Re: [PATCH] of/base: Fix PowerPC address parsing hack
On Thu, Nov 13, 2014 at 12:45 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: What about this one instead ? I want to cache it because that function can be called quite a while and doing two additional property lookup and string compares every time might hurt some platforms. We have a historical hack that treats missing ranges properties as the equivalent of an empty one. This is needed for ancient PowerMac bad device-trees, and shouldn't be enabled for any other PowerPC platform, otherwise we get some nasty layout of devices in sysfs or even duplication when a set of otherwise identically named devices is created multiple times under a different parent node with no ranges property. This fix is needed for the PowerNV i2c busses to be exposed properly and will fix a number of other embedded cases. Signed-off-by: Benjamin Herrenschmidt b...@kernel.crashing.org CC: sta...@vger.kernel.org How far back does this need to go? I assume I need to get this in for v3.18. diff --git a/drivers/of/address.c b/drivers/of/address.c index e371825..5eae0cd 100644 --- a/drivers/of/address.c +++ b/drivers/of/address.c @@ -403,6 +403,17 @@ static struct of_bus *of_match_bus(struct device_node *np) return NULL; } +static int of_empty_ranges_quirk(void) +{ + /* To save cycles, we cache the result */ + static int quirk_state = -1; + if (IS_ENABLED(CONFIG_POWERPC)) { + if (quirk_state 0) + quirk_state = of_machine_is_compatible(Power Macintosh) || + of_machine_is_compatible(MacRISC); + return quirk_state; } return 0; So it gets compiled out for non powerpc. +} + static int of_translate_one(struct device_node *parent, struct of_bus *bus, struct of_bus *pbus, __be32 *addr, int na, int ns, int pna, const char *rprop) @@ -428,12 +439,10 @@ static int of_translate_one(struct device_node *parent, struct of_bus *bus, * This code is only enabled on powerpc. --gcl */ ranges = of_get_property(parent, rprop, rlen); -#if !defined(CONFIG_PPC) - if (ranges == NULL) { + if (ranges == NULL !of_empty_ranges_quirk()) { pr_err(OF: no ranges; cannot translate\n); return 1; } -#endif /* !defined(CONFIG_PPC) */ if (ranges == NULL || rlen == 0) { offset = of_read_number(addr, na); memset(addr, 0, pna * 4); ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH] of/base: Fix PowerPC address parsing hack
On Thu, 2014-11-13 at 12:44 +, Grant Likely wrote: On Thu, Nov 13, 2014 at 12:45 AM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: What about this one instead ? I want to cache it because that function can be called quite a while and doing two additional property lookup and string compares every time might hurt some platforms. We have a historical hack that treats missing ranges properties as the equivalent of an empty one. This is needed for ancient PowerMac bad device-trees, and shouldn't be enabled for any other PowerPC platform, otherwise we get some nasty layout of devices in sysfs or even duplication when a set of otherwise identically named devices is created multiple times under a different parent node with no ranges property. This fix is needed for the PowerNV i2c busses to be exposed properly and will fix a number of other embedded cases. Signed-off-by: Benjamin Herrenschmidt b...@kernel.crashing.org CC: sta...@vger.kernel.org How far back does this need to go? I assume I need to get this in for v3.18. I'd like some distro to pick it up in 3.16 diff --git a/drivers/of/address.c b/drivers/of/address.c index e371825..5eae0cd 100644 --- a/drivers/of/address.c +++ b/drivers/of/address.c @@ -403,6 +403,17 @@ static struct of_bus *of_match_bus(struct device_node *np) return NULL; } +static int of_empty_ranges_quirk(void) +{ + /* To save cycles, we cache the result */ + static int quirk_state = -1; + if (IS_ENABLED(CONFIG_POWERPC)) { + if (quirk_state 0) + quirk_state = of_machine_is_compatible(Power Macintosh) || + of_machine_is_compatible(MacRISC); + return quirk_state; } return 0; So it gets compiled out for non powerpc. Yeah, I'm set in my ways, keep forgetting about all the shiny new stuff. +} + static int of_translate_one(struct device_node *parent, struct of_bus *bus, struct of_bus *pbus, __be32 *addr, int na, int ns, int pna, const char *rprop) @@ -428,12 +439,10 @@ static int of_translate_one(struct device_node *parent, struct of_bus *bus, * This code is only enabled on powerpc. --gcl */ ranges = of_get_property(parent, rprop, rlen); -#if !defined(CONFIG_PPC) - if (ranges == NULL) { + if (ranges == NULL !of_empty_ranges_quirk()) { pr_err(OF: no ranges; cannot translate\n); return 1; } -#endif /* !defined(CONFIG_PPC) */ if (ranges == NULL || rlen == 0) { offset = of_read_number(addr, na); memset(addr, 0, pna * 4); -- To unsubscribe from this list: send the line unsubscribe devicetree in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH] of/base: Fix PowerPC address parsing hack
On Tue, Nov 11, 2014 at 11:51 PM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: We have a historical hack that treats missing ranges properties as the equivalent of an empty one. This is needed for ancient PowerMac bad device-trees, and shouldn't be enabled for any other PowerPC platform, otherwise we get some nasty layout of devices in sysfs or even duplication when a set of otherwise identically named devices is created multiple times under a different parent node with no ranges property. This fix is needed for the PowerNV i2c busses to be exposed properly and will fix a number of other embedded cases. Signed-off-by: Benjamin Herrenschmidt b...@kernel.crashing.org CC: sta...@vger.kernel.org --- diff --git a/drivers/of/address.c b/drivers/of/address.c index e371825..e37f017 100644 --- a/drivers/of/address.c +++ b/drivers/of/address.c @@ -7,6 +7,10 @@ #include linux/pci_regs.h #include linux/string.h +#ifdef CONFIG_PPC +#include asm/machdep.h +#endif + /* Max address size we deal with */ #define OF_MAX_ADDR_CELLS 4 #define OF_CHECK_ADDR_COUNT(na)((na) 0 (na) = OF_MAX_ADDR_CELLS) @@ -428,12 +432,13 @@ static int of_translate_one(struct device_node *parent, struct of_bus *bus, * This code is only enabled on powerpc. --gcl */ ranges = of_get_property(parent, rprop, rlen); -#if !defined(CONFIG_PPC) +#if defined(CONFIG_PPC) + if (!machine_is(powermac)) Can we use a machine compatible here or something not PPC specific? Then we can use IS_ENABLED(CONFIG_PPC) instead of ifdefs. Rob +#endif /* defined(CONFIG_PPC) */ if (ranges == NULL) { - pr_err(OF: no ranges; cannot translate\n); + pr_debug(OF: no ranges; cannot translate\n); return 1; } -#endif /* !defined(CONFIG_PPC) */ if (ranges == NULL || rlen == 0) { offset = of_read_number(addr, na); memset(addr, 0, pna * 4); ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH] of/base: Fix PowerPC address parsing hack
On Wed, 2014-11-12 at 08:39 -0600, Rob Herring wrote: On Tue, Nov 11, 2014 at 11:51 PM, Benjamin Herrenschmidt b...@kernel.crashing.org wrote: We have a historical hack that treats missing ranges properties as the equivalent of an empty one. This is needed for ancient PowerMac bad device-trees, and shouldn't be enabled for any other PowerPC platform, otherwise we get some nasty layout of devices in sysfs or even duplication when a set of otherwise identically named devices is created multiple times under a different parent node with no ranges property. This fix is needed for the PowerNV i2c busses to be exposed properly and will fix a number of other embedded cases. Signed-off-by: Benjamin Herrenschmidt b...@kernel.crashing.org CC: sta...@vger.kernel.org --- diff --git a/drivers/of/address.c b/drivers/of/address.c index e371825..e37f017 100644 --- a/drivers/of/address.c +++ b/drivers/of/address.c @@ -7,6 +7,10 @@ #include linux/pci_regs.h #include linux/string.h +#ifdef CONFIG_PPC +#include asm/machdep.h +#endif + /* Max address size we deal with */ #define OF_MAX_ADDR_CELLS 4 #define OF_CHECK_ADDR_COUNT(na)((na) 0 (na) = OF_MAX_ADDR_CELLS) @@ -428,12 +432,13 @@ static int of_translate_one(struct device_node *parent, struct of_bus *bus, * This code is only enabled on powerpc. --gcl */ ranges = of_get_property(parent, rprop, rlen); -#if !defined(CONFIG_PPC) +#if defined(CONFIG_PPC) + if (!machine_is(powermac)) Can we use a machine compatible here or something not PPC specific? Then we can use IS_ENABLED(CONFIG_PPC) instead of ifdefs. We could, we'd have to use a pair of machine compatible, I'll spin a new patch later today. Ben. Rob +#endif /* defined(CONFIG_PPC) */ if (ranges == NULL) { - pr_err(OF: no ranges; cannot translate\n); + pr_debug(OF: no ranges; cannot translate\n); return 1; } -#endif /* !defined(CONFIG_PPC) */ if (ranges == NULL || rlen == 0) { offset = of_read_number(addr, na); memset(addr, 0, pna * 4); -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH] of/base: Fix PowerPC address parsing hack
On Thu, 2014-11-13 at 09:10 +1100, Stephen Rothwell wrote: Hi Ben, Urk! :-) How about: On Wed, 12 Nov 2014 16:51:01 +1100 Benjamin Herrenschmidt b...@kernel.crashing.org wrote: diff --git a/drivers/of/address.c b/drivers/of/address.c index e371825..e37f017 100644 --- a/drivers/of/address.c +++ b/drivers/of/address.c @@ -7,6 +7,10 @@ #include linux/pci_regs.h #include linux/string.h +#ifdef CONFIG_PPC +#include asm/machdep.h #define IS_PMAC machine_is(pmac) #else #define IS_PMAC (0) I'll just do an of machine compatible check instead, so there's no ifdef at all, I'll send a new patch later today. +#endif + /* Max address size we deal with */ #define OF_MAX_ADDR_CELLS 4 #define OF_CHECK_ADDR_COUNT(na)((na) 0 (na) = OF_MAX_ADDR_CELLS) @@ -428,12 +432,13 @@ static int of_translate_one(struct device_node *parent, struct of_bus *bus, * This code is only enabled on powerpc. --gcl */ ranges = of_get_property(parent, rprop, rlen); -#if !defined(CONFIG_PPC) +#if defined(CONFIG_PPC) + if (!machine_is(powermac)) +#endif /* defined(CONFIG_PPC) */ if (ranges == NULL) { if ((!IS_PMAC) (ranges == NULL)) { - pr_err(OF: no ranges; cannot translate\n); + pr_debug(OF: no ranges; cannot translate\n); return 1; } -#endif /* !defined(CONFIG_PPC) */ if (ranges == NULL || rlen == 0) { offset = of_read_number(addr, na); memset(addr, 0, pna * 4); There might be a better identifier than IS_PMAC ... ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH] of/base: Fix PowerPC address parsing hack
What about this one instead ? I want to cache it because that function can be called quite a while and doing two additional property lookup and string compares every time might hurt some platforms. We have a historical hack that treats missing ranges properties as the equivalent of an empty one. This is needed for ancient PowerMac bad device-trees, and shouldn't be enabled for any other PowerPC platform, otherwise we get some nasty layout of devices in sysfs or even duplication when a set of otherwise identically named devices is created multiple times under a different parent node with no ranges property. This fix is needed for the PowerNV i2c busses to be exposed properly and will fix a number of other embedded cases. Signed-off-by: Benjamin Herrenschmidt b...@kernel.crashing.org CC: sta...@vger.kernel.org diff --git a/drivers/of/address.c b/drivers/of/address.c index e371825..5eae0cd 100644 --- a/drivers/of/address.c +++ b/drivers/of/address.c @@ -403,6 +403,17 @@ static struct of_bus *of_match_bus(struct device_node *np) return NULL; } +static int of_empty_ranges_quirk(void) +{ + /* To save cycles, we cache the result */ + static int quirk_state = -1; + + if (quirk_state 0) + quirk_state = of_machine_is_compatible(Power Macintosh) || + of_machine_is_compatible(MacRISC); + return quirk_state; +} + static int of_translate_one(struct device_node *parent, struct of_bus *bus, struct of_bus *pbus, __be32 *addr, int na, int ns, int pna, const char *rprop) @@ -428,12 +439,10 @@ static int of_translate_one(struct device_node *parent, struct of_bus *bus, * This code is only enabled on powerpc. --gcl */ ranges = of_get_property(parent, rprop, rlen); -#if !defined(CONFIG_PPC) - if (ranges == NULL) { + if (ranges == NULL !of_empty_ranges_quirk()) { pr_err(OF: no ranges; cannot translate\n); return 1; } -#endif /* !defined(CONFIG_PPC) */ if (ranges == NULL || rlen == 0) { offset = of_read_number(addr, na); memset(addr, 0, pna * 4); ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH] of/base: Fix PowerPC address parsing hack
Hi Ben, On Thu, 13 Nov 2014 11:45:22 +1100 Benjamin Herrenschmidt b...@kernel.crashing.org wrote: What about this one instead ? I want to cache it because that function can be called quite a while and doing two additional property lookup and string compares every time might hurt some platforms. Looks good to me. -- Cheers, Stephen Rothwells...@canb.auug.org.au pgpVbekkqMXtY.pgp Description: OpenPGP digital signature ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
[PATCH] of/base: Fix PowerPC address parsing hack
We have a historical hack that treats missing ranges properties as the equivalent of an empty one. This is needed for ancient PowerMac bad device-trees, and shouldn't be enabled for any other PowerPC platform, otherwise we get some nasty layout of devices in sysfs or even duplication when a set of otherwise identically named devices is created multiple times under a different parent node with no ranges property. This fix is needed for the PowerNV i2c busses to be exposed properly and will fix a number of other embedded cases. Signed-off-by: Benjamin Herrenschmidt b...@kernel.crashing.org CC: sta...@vger.kernel.org --- diff --git a/drivers/of/address.c b/drivers/of/address.c index e371825..e37f017 100644 --- a/drivers/of/address.c +++ b/drivers/of/address.c @@ -7,6 +7,10 @@ #include linux/pci_regs.h #include linux/string.h +#ifdef CONFIG_PPC +#include asm/machdep.h +#endif + /* Max address size we deal with */ #define OF_MAX_ADDR_CELLS 4 #define OF_CHECK_ADDR_COUNT(na)((na) 0 (na) = OF_MAX_ADDR_CELLS) @@ -428,12 +432,13 @@ static int of_translate_one(struct device_node *parent, struct of_bus *bus, * This code is only enabled on powerpc. --gcl */ ranges = of_get_property(parent, rprop, rlen); -#if !defined(CONFIG_PPC) +#if defined(CONFIG_PPC) + if (!machine_is(powermac)) +#endif /* defined(CONFIG_PPC) */ if (ranges == NULL) { - pr_err(OF: no ranges; cannot translate\n); + pr_debug(OF: no ranges; cannot translate\n); return 1; } -#endif /* !defined(CONFIG_PPC) */ if (ranges == NULL || rlen == 0) { offset = of_read_number(addr, na); memset(addr, 0, pna * 4); ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev