drivers/net/e1000/e1000.h | 2 -- drivers/net/ixgb/ixgb.h | 2 -- include/linux/pci.h | 3 +++ 4 files changed, 11 insertions(+), 12 deletions(-)
through these ChangeSets:
<[email protected]> (04/03/23 1.1849)
Create PCI_DMA_{64,32]BIT constants, for use in passing to pci_set_{consistent_}dma_mask().
Use them in e1000 and ixgb.
Documentation/DMA-mapping.txt +8 -8
@@ -132,7 +132,7 @@ The standard 32-bit addressing PCI device would do something like this: - if (pci_set_dma_mask(pdev, 0xffffffff)) { + if (pci_set_dma_mask(pdev, PCI_DMA_32BIT)) { printk(KERN_WARNING "mydev: No suitable DMA available.\n"); goto ignore_this_device;
@@ -151,9 +151,9 @@ int using_dac; - if (!pci_set_dma_mask(pdev, 0xffffffffffffffff)) { + if (!pci_set_dma_mask(pdev, PCI_DMA_64BIT)) { using_dac = 1; - } else if (!pci_set_dma_mask(pdev, 0xffffffff)) { + } else if (!pci_set_dma_mask(pdev, PCI_DMA_32BIT)) { using_dac = 0; } else { printk(KERN_WARNING
@@ -166,14 +166,14 @@ int using_dac, consistent_using_dac; - if (!pci_set_dma_mask(pdev, 0xffffffffffffffff)) { + if (!pci_set_dma_mask(pdev, PCI_DMA_64BIT)) { using_dac = 1; consistent_using_dac = 1; - pci_set_consistent_dma_mask(pdev, 0xffffffffffffffff) - } else if (!pci_set_dma_mask(pdev, 0xffffffff)) { + pci_set_consistent_dma_mask(pdev, PCI_DMA_64BIT); + } else if (!pci_set_dma_mask(pdev, PCI_DMA_32BIT)) { using_dac = 0; consistent_using_dac = 0; - pci_set_consistent_dma_mask(pdev, 0xffffffff) + pci_set_consistent_dma_mask(pdev, PCI_DMA_32BIT); } else { printk(KERN_WARNING "mydev: No suitable DMA available.\n");
@@ -215,7 +215,7 @@ Here is pseudo-code showing how this might be done: - #define PLAYBACK_ADDRESS_BITS 0xffffffff + #define PLAYBACK_ADDRESS_BITS PCI_DMA_32BIT #define RECORD_ADDRESS_BITS 0x00ffffff struct my_sound_card *card;
drivers/net/e1000/e1000.h +0 -2
@@ -74,8 +74,6 @@#define BAR_0 0#define BAR_1 1#define BAR_5 5 -#define PCI_DMA_64BIT 0xffffffffffffffffULL -#define PCI_DMA_32BIT 0x00000000ffffffffULLstruct e1000_adapter ;
drivers/net/ixgb/ixgb.h +0 -2
@@ -65,8 +65,6 @@#define BAR_0 0#define BAR_1 1#define BAR_5 5 -#define PCI_DMA_64BIT 0xffffffffffffffffULL -#define PCI_DMA_32BIT 0x00000000ffffffffULL#include "ixgb_hw.h" #include "ixgb_ee.h"
include/linux/pci.h +3 -0
@@ -362,6 +362,9 @@#define PCI_DMA_FROMDEVICE 2#define PCI_DMA_NONE 3 +#define PCI_DMA_64BIT 0xffffffffffffffffULL +#define PCI_DMA_32BIT 0x00000000ffffffffULL +#define DEVICE_COUNT_COMPATIBLE 4#define DEVICE_COUNT_RESOURCE 12
Quoted 10 lines · Jeff Garzik
Been meaning to do this for ages...
Another one for the janitors.
Please do a
bk pull bk://kernel.bkbits.net/jgarzik/pci-dma-mask-2.6
This will update the following files:
Nice, I've pulled this to my pci tree and will forward it on to Linus in the next round of pci patches after 2.6.5 is out.
thanks,
greg k-h
Quoted 8 lines · Greg KH
Quoted 6 lines · Jeff Garzik
Been meaning to do this for ages...
Another one for the janitors.
Please do a
Quoted 7 lines
Quoted 3 lines
bk pull bk://kernel.bkbits.net/jgarzik/pci-dma-mask-2.6
This will update the following files:
Nice, I've pulled this to my pci tree and will forward it on to Linus in the next round of pci patches after 2.6.5 is out.
Yeah well... in the intervening time, somebody on IRC commented
"so what is so PCI-specific about those constants?"
They probably ought to be DMA_{32,64}BIT_MASK or somesuch.
Jeff
Quoted 14 lines · Jeff Garzik
Quoted 13 lines · Greg KH
Quoted 9 lines · Jeff Garzik
Been meaning to do this for ages...
Another one for the janitors.
Please do a
bk pull bk://kernel.bkbits.net/jgarzik/pci-dma-mask-2.6
This will update the following files:
Quoted 8 lines
Quoted 2 lines
Nice, I've pulled this to my pci tree and will forward it on to Linus in the next round of pci patches after 2.6.5 is out.
Yeah well... in the intervening time, somebody on IRC commented
"so what is so PCI-specific about those constants?"
They probably ought to be DMA_{32,64}BIT_MASK or somesuch.
Heh, ok, care to make up another patch for this? :)
greg k-h
Quoted 21 lines · Greg KH
Quoted 19 lines · Jeff Garzik
Quoted 16 lines · Greg KH
Quoted 9 lines · Jeff Garzik
Been meaning to do this for ages...
Another one for the janitors.
Please do a
bk pull bk://kernel.bkbits.net/jgarzik/pci-dma-mask-2.6
This will update the following files:
Nice, I've pulled this to my pci tree and will forward it on to Linus in the next round of pci patches after 2.6.5 is out.
Quoted 8 lines
Quoted 5 lines
Yeah well... in the intervening time, somebody on IRC commented
"so what is so PCI-specific about those constants?"
They probably ought to be DMA_{32,64}BIT_MASK or somesuch.
Heh, ok, care to make up another patch for this? :)
Yeah, but not at the moment :) So just drop it for now, I suppose.
Jeff
On Tue, 23 Mar 2004 00:23:05 -0500 Jeff Garzik <[email protected]> wrote:
| | Been meaning to do this for ages... | | Another one for the janitors.
Quoted 11 lines
Quoted 8 lines
Quoted 2 lines
Nice, I've pulled this to my pci tree and will forward it on to Linus in the next round of pci patches after 2.6.5 is out.
Yeah well... in the intervening time, somebody on IRC commented
"so what is so PCI-specific about those constants?"
They probably ought to be DMA_{32,64}BIT_MASK or somesuch.
Heh, ok, care to make up another patch for this? :)
Here's an updated patch, applies to 2.6.5-rc2-bk9. I left the DMA_xxBIT_MASK defines in linux/pci.h, although they aren't necessarily PCI-specific. Would we prefer to have them in linux/dma-mapping.h ?
--
~Randy
// linux-2.6.5-rc2-bk9
Quoted 1 line
From <[email protected]>
Create DMA_{64,32]BIT_MASK constants, for use in passing to pci_set_{consistent_}dma_mask().
Use them in e1000 and ixgb.
diffstat:=
Documentation/DMA-mapping.txt | 16 ++++++++-------- drivers/net/e1000/e1000.h | 2 -- drivers/net/e1000/e1000_main.c | 4 ++-- drivers/net/ixgb/ixgb.h | 2 -- drivers/net/ixgb/ixgb_main.c | 4 ++-- include/linux/pci.h | 3 +++ 6 files changed, 15 insertions(+), 16 deletions(-)
drivers/net/ixgb/ixgb.h renamed from drivers/net/ixgb/ixgb.h~dma_mask +0 -2
@@ -65,8 +65,6 @@ struct ixgb_adapter;#define BAR_0 0#define BAR_1 1#define BAR_5 5 -#define PCI_DMA_64BIT 0xffffffffffffffffULL -#define PCI_DMA_32BIT 0x00000000ffffffffULL#include "ixgb_hw.h" #include "ixgb_ee.h"
drivers/net/ixgb/ixgb_main.c renamed from drivers/net/ixgb/ixgb_main.c~dma_mask +2 -2
@@ -308,10 +308,10 @@ ixgb_probe(struct pci_dev *pdev, const sreturn i; } -if (!(i =pci_set_dma_mask (pdev,PCI_DMA_64BIT ))) { +if (!(i =pci_set_dma_mask (pdev,DMA_64BIT_MASK ))) { pci_using_dac =1 ; }else { -if ((i =pci_set_dma_mask (pdev,PCI_DMA_32BIT ))) { +if ((i =pci_set_dma_mask (pdev,DMA_32BIT_MASK ))) {IXGB_ERR ("No usable DMA configuration, aborting\n" );return i; }
drivers/net/e1000/e1000_main.c renamed from drivers/net/e1000/e1000_main.c~dma_mask +2 -2
@@ -383,10 +383,10 @@ e1000_probe(struct pci_dev *pdev,if ((err =pci_enable_device (pdev)))return err; -if (!(err =pci_set_dma_mask (pdev,PCI_DMA_64BIT ))) { +if (!(err =pci_set_dma_mask (pdev,DMA_64BIT_MASK ))) { pci_using_dac =1 ; }else { -if ((err =pci_set_dma_mask (pdev,PCI_DMA_32BIT ))) { +if ((err =pci_set_dma_mask (pdev,DMA_32BIT_MASK ))) {E1000_ERR ("No usable DMA configuration, aborting\n" );return err; }
drivers/net/e1000/e1000.h renamed from drivers/net/e1000/e1000.h~dma_mask +0 -2
@@ -74,8 +74,6 @@#define BAR_0 0#define BAR_1 1#define BAR_5 5 -#define PCI_DMA_64BIT 0xffffffffffffffffULL -#define PCI_DMA_32BIT 0x00000000ffffffffULLstruct e1000_adapter ;
include/linux/pci.h renamed from include/linux/pci.h~dma_mask +3 -0
@@ -362,6 +362,9 @@ enum pci_mmap_state {#define PCI_DMA_FROMDEVICE 2#define PCI_DMA_NONE 3 +#define DMA_64BIT_MASK 0xffffffffffffffffULL +#define DMA_32BIT_MASK 0x00000000ffffffffULL +#define DEVICE_COUNT_COMPATIBLE 4#define DEVICE_COUNT_RESOURCE 12
Documentation/DMA-mapping.txt renamed from Documentation/DMA-mapping.txt~dma_mask +8 -8
@@ -132,7 +132,7 @@ exactly why. The standard 32-bit addressing PCI device would do something like this: - if (pci_set_dma_mask(pdev, 0xffffffff)) { + if (pci_set_dma_mask(pdev, DMA_32BIT_MASK)) { printk(KERN_WARNING "mydev: No suitable DMA available.\n"); goto ignore_this_device;
@@ -151,9 +151,9 @@ all 64-bits when accessing streaming DMA int using_dac; - if (!pci_set_dma_mask(pdev, 0xffffffffffffffff)) { + if (!pci_set_dma_mask(pdev, DMA_64BIT_MASK)) { using_dac = 1; - } else if (!pci_set_dma_mask(pdev, 0xffffffff)) { + } else if (!pci_set_dma_mask(pdev, DMA_32BIT_MASK)) { using_dac = 0; } else { printk(KERN_WARNING
@@ -166,14 +166,14 @@ the case would look like this: int using_dac, consistent_using_dac; - if (!pci_set_dma_mask(pdev, 0xffffffffffffffff)) { + if (!pci_set_dma_mask(pdev, DMA_64BIT_MASK)) { using_dac = 1; consistent_using_dac = 1; - pci_set_consistent_dma_mask(pdev, 0xffffffffffffffff) - } else if (!pci_set_dma_mask(pdev, 0xffffffff)) { + pci_set_consistent_dma_mask(pdev, DMA_64BIT_MASK); + } else if (!pci_set_dma_mask(pdev, DMA_32BIT_MASK)) { using_dac = 0; consistent_using_dac = 0; - pci_set_consistent_dma_mask(pdev, 0xffffffff) + pci_set_consistent_dma_mask(pdev, DMA_32BIT_MASK); } else { printk(KERN_WARNING "mydev: No suitable DMA available.\n");
@@ -215,7 +215,7 @@ most specific mask. Here is pseudo-code showing how this might be done: - #define PLAYBACK_ADDRESS_BITS 0xffffffff + #define PLAYBACK_ADDRESS_BITS DMA_32BIT_MASK #define RECORD_ADDRESS_BITS 0x00ffffff struct my_sound_card *card;
Quoted 16 lines · Randy.Dunlap
On Tue, 23 Mar 2004 00:23:05 -0500 Jeff Garzik <[email protected]> wrote:
| | Been meaning to do this for ages... | | Another one for the janitors.Quoted 10 lines
Quoted 8 lines
Quoted 2 lines
Nice, I've pulled this to my pci tree and will forward it on to Linus in the next round of pci patches after 2.6.5 is out.
Yeah well... in the intervening time, somebody on IRC commented
"so what is so PCI-specific about those constants?"
They probably ought to be DMA_{32,64}BIT_MASK or somesuch.
Quoted 8 lines
Quoted 1 line
Heh, ok, care to make up another patch for this? :)
Here's an updated patch, applies to 2.6.5-rc2-bk9. I left the
DMA_xxBIT_MASKdefines inlinux/pci.h, although they aren't necessarily PCI-specific. Would we prefer to have them inlinux/dma-mapping.h?
Put them whereever the DMA direction constants are, I suppose...
Jeff
Quoted 17 lines · Randy.Dunlap
On Tue, 23 Mar 2004 00:23:05 -0500 Jeff Garzik <[email protected]> wrote:
| | Been meaning to do this for ages... | | Another one for the janitors.Quoted 10 lines
Quoted 8 lines
Quoted 2 lines
Nice, I've pulled this to my pci tree and will forward it on to Linus in the next round of pci patches after 2.6.5 is out.
Yeah well... in the intervening time, somebody on IRC commented
"so what is so PCI-specific about those constants?"
They probably ought to be DMA_{32,64}BIT_MASK or somesuch.
Quoted 7 lines
Quoted 1 line
Heh, ok, care to make up another patch for this? :)
Here's an updated patch, applies to 2.6.5-rc2-bk9. I left the
DMA_xxBIT_MASKdefines inlinux/pci.h, although they aren't necessarily PCI-specific. Would we prefer to have them inlinux/dma-mapping.h?
Ok, this looks good. I've backed out Jeff's patch and applied this one instead to my tree. If you want to move the defines, please send me a patch relative to this one.
thanks,
greg k-h
On Tue, 30 Mar 2004 17:26:15 -0800 Greg KH <[email protected]> wrote:
| On Mon, Mar 29, 2004 at 10:36:04PM -0800, Randy.Dunlap wrote: | > On Tue, 23 Mar 2004 00:23:05 -0500 Jeff Garzik <[email protected]> wrote: | > | > | | > | Been meaning to do this for ages... | > | | > | Another one for the janitors. | > | > >>>Nice, I've pulled this to my pci tree and will forward it on to Linus in | > >>>the next round of pci patches after 2.6.5 is out. | > >> | > >>Yeah well... in the intervening time, somebody on IRC commented | > >> | > >>"so what is so PCI-specific about those constants?" | > >> | > >>They probably ought to be DMA_{32,64}BIT_MASK or somesuch. | > > | > > | > > Heh, ok, care to make up another patch for this? :) | > | > | > Here's an updated patch, applies to 2.6.5-rc2-bk9. | > I left the DMA_xxBIT_MASK defines in linux/pci.h, although | > they aren't necessarily PCI-specific. Would we prefer to | > have them in linux/dma-mapping.h ? | | Ok, this looks good. I've backed out Jeff's patch and applied this one | instead to my tree. If you want to move the defines, please send me a | patch relative to this one.
Yes, let's move the defines like Jeff suggested.
Additional (relative) patch is below.
Thanks,
--
~Randy
// linux-2.6.5-rc3
// move DMA_nnBIT_MASK to linux/dma-mapping.h;
diffstat:=
include/linux/dma-mapping.h | 3 +++ include/linux/pci.h | 3 --- 2 files changed, 3 insertions(+), 3 deletions(-)
include/linux/pci.h renamed from include/linux/pci.h~dma_bits +0 -3
@@ -362,9 +362,6 @@ enum pci_mmap_state {#define PCI_DMA_FROMDEVICE 2#define PCI_DMA_NONE 3 -#define DMA_64BIT_MASK 0xffffffffffffffffULL -#define DMA_32BIT_MASK 0x00000000ffffffffULL -#define DEVICE_COUNT_COMPATIBLE 4#define DEVICE_COUNT_RESOURCE 12
include/linux/dma-mapping.h renamed from include/linux/dma-mapping.h~dma_bits +3 -0
@@ -10,6 +10,9 @@ enum dma_data_direction { DMA_NONE = 3, }; +#define DMA_64BIT_MASK 0xffffffffffffffffULL +#define DMA_32BIT_MASK 0x00000000ffffffffULL +#include <asm/dma-mapping.h> /* Backwards compat, remove in 2.7.x */
Greg KH
AppliedQuoted 3 lines · Randy.Dunlap
Yes, let's move the defines like Jeff suggested.
Additional (relative) patch is below.
Applied, thanks.
greg k-h