OF/PCI: Export inbound memory interface to PCI RC drivers.
Subsystems (2)
PATCH 1/2 OF/PCI: Export inbound memory interface to PCI RC drivers.
From: Oza Pawandeep <[email protected]>
The patch exports interface to PCIe RC drivers so that, Drivers can get their inbound memory configuration.
It provides basis for IOVA reservations for inbound memory holes, if RC is not capable of addressing all the host memory, Specifically when IOMMU is enabled and on ARMv8 where 64bit IOVA could be allocated.
It handles multiple inbound windows, and returns resources, and is left to the caller, how it wants to use them.
--- drivers/pci/of.c | 96 ++++++++++++++++++++++++++++++++++++++++++ include/linux/of_pci.h | 7 +++ 2 files changed, 103 insertions(+)
drivers/pci/of.c +96 -0
@@ -347,6 +347,102 @@ int devm_of_pci_get_host_bridge_resources(struct device *dev,return err; }EXPORT_SYMBOL_GPL (devm_of_pci_get_host_bridge_resources); + +/** +* of_pci_get_dma_ranges - Parse PCI host bridge inbound resources from DT +* @np: device node of the host bridge having the dma-ranges property +* @resources: list where the range of resources will be added after DT parsing +* +* It is the caller's job to free the @resources list. +* +* This function will parse the "dma-ranges" property of a +* PCI host bridge device node and setup the resource mapping based +* on its content. +* +* It returns zero if the range parsing has been successful or a standard error +* value if it failed. +*/ + +int of_pci_get_dma_ranges (struct device_node *dn,struct list_head *resources) +{ +struct device_node *node =of_node_get (dn); +int rlen; +int pna =of_n_addr_cells (node); +const int na =3 , ns =2 ; +int np = pna + na + ns; +int ret =0 ; +struct resource *res; +const u32 *dma_ranges; +struct of_pci_range range; + +if (!node) +return -EINVAL ; + +while (1 ) { + dma_ranges =of_get_property (node,"dma-ranges" , &rlen); + +/* Ignore empty ranges, they imply no translation required. */ +if (dma_ranges && rlen >0 ) +break ; + +/* no dma-ranges, they imply no translation required. */ +if (!dma_ranges) +break ; + + node =of_get_next_parent (node); + +if (!node) +break ; + } + +if (!dma_ranges) { +pr_debug ("pcie device has no dma-ranges defined for node(%s)\n" , + dn->full_name); + ret = -EINVAL ; + goto out; + } + +while ((rlen -= np *4 ) >=0 ) { + range.pci_space =be32_to_cpup ((const __be32 *) &dma_ranges[0 ]); + range.pci_addr =of_read_number (dma_ranges +1 , ns); + range.cpu_addr =of_translate_dma_address (node, + dma_ranges + na); + range.size =of_read_number (dma_ranges + pna + na, ns); + range.flags =IORESOURCE_MEM ; + + dma_ranges += np; + +/* +* If we failed translation or got a zero-sized region +* then skip this range. +*/ +if (range.cpu_addr ==OF_BAD_ADDR || range.size ==0 ) +continue ; + + res =kzalloc (sizeof (struct resource ),GFP_KERNEL ); +if (!res) { + ret = -ENOMEM ; + goto parse_failed; + } + + ret =of_pci_range_to_resource (&range, dn, res); +if (ret) { +kfree (res); +continue ; + } + +pci_add_resource_offset (resources, res, + res->start - range.pci_addr); + } +return ret; + +parse_failed: +pci_free_resource_list (resources); +out: +of_node_put (node); +return ret; +} +EXPORT_SYMBOL_GPL (of_pci_get_dma_ranges);#endif /* CONFIG_OF_ADDRESS */ #if IS_ENABLED (CONFIG_OF_IRQ )
include/linux/of_pci.h +7 -0
@@ -13,6 +13,7 @@ struct device_node;struct device_node *of_pci_find_child_device (struct device_node *parent,unsigned int devfn);int of_pci_get_devfn (struct device_node *np); +int of_pci_get_dma_ranges (struct device_node *np,struct list_head *resources);void of_pci_check_probe_only (void );#else static inline struct device_node *of_pci_find_child_device(struct device_node *parent,
@@ -26,6 +27,12 @@ static inline int of_pci_get_devfn(struct device_node *np) return -EINVAL; } +static inline int of_pci_get_dma_ranges (struct device_node *np, +struct list_head *resources) +{ +return -EINVAL ; +} +static inline void of_pci_check_probe_only (void ) { }#endif
Signature · 1 line
2.20.1
PATCH 2/2 PCI/of fix of_dma_get_range; get PCI specific dma-ranges
From: Oza Pawandeep <[email protected]>
current device framework and OF framework integration assumes dma-ranges in a way where memory-mapped devices define their dma-ranges. (child-bus-address, parent-bus-address, length).
of_dma_configure is specifically written to take care of memory mapped devices. but no implementation exists for pci to take care of pcie based memory ranges.
for e.g. iproc based SOCs and other SOCs(such as rcar) have PCI world dma-ranges.
dma-ranges = <0x43000000 0x00 0x00 0x00 0x00 0x80 0x00>;
this patch fixes the following problems of_dma_get_range.
1) return of wrong size as 0.
2) not handling absence of dma-ranges which is valid for PCI master.
3) not handling multipe inbound windows.
4) in order to get largest possible dma_mask. this patch also retuns the largest possible size based on dma-ranges,
for e.g.
dma-ranges = <0x43000000 0x00 0x00 0x00 0x00 0x80 0x00>; we should get dev->coherent_dma_mask=0x7fffffffff.
based on which IOVA allocation space will honour PCI host bridge limitations.
the implementation hooks bus specific callbacks for getting dma-ranges.
--- drivers/of/address.c | 220 +++++++++++++++++++++++++++++-------------- 1 file changed, 150 insertions(+), 70 deletions(-)
drivers/of/address.c +150 -70
@@ -8,6 +8,7 @@#include <linux/logic_pio.h> #include <linux/module.h> #include <linux/of_address.h> +#include <linux/of_pci.h> #include <linux/pci.h> #include <linux/pci_regs.h> #include <linux/sizes.h>
@@ -48,6 +49,8 @@ struct of_bus { int na, int ns, int pna);int (*translate)(__be32 *addr,u64 offset,int na);unsigned int (*get_flags)(const __be32 *addr); +int (*get_dma_ranges)(struct device_node *np, +u64 *dma_addr,u64 *paddr,u64 *size); };/*
@@ -174,6 +177,143 @@ static int of_bus_pci_translate(__be32 *addr, u64 offset, int na)return of_bus_default_translate (addr +1 , offset, na -1 ); } +static int of_bus_pci_get_dma_ranges (struct device_node *np,u64 *dma_addr, +u64 *paddr,u64 *size) +{ +struct device_node *node =of_node_get (np); +int ret =0 ; +struct resource_entry *window; +LIST_HEAD (res); + +if (!node) +return -EINVAL ; + + *size =0 ; +/* +* PCI dma-ranges is not mandatory property. +* many devices do no need to have it, since +* host bridge does not require inbound memory +* configuration or rather have design limitations. +* so we look for dma-ranges, if missing we +* just return the caller full size, and also +* no dma-ranges suggests that, host bridge allows +* whatever comes in, so we set dma_addr to 0. +*/ + ret =of_pci_get_dma_ranges (np, &res); +if (!ret) { +resource_list_for_each_entry (window, &res) { +struct resource *res_dma = window->res; + +if (*size <resource_size (res_dma)) { + *dma_addr = res_dma->start - window->offset; + *paddr = res_dma->start; + *size =resource_size (res_dma); + } + } + } +pci_free_resource_list (&res); + +/* +* return the largest possible size, +* since PCI master allows everything. +*/ +if (*size ==0 ) { +pr_debug ("empty/zero size dma-ranges found for node(%s)\n" , + np->full_name); + *size =DMA_BIT_MASK (sizeof (dma_addr_t) *8 ) -1 ; + *dma_addr = *paddr =0 ; + ret =0 ; + } + +pr_debug ("dma_addr(%llx) cpu_addr(%llx) size(%llx)\n" , + *dma_addr, *paddr, *size); + +of_node_put (node); + +return ret; +} + +static int get_dma_ranges (struct device_node *np,u64 *dma_addr, +u64 *paddr,u64 *size) +{ +struct device_node *node =of_node_get (np); +const __be32 *ranges =NULL ; +int len, naddr, nsize, pna; +int ret =0 ; +u64 dmaaddr; + +if (!node) +return -EINVAL ; + +while (1 ) { + naddr =of_n_addr_cells (node); + nsize =of_n_size_cells (node); + node =of_get_next_parent (node); +if (!node) +break ; + + ranges =of_get_property (node,"dma-ranges" , &len); + +/* Ignore empty ranges, they imply no translation required */ +if (ranges && len >0 ) +break ; + +/* +* At least empty ranges has to be defined for parent node if +* DMA is supported +*/ +if (!ranges) +break ; + } + +if (!ranges) { +pr_debug ("no dma-ranges found for node(%s)\n" , np->full_name); + ret = -ENODEV ; + goto out; + } + + len /=sizeof (u32); + + pna =of_n_addr_cells (node); + +/* dma-ranges format: +* DMA addr : naddr cells +* CPU addr : pna cells +* size : nsize cells +*/ + dmaaddr =of_read_number (ranges, naddr); + *paddr =of_translate_dma_address (np, ranges); +if (*paddr ==OF_BAD_ADDR ) { +pr_err ("translation of DMA address(%pad) to CPU address failed node(%s)\n" , + dma_addr, np->full_name); + ret = -EINVAL ; + goto out; + } + *dma_addr = dmaaddr; + + *size =of_read_number (ranges + naddr + pna, nsize); + +pr_debug ("dma_addr(%llx) cpu_addr(%llx) size(%llx)\n" , + *dma_addr, *paddr, *size); + +out: +of_node_put (node); + +return ret; +} + +static int of_bus_isa_get_dma_ranges (struct device_node *np,u64 *dma_addr, +u64 *paddr,u64 *size) +{ +return get_dma_ranges (np, dma_addr, paddr, size); +} + +static int of_bus_default_get_dma_ranges (struct device_node *np,u64 *dma_addr, +u64 *paddr,u64 *size) +{ +return get_dma_ranges (np, dma_addr, paddr, size); +} +const __be32 *of_get_pci_address (struct device_node *dev,int bar_no,u64 *size,unsigned int *flags) {
@@ -438,6 +578,7 @@ static struct of_bus of_busses[] = { .map = of_bus_pci_map, .translate = of_bus_pci_translate, .get_flags = of_bus_pci_get_flags, + .get_dma_ranges = of_bus_pci_get_dma_ranges, },#endif /* CONFIG_PCI */ /* ISA */
@@ -449,6 +590,7 @@ static struct of_bus of_busses[] = { .map = of_bus_isa_map, .translate = of_bus_isa_translate, .get_flags = of_bus_isa_get_flags, + .get_dma_ranges = of_bus_isa_get_dma_ranges, },/* Default */ {
@@ -459,6 +601,7 @@ static struct of_bus of_busses[] = { .map = of_bus_default_map, .translate = of_bus_default_translate, .get_flags = of_bus_default_get_flags, + .get_dma_ranges = of_bus_default_get_dma_ranges, }, };
@@ -917,80 +1060,17 @@ EXPORT_SYMBOL(of_io_request_and_map);* size : nsize cells * * It returns -ENODEV if "dma-ranges" property was not found -* for this device in DT. +* for this device in DT, except if PCI device then, dma-ranges +* can be optional property, and in that case returns size with +* entire host memory. */ int of_dma_get_range (struct device_node *np,u64 *dma_addr,u64 *paddr,u64 *size) { -struct device_node *node =of_node_get (np); -const __be32 *ranges =NULL ; -int len, naddr, nsize, pna; -int ret =0 ; -u64 dmaaddr; - -if (!node) -return -EINVAL ; - -while (1 ) { -struct device_node *parent; - - naddr =of_n_addr_cells (node); - nsize =of_n_size_cells (node); - - parent =__of_get_dma_parent (node); -of_node_put (node); - - node = parent; -if (!node) -break ; - - ranges =of_get_property (node,"dma-ranges" , &len); - -/* Ignore empty ranges, they imply no translation required */ -if (ranges && len >0 ) -break ; - -/* -* At least empty ranges has to be defined for parent node if -* DMA is supported -*/ -if (!ranges) -break ; - } - -if (!ranges) { -pr_debug ("no dma-ranges found for node(%pOF)\n" , np); - ret = -ENODEV ; - goto out; - } - - len /=sizeof (u32); - - pna =of_n_addr_cells (node); - -/* dma-ranges format: -* DMA addr : naddr cells -* CPU addr : pna cells -* size : nsize cells -*/ - dmaaddr =of_read_number (ranges, naddr); - *paddr =of_translate_dma_address (np, ranges); -if (*paddr ==OF_BAD_ADDR ) { -pr_err ("translation of DMA address(%pad) to CPU address failed node(%pOF)\n" , - dma_addr, np); - ret = -EINVAL ; - goto out; - } - *dma_addr = dmaaddr; - - *size =of_read_number (ranges + naddr + pna, nsize); - -pr_debug ("dma_addr(%llx) cpu_addr(%llx) size(%llx)\n" , - *dma_addr, *paddr, *size); - -out: -of_node_put (node); +struct of_bus *bus; -return ret; +/* get bus specific dma-ranges. */ + bus =of_match_bus (np); +return bus->get_dma_ranges (np, dma_addr, paddr, size); }EXPORT_SYMBOL_GPL (of_dma_get_range);
Signature · 1 line
2.20.1
Re: [PATCH 2/2] PCI/of fix of_dma_get_range; get PCI specific dma-ranges
Quoted 85 lines · On Fri, Aug 09, 2019 at 07:34:49PM +0200, [email protected] wrote:
From: Oza Pawandeep <[email protected]>
current device framework and OF framework integration assumes dma-ranges in a way where memory-mapped devices define their dma-ranges. (child-bus-address, parent-bus-address, length).
of_dma_configureis specifically written to take care of memory mapped devices. but no implementation exists for pci to take care of pcie based memory ranges.for e.g. iproc based SOCs and other SOCs(such as rcar) have PCI world dma-ranges.
dma-ranges = <0x430000000x000x000x000x000x800x00>;this patch fixes the following problems
of_dma_get_range.
1) return of wrong size as 0.
2) not handling absence of dma-ranges which is valid for PCI master.
3) not handling multipe inbound windows.
4) in order to get largest possibledma_mask. this patch also retuns the largest possible size based on dma-ranges,for e.g.
dma-ranges = <0x430000000x000x000x000x000x800x00>; we should get dev->coherent_dma_mask=0x7fffffffff.based on which IOVA allocation space will honour PCI host bridge limitations.
the implementation hooks bus specific callbacks for getting dma-ranges.
Signed-off-by Oza Pawandeep <[email protected]>Signed-off-by Marek Vasut <[email protected]>Cc Arnd Bergmann <[email protected]>Cc Geert Uytterhoeven <[email protected]>Cc Lorenzo Pieralisi <[email protected]>Cc Robin Murphy <[email protected]>Cc Simon Horman <[email protected]>Cc Wolfram Sang <[email protected]>--- drivers/of/address.c | 220 +++++++++++++++++++++++++++++-------------- 1 file changed, 150 insertions(+), 70 deletions(-)drivers/of/address.c +18
-0@@ -8,6 +8,7 @@#include <linux/logic_pio.h> #include <linux/module.h> #include <linux/of_address.h> +#include <linux/of_pci.h> #include <linux/pci.h> #include <linux/pci_regs.h> #include <linux/sizes.h> @@ -48,6 +49,8 @@ struct of_bus { int na, int ns, int pna);int (*translate)(__be32 *addr,u64 offset,int na);unsigned int (*get_flags)(const __be32 *addr); +int (*get_dma_ranges)(struct device_node *np, +u64 *dma_addr,u64 *paddr,u64 *size); };/* @@ -174,6 +177,143 @@ static int of_bus_pci_translate(__be32 *addr, u64 offset, int na)return of_bus_default_translate (addr +1 , offset, na -1 ); } +static int of_bus_pci_get_dma_ranges (struct device_node *np,u64 *dma_addr, +u64 *paddr,u64 *size) +{ +struct device_node *node =of_node_get (np); +int ret =0 ; +struct resource_entry *window; +LIST_HEAD (res); + +if (!node) +return -EINVAL ; + + *size =0 ; +/* +* PCI dma-ranges is not mandatory property. +* many devices do no need to have it, since
drivers/of/address.c static int of_bus_pci_translate(__be32 *addr, u64 offset, int na) · line 204
Quoted 10 lines
drivers/of/address.c +10
-0+* host bridge does not require inbound memory +* configuration or rather have design limitations. +* so we look for dma-ranges, if missing we +* just return the caller full size, and also +* no dma-ranges suggests that, host bridge allows +* whatever comes in, so we set dma_addr to 0. +*/ + ret =of_pci_get_dma_ranges (np, &res); +if (!ret) { +resource_list_for_each_entry (window, &res) {
The { should appear on a newline and This block should be indented by one more tab. Or the scope of res_dma could be moved and this extra block avoided.
Quoted 230 lines
drivers/of/address.c +122
-70+struct resource *res_dma = window->res; + +if (*size <resource_size (res_dma)) { + *dma_addr = res_dma->start - window->offset; + *paddr = res_dma->start; + *size =resource_size (res_dma); + } + } + } +pci_free_resource_list (&res); + +/* +* return the largest possible size, +* since PCI master allows everything. +*/ +if (*size ==0 ) { +pr_debug ("empty/zero size dma-ranges found for node(%s)\n" , + np->full_name); + *size =DMA_BIT_MASK (sizeof (dma_addr_t) *8 ) -1 ; + *dma_addr = *paddr =0 ; + ret =0 ; + } + +pr_debug ("dma_addr(%llx) cpu_addr(%llx) size(%llx)\n" , + *dma_addr, *paddr, *size); + +of_node_put (node); + +return ret; +} + +static int get_dma_ranges (struct device_node *np,u64 *dma_addr, +u64 *paddr,u64 *size) +{ +struct device_node *node =of_node_get (np); +const __be32 *ranges =NULL ; +int len, naddr, nsize, pna; +int ret =0 ; +u64 dmaaddr; + +if (!node) +return -EINVAL ; + +while (1 ) { + naddr =of_n_addr_cells (node); + nsize =of_n_size_cells (node); + node =of_get_next_parent (node); +if (!node) +break ; + + ranges =of_get_property (node,"dma-ranges" , &len); + +/* Ignore empty ranges, they imply no translation required */ +if (ranges && len >0 ) +break ; + +/* +* At least empty ranges has to be defined for parent node if +* DMA is supported +*/ +if (!ranges) +break ; + } + +if (!ranges) { +pr_debug ("no dma-ranges found for node(%s)\n" , np->full_name); + ret = -ENODEV ; + goto out; + } + + len /=sizeof (u32); + + pna =of_n_addr_cells (node); + +/* dma-ranges format: +* DMA addr : naddr cells +* CPU addr : pna cells +* size : nsize cells +*/ + dmaaddr =of_read_number (ranges, naddr); + *paddr =of_translate_dma_address (np, ranges); +if (*paddr ==OF_BAD_ADDR ) { +pr_err ("translation of DMA address(%pad) to CPU address failed node(%s)\n" , + dma_addr, np->full_name); + ret = -EINVAL ; + goto out; + } + *dma_addr = dmaaddr; + + *size =of_read_number (ranges + naddr + pna, nsize); + +pr_debug ("dma_addr(%llx) cpu_addr(%llx) size(%llx)\n" , + *dma_addr, *paddr, *size); + +out: +of_node_put (node); + +return ret; +} + +static int of_bus_isa_get_dma_ranges (struct device_node *np,u64 *dma_addr, +u64 *paddr,u64 *size) +{ +return get_dma_ranges (np, dma_addr, paddr, size); +} + +static int of_bus_default_get_dma_ranges (struct device_node *np,u64 *dma_addr, +u64 *paddr,u64 *size) +{ +return get_dma_ranges (np, dma_addr, paddr, size); +} +const __be32 *of_get_pci_address (struct device_node *dev,int bar_no,u64 *size,unsigned int *flags) {@@ -438,6 +578,7 @@ static struct of_bus of_busses[] = { .map = of_bus_pci_map, .translate = of_bus_pci_translate, .get_flags = of_bus_pci_get_flags, + .get_dma_ranges = of_bus_pci_get_dma_ranges, },#endif /* CONFIG_PCI */ /* ISA */ @@ -449,6 +590,7 @@ static struct of_bus of_busses[] = { .map = of_bus_isa_map, .translate = of_bus_isa_translate, .get_flags = of_bus_isa_get_flags, + .get_dma_ranges = of_bus_isa_get_dma_ranges, },/* Default */ {@@ -459,6 +601,7 @@ static struct of_bus of_busses[] = { .map = of_bus_default_map, .translate = of_bus_default_translate, .get_flags = of_bus_default_get_flags, + .get_dma_ranges = of_bus_default_get_dma_ranges, }, };@@ -917,80 +1060,17 @@ EXPORT_SYMBOL(of_io_request_and_map);* size : nsize cells * * It returns -ENODEV if "dma-ranges" property was not found -* for this device in DT. +* for this device in DT, except if PCI device then, dma-ranges +* can be optional property, and in that case returns size with +* entire host memory. */ int of_dma_get_range (struct device_node *np,u64 *dma_addr,u64 *paddr,u64 *size) { -struct device_node *node =of_node_get (np); -const __be32 *ranges =NULL ; -int len, naddr, nsize, pna; -int ret =0 ; -u64 dmaaddr; - -if (!node) -return -EINVAL ; - -while (1 ) { -struct device_node *parent; - - naddr =of_n_addr_cells (node); - nsize =of_n_size_cells (node); - - parent =__of_get_dma_parent (node); -of_node_put (node); - - node = parent; -if (!node) -break ; - - ranges =of_get_property (node,"dma-ranges" , &len); - -/* Ignore empty ranges, they imply no translation required */ -if (ranges && len >0 ) -break ; - -/* -* At least empty ranges has to be defined for parent node if -* DMA is supported -*/ -if (!ranges) -break ; - } - -if (!ranges) { -pr_debug ("no dma-ranges found for node(%pOF)\n" , np); - ret = -ENODEV ; - goto out; - } - - len /=sizeof (u32); - - pna =of_n_addr_cells (node); - -/* dma-ranges format: -* DMA addr : naddr cells -* CPU addr : pna cells -* size : nsize cells -*/ - dmaaddr =of_read_number (ranges, naddr); - *paddr =of_translate_dma_address (np, ranges); -if (*paddr ==OF_BAD_ADDR ) { -pr_err ("translation of DMA address(%pad) to CPU address failed node(%pOF)\n" , - dma_addr, np); - ret = -EINVAL ; - goto out; - } - *dma_addr = dmaaddr; - - *size =of_read_number (ranges + naddr + pna, nsize); - -pr_debug ("dma_addr(%llx) cpu_addr(%llx) size(%llx)\n" , - *dma_addr, *paddr, *size); - -out: -of_node_put (node); +struct of_bus *bus; -return ret; +/* get bus specific dma-ranges. */ + bus =of_match_bus (np); +return bus->get_dma_ranges (np, dma_addr, paddr, size); }EXPORT_SYMBOL_GPL (of_dma_get_range);Signature · 1 line
2.20.1
Re: [PATCH 2/2] PCI/of fix of_dma_get_range; get PCI specific dma-ranges
In reply to [email protected]Quoted 5 lines · On Fri, Aug 09, 2019 at 07:34:49PM +0200, [email protected] wrote:
From: Oza Pawandeep <[email protected]>
current device framework and OF framework integration assumes dma-ranges in a way where memory-mapped devices define their dma-ranges. (child-bus-address, parent-bus-address, length).
"Memory-mapped" means device memory and registers are mapped into the CPU address space so reads/writes from the CPU can access the device. IIUC, the issue you're solving is for DMA, which is the other direction -- DMA reads/writes from the *device* can access system memory.
Quoted 3 lines
of_dma_configureis specifically written to take care of memory mapped devices. but no implementation exists for pci to take care of pcie based memory ranges.
s/pci/PCI/
There's nothing PCIe-specific here; the same issue could occur with either PCI or PCIe, so just say "PCI".
Quoted 8 lines
for e.g. iproc based SOCs and other SOCs(such as rcar) have PCI world dma-ranges.
dma-ranges = <0x430000000x000x000x000x000x800x00>;this patch fixes the following problems
of_dma_get_range.
1) return of wrong size as 0.
2) not handling absence of dma-ranges which is valid for PCI master.
3) not handling multipe inbound windows.
s/multipe/multiple/
Quoted 63 lines
4) in order to get largest possible
dma_mask. this patch also retuns the largest possible size based on dma-ranges,for e.g.
dma-ranges = <0x430000000x000x000x000x000x800x00>; we should get dev->coherent_dma_mask=0x7fffffffff.based on which IOVA allocation space will honour PCI host bridge limitations.
the implementation hooks bus specific callbacks for getting dma-ranges.
Signed-off-by Oza Pawandeep <[email protected]>Signed-off-by Marek Vasut <[email protected]>Cc Arnd Bergmann <[email protected]>Cc Geert Uytterhoeven <[email protected]>Cc Lorenzo Pieralisi <[email protected]>Cc Robin Murphy <[email protected]>Cc Simon Horman <[email protected]>Cc Wolfram Sang <[email protected]>--- drivers/of/address.c | 220 +++++++++++++++++++++++++++++-------------- 1 file changed, 150 insertions(+), 70 deletions(-)drivers/of/address.c +14
-0@@ -8,6 +8,7 @@#include <linux/logic_pio.h> #include <linux/module.h> #include <linux/of_address.h> +#include <linux/of_pci.h> #include <linux/pci.h> #include <linux/pci_regs.h> #include <linux/sizes.h> @@ -48,6 +49,8 @@ struct of_bus { int na, int ns, int pna);int (*translate)(__be32 *addr,u64 offset,int na);unsigned int (*get_flags)(const __be32 *addr); +int (*get_dma_ranges)(struct device_node *np, +u64 *dma_addr,u64 *paddr,u64 *size); };/* @@ -174,6 +177,143 @@ static int of_bus_pci_translate(__be32 *addr, u64 offset, int na)return of_bus_default_translate (addr +1 , offset, na -1 ); } +static int of_bus_pci_get_dma_ranges (struct device_node *np,u64 *dma_addr, +u64 *paddr,u64 *size) +{ +struct device_node *node =of_node_get (np); +int ret =0 ; +struct resource_entry *window; +LIST_HEAD (res); + +if (!node) +return -EINVAL ; +
drivers/of/address.c static int of_bus_pci_translate(__be32 *addr, u64 offset, int na) · line 200
Quoted 10 lines
drivers/of/address.c +10
-0+ *size =0 ; +/* +* PCI dma-ranges is not mandatory property. +* many devices do no need to have it, since +* host bridge does not require inbound memory +* configuration or rather have design limitations. +* so we look for dma-ranges, if missing we +* just return the caller full size, and also +* no dma-ranges suggests that, host bridge allows +* whatever comes in, so we set dma_addr to 0.
Wrap to fill the usual line length in this file.
Quoted 8 lines
drivers/of/address.c +8
-0+*/ + ret =of_pci_get_dma_ranges (np, &res); +if (!ret) { +resource_list_for_each_entry (window, &res) { +struct resource *res_dma = window->res; + +if (*size <resource_size (res_dma)) { + *dma_addr = res_dma->start - window->offset;
drivers/of/address.c static int of_bus_pci_translate(__be32 *addr, u64 offset, int na) · line 218
Quoted 10 lines
drivers/of/address.c +10
-0+ *paddr = res_dma->start; + *size =resource_size (res_dma); + } + } + } +pci_free_resource_list (&res); + +/* +* return the largest possible size, +* since PCI master allows everything.
Wrap.
Quoted 60 lines
drivers/of/address.c +60
-0+*/ +if (*size ==0 ) { +pr_debug ("empty/zero size dma-ranges found for node(%s)\n" , + np->full_name); + *size =DMA_BIT_MASK (sizeof (dma_addr_t) *8 ) -1 ; + *dma_addr = *paddr =0 ; + ret =0 ; + } + +pr_debug ("dma_addr(%llx) cpu_addr(%llx) size(%llx)\n" , + *dma_addr, *paddr, *size); + +of_node_put (node); + +return ret; +} + +static int get_dma_ranges (struct device_node *np,u64 *dma_addr, +u64 *paddr,u64 *size) +{ +struct device_node *node =of_node_get (np); +const __be32 *ranges =NULL ; +int len, naddr, nsize, pna; +int ret =0 ; +u64 dmaaddr; + +if (!node) +return -EINVAL ; + +while (1 ) { + naddr =of_n_addr_cells (node); + nsize =of_n_size_cells (node); + node =of_get_next_parent (node); +if (!node) +break ; + + ranges =of_get_property (node,"dma-ranges" , &len); + +/* Ignore empty ranges, they imply no translation required */ +if (ranges && len >0 ) +break ; + +/* +* At least empty ranges has to be defined for parent node if +* DMA is supported +*/ +if (!ranges) +break ; + } + +if (!ranges) { +pr_debug ("no dma-ranges found for node(%s)\n" , np->full_name); + ret = -ENODEV ; + goto out; + } + + len /=sizeof (u32); + + pna =of_n_addr_cells (node); +
drivers/of/address.c static int of_bus_pci_translate(__be32 *addr, u64 offset, int na) · line 288
Quoted 10 lines
drivers/of/address.c +10
-0+/* dma-ranges format: +* DMA addr : naddr cells +* CPU addr : pna cells +* size : nsize cells +*/ + dmaaddr =of_read_number (ranges, naddr); + *paddr =of_translate_dma_address (np, ranges); +if (*paddr ==OF_BAD_ADDR ) { +pr_err ("translation of DMA address(%pad) to CPU address failed node(%s)\n" , + dma_addr, np->full_name);
The old code used %pOF; why the change to %s?
drivers/of/address.c static int of_bus_pci_translate(__be32 *addr, u64 offset, int na) · line 297
Quoted 9 lines
drivers/of/address.c +9
-0+ ret = -EINVAL ; + goto out; + } + *dma_addr = dmaaddr; + + *size =of_read_number (ranges + naddr + pna, nsize); + +pr_debug ("dma_addr(%llx) cpu_addr(%llx) size(%llx)\n" , + *dma_addr, *paddr, *size);
I'm a little confused about the usage of %pad above and %llx here. Are these printing different things, or could they both be done the same way? (There are other occurrences of %llx earlier, too.)
Quoted 137 lines
drivers/of/address.c +29
-70+ +out: +of_node_put (node); + +return ret; +} + +static int of_bus_isa_get_dma_ranges (struct device_node *np,u64 *dma_addr, +u64 *paddr,u64 *size) +{ +return get_dma_ranges (np, dma_addr, paddr, size); +} + +static int of_bus_default_get_dma_ranges (struct device_node *np,u64 *dma_addr, +u64 *paddr,u64 *size) +{ +return get_dma_ranges (np, dma_addr, paddr, size); +} +const __be32 *of_get_pci_address (struct device_node *dev,int bar_no,u64 *size,unsigned int *flags) {@@ -438,6 +578,7 @@ static struct of_bus of_busses[] = { .map = of_bus_pci_map, .translate = of_bus_pci_translate, .get_flags = of_bus_pci_get_flags, + .get_dma_ranges = of_bus_pci_get_dma_ranges, },#endif /* CONFIG_PCI */ /* ISA */ @@ -449,6 +590,7 @@ static struct of_bus of_busses[] = { .map = of_bus_isa_map, .translate = of_bus_isa_translate, .get_flags = of_bus_isa_get_flags, + .get_dma_ranges = of_bus_isa_get_dma_ranges, },/* Default */ {@@ -459,6 +601,7 @@ static struct of_bus of_busses[] = { .map = of_bus_default_map, .translate = of_bus_default_translate, .get_flags = of_bus_default_get_flags, + .get_dma_ranges = of_bus_default_get_dma_ranges, }, };@@ -917,80 +1060,17 @@ EXPORT_SYMBOL(of_io_request_and_map);* size : nsize cells * * It returns -ENODEV if "dma-ranges" property was not found -* for this device in DT. +* for this device in DT, except if PCI device then, dma-ranges +* can be optional property, and in that case returns size with +* entire host memory. */ int of_dma_get_range (struct device_node *np,u64 *dma_addr,u64 *paddr,u64 *size) { -struct device_node *node =of_node_get (np); -const __be32 *ranges =NULL ; -int len, naddr, nsize, pna; -int ret =0 ; -u64 dmaaddr; - -if (!node) -return -EINVAL ; - -while (1 ) { -struct device_node *parent; - - naddr =of_n_addr_cells (node); - nsize =of_n_size_cells (node); - - parent =__of_get_dma_parent (node); -of_node_put (node); - - node = parent; -if (!node) -break ; - - ranges =of_get_property (node,"dma-ranges" , &len); - -/* Ignore empty ranges, they imply no translation required */ -if (ranges && len >0 ) -break ; - -/* -* At least empty ranges has to be defined for parent node if -* DMA is supported -*/ -if (!ranges) -break ; - } - -if (!ranges) { -pr_debug ("no dma-ranges found for node(%pOF)\n" , np); - ret = -ENODEV ; - goto out; - } - - len /=sizeof (u32); - - pna =of_n_addr_cells (node); - -/* dma-ranges format: -* DMA addr : naddr cells -* CPU addr : pna cells -* size : nsize cells -*/ - dmaaddr =of_read_number (ranges, naddr); - *paddr =of_translate_dma_address (np, ranges); -if (*paddr ==OF_BAD_ADDR ) { -pr_err ("translation of DMA address(%pad) to CPU address failed node(%pOF)\n" , - dma_addr, np); - ret = -EINVAL ; - goto out; - } - *dma_addr = dmaaddr; - - *size =of_read_number (ranges + naddr + pna, nsize); - -pr_debug ("dma_addr(%llx) cpu_addr(%llx) size(%llx)\n" , - *dma_addr, *paddr, *size); - -out: -of_node_put (node); +struct of_bus *bus; -return ret; +/* get bus specific dma-ranges. */ + bus =of_match_bus (np); +return bus->get_dma_ranges (np, dma_addr, paddr, size); }EXPORT_SYMBOL_GPL (of_dma_get_range);Signature · 1 line
2.20.1
Quoted 52 lines · On Fri, Aug 09, 2019 at 07:34:48PM +0200, [email protected] wrote:
From: Oza Pawandeep <[email protected]>
The patch exports interface to PCIe RC drivers so that, Drivers can get their inbound memory configuration.
It provides basis for IOVA reservations for inbound memory holes, if RC is not capable of addressing all the host memory, Specifically when IOMMU is enabled and on ARMv8 where 64bit IOVA could be allocated.
It handles multiple inbound windows, and returns resources, and is left to the caller, how it wants to use them.
Signed-off-by Oza Pawandeep <[email protected]>Signed-off-by Marek Vasut <[email protected]>Cc Arnd Bergmann <[email protected]>Cc Geert Uytterhoeven <[email protected]>Cc Lorenzo Pieralisi <[email protected]>Cc Robin Murphy <[email protected]>Cc Simon Horman <[email protected]>Cc Wolfram Sang <[email protected]>--- drivers/pci/of.c | 96 ++++++++++++++++++++++++++++++++++++++++++ include/linux/of_pci.h | 7 +++ 2 files changed, 103 insertions(+)drivers/pci/of.c +16
-0@@ -347,6 +347,102 @@ int devm_of_pci_get_host_bridge_resources(struct device *dev,return err; }EXPORT_SYMBOL_GPL (devm_of_pci_get_host_bridge_resources); + +/** +* of_pci_get_dma_ranges - Parse PCI host bridge inbound resources from DT +* @np: device node of the host bridge having the dma-ranges property +* @resources: list where the range of resources will be added after DT parsing +* +* It is the caller's job to free the @resources list. +* +* This function will parse the "dma-ranges" property of a +* PCI host bridge device node and setup the resource mapping based +* on its content. +* +* It returns zero if the range parsing has been successful or a standard error +* value if it failed. +*/ +
drivers/pci/of.c int devm_of_pci_get_host_bridge_resources(struct device *dev, · line 375
Quoted 10 lines
drivers/pci/of.c +10
-0+int of_pci_get_dma_ranges (struct device_node *dn,struct list_head *resources) +{ +struct device_node *node =of_node_get (dn); +int rlen; +int pna =of_n_addr_cells (node); +const int na =3 , ns =2 ; +int np = pna + na + ns; +int ret =0 ; +struct resource *res; +const u32 *dma_ranges;
It seems that const __be32 * would be a more appropriate type for dma_ranges as that seems to be the expected type in all uses below.
Quoted 21 lines
drivers/pci/of.c +21
-0+struct of_pci_range range; + +if (!node) +return -EINVAL ; + +while (1 ) { + dma_ranges =of_get_property (node,"dma-ranges" , &rlen); + +/* Ignore empty ranges, they imply no translation required. */ +if (dma_ranges && rlen >0 ) +break ; + +/* no dma-ranges, they imply no translation required. */ +if (!dma_ranges) +break ; + + node =of_get_next_parent (node); + +if (!node) +break ; + }
drivers/pci/of.c int devm_of_pci_get_host_bridge_resources(struct device *dev, · line 406
Quoted 10 lines
drivers/pci/of.c +10
-0+ +if (!dma_ranges) { +pr_debug ("pcie device has no dma-ranges defined for node(%s)\n" , + dn->full_name); + ret = -EINVAL ; + goto out; + } + +while ((rlen -= np *4 ) >=0 ) { + range.pci_space =be32_to_cpup ((const __be32 *) &dma_ranges[0 ]);
Can &dma_ranges[0] be more succinctly written as dma_ranges ?
Quoted 70 lines
drivers/pci/of.c +39
-0+ range.pci_addr =of_read_number (dma_ranges +1 , ns); + range.cpu_addr =of_translate_dma_address (node, + dma_ranges + na); + range.size =of_read_number (dma_ranges + pna + na, ns); + range.flags =IORESOURCE_MEM ; + + dma_ranges += np; + +/* +* If we failed translation or got a zero-sized region +* then skip this range. +*/ +if (range.cpu_addr ==OF_BAD_ADDR || range.size ==0 ) +continue ; + + res =kzalloc (sizeof (struct resource ),GFP_KERNEL ); +if (!res) { + ret = -ENOMEM ; + goto parse_failed; + } + + ret =of_pci_range_to_resource (&range, dn, res); +if (ret) { +kfree (res); +continue ; + } + +pci_add_resource_offset (resources, res, + res->start - range.pci_addr); + } +return ret; + +parse_failed: +pci_free_resource_list (resources); +out: +of_node_put (node); +return ret; +} +EXPORT_SYMBOL_GPL (of_pci_get_dma_ranges);#endif /* CONFIG_OF_ADDRESS */ #if IS_ENABLED (CONFIG_OF_IRQ )include/linux/of_pci.h +7
-0@@ -13,6 +13,7 @@ struct device_node;struct device_node *of_pci_find_child_device (struct device_node *parent,unsigned int devfn);int of_pci_get_devfn (struct device_node *np); +int of_pci_get_dma_ranges (struct device_node *np,struct list_head *resources);void of_pci_check_probe_only (void );#else static inline struct device_node *of_pci_find_child_device(struct device_node *parent,@@ -26,6 +27,12 @@ static inline int of_pci_get_devfn(struct device_node *np) return -EINVAL; } +static inline int of_pci_get_dma_ranges (struct device_node *np, +struct list_head *resources) +{ +return -EINVAL ; +} +static inline void of_pci_check_probe_only (void ) { }#endif Signature · 1 line
2.20.1
Quoted 59 lines · Simon Horman
Quoted 58 lines · On Fri, Aug 09, 2019 at 07:34:48PM +0200, [email protected] wrote:
From: Oza Pawandeep <[email protected]>
The patch exports interface to PCIe RC drivers so that, Drivers can get their inbound memory configuration.
It provides basis for IOVA reservations for inbound memory holes, if RC is not capable of addressing all the host memory, Specifically when IOMMU is enabled and on ARMv8 where 64bit IOVA could be allocated.
It handles multiple inbound windows, and returns resources, and is left to the caller, how it wants to use them.
Signed-off-by Oza Pawandeep <[email protected]>Signed-off-by Marek Vasut <[email protected]>Cc Arnd Bergmann <[email protected]>Cc Geert Uytterhoeven <[email protected]>Cc Lorenzo Pieralisi <[email protected]>Cc Robin Murphy <[email protected]>Cc Simon Horman <[email protected]>Cc Wolfram Sang <[email protected]>--- drivers/pci/of.c | 96 ++++++++++++++++++++++++++++++++++++++++++ include/linux/of_pci.h | 7 +++ 2 files changed, 103 insertions(+)drivers/pci/of.c +22
-0@@ -347,6 +347,102 @@ int devm_of_pci_get_host_bridge_resources(struct device *dev,return err; }EXPORT_SYMBOL_GPL (devm_of_pci_get_host_bridge_resources); + +/** +* of_pci_get_dma_ranges - Parse PCI host bridge inbound resources from DT +* @np: device node of the host bridge having the dma-ranges property +* @resources: list where the range of resources will be added after DT parsing +* +* It is the caller's job to free the @resources list. +* +* This function will parse the "dma-ranges" property of a +* PCI host bridge device node and setup the resource mapping based +* on its content. +* +* It returns zero if the range parsing has been successful or a standard error +* value if it failed. +*/ + +int of_pci_get_dma_ranges (struct device_node *dn,struct list_head *resources) +{ +struct device_node *node =of_node_get (dn); +int rlen; +int pna =of_n_addr_cells (node); +const int na =3 , ns =2 ;
Quoted 8 lines
drivers/pci/of.c int devm_of_pci_get_host_bridge_resources(struct device *dev, · line 375
Quoted 4 lines
drivers/pci/of.c +4
-0+int np = pna + na + ns; +int ret =0 ; +struct resource *res; +const u32 *dma_ranges;It seems that const
__be32* would be a more appropriate type fordma_rangesas that seems to be the expected type in all uses below.
Noted, but I'm more interested in some conceptual input about this -- whether this is the right approach or not.
Seems like we now have iMX8, R-Car3 and some Broadcom SoCs affected by this issue, with no solution in sight for years, so I think something has to be done here.
Signature · 2 lines
Best regards, Marek Vasut
This would follow the convention for subject lines:
PCI: OF: Add of_pci_get_dma_ranges() for inbound DMA restrictions
Quoted 4 lines · On Fri, Aug 09, 2019 at 07:34:48PM +0200, [email protected] wrote:
From: Oza Pawandeep <[email protected]>
The patch exports interface to PCIe RC drivers so that, Drivers can get their inbound memory configuration.
IIUC this interface (of_pci_get_dma_ranges()) is not used directly by *drivers*; it is used by of_bus_pci_get_dma_ranges() in the next patch, and it's called by the driver core via this path:
really_probe
pci_dma_configure # pci_bus_type.dma_configure
of_dma_configure
of_dma_get_range
bus->get_dma_ranges
of_bus_pci_get_dma_ranges # of_busses[0].get_dma_ranges
of_pci_get_dma_ranges
Quoted 7 lines
It provides basis for IOVA reservations for inbound memory holes, if RC is not capable of addressing all the host memory, Specifically when IOMMU is enabled and on ARMv8 where 64bit IOVA could be allocated.
It handles multiple inbound windows, and returns resources, and is left to the caller, how it wants to use them.
This should say exactly what the problem is, maybe even with a link to a problem report. I assume it is something like "RC <X> cannot address all of host memory, and if we happen to allocate a buffer that's not addressable, DMA to the buffer fails". It'd be nice to know what the failure looks like (SERR# asserted, panic, reboot, etc).
Quoted 24 lines
Signed-off-by Oza Pawandeep <[email protected]>Signed-off-by Marek Vasut <[email protected]>Cc Arnd Bergmann <[email protected]>Cc Geert Uytterhoeven <[email protected]>Cc Lorenzo Pieralisi <[email protected]>Cc Robin Murphy <[email protected]>Cc Simon Horman <[email protected]>Cc Wolfram Sang <[email protected]>--- drivers/pci/of.c | 96 ++++++++++++++++++++++++++++++++++++++++++ include/linux/of_pci.h | 7 +++ 2 files changed, 103 insertions(+)drivers/pci/of.c +1
-0@@ -347,6 +347,102 @@ int devm_of_pci_get_host_bridge_resources(struct device *dev,return err; }EXPORT_SYMBOL_GPL (devm_of_pci_get_host_bridge_resources); +
drivers/pci/of.c int devm_of_pci_get_host_bridge_resources(struct device *dev, · line 360
Quoted 10 lines
drivers/pci/of.c +10
-0+/** +* of_pci_get_dma_ranges - Parse PCI host bridge inbound resources from DT +* @np: device node of the host bridge having the dma-ranges property +* @resources: list where the range of resources will be added after DT parsing +* +* It is the caller's job to free the @resources list. +* +* This function will parse the "dma-ranges" property of a +* PCI host bridge device node and setup the resource mapping based +* on its content.
Rewrap this so it uses the whole width (75-78 columns).
Quoted 17 lines
drivers/pci/of.c +17
-0+* It returns zero if the range parsing has been successful or a standard error +* value if it failed. +*/ + +int of_pci_get_dma_ranges (struct device_node *dn,struct list_head *resources) +{ +struct device_node *node =of_node_get (dn); +int rlen; +int pna =of_n_addr_cells (node); +const int na =3 , ns =2 ; +int np = pna + na + ns; +int ret =0 ; +struct resource *res; +const u32 *dma_ranges; +struct of_pci_range range; + +if (!node)
drivers/pci/of.c int devm_of_pci_get_host_bridge_resources(struct device *dev, · line 388
Quoted 10 lines
drivers/pci/of.c +10
-0+return -EINVAL ; + +while (1 ) { + dma_ranges =of_get_property (node,"dma-ranges" , &rlen); + +/* Ignore empty ranges, they imply no translation required. */ +if (dma_ranges && rlen >0 ) +break ; + +/* no dma-ranges, they imply no translation required. */
Capitalize as you do other comments (s/no/No/).
drivers/pci/of.c int devm_of_pci_get_host_bridge_resources(struct device *dev, · line 399
Quoted 11 lines
drivers/pci/of.c +11
-0+if (!dma_ranges) +break ; + + node =of_get_next_parent (node); + +if (!node) +break ; + } + +if (!dma_ranges) { + pr_debug("pcie device has no dma-ranges defined for node(%s)\n",
This isn't PCIe-specific, so it should say "PCI".
Quoted 36 lines
drivers/pci/of.c +36
-0+ dn->full_name); + ret = -EINVAL ; + goto out; + } + +while ((rlen -= np *4 ) >=0 ) { + range.pci_space =be32_to_cpup ((const __be32 *) &dma_ranges[0 ]); + range.pci_addr =of_read_number (dma_ranges +1 , ns); + range.cpu_addr =of_translate_dma_address (node, + dma_ranges + na); + range.size =of_read_number (dma_ranges + pna + na, ns); + range.flags =IORESOURCE_MEM ; + + dma_ranges += np; + +/* +* If we failed translation or got a zero-sized region +* then skip this range. +*/ +if (range.cpu_addr ==OF_BAD_ADDR || range.size ==0 ) +continue ; + + res =kzalloc (sizeof (struct resource ),GFP_KERNEL ); +if (!res) { + ret = -ENOMEM ; + goto parse_failed; + } + + ret =of_pci_range_to_resource (&range, dn, res); +if (ret) { +kfree (res); +continue ; + } + +pci_add_resource_offset (resources, res, + res->start - range.pci_addr);
drivers/pci/of.c int devm_of_pci_get_host_bridge_resources(struct device *dev, · line 445
Quoted 10 lines
drivers/pci/of.c +10
-0+ } +return ret; + +parse_failed: +pci_free_resource_list (resources); +out: +of_node_put (node); +return ret; +} +EXPORT_SYMBOL_GPL (of_pci_get_dma_ranges);
I don't think this needs to be exported, does it? The only caller I see (of_bus_pci_get_dma_ranges()) cannot be in a module.
Quoted 31 lines
drivers/pci/of.c +0
-0#endif /* CONFIG_OF_ADDRESS */ #if IS_ENABLED (CONFIG_OF_IRQ )include/linux/of_pci.h +7
-0@@ -13,6 +13,7 @@ struct device_node;struct device_node *of_pci_find_child_device (struct device_node *parent,unsigned int devfn);int of_pci_get_devfn (struct device_node *np); +int of_pci_get_dma_ranges (struct device_node *np,struct list_head *resources);void of_pci_check_probe_only (void );#else static inline struct device_node *of_pci_find_child_device(struct device_node *parent,@@ -26,6 +27,12 @@ static inline int of_pci_get_devfn(struct device_node *np) return -EINVAL; } +static inline int of_pci_get_dma_ranges (struct device_node *np, +struct list_head *resources) +{ +return -EINVAL ; +} +static inline void of_pci_check_probe_only (void ) { }#endif Signature · 1 line
2.20.1
Quoted 27 lines · Bjorn Helgaas
This would follow the convention for subject lines:
PCI: OF: Add
of_pci_get_dma_ranges()for inbound DMA restrictionsQuoted 4 lines · On Fri, Aug 09, 2019 at 07:34:48PM +0200 wrote:
From: Oza Pawandeep
The patch exports interface to PCIe RC drivers so that, Drivers can get their inbound memory configuration.
IIUC this interface (
of_pci_get_dma_ranges()) is not used directly by *drivers*; it is used byof_bus_pci_get_dma_ranges()in the next patch, and it's called by the driver core via this path:
really_probe
pci_dma_configure# pci_bus_type.dma_configure
of_dma_configure
of_dma_get_range
bus->get_dma_ranges
of_bus_pci_get_dma_ranges# of_busses[0].get_dma_ranges
of_pci_get_dma_rangesQuoted 5 lines
It provides basis for IOVA reservations for inbound memory holes, if RC is not capable of addressing all the host memory, Specifically when IOMMU is enabled and on ARMv8 where 64bit IOVA could be allocated.
Quoted 8 lines
Quoted 2 lines
It handles multiple inbound windows, and returns resources, and is left to the caller, how it wants to use them.
This should say exactly what the problem is, maybe even with a link to a problem report. I assume it is something like "RC <X> cannot address all of host memory, and if we happen to allocate a buffer that's not addressable, DMA to the buffer fails". It'd be nice to know what the failure looks like (SERR# asserted, panic, reboot, etc).
[...]
While it's good that someone finally looked at these patches, these were since superseded by the following series:
https://patchwork.ozlabs.org/cover/1168166/
Signature · 2 lines
Best regards, Marek Vasut
[-cc Oza, bounced]
Quoted 37 lines · Marek Vasut
Quoted 27 lines · Bjorn Helgaas
This would follow the convention for subject lines:
PCI: OF: Add
of_pci_get_dma_ranges()for inbound DMA restrictionsQuoted 4 lines · On Fri, Aug 09, 2019 at 07:34:48PM +0200 wrote:
From: Oza Pawandeep
The patch exports interface to PCIe RC drivers so that, Drivers can get their inbound memory configuration.
IIUC this interface (
of_pci_get_dma_ranges()) is not used directly by *drivers*; it is used byof_bus_pci_get_dma_ranges()in the next patch, and it's called by the driver core via this path:
really_probe
pci_dma_configure# pci_bus_type.dma_configure
of_dma_configure
of_dma_get_range
bus->get_dma_ranges
of_bus_pci_get_dma_ranges# of_busses[0].get_dma_ranges
of_pci_get_dma_rangesQuoted 5 lines
It provides basis for IOVA reservations for inbound memory holes, if RC is not capable of addressing all the host memory, Specifically when IOMMU is enabled and on ARMv8 where 64bit IOVA could be allocated.
Quoted 8 lines
Quoted 2 lines
It handles multiple inbound windows, and returns resources, and is left to the caller, how it wants to use them.
This should say exactly what the problem is, maybe even with a link to a problem report. I assume it is something like "RC <X> cannot address all of host memory, and if we happen to allocate a buffer that's not addressable, DMA to the buffer fails". It'd be nice to know what the failure looks like (SERR# asserted, panic, reboot, etc).
Quoted 5 lines
[...]
While it's good that someone finally looked at these patches, these were since superseded by the following series:
https://patchwork.ozlabs.org/cover/1168166/
Good to know, thanks :)