diff mbox series

[5/4] ARM: footbridge: remove custom DMA address handling

Message ID 20220825091752.3535657-1-arnd@kernel.org (mailing list archive)
State New, archived
Headers show
Series None | expand

Commit Message

Arnd Bergmann Aug. 25, 2022, 9:17 a.m. UTC 
From: Arnd Bergmann <arnd@arndb.de>

Footbridge is the last Arm platform that has its own
__virt_to_bus()/__bus_to_virt()/phys_to_dma()/dma_to_phys() abstraction,
but this is just a simple offset now.

For ISA DMA (floppy etc), the offset is actually zero, so the special
hack to subtract the bus offset can be removed from common code.

For PCI devices, the offset that is programmed into the PCI bridge
must also be set in each device using dma_direct_set_offset().
As Arm does not have a pcibios_bus_add_device() helper yet, just
use a bus notifier for this.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
I added this patch on top of the footbridge series,
the other four are unmodified.

 arch/arm/include/asm/dma-direct.h             |  1 -
 arch/arm/include/asm/dma.h                    |  2 +-
 arch/arm/include/asm/memory.h                 | 11 --------
 arch/arm/mach-footbridge/Kconfig              |  1 -
 arch/arm/mach-footbridge/common.c             | 10 -------
 arch/arm/mach-footbridge/dc21285.c            | 28 ++++++++++++++++++-
 .../mach-footbridge/include/mach/dma-direct.h |  8 ------
 .../arm/mach-footbridge/include/mach/memory.h |  9 ------
 8 files changed, 28 insertions(+), 42 deletions(-)
 delete mode 100644 arch/arm/include/asm/dma-direct.h
 delete mode 100644 arch/arm/mach-footbridge/include/mach/dma-direct.h

Comments

Russell King (Oracle) Aug. 25, 2022, 10:34 a.m. UTC | #1 
On Thu, Aug 25, 2022 at 11:17:52AM +0200, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@arndb.de>
> 
> Footbridge is the last Arm platform that has its own
> __virt_to_bus()/__bus_to_virt()/phys_to_dma()/dma_to_phys() abstraction,
> but this is just a simple offset now.
> 
> For ISA DMA (floppy etc), the offset is actually zero, so the special
> hack to subtract the bus offset can be removed from common code.

No, the offset is definitely not zero for ISA devices. ISA devices have
the same view of system memory as PCI devices, since ISA devices go
through a southbridge and on to the PCI bus. So their view of the system
is exactly the same as any other PCI device.
Russell King (Oracle) Aug. 25, 2022, 10:35 a.m. UTC | #2 
On Thu, Aug 25, 2022 at 11:17:52AM +0200, Arnd Bergmann wrote:
> -/*
> - * The footbridge is programmed to expose the system RAM at 0xe0000000.
> - * The requirement is that the RAM isn't placed at bus address 0, which
> - * would clash with VGA cards.
> - */
> -#define BUS_OFFSET		0xe0000000
> -#define __virt_to_bus(x)	((x) + (BUS_OFFSET - PAGE_OFFSET))
> -#define __bus_to_virt(x)	((x) - (BUS_OFFSET - PAGE_OFFSET))

... and this tells you why that is with ISA devices.

So, one hell of a big NAK to this patch, sorry.
Arnd Bergmann Aug. 25, 2022, 5:17 p.m. UTC | #3 
On Thu, Aug 25, 2022 at 12:35 PM Russell King (Oracle)
<linux@armlinux.org.uk> wrote:
>
> On Thu, Aug 25, 2022 at 11:17:52AM +0200, Arnd Bergmann wrote:
> > -/*
> > - * The footbridge is programmed to expose the system RAM at 0xe0000000.
> > - * The requirement is that the RAM isn't placed at bus address 0, which
> > - * would clash with VGA cards.
> > - */
> > -#define BUS_OFFSET           0xe0000000
> > -#define __virt_to_bus(x)     ((x) + (BUS_OFFSET - PAGE_OFFSET))
> > -#define __bus_to_virt(x)     ((x) - (BUS_OFFSET - PAGE_OFFSET))
>
> ... and this tells you why that is with ISA devices.
>
> So, one hell of a big NAK to this patch, sorry.

Thanks for taking a look. Apparently I misread the interface, and I assumed
that the __bus_to_virt() BUS_OFFSET in set_dma_addr() would
just cancel out the dma_map_single() in isa_enable_dma().

It turns out the logic is actually simpler and we just convert between
virtual address and bus address three times. for most drivers.

I found one driver (cs89x0-isa) that uses isa_virt_to_bus()/virt_to_phys()
instead of dma_map_single(), but that is probably never used here.
Once we move the DMA offset into the device, all other ISA devices will
also get the default zero-offset, but I think that will be fine as long as the
"isa_dma_dev" that is used for the final translation has the correct offset.

I'll try to come up with a new version tomorrow.

       Arnd
diff mbox series

Patch

diff --git a/arch/arm/include/asm/dma-direct.h b/arch/arm/include/asm/dma-direct.h
deleted file mode 100644
index 4f7bcde03abb..000000000000
--- a/arch/arm/include/asm/dma-direct.h
+++ /dev/null
@@ -1 +0,0 @@ 
-#include <mach/dma-direct.h>
diff --git a/arch/arm/include/asm/dma.h b/arch/arm/include/asm/dma.h
index 05f29a72150b..ab3863a3cdf2 100644
--- a/arch/arm/include/asm/dma.h
+++ b/arch/arm/include/asm/dma.h
@@ -106,7 +106,7 @@  extern void set_dma_sg(unsigned int chan, struct scatterlist *sg, int nr_sg);
  */
 extern void __set_dma_addr(unsigned int chan, void *addr);
 #define set_dma_addr(chan, addr)				\
-	__set_dma_addr(chan, (void *)__bus_to_virt(addr))
+	__set_dma_addr(chan, (void *)(addr))
 
 /* Set the DMA byte count for this channel
  *
diff --git a/arch/arm/include/asm/memory.h b/arch/arm/include/asm/memory.h
index a55a9038abc8..d8eef4bd8c71 100644
--- a/arch/arm/include/asm/memory.h
+++ b/arch/arm/include/asm/memory.h
@@ -369,17 +369,6 @@  static inline unsigned long __virt_to_idmap(unsigned long x)
 
 #define virt_to_idmap(x)	__virt_to_idmap((unsigned long)(x))
 
-/*
- * Virtual <-> DMA view memory address translations
- * Again, these are *only* valid on the kernel direct mapped RAM
- * memory.  Use of these is *deprecated* (and that doesn't mean
- * use the __ prefixed forms instead.)  See dma-mapping.h.
- */
-#ifndef __virt_to_bus
-#define __virt_to_bus	__virt_to_phys
-#define __bus_to_virt	__phys_to_virt
-#endif
-
 /*
  * Conversion between a struct page and a physical address.
  *
diff --git a/arch/arm/mach-footbridge/Kconfig b/arch/arm/mach-footbridge/Kconfig
index cd18871fd03f..f452663576d1 100644
--- a/arch/arm/mach-footbridge/Kconfig
+++ b/arch/arm/mach-footbridge/Kconfig
@@ -44,7 +44,6 @@  endmenu
 # Footbridge support
 config FOOTBRIDGE
 	def_bool y
-	select ARCH_HAS_PHYS_TO_DMA
 	select ARCH_MIGHT_HAVE_PC_SERIO
 	select ISA_DMA_API
 
diff --git a/arch/arm/mach-footbridge/common.c b/arch/arm/mach-footbridge/common.c
index 9483eccea5ae..629e4676ed77 100644
--- a/arch/arm/mach-footbridge/common.c
+++ b/arch/arm/mach-footbridge/common.c
@@ -281,13 +281,3 @@  void footbridge_restart(enum reboot_mode mode, const char *cmd)
 		*CSR_SA110_CNTL |= (1 << 13);
 	}
 }
-
-dma_addr_t phys_to_dma(struct device *dev, phys_addr_t paddr)
-{
-	return paddr + (BUS_OFFSET - PHYS_OFFSET);
-}
-
-phys_addr_t dma_to_phys(struct device *dev, dma_addr_t dev_addr)
-{
-	return dev_addr - (BUS_OFFSET - PHYS_OFFSET);
-}
diff --git a/arch/arm/mach-footbridge/dc21285.c b/arch/arm/mach-footbridge/dc21285.c
index 372600bc15c8..e34b33a955e2 100644
--- a/arch/arm/mach-footbridge/dc21285.c
+++ b/arch/arm/mach-footbridge/dc21285.c
@@ -5,6 +5,7 @@ 
  *  Copyright (C) 1998-2001 Russell King
  *  Copyright (C) 1998-2000 Phil Blundell
  */
+#include <linux/dma-map-ops.h>
 #include <linux/kernel.h>
 #include <linux/pci.h>
 #include <linux/interrupt.h>
@@ -241,6 +242,29 @@  static irqreturn_t dc21285_parity_irq(int irq, void *dev_id)
 	return IRQ_HANDLED;
 }
 
+/*
+ * The footbridge is programmed to expose the system RAM at 0xe0000000.
+ * The requirement is that the RAM isn't placed at bus address 0, which
+ * would clash with VGA cards.
+ */
+#define BUS_OFFSET 0xe0000000
+
+static int dc21285_pci_bus_notifier(struct notifier_block *nb,
+				    unsigned long action,
+				    void *data)
+{
+	if (action != BUS_NOTIFY_ADD_DEVICE)
+		return NOTIFY_DONE;
+
+	dma_direct_set_offset(data, PHYS_OFFSET, BUS_OFFSET, SZ_256M);
+
+	return NOTIFY_OK;
+}
+
+static struct notifier_block dc21285_pci_bus_nb = {
+	.notifier_call = dc21285_pci_bus_notifier,
+};
+
 int __init dc21285_setup(int nr, struct pci_sys_data *sys)
 {
 	struct resource *res;
@@ -266,6 +290,8 @@  int __init dc21285_setup(int nr, struct pci_sys_data *sys)
 	pci_add_resource_offset(&sys->resources, &res[0], sys->mem_offset);
 	pci_add_resource_offset(&sys->resources, &res[1], sys->mem_offset);
 
+	bus_register_notifier(&pci_bus_type, &dc21285_pci_bus_nb);
+
 	return 1;
 }
 
@@ -329,7 +355,7 @@  void __init dc21285_preinit(void)
 	 */
 	*CSR_PCICSRBASE       = 0xf4000000;
 	*CSR_PCICSRIOBASE     = 0;
-	*CSR_PCISDRAMBASE     = __virt_to_bus(PAGE_OFFSET);
+	*CSR_PCISDRAMBASE     = BUS_OFFSET;
 	*CSR_PCIROMBASE       = 0;
 	*CSR_PCICMD = PCI_COMMAND_MEMORY | PCI_COMMAND_MASTER |
 		      PCI_COMMAND_INVALIDATE | PCICMD_ERROR_BITS;
diff --git a/arch/arm/mach-footbridge/include/mach/dma-direct.h b/arch/arm/mach-footbridge/include/mach/dma-direct.h
deleted file mode 100644
index 01f9e8367c00..000000000000
--- a/arch/arm/mach-footbridge/include/mach/dma-direct.h
+++ /dev/null
@@ -1,8 +0,0 @@ 
-/* SPDX-License-Identifier: GPL-2.0 */
-#ifndef MACH_FOOTBRIDGE_DMA_DIRECT_H
-#define MACH_FOOTBRIDGE_DMA_DIRECT_H 1
-
-dma_addr_t phys_to_dma(struct device *dev, phys_addr_t paddr);
-phys_addr_t dma_to_phys(struct device *dev, dma_addr_t dev_addr);
-
-#endif /* MACH_FOOTBRIDGE_DMA_DIRECT_H */
diff --git a/arch/arm/mach-footbridge/include/mach/memory.h b/arch/arm/mach-footbridge/include/mach/memory.h
index 8f0ca8e8b37e..9516877667d7 100644
--- a/arch/arm/mach-footbridge/include/mach/memory.h
+++ b/arch/arm/mach-footbridge/include/mach/memory.h
@@ -16,15 +16,6 @@ 
 #ifndef __ASM_ARCH_MEMORY_H
 #define __ASM_ARCH_MEMORY_H
 
-/*
- * The footbridge is programmed to expose the system RAM at 0xe0000000.
- * The requirement is that the RAM isn't placed at bus address 0, which
- * would clash with VGA cards.
- */
-#define BUS_OFFSET		0xe0000000
-#define __virt_to_bus(x)	((x) + (BUS_OFFSET - PAGE_OFFSET))
-#define __bus_to_virt(x)	((x) - (BUS_OFFSET - PAGE_OFFSET))
-
 /*
  * Cache flushing area.
  */