* [PATCH v2 0/5] Removing full paths from DT full_name @ 2017-08-21 15:16 ` Rob Herring 0 siblings, 0 replies; 53+ messages in thread From: Rob Herring @ 2017-08-21 15:16 UTC (permalink / raw) To: devicetree, Michael Ellerman, linuxppc-dev Cc: linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight This series is the last steps to remove storing the full path for every DT node. Instead, we can create full path strings dynamically as needed with printf %pOF specifiers (commit ce4fecf1fe15). There are a number of remaining direct users of full_name after this series. I don't believe there should be any functional impact for those users with the change to only the node name (+unit-address). The majority are for struct resource.name. This should only affect /proc/iomem display. Michael, Please apply patches 1-3 for 4.14. Patches 4 and 5 are intended for 4.15. Testing of those in particular would be helpful. Rob Rob Herring (5): powerpc: Convert to using %pOF instead of full_name powerpc: pseries: vio: match parent nodes with of_find_node_by_path powerpc: pseries: remove dlpar_attach_node dependency on full path powerpc: pseries: only store the device node basename in full_name of/fdt: only store the device node basename in full_name arch/powerpc/kernel/btext.c | 2 +- arch/powerpc/kernel/cacheinfo.c | 34 ++++--- arch/powerpc/kernel/io-workarounds.c | 4 +- arch/powerpc/kernel/isa-bridge.c | 32 +++---- arch/powerpc/kernel/legacy_serial.c | 12 +-- arch/powerpc/kernel/of_platform.c | 2 +- arch/powerpc/kernel/pci-common.c | 15 ++- arch/powerpc/kernel/pci_32.c | 4 +- arch/powerpc/kernel/pci_64.c | 4 +- arch/powerpc/kernel/pci_of_scan.c | 24 ++--- arch/powerpc/kernel/setup-common.c | 2 +- arch/powerpc/platforms/4xx/cpm.c | 8 +- arch/powerpc/platforms/4xx/gpio.c | 3 +- arch/powerpc/platforms/4xx/msi.c | 3 +- arch/powerpc/platforms/4xx/pci.c | 116 ++++++++++------------- arch/powerpc/platforms/4xx/soc.c | 5 +- arch/powerpc/platforms/4xx/uic.c | 14 +-- arch/powerpc/platforms/512x/mpc512x_shared.c | 12 +-- arch/powerpc/platforms/52xx/efika.c | 8 +- arch/powerpc/platforms/52xx/media5200.c | 2 +- arch/powerpc/platforms/52xx/mpc52xx_gpt.c | 4 +- arch/powerpc/platforms/52xx/mpc52xx_pci.c | 8 +- arch/powerpc/platforms/83xx/mcu_mpc8349emitx.c | 3 +- arch/powerpc/platforms/83xx/mpc832x_rdb.c | 2 +- arch/powerpc/platforms/85xx/p1022_ds.c | 8 +- arch/powerpc/platforms/85xx/xes_mpc85xx.c | 4 +- arch/powerpc/platforms/amigaone/setup.c | 6 +- arch/powerpc/platforms/cell/axon_msi.c | 36 +++---- arch/powerpc/platforms/cell/interrupt.c | 4 +- arch/powerpc/platforms/cell/iommu.c | 24 ++--- arch/powerpc/platforms/cell/ras.c | 4 +- arch/powerpc/platforms/cell/spider-pci.c | 4 +- arch/powerpc/platforms/cell/spider-pic.c | 4 +- arch/powerpc/platforms/cell/spu_manage.c | 26 ++--- arch/powerpc/platforms/chrp/pci.c | 18 ++-- arch/powerpc/platforms/embedded6xx/linkstation.c | 6 +- arch/powerpc/platforms/embedded6xx/mvme5100.c | 2 +- arch/powerpc/platforms/embedded6xx/storcenter.c | 2 +- arch/powerpc/platforms/maple/pci.c | 10 +- arch/powerpc/platforms/pasemi/pci.c | 2 +- arch/powerpc/platforms/powermac/feature.c | 14 +-- arch/powerpc/platforms/powermac/low_i2c.c | 50 +++++----- arch/powerpc/platforms/powermac/pci.c | 6 +- arch/powerpc/platforms/powermac/pfunc_base.c | 24 ++--- arch/powerpc/platforms/powermac/pfunc_core.c | 6 +- arch/powerpc/platforms/powermac/pic.c | 8 +- arch/powerpc/platforms/powermac/setup.c | 2 +- arch/powerpc/platforms/powernv/opal-async.c | 4 +- arch/powerpc/platforms/powernv/opal-xscom.c | 8 +- arch/powerpc/platforms/powernv/pci-ioda.c | 15 ++- arch/powerpc/platforms/powernv/rng.c | 6 +- arch/powerpc/platforms/pseries/dlpar.c | 39 ++------ arch/powerpc/platforms/pseries/event_sources.c | 6 +- arch/powerpc/platforms/pseries/hotplug-cpu.c | 6 +- arch/powerpc/platforms/pseries/ibmebus.c | 5 +- arch/powerpc/platforms/pseries/iommu.c | 58 ++++++------ arch/powerpc/platforms/pseries/mobility.c | 2 +- arch/powerpc/platforms/pseries/msi.c | 12 +-- arch/powerpc/platforms/pseries/pci_dlpar.c | 2 +- arch/powerpc/platforms/pseries/pseries.h | 2 +- arch/powerpc/platforms/pseries/reconfig.c | 2 +- arch/powerpc/platforms/pseries/vio.c | 10 +- arch/powerpc/sysdev/axonram.c | 4 +- arch/powerpc/sysdev/dcr.c | 4 +- arch/powerpc/sysdev/fsl_85xx_cache_sram.c | 12 +-- arch/powerpc/sysdev/fsl_gtm.c | 14 +-- arch/powerpc/sysdev/fsl_msi.c | 16 ++-- arch/powerpc/sysdev/fsl_pci.c | 47 +++++---- arch/powerpc/sysdev/fsl_rio.c | 36 +++---- arch/powerpc/sysdev/fsl_rmu.c | 12 +-- arch/powerpc/sysdev/mpic.c | 4 +- arch/powerpc/sysdev/mpic_msgr.c | 2 +- arch/powerpc/sysdev/mpic_msi.c | 2 +- arch/powerpc/sysdev/mpic_timer.c | 19 ++-- arch/powerpc/sysdev/msi_bitmap.c | 4 +- arch/powerpc/sysdev/mv64x60_dev.c | 20 ++-- arch/powerpc/sysdev/mv64x60_pci.c | 4 +- arch/powerpc/sysdev/of_rtc.c | 12 +-- arch/powerpc/sysdev/scom.c | 5 +- arch/powerpc/sysdev/simple_gpio.c | 3 +- arch/powerpc/sysdev/tsi108_pci.c | 4 +- arch/powerpc/sysdev/xive/native.c | 2 +- drivers/of/fdt.c | 69 +++----------- 83 files changed, 487 insertions(+), 578 deletions(-) -- 2.11.0 ^ permalink raw reply [flat|nested] 53+ messages in thread
* [PATCH v2 0/5] Removing full paths from DT full_name @ 2017-08-21 15:16 ` Rob Herring 0 siblings, 0 replies; 53+ messages in thread From: Rob Herring @ 2017-08-21 15:16 UTC (permalink / raw) To: devicetree-u79uwXL29TY76Z2rM5mHXA, Michael Ellerman, linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ Cc: linux-kernel-u79uwXL29TY76Z2rM5mHXA, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight This series is the last steps to remove storing the full path for every DT node. Instead, we can create full path strings dynamically as needed with printf %pOF specifiers (commit ce4fecf1fe15). There are a number of remaining direct users of full_name after this series. I don't believe there should be any functional impact for those users with the change to only the node name (+unit-address). The majority are for struct resource.name. This should only affect /proc/iomem display. Michael, Please apply patches 1-3 for 4.14. Patches 4 and 5 are intended for 4.15. Testing of those in particular would be helpful. Rob Rob Herring (5): powerpc: Convert to using %pOF instead of full_name powerpc: pseries: vio: match parent nodes with of_find_node_by_path powerpc: pseries: remove dlpar_attach_node dependency on full path powerpc: pseries: only store the device node basename in full_name of/fdt: only store the device node basename in full_name arch/powerpc/kernel/btext.c | 2 +- arch/powerpc/kernel/cacheinfo.c | 34 ++++--- arch/powerpc/kernel/io-workarounds.c | 4 +- arch/powerpc/kernel/isa-bridge.c | 32 +++---- arch/powerpc/kernel/legacy_serial.c | 12 +-- arch/powerpc/kernel/of_platform.c | 2 +- arch/powerpc/kernel/pci-common.c | 15 ++- arch/powerpc/kernel/pci_32.c | 4 +- arch/powerpc/kernel/pci_64.c | 4 +- arch/powerpc/kernel/pci_of_scan.c | 24 ++--- arch/powerpc/kernel/setup-common.c | 2 +- arch/powerpc/platforms/4xx/cpm.c | 8 +- arch/powerpc/platforms/4xx/gpio.c | 3 +- arch/powerpc/platforms/4xx/msi.c | 3 +- arch/powerpc/platforms/4xx/pci.c | 116 ++++++++++------------- arch/powerpc/platforms/4xx/soc.c | 5 +- arch/powerpc/platforms/4xx/uic.c | 14 +-- arch/powerpc/platforms/512x/mpc512x_shared.c | 12 +-- arch/powerpc/platforms/52xx/efika.c | 8 +- arch/powerpc/platforms/52xx/media5200.c | 2 +- arch/powerpc/platforms/52xx/mpc52xx_gpt.c | 4 +- arch/powerpc/platforms/52xx/mpc52xx_pci.c | 8 +- arch/powerpc/platforms/83xx/mcu_mpc8349emitx.c | 3 +- arch/powerpc/platforms/83xx/mpc832x_rdb.c | 2 +- arch/powerpc/platforms/85xx/p1022_ds.c | 8 +- arch/powerpc/platforms/85xx/xes_mpc85xx.c | 4 +- arch/powerpc/platforms/amigaone/setup.c | 6 +- arch/powerpc/platforms/cell/axon_msi.c | 36 +++---- arch/powerpc/platforms/cell/interrupt.c | 4 +- arch/powerpc/platforms/cell/iommu.c | 24 ++--- arch/powerpc/platforms/cell/ras.c | 4 +- arch/powerpc/platforms/cell/spider-pci.c | 4 +- arch/powerpc/platforms/cell/spider-pic.c | 4 +- arch/powerpc/platforms/cell/spu_manage.c | 26 ++--- arch/powerpc/platforms/chrp/pci.c | 18 ++-- arch/powerpc/platforms/embedded6xx/linkstation.c | 6 +- arch/powerpc/platforms/embedded6xx/mvme5100.c | 2 +- arch/powerpc/platforms/embedded6xx/storcenter.c | 2 +- arch/powerpc/platforms/maple/pci.c | 10 +- arch/powerpc/platforms/pasemi/pci.c | 2 +- arch/powerpc/platforms/powermac/feature.c | 14 +-- arch/powerpc/platforms/powermac/low_i2c.c | 50 +++++----- arch/powerpc/platforms/powermac/pci.c | 6 +- arch/powerpc/platforms/powermac/pfunc_base.c | 24 ++--- arch/powerpc/platforms/powermac/pfunc_core.c | 6 +- arch/powerpc/platforms/powermac/pic.c | 8 +- arch/powerpc/platforms/powermac/setup.c | 2 +- arch/powerpc/platforms/powernv/opal-async.c | 4 +- arch/powerpc/platforms/powernv/opal-xscom.c | 8 +- arch/powerpc/platforms/powernv/pci-ioda.c | 15 ++- arch/powerpc/platforms/powernv/rng.c | 6 +- arch/powerpc/platforms/pseries/dlpar.c | 39 ++------ arch/powerpc/platforms/pseries/event_sources.c | 6 +- arch/powerpc/platforms/pseries/hotplug-cpu.c | 6 +- arch/powerpc/platforms/pseries/ibmebus.c | 5 +- arch/powerpc/platforms/pseries/iommu.c | 58 ++++++------ arch/powerpc/platforms/pseries/mobility.c | 2 +- arch/powerpc/platforms/pseries/msi.c | 12 +-- arch/powerpc/platforms/pseries/pci_dlpar.c | 2 +- arch/powerpc/platforms/pseries/pseries.h | 2 +- arch/powerpc/platforms/pseries/reconfig.c | 2 +- arch/powerpc/platforms/pseries/vio.c | 10 +- arch/powerpc/sysdev/axonram.c | 4 +- arch/powerpc/sysdev/dcr.c | 4 +- arch/powerpc/sysdev/fsl_85xx_cache_sram.c | 12 +-- arch/powerpc/sysdev/fsl_gtm.c | 14 +-- arch/powerpc/sysdev/fsl_msi.c | 16 ++-- arch/powerpc/sysdev/fsl_pci.c | 47 +++++---- arch/powerpc/sysdev/fsl_rio.c | 36 +++---- arch/powerpc/sysdev/fsl_rmu.c | 12 +-- arch/powerpc/sysdev/mpic.c | 4 +- arch/powerpc/sysdev/mpic_msgr.c | 2 +- arch/powerpc/sysdev/mpic_msi.c | 2 +- arch/powerpc/sysdev/mpic_timer.c | 19 ++-- arch/powerpc/sysdev/msi_bitmap.c | 4 +- arch/powerpc/sysdev/mv64x60_dev.c | 20 ++-- arch/powerpc/sysdev/mv64x60_pci.c | 4 +- arch/powerpc/sysdev/of_rtc.c | 12 +-- arch/powerpc/sysdev/scom.c | 5 +- arch/powerpc/sysdev/simple_gpio.c | 3 +- arch/powerpc/sysdev/tsi108_pci.c | 4 +- arch/powerpc/sysdev/xive/native.c | 2 +- drivers/of/fdt.c | 69 +++----------- 83 files changed, 487 insertions(+), 578 deletions(-) -- 2.11.0 -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* [PATCH v2 1/5] powerpc: Convert to using %pOF instead of full_name 2017-08-21 15:16 ` Rob Herring (?) @ 2017-08-21 15:16 ` Rob Herring 2017-08-24 12:37 ` Michael Ellerman -1 siblings, 1 reply; 53+ messages in thread From: Rob Herring @ 2017-08-21 15:16 UTC (permalink / raw) To: devicetree, Michael Ellerman, linuxppc-dev Cc: linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight, Anatolij Gustschin, Scott Wood, Kumar Gala, Arnd Bergmann Now that we have a custom printf format specifier, convert users of full_name to use %pOF instead. This is preparation to remove storing of the full path string for each node. Signed-off-by: Rob Herring <robh@kernel.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Anatolij Gustschin <agust@denx.de> Cc: Scott Wood <oss@buserror.net> Cc: Kumar Gala <galak@kernel.crashing.org> Cc: Arnd Bergmann <arnd@arndb.de> Cc: linuxppc-dev@lists.ozlabs.org Reviewed-by: Tyrel Datwyler <tyreld@linux.vnet.ibm.com> --- v2: - drop spurious vio.c vio_register_device_node() change arch/powerpc/kernel/btext.c | 2 +- arch/powerpc/kernel/cacheinfo.c | 34 ++++--- arch/powerpc/kernel/io-workarounds.c | 4 +- arch/powerpc/kernel/isa-bridge.c | 32 +++---- arch/powerpc/kernel/legacy_serial.c | 12 +-- arch/powerpc/kernel/of_platform.c | 2 +- arch/powerpc/kernel/pci-common.c | 15 ++- arch/powerpc/kernel/pci_32.c | 4 +- arch/powerpc/kernel/pci_64.c | 4 +- arch/powerpc/kernel/pci_of_scan.c | 24 ++--- arch/powerpc/kernel/setup-common.c | 2 +- arch/powerpc/platforms/4xx/cpm.c | 8 +- arch/powerpc/platforms/4xx/gpio.c | 3 +- arch/powerpc/platforms/4xx/msi.c | 3 +- arch/powerpc/platforms/4xx/pci.c | 116 ++++++++++------------- arch/powerpc/platforms/4xx/soc.c | 5 +- arch/powerpc/platforms/4xx/uic.c | 14 +-- arch/powerpc/platforms/512x/mpc512x_shared.c | 12 +-- arch/powerpc/platforms/52xx/efika.c | 8 +- arch/powerpc/platforms/52xx/media5200.c | 2 +- arch/powerpc/platforms/52xx/mpc52xx_gpt.c | 4 +- arch/powerpc/platforms/52xx/mpc52xx_pci.c | 8 +- arch/powerpc/platforms/83xx/mcu_mpc8349emitx.c | 3 +- arch/powerpc/platforms/83xx/mpc832x_rdb.c | 2 +- arch/powerpc/platforms/85xx/p1022_ds.c | 8 +- arch/powerpc/platforms/85xx/xes_mpc85xx.c | 4 +- arch/powerpc/platforms/amigaone/setup.c | 6 +- arch/powerpc/platforms/cell/axon_msi.c | 36 +++---- arch/powerpc/platforms/cell/interrupt.c | 4 +- arch/powerpc/platforms/cell/iommu.c | 24 ++--- arch/powerpc/platforms/cell/ras.c | 4 +- arch/powerpc/platforms/cell/spider-pci.c | 4 +- arch/powerpc/platforms/cell/spider-pic.c | 4 +- arch/powerpc/platforms/cell/spu_manage.c | 26 ++--- arch/powerpc/platforms/chrp/pci.c | 18 ++-- arch/powerpc/platforms/embedded6xx/linkstation.c | 6 +- arch/powerpc/platforms/embedded6xx/mvme5100.c | 2 +- arch/powerpc/platforms/embedded6xx/storcenter.c | 2 +- arch/powerpc/platforms/maple/pci.c | 10 +- arch/powerpc/platforms/pasemi/pci.c | 2 +- arch/powerpc/platforms/powermac/feature.c | 14 +-- arch/powerpc/platforms/powermac/low_i2c.c | 50 +++++----- arch/powerpc/platforms/powermac/pci.c | 6 +- arch/powerpc/platforms/powermac/pfunc_base.c | 24 ++--- arch/powerpc/platforms/powermac/pfunc_core.c | 6 +- arch/powerpc/platforms/powermac/pic.c | 8 +- arch/powerpc/platforms/powermac/setup.c | 2 +- arch/powerpc/platforms/powernv/opal-async.c | 4 +- arch/powerpc/platforms/powernv/opal-xscom.c | 8 +- arch/powerpc/platforms/powernv/pci-ioda.c | 15 ++- arch/powerpc/platforms/powernv/rng.c | 6 +- arch/powerpc/platforms/pseries/dlpar.c | 3 +- arch/powerpc/platforms/pseries/event_sources.c | 6 +- arch/powerpc/platforms/pseries/hotplug-cpu.c | 4 +- arch/powerpc/platforms/pseries/ibmebus.c | 5 +- arch/powerpc/platforms/pseries/iommu.c | 58 ++++++------ arch/powerpc/platforms/pseries/msi.c | 12 +-- arch/powerpc/platforms/pseries/pci_dlpar.c | 2 +- arch/powerpc/platforms/pseries/vio.c | 6 +- arch/powerpc/sysdev/axonram.c | 4 +- arch/powerpc/sysdev/dcr.c | 4 +- arch/powerpc/sysdev/fsl_85xx_cache_sram.c | 12 +-- arch/powerpc/sysdev/fsl_gtm.c | 14 +-- arch/powerpc/sysdev/fsl_msi.c | 16 ++-- arch/powerpc/sysdev/fsl_pci.c | 47 +++++---- arch/powerpc/sysdev/fsl_rio.c | 36 +++---- arch/powerpc/sysdev/fsl_rmu.c | 12 +-- arch/powerpc/sysdev/mpic.c | 4 +- arch/powerpc/sysdev/mpic_msgr.c | 2 +- arch/powerpc/sysdev/mpic_msi.c | 2 +- arch/powerpc/sysdev/mpic_timer.c | 19 ++-- arch/powerpc/sysdev/msi_bitmap.c | 4 +- arch/powerpc/sysdev/mv64x60_dev.c | 20 ++-- arch/powerpc/sysdev/mv64x60_pci.c | 4 +- arch/powerpc/sysdev/of_rtc.c | 12 +-- arch/powerpc/sysdev/scom.c | 5 +- arch/powerpc/sysdev/simple_gpio.c | 3 +- arch/powerpc/sysdev/tsi108_pci.c | 4 +- arch/powerpc/sysdev/xive/native.c | 2 +- 79 files changed, 461 insertions(+), 487 deletions(-) diff --git a/arch/powerpc/kernel/btext.c b/arch/powerpc/kernel/btext.c index 8275858a434d..3f46ca1c59f9 100644 --- a/arch/powerpc/kernel/btext.c +++ b/arch/powerpc/kernel/btext.c @@ -253,7 +253,7 @@ int __init btext_find_display(int allow_nonstdout) for_each_node_by_type(np, "display") { if (of_get_property(np, "linux,opened", NULL)) { - printk("trying %s ...\n", np->full_name); + printk("trying %pOF ...\n", np); rc = btext_initialize(np); printk("result: %d\n", rc); } diff --git a/arch/powerpc/kernel/cacheinfo.c b/arch/powerpc/kernel/cacheinfo.c index c641983bbdd6..a8f20e5928e1 100644 --- a/arch/powerpc/kernel/cacheinfo.c +++ b/arch/powerpc/kernel/cacheinfo.c @@ -167,10 +167,10 @@ static void release_cache_debugcheck(struct cache *cache) list_for_each_entry(iter, &cache_list, list) WARN_ONCE(iter->next_local == cache, - "cache for %s(%s) refers to cache for %s(%s)\n", - iter->ofnode->full_name, + "cache for %pOF(%s) refers to cache for %pOF(%s)\n", + iter->ofnode, cache_type_string(iter), - cache->ofnode->full_name, + cache->ofnode, cache_type_string(cache)); } @@ -179,8 +179,8 @@ static void release_cache(struct cache *cache) if (!cache) return; - pr_debug("freeing L%d %s cache for %s\n", cache->level, - cache_type_string(cache), cache->ofnode->full_name); + pr_debug("freeing L%d %s cache for %pOF\n", cache->level, + cache_type_string(cache), cache->ofnode); release_cache_debugcheck(cache); list_del(&cache->list); @@ -194,8 +194,8 @@ static void cache_cpu_set(struct cache *cache, int cpu) while (next) { WARN_ONCE(cpumask_test_cpu(cpu, &next->shared_cpu_map), - "CPU %i already accounted in %s(%s)\n", - cpu, next->ofnode->full_name, + "CPU %i already accounted in %pOF(%s)\n", + cpu, next->ofnode, cache_type_string(next)); cpumask_set_cpu(cpu, &next->shared_cpu_map); next = next->next_local; @@ -355,7 +355,7 @@ static int cache_is_unified_d(const struct device_node *np) */ static struct cache *cache_do_one_devnode_unified(struct device_node *node, int level) { - pr_debug("creating L%d ucache for %s\n", level, node->full_name); + pr_debug("creating L%d ucache for %pOF\n", level, node); return new_cache(cache_is_unified_d(node), level, node); } @@ -365,8 +365,8 @@ static struct cache *cache_do_one_devnode_split(struct device_node *node, { struct cache *dcache, *icache; - pr_debug("creating L%d dcache and icache for %s\n", level, - node->full_name); + pr_debug("creating L%d dcache and icache for %pOF\n", level, + node); dcache = new_cache(CACHE_TYPE_DATA, level, node); icache = new_cache(CACHE_TYPE_INSTRUCTION, level, node); @@ -679,7 +679,6 @@ static struct kobj_type cache_index_type = { static void cacheinfo_create_index_opt_attrs(struct cache_index_dir *dir) { - const char *cache_name; const char *cache_type; struct cache *cache; char *buf; @@ -690,7 +689,6 @@ static void cacheinfo_create_index_opt_attrs(struct cache_index_dir *dir) return; cache = dir->cache; - cache_name = cache->ofnode->full_name; cache_type = cache_type_string(cache); /* We don't want to create an attribute that can't provide a @@ -707,14 +705,14 @@ static void cacheinfo_create_index_opt_attrs(struct cache_index_dir *dir) rc = attr->show(&dir->kobj, attr, buf); if (rc <= 0) { pr_debug("not creating %s attribute for " - "%s(%s) (rc = %zd)\n", - attr->attr.name, cache_name, + "%pOF(%s) (rc = %zd)\n", + attr->attr.name, cache->ofnode, cache_type, rc); continue; } if (sysfs_create_file(&dir->kobj, &attr->attr)) - pr_debug("could not create %s attribute for %s(%s)\n", - attr->attr.name, cache_name, cache_type); + pr_debug("could not create %s attribute for %pOF(%s)\n", + attr->attr.name, cache->ofnode, cache_type); } kfree(buf); @@ -831,8 +829,8 @@ static void cache_cpu_clear(struct cache *cache, int cpu) struct cache *next = cache->next_local; WARN_ONCE(!cpumask_test_cpu(cpu, &cache->shared_cpu_map), - "CPU %i not accounted in %s(%s)\n", - cpu, cache->ofnode->full_name, + "CPU %i not accounted in %pOF(%s)\n", + cpu, cache->ofnode, cache_type_string(cache)); cpumask_clear_cpu(cpu, &cache->shared_cpu_map); diff --git a/arch/powerpc/kernel/io-workarounds.c b/arch/powerpc/kernel/io-workarounds.c index a582e0d42525..9fd7bf1481bf 100644 --- a/arch/powerpc/kernel/io-workarounds.c +++ b/arch/powerpc/kernel/io-workarounds.c @@ -192,7 +192,7 @@ void iowa_register_bus(struct pci_controller *phb, struct ppc_pci_io *ops, if (iowa_bus_count >= IOWA_MAX_BUS) { pr_err("IOWA:Too many pci bridges, " - "workarounds disabled for %s\n", np->full_name); + "workarounds disabled for %pOF\n", np); return; } @@ -207,6 +207,6 @@ void iowa_register_bus(struct pci_controller *phb, struct ppc_pci_io *ops, iowa_bus_count++; - pr_debug("IOWA:[%d]Add bus, %s.\n", iowa_bus_count-1, np->full_name); + pr_debug("IOWA:[%d]Add bus, %pOF.\n", iowa_bus_count-1, np); } diff --git a/arch/powerpc/kernel/isa-bridge.c b/arch/powerpc/kernel/isa-bridge.c index bb6f8993412e..1df6c74aa731 100644 --- a/arch/powerpc/kernel/isa-bridge.c +++ b/arch/powerpc/kernel/isa-bridge.c @@ -164,7 +164,7 @@ void __init isa_bridge_find_early(struct pci_controller *hose) /* Set the global ISA io base to indicate we have an ISA bridge */ isa_io_base = ISA_IO_BASE; - pr_debug("ISA bridge (early) is %s\n", np->full_name); + pr_debug("ISA bridge (early) is %pOF\n", np); } /** @@ -187,15 +187,15 @@ void __init isa_bridge_init_non_pci(struct device_node *np) pna = of_n_addr_cells(np); if (of_property_read_u32(np, "#address-cells", &na) || of_property_read_u32(np, "#size-cells", &ns)) { - pr_warn("ISA: Non-PCI bridge %s is missing address format\n", - np->full_name); + pr_warn("ISA: Non-PCI bridge %pOF is missing address format\n", + np); return; } /* Check it's a supported address format */ if (na != 2 || ns != 1) { - pr_warn("ISA: Non-PCI bridge %s has unsupported address format\n", - np->full_name); + pr_warn("ISA: Non-PCI bridge %pOF has unsupported address format\n", + np); return; } rs = na + ns + pna; @@ -203,8 +203,8 @@ void __init isa_bridge_init_non_pci(struct device_node *np) /* Grab the ranges property */ ranges = of_get_property(np, "ranges", &rlen); if (ranges == NULL || rlen < rs) { - pr_warn("ISA: Non-PCI bridge %s has absent or invalid ranges\n", - np->full_name); + pr_warn("ISA: Non-PCI bridge %pOF has absent or invalid ranges\n", + np); return; } @@ -220,8 +220,8 @@ void __init isa_bridge_init_non_pci(struct device_node *np) /* Got something ? */ if (!size || !pbasep) { - pr_warn("ISA: Non-PCI bridge %s has no usable IO range\n", - np->full_name); + pr_warn("ISA: Non-PCI bridge %pOF has no usable IO range\n", + np); return; } @@ -233,15 +233,15 @@ void __init isa_bridge_init_non_pci(struct device_node *np) /* Map pbase */ pbase = of_translate_address(np, pbasep); if (pbase == OF_BAD_ADDR) { - pr_warn("ISA: Non-PCI bridge %s failed to translate IO base\n", - np->full_name); + pr_warn("ISA: Non-PCI bridge %pOF failed to translate IO base\n", + np); return; } /* We need page alignment */ if ((cbase & ~PAGE_MASK) || (pbase & ~PAGE_MASK)) { - pr_warn("ISA: Non-PCI bridge %s has non aligned IO range\n", - np->full_name); + pr_warn("ISA: Non-PCI bridge %pOF has non aligned IO range\n", + np); return; } @@ -255,7 +255,7 @@ void __init isa_bridge_init_non_pci(struct device_node *np) __ioremap_at(pbase, (void *)ISA_IO_BASE, size, pgprot_val(pgprot_noncached(__pgprot(0)))); - pr_debug("ISA: Non-PCI bridge is %s\n", np->full_name); + pr_debug("ISA: Non-PCI bridge is %pOF\n", np); } /** @@ -277,8 +277,8 @@ static void isa_bridge_find_late(struct pci_dev *pdev, /* Set the global ISA io base to indicate we have an ISA bridge */ isa_io_base = ISA_IO_BASE; - pr_debug("ISA bridge (late) is %s on %s\n", - devnode->full_name, pci_name(pdev)); + pr_debug("ISA bridge (late) is %pOF on %s\n", + devnode, pci_name(pdev)); } /** diff --git a/arch/powerpc/kernel/legacy_serial.c b/arch/powerpc/kernel/legacy_serial.c index 0694d20f85b6..5e5a64a8b4e4 100644 --- a/arch/powerpc/kernel/legacy_serial.c +++ b/arch/powerpc/kernel/legacy_serial.c @@ -147,8 +147,8 @@ static int __init add_legacy_port(struct device_node *np, int want_index, legacy_serial_ports[index].serial_out = tsi_serial_out; } - printk(KERN_DEBUG "Found legacy serial port %d for %s\n", - index, np->full_name); + printk(KERN_DEBUG "Found legacy serial port %d for %pOF\n", + index, np); printk(KERN_DEBUG " %s=%llx, taddr=%llx, irq=%lx, clk=%d, speed=%d\n", (iotype == UPIO_PORT) ? "port" : "mem", (unsigned long long)base, (unsigned long long)taddr, irq, @@ -207,7 +207,7 @@ static int __init add_legacy_isa_port(struct device_node *np, int index = -1; u64 taddr; - DBG(" -> add_legacy_isa_port(%s)\n", np->full_name); + DBG(" -> add_legacy_isa_port(%pOF)\n", np); /* Get the ISA port number */ reg = of_get_property(np, "reg", NULL); @@ -256,7 +256,7 @@ static int __init add_legacy_pci_port(struct device_node *np, unsigned int flags; int iotype, index = -1, lindex = 0; - DBG(" -> add_legacy_pci_port(%s)\n", np->full_name); + DBG(" -> add_legacy_pci_port(%pOF)\n", np); /* We only support ports that have a clock frequency properly * encoded in the device-tree (that is have an fcode). Anything @@ -374,7 +374,7 @@ void __init find_legacy_serial_ports(void) if (path != NULL) { stdout = of_find_node_by_path(path); if (stdout) - DBG("stdout is %s\n", stdout->full_name); + DBG("stdout is %pOF\n", stdout); } else { DBG(" no linux,stdout-path !\n"); } @@ -603,7 +603,7 @@ static int __init check_legacy_serial_console(void) DBG(" can't find stdout package %s !\n", name); return -ENODEV; } - DBG("stdout is %s\n", prom_stdout->full_name); + DBG("stdout is %pOF\n", prom_stdout); name = of_get_property(prom_stdout, "name", NULL); if (!name) { diff --git a/arch/powerpc/kernel/of_platform.c b/arch/powerpc/kernel/of_platform.c index 34aeac54f120..becaec990140 100644 --- a/arch/powerpc/kernel/of_platform.c +++ b/arch/powerpc/kernel/of_platform.c @@ -45,7 +45,7 @@ static int of_pci_phb_probe(struct platform_device *dev) if (ppc_md.pci_setup_phb == NULL) return -ENODEV; - pr_info("Setting up PCI bus %s\n", dev->dev.of_node->full_name); + pr_info("Setting up PCI bus %pOF\n", dev->dev.of_node); /* Alloc and setup PHB data structure */ phb = pcibios_alloc_controller(dev->dev.of_node); diff --git a/arch/powerpc/kernel/pci-common.c b/arch/powerpc/kernel/pci-common.c index 341a7469cab8..02831a396419 100644 --- a/arch/powerpc/kernel/pci-common.c +++ b/arch/powerpc/kernel/pci-common.c @@ -373,9 +373,8 @@ static int pci_read_irq_line(struct pci_dev *pci_dev) if (virq) irq_set_irq_type(virq, IRQ_TYPE_LEVEL_LOW); } else { - pr_debug(" Got one, spec %d cells (0x%08x 0x%08x...) on %s\n", - oirq.args_count, oirq.args[0], oirq.args[1], - of_node_full_name(oirq.np)); + pr_debug(" Got one, spec %d cells (0x%08x 0x%08x...) on %pOF\n", + oirq.args_count, oirq.args[0], oirq.args[1], oirq.np); virq = irq_create_of_mapping(&oirq); } @@ -741,8 +740,8 @@ void pci_process_bridge_OF_ranges(struct pci_controller *hose, struct of_pci_range range; struct of_pci_range_parser parser; - printk(KERN_INFO "PCI host bridge %s %s ranges:\n", - dev->full_name, primary ? "(primary)" : ""); + printk(KERN_INFO "PCI host bridge %pOF %s ranges:\n", + dev, primary ? "(primary)" : ""); /* Check for ranges property */ if (of_pci_range_parser_init(&parser, dev)) @@ -1556,8 +1555,8 @@ static void pcibios_setup_phb_resources(struct pci_controller *hose, if (!res->flags) { pr_debug("PCI: I/O resource not set for host" - " bridge %s (domain %d)\n", - hose->dn->full_name, hose->global_number); + " bridge %pOF (domain %d)\n", + hose->dn, hose->global_number); } else { offset = pcibios_io_space_offset(hose); @@ -1668,7 +1667,7 @@ void pcibios_scan_phb(struct pci_controller *hose) struct device_node *node = hose->dn; int mode; - pr_debug("PCI: Scanning PHB %s\n", of_node_full_name(node)); + pr_debug("PCI: Scanning PHB %pOF\n", node); /* Get some IO space for the new PHB */ pcibios_setup_phb_io_space(hose); diff --git a/arch/powerpc/kernel/pci_32.c b/arch/powerpc/kernel/pci_32.c index 41c86c6b6e4d..1d817f4d97d9 100644 --- a/arch/powerpc/kernel/pci_32.c +++ b/arch/powerpc/kernel/pci_32.c @@ -79,8 +79,8 @@ make_one_node_map(struct device_node* node, u8 pci_bus) return; bus_range = of_get_property(node, "bus-range", &len); if (bus_range == NULL || len < 2 * sizeof(int)) { - printk(KERN_WARNING "Can't get bus-range for %s, " - "assuming it starts at 0\n", node->full_name); + printk(KERN_WARNING "Can't get bus-range for %pOF, " + "assuming it starts at 0\n", node); pci_to_OF_bus_map[pci_bus] = 0; } else pci_to_OF_bus_map[pci_bus] = bus_range[0]; diff --git a/arch/powerpc/kernel/pci_64.c b/arch/powerpc/kernel/pci_64.c index ed5e9ff61a68..932b9741aa8f 100644 --- a/arch/powerpc/kernel/pci_64.c +++ b/arch/powerpc/kernel/pci_64.c @@ -111,7 +111,7 @@ int pcibios_unmap_io_space(struct pci_bus *bus) if (hose->io_base_alloc == NULL) return 0; - pr_debug("IO unmapping for PHB %s\n", hose->dn->full_name); + pr_debug("IO unmapping for PHB %pOF\n", hose->dn); pr_debug(" alloc=0x%p\n", hose->io_base_alloc); /* This is a PHB, we fully unmap the IO area */ @@ -151,7 +151,7 @@ static int pcibios_map_phb_io_space(struct pci_controller *hose) hose->io_base_virt = (void __iomem *)(area->addr + hose->io_base_phys - phys_page); - pr_debug("IO mapping for PHB %s\n", hose->dn->full_name); + pr_debug("IO mapping for PHB %pOF\n", hose->dn); pr_debug(" phys=0x%016llx, virt=0x%p (alloc=0x%p)\n", hose->io_base_phys, hose->io_base_virt, hose->io_base_alloc); pr_debug(" size=0x%016llx (alloc=0x%016lx)\n", diff --git a/arch/powerpc/kernel/pci_of_scan.c b/arch/powerpc/kernel/pci_of_scan.c index ea3d98115b88..0d790f8432d2 100644 --- a/arch/powerpc/kernel/pci_of_scan.c +++ b/arch/powerpc/kernel/pci_of_scan.c @@ -211,19 +211,19 @@ void of_scan_pci_bridge(struct pci_dev *dev) unsigned int flags; u64 size; - pr_debug("of_scan_pci_bridge(%s)\n", node->full_name); + pr_debug("of_scan_pci_bridge(%pOF)\n", node); /* parse bus-range property */ busrange = of_get_property(node, "bus-range", &len); if (busrange == NULL || len != 8) { - printk(KERN_DEBUG "Can't get bus-range for PCI-PCI bridge %s\n", - node->full_name); + printk(KERN_DEBUG "Can't get bus-range for PCI-PCI bridge %pOF\n", + node); return; } ranges = of_get_property(node, "ranges", &len); if (ranges == NULL) { - printk(KERN_DEBUG "Can't get ranges for PCI-PCI bridge %s\n", - node->full_name); + printk(KERN_DEBUG "Can't get ranges for PCI-PCI bridge %pOF\n", + node); return; } @@ -233,8 +233,8 @@ void of_scan_pci_bridge(struct pci_dev *dev) bus = pci_add_new_bus(dev->bus, dev, of_read_number(busrange, 1)); if (!bus) { - printk(KERN_ERR "Failed to create pci bus for %s\n", - node->full_name); + printk(KERN_ERR "Failed to create pci bus for %pOF\n", + node); return; } } @@ -262,13 +262,13 @@ void of_scan_pci_bridge(struct pci_dev *dev) res = bus->resource[0]; if (res->flags) { printk(KERN_ERR "PCI: ignoring extra I/O range" - " for bridge %s\n", node->full_name); + " for bridge %pOF\n", node); continue; } } else { if (i >= PCI_NUM_RESOURCES - PCI_BRIDGE_RESOURCES) { printk(KERN_ERR "PCI: too many memory ranges" - " for bridge %s\n", node->full_name); + " for bridge %pOF\n", node); continue; } res = bus->resource[i]; @@ -307,7 +307,7 @@ static struct pci_dev *of_scan_pci_dev(struct pci_bus *bus, struct eeh_dev *edev = pdn_to_eeh_dev(PCI_DN(dn)); #endif - pr_debug(" * %s\n", dn->full_name); + pr_debug(" * %pOF\n", dn); if (!of_device_is_available(dn)) return NULL; @@ -350,8 +350,8 @@ static void __of_scan_bus(struct device_node *node, struct pci_bus *bus, struct device_node *child; struct pci_dev *dev; - pr_debug("of_scan_bus(%s) bus no %d...\n", - node->full_name, bus->number); + pr_debug("of_scan_bus(%pOF) bus no %d...\n", + node, bus->number); /* Scan direct children */ for_each_child_of_node(node, child) { diff --git a/arch/powerpc/kernel/setup-common.c b/arch/powerpc/kernel/setup-common.c index 94a948207cd2..417871054006 100644 --- a/arch/powerpc/kernel/setup-common.c +++ b/arch/powerpc/kernel/setup-common.c @@ -481,7 +481,7 @@ void __init smp_setup_cpu_maps(void) __be32 cpu_be; int j, len; - DBG(" * %s...\n", dn->full_name); + DBG(" * %pOF...\n", dn); intserv = of_get_property(dn, "ibm,ppc-interrupt-server#s", &len); diff --git a/arch/powerpc/platforms/4xx/cpm.c b/arch/powerpc/platforms/4xx/cpm.c index ba95adf81d8d..ac7af8b5ffa1 100644 --- a/arch/powerpc/platforms/4xx/cpm.c +++ b/arch/powerpc/platforms/4xx/cpm.c @@ -278,8 +278,8 @@ static int __init cpm_init(void) dcr_len = dcr_resource_len(np, 0); if (dcr_base == 0 || dcr_len == 0) { - printk(KERN_ERR "cpm: could not parse dcr property for %s\n", - np->full_name); + printk(KERN_ERR "cpm: could not parse dcr property for %pOF\n", + np); ret = -EINVAL; goto node_put; } @@ -287,8 +287,8 @@ static int __init cpm_init(void) cpm.dcr_host = dcr_map(np, dcr_base, dcr_len); if (!DCR_MAP_OK(cpm.dcr_host)) { - printk(KERN_ERR "cpm: failed to map dcr property for %s\n", - np->full_name); + printk(KERN_ERR "cpm: failed to map dcr property for %pOF\n", + np); ret = -EINVAL; goto node_put; } diff --git a/arch/powerpc/platforms/4xx/gpio.c b/arch/powerpc/platforms/4xx/gpio.c index 5382d04dd872..2238e369cde4 100644 --- a/arch/powerpc/platforms/4xx/gpio.c +++ b/arch/powerpc/platforms/4xx/gpio.c @@ -198,8 +198,7 @@ static int __init ppc4xx_add_gpiochips(void) goto err; continue; err: - pr_err("%s: registration failed with status %d\n", - np->full_name, ret); + pr_err("%pOF: registration failed with status %d\n", np, ret); kfree(ppc4xx_gc); /* try others anyway */ } diff --git a/arch/powerpc/platforms/4xx/msi.c b/arch/powerpc/platforms/4xx/msi.c index 590dab4f47d6..d50417e23add 100644 --- a/arch/powerpc/platforms/4xx/msi.c +++ b/arch/powerpc/platforms/4xx/msi.c @@ -233,8 +233,7 @@ static int ppc4xx_msi_probe(struct platform_device *dev) /* Get MSI ranges */ err = of_address_to_resource(dev->dev.of_node, 0, &res); if (err) { - dev_err(&dev->dev, "%s resource error!\n", - dev->dev.of_node->full_name); + dev_err(&dev->dev, "%pOF resource error!\n", dev->dev.of_node); goto error_out; } diff --git a/arch/powerpc/platforms/4xx/pci.c b/arch/powerpc/platforms/4xx/pci.c index 6713edc93a55..73e6b36bcd51 100644 --- a/arch/powerpc/platforms/4xx/pci.c +++ b/arch/powerpc/platforms/4xx/pci.c @@ -127,9 +127,9 @@ static int __init ppc4xx_parse_dma_ranges(struct pci_controller *hose, * within 32 bits space */ if (cpu_addr != 0 || pci_addr > 0xffffffff) { - printk(KERN_WARNING "%s: Ignored unsupported dma range" + printk(KERN_WARNING "%pOF: Ignored unsupported dma range" " 0x%016llx...0x%016llx -> 0x%016llx\n", - hose->dn->full_name, + hose->dn, pci_addr, pci_addr + size - 1, cpu_addr); continue; } @@ -152,8 +152,7 @@ static int __init ppc4xx_parse_dma_ranges(struct pci_controller *hose, /* We only support one global DMA offset */ if (dma_offset_set && pci_dram_offset != res->start) { - printk(KERN_ERR "%s: dma-ranges(s) mismatch\n", - hose->dn->full_name); + printk(KERN_ERR "%pOF: dma-ranges(s) mismatch\n", hose->dn); return -ENXIO; } @@ -161,17 +160,16 @@ static int __init ppc4xx_parse_dma_ranges(struct pci_controller *hose, * DMA bounce buffers */ if (size < total_memory) { - printk(KERN_ERR "%s: dma-ranges too small " + printk(KERN_ERR "%pOF: dma-ranges too small " "(size=%llx total_memory=%llx)\n", - hose->dn->full_name, size, (u64)total_memory); + hose->dn, size, (u64)total_memory); return -ENXIO; } /* Check we are a power of 2 size and that base is a multiple of size*/ if ((size & (size - 1)) != 0 || (res->start & (size - 1)) != 0) { - printk(KERN_ERR "%s: dma-ranges unaligned\n", - hose->dn->full_name); + printk(KERN_ERR "%pOF: dma-ranges unaligned\n", hose->dn); return -ENXIO; } @@ -181,8 +179,8 @@ static int __init ppc4xx_parse_dma_ranges(struct pci_controller *hose, if (res->end > 0xffffffff && !(of_device_is_compatible(hose->dn, "ibm,plb-pciex-460sx") || of_device_is_compatible(hose->dn, "ibm,plb-pciex-476fpe"))) { - printk(KERN_ERR "%s: dma-ranges outside of 32 bits space\n", - hose->dn->full_name); + printk(KERN_ERR "%pOF: dma-ranges outside of 32 bits space\n", + hose->dn); return -ENXIO; } out: @@ -233,8 +231,7 @@ static int __init ppc4xx_setup_one_pci_PMM(struct pci_controller *hose, */ if ((plb_addr + size) > 0xffffffffull || !is_power_of_2(size) || size < 0x1000 || (plb_addr & (size - 1)) != 0) { - printk(KERN_WARNING "%s: Resource out of range\n", - hose->dn->full_name); + printk(KERN_WARNING "%pOF: Resource out of range\n", hose->dn); return -1; } ma = (0xffffffffu << ilog2(size)) | 1; @@ -266,8 +263,7 @@ static void __init ppc4xx_configure_pci_PMMs(struct pci_controller *hose, if (!(res->flags & IORESOURCE_MEM)) continue; if (j > 2) { - printk(KERN_WARNING "%s: Too many ranges\n", - hose->dn->full_name); + printk(KERN_WARNING "%pOF: Too many ranges\n", hose->dn); break; } @@ -292,8 +288,8 @@ static void __init ppc4xx_configure_pci_PMMs(struct pci_controller *hose, if (j <= 2 && !found_isa_hole && hose->isa_mem_size) if (ppc4xx_setup_one_pci_PMM(hose, reg, hose->isa_mem_phys, 0, hose->isa_mem_size, 0, j) == 0) - printk(KERN_INFO "%s: Legacy ISA memory support enabled\n", - hose->dn->full_name); + printk(KERN_INFO "%pOF: Legacy ISA memory support enabled\n", + hose->dn); } static void __init ppc4xx_configure_pci_PTMs(struct pci_controller *hose, @@ -333,21 +329,20 @@ static void __init ppc4xx_probe_pci_bridge(struct device_node *np) /* Check if device is enabled */ if (!of_device_is_available(np)) { - printk(KERN_INFO "%s: Port disabled via device-tree\n", - np->full_name); + printk(KERN_INFO "%pOF: Port disabled via device-tree\n", np); return; } /* Fetch config space registers address */ if (of_address_to_resource(np, 0, &rsrc_cfg)) { - printk(KERN_ERR "%s: Can't get PCI config register base !", - np->full_name); + printk(KERN_ERR "%pOF: Can't get PCI config register base !", + np); return; } /* Fetch host bridge internal registers address */ if (of_address_to_resource(np, 3, &rsrc_reg)) { - printk(KERN_ERR "%s: Can't get PCI internal register base !", - np->full_name); + printk(KERN_ERR "%pOF: Can't get PCI internal register base !", + np); return; } @@ -361,7 +356,7 @@ static void __init ppc4xx_probe_pci_bridge(struct device_node *np) /* Map registers */ reg = ioremap(rsrc_reg.start, resource_size(&rsrc_reg)); if (reg == NULL) { - printk(KERN_ERR "%s: Can't map registers !", np->full_name); + printk(KERN_ERR "%pOF: Can't map registers !", np); goto fail; } @@ -423,8 +418,8 @@ static int __init ppc4xx_setup_one_pcix_POM(struct pci_controller *hose, if (!is_power_of_2(size) || size < 0x1000 || (plb_addr & (size - 1)) != 0) { - printk(KERN_WARNING "%s: Resource out of range\n", - hose->dn->full_name); + printk(KERN_WARNING "%pOF: Resource out of range\n", + hose->dn); return -1; } @@ -467,8 +462,7 @@ static void __init ppc4xx_configure_pcix_POMs(struct pci_controller *hose, if (!(res->flags & IORESOURCE_MEM)) continue; if (j > 1) { - printk(KERN_WARNING "%s: Too many ranges\n", - hose->dn->full_name); + printk(KERN_WARNING "%pOF: Too many ranges\n", hose->dn); break; } @@ -493,8 +487,8 @@ static void __init ppc4xx_configure_pcix_POMs(struct pci_controller *hose, if (j <= 1 && !found_isa_hole && hose->isa_mem_size) if (ppc4xx_setup_one_pcix_POM(hose, reg, hose->isa_mem_phys, 0, hose->isa_mem_size, 0, j) == 0) - printk(KERN_INFO "%s: Legacy ISA memory support enabled\n", - hose->dn->full_name); + printk(KERN_INFO "%pOF: Legacy ISA memory support enabled\n", + hose->dn); } static void __init ppc4xx_configure_pcix_PIMs(struct pci_controller *hose, @@ -539,14 +533,14 @@ static void __init ppc4xx_probe_pcix_bridge(struct device_node *np) /* Fetch config space registers address */ if (of_address_to_resource(np, 0, &rsrc_cfg)) { - printk(KERN_ERR "%s:Can't get PCI-X config register base !", - np->full_name); + printk(KERN_ERR "%pOF: Can't get PCI-X config register base !", + np); return; } /* Fetch host bridge internal registers address */ if (of_address_to_resource(np, 3, &rsrc_reg)) { - printk(KERN_ERR "%s: Can't get PCI-X internal register base !", - np->full_name); + printk(KERN_ERR "%pOF: Can't get PCI-X internal register base !", + np); return; } @@ -568,7 +562,7 @@ static void __init ppc4xx_probe_pcix_bridge(struct device_node *np) /* Map registers */ reg = ioremap(rsrc_reg.start, resource_size(&rsrc_reg)); if (reg == NULL) { - printk(KERN_ERR "%s: Can't map registers !", np->full_name); + printk(KERN_ERR "%pOF: Can't map registers !", np); goto fail; } @@ -1246,8 +1240,8 @@ static void __init ppc460sx_pciex_check_link(struct ppc4xx_pciex_port *port) mbase = ioremap(port->cfg_space.start + 0x10000000, 0x1000); if (mbase == NULL) { - printk(KERN_ERR "%s: Can't map internal config space !", - port->node->full_name); + printk(KERN_ERR "%pOF: Can't map internal config space !", + port->node); goto done; } @@ -1389,7 +1383,7 @@ static void __init ppc_476fpe_pciex_check_link(struct ppc4xx_pciex_port *port) port->index); return; } - + while (timeout_ms--) { val = in_le32(mbase + PECFG_TLDLP); @@ -1448,8 +1442,7 @@ static int __init ppc4xx_pciex_check_core_init(struct device_node *np) ppc4xx_pciex_hwops = &ppc_476fpe_pcie_hwops; #endif if (ppc4xx_pciex_hwops == NULL) { - printk(KERN_WARNING "PCIE: unknown host type %s\n", - np->full_name); + printk(KERN_WARNING "PCIE: unknown host type %pOF\n", np); return -ENODEV; } @@ -1730,8 +1723,7 @@ static int __init ppc4xx_setup_one_pciex_POM(struct ppc4xx_pciex_port *port, (index < 2 && size < 0x100000) || (index == 2 && size < 0x100) || (plb_addr & (size - 1)) != 0) { - printk(KERN_WARNING "%s: Resource out of range\n", - hose->dn->full_name); + printk(KERN_WARNING "%pOF: Resource out of range\n", hose->dn); return -1; } @@ -1807,8 +1799,8 @@ static void __init ppc4xx_configure_pciex_POMs(struct ppc4xx_pciex_port *port, if (!(res->flags & IORESOURCE_MEM)) continue; if (j > 1) { - printk(KERN_WARNING "%s: Too many ranges\n", - port->node->full_name); + printk(KERN_WARNING "%pOF: Too many ranges\n", + port->node); break; } @@ -1834,8 +1826,8 @@ static void __init ppc4xx_configure_pciex_POMs(struct ppc4xx_pciex_port *port, if (ppc4xx_setup_one_pciex_POM(port, hose, mbase, hose->isa_mem_phys, 0, hose->isa_mem_size, 0, j) == 0) - printk(KERN_INFO "%s: Legacy ISA memory support enabled\n", - hose->dn->full_name); + printk(KERN_INFO "%pOF: Legacy ISA memory support enabled\n", + hose->dn); /* Configure IO, always 64K starting at 0. We hard wire it to 64K ! * Note also that it -has- to be region index 2 on this HW @@ -1970,8 +1962,8 @@ static void __init ppc4xx_pciex_port_setup_hose(struct ppc4xx_pciex_port *port) (hose->first_busno + 1) * 0x100000, busses * 0x100000); if (cfg_data == NULL) { - printk(KERN_ERR "%s: Can't map external config space !", - port->node->full_name); + printk(KERN_ERR "%pOF: Can't map external config space !", + port->node); goto fail; } hose->cfg_data = cfg_data; @@ -1982,13 +1974,13 @@ static void __init ppc4xx_pciex_port_setup_hose(struct ppc4xx_pciex_port *port) */ mbase = ioremap(port->cfg_space.start + 0x10000000, 0x1000); if (mbase == NULL) { - printk(KERN_ERR "%s: Can't map internal config space !", - port->node->full_name); + printk(KERN_ERR "%pOF: Can't map internal config space !", + port->node); goto fail; } hose->cfg_addr = mbase; - pr_debug("PCIE %s, bus %d..%d\n", port->node->full_name, + pr_debug("PCIE %pOF, bus %d..%d\n", port->node, hose->first_busno, hose->last_busno); pr_debug(" config space mapped at: root @0x%p, other @0x%p\n", hose->cfg_addr, hose->cfg_data); @@ -2100,14 +2092,13 @@ static void __init ppc4xx_probe_pciex_bridge(struct device_node *np) /* Get the port number from the device-tree */ pval = of_get_property(np, "port", NULL); if (pval == NULL) { - printk(KERN_ERR "PCIE: Can't find port number for %s\n", - np->full_name); + printk(KERN_ERR "PCIE: Can't find port number for %pOF\n", np); return; } portno = *pval; if (portno >= ppc4xx_pciex_port_count) { - printk(KERN_ERR "PCIE: port number out of range for %s\n", - np->full_name); + printk(KERN_ERR "PCIE: port number out of range for %pOF\n", + np); return; } port = &ppc4xx_pciex_ports[portno]; @@ -2125,8 +2116,8 @@ static void __init ppc4xx_probe_pciex_bridge(struct device_node *np) if (ppc4xx_pciex_hwops->want_sdr) { pval = of_get_property(np, "sdr-base", NULL); if (pval == NULL) { - printk(KERN_ERR "PCIE: missing sdr-base for %s\n", - np->full_name); + printk(KERN_ERR "PCIE: missing sdr-base for %pOF\n", + np); return; } port->sdr_base = *pval; @@ -2142,29 +2133,26 @@ static void __init ppc4xx_probe_pciex_bridge(struct device_node *np) } else if (!strcmp(val, "pci")) { port->endpoint = 0; } else { - printk(KERN_ERR "PCIE: missing or incorrect device_type for %s\n", - np->full_name); + printk(KERN_ERR "PCIE: missing or incorrect device_type for %pOF\n", + np); return; } /* Fetch config space registers address */ if (of_address_to_resource(np, 0, &port->cfg_space)) { - printk(KERN_ERR "%s: Can't get PCI-E config space !", - np->full_name); + printk(KERN_ERR "%pOF: Can't get PCI-E config space !", np); return; } /* Fetch host bridge internal registers address */ if (of_address_to_resource(np, 1, &port->utl_regs)) { - printk(KERN_ERR "%s: Can't get UTL register base !", - np->full_name); + printk(KERN_ERR "%pOF: Can't get UTL register base !", np); return; } /* Map DCRs */ dcrs = dcr_resource_start(np, 0); if (dcrs == 0) { - printk(KERN_ERR "%s: Can't get DCR register base !", - np->full_name); + printk(KERN_ERR "%pOF: Can't get DCR register base !", np); return; } port->dcrs = dcr_map(np, dcrs, dcr_resource_len(np, 0)); diff --git a/arch/powerpc/platforms/4xx/soc.c b/arch/powerpc/platforms/4xx/soc.c index d41134d2f786..5e36508b2a70 100644 --- a/arch/powerpc/platforms/4xx/soc.c +++ b/arch/powerpc/platforms/4xx/soc.c @@ -90,7 +90,7 @@ static int __init ppc4xx_l2c_probe(void) /* Get l2 cache size */ prop = of_get_property(np, "cache-size", NULL); if (prop == NULL) { - printk(KERN_ERR "%s: Can't get cache-size!\n", np->full_name); + printk(KERN_ERR "%pOF: Can't get cache-size!\n", np); of_node_put(np); return -ENODEV; } @@ -99,8 +99,7 @@ static int __init ppc4xx_l2c_probe(void) /* Map DCRs */ dcrreg = of_get_property(np, "dcr-reg", &len); if (!dcrreg || (len != 4 * sizeof(u32))) { - printk(KERN_ERR "%s: Can't get DCR register base !", - np->full_name); + printk(KERN_ERR "%pOF: Can't get DCR register base !", np); of_node_put(np); return -ENODEV; } diff --git a/arch/powerpc/platforms/4xx/uic.c b/arch/powerpc/platforms/4xx/uic.c index a00949f3e378..8b4dd0da0839 100644 --- a/arch/powerpc/platforms/4xx/uic.c +++ b/arch/powerpc/platforms/4xx/uic.c @@ -243,16 +243,16 @@ static struct uic * __init uic_init_one(struct device_node *node) raw_spin_lock_init(&uic->lock); indexp = of_get_property(node, "cell-index", &len); if (!indexp || (len != sizeof(u32))) { - printk(KERN_ERR "uic: Device node %s has missing or invalid " - "cell-index property\n", node->full_name); + printk(KERN_ERR "uic: Device node %pOF has missing or invalid " + "cell-index property\n", node); return NULL; } uic->index = *indexp; dcrreg = of_get_property(node, "dcr-reg", &len); if (!dcrreg || (len != 2*sizeof(u32))) { - printk(KERN_ERR "uic: Device node %s has missing or invalid " - "dcr-reg property\n", node->full_name); + printk(KERN_ERR "uic: Device node %pOF has missing or invalid " + "dcr-reg property\n", node); return NULL; } uic->dcrbase = *dcrreg; @@ -292,7 +292,7 @@ void __init uic_init_tree(void) * top-level interrupt controller */ primary_uic = uic_init_one(np); if (!primary_uic) - panic("Unable to initialize primary UIC %s\n", np->full_name); + panic("Unable to initialize primary UIC %pOF\n", np); irq_set_default_host(primary_uic->irqhost); of_node_put(np); @@ -306,8 +306,8 @@ void __init uic_init_tree(void) uic = uic_init_one(np); if (! uic) - panic("Unable to initialize a secondary UIC %s\n", - np->full_name); + panic("Unable to initialize a secondary UIC %pOF\n", + np); cascade_virq = irq_of_parse_and_map(np, 0); diff --git a/arch/powerpc/platforms/512x/mpc512x_shared.c b/arch/powerpc/platforms/512x/mpc512x_shared.c index 6b4f4cb7009a..f99e79ee060e 100644 --- a/arch/powerpc/platforms/512x/mpc512x_shared.c +++ b/arch/powerpc/platforms/512x/mpc512x_shared.c @@ -387,8 +387,8 @@ static unsigned int __init get_fifo_size(struct device_node *np, if (fp) return *fp; - pr_warning("no %s property in %s node, defaulting to %d\n", - prop_name, np->full_name, DEFAULT_FIFO_SIZE); + pr_warning("no %s property in %pOF node, defaulting to %d\n", + prop_name, np, DEFAULT_FIFO_SIZE); return DEFAULT_FIFO_SIZE; } @@ -426,15 +426,15 @@ static void __init mpc512x_psc_fifo_init(void) psc = of_iomap(np, 0); if (!psc) { - pr_err("%s: Can't map %s device\n", - __func__, np->full_name); + pr_err("%s: Can't map %pOF device\n", + __func__, np); continue; } /* FIFO space is 4KiB, check if requested size is available */ if ((fifobase + tx_fifo_size + rx_fifo_size) > 0x1000) { - pr_err("%s: no fifo space available for %s\n", - __func__, np->full_name); + pr_err("%s: no fifo space available for %pOF\n", + __func__, np); iounmap(psc); /* * chances are that another device requests less diff --git a/arch/powerpc/platforms/52xx/efika.c b/arch/powerpc/platforms/52xx/efika.c index 39b49822ace1..1ecbf176d35a 100644 --- a/arch/powerpc/platforms/52xx/efika.c +++ b/arch/powerpc/platforms/52xx/efika.c @@ -99,7 +99,7 @@ static void __init efika_pcisetup(void) bus_range = of_get_property(pcictrl, "bus-range", &len); if (bus_range == NULL || len < 2 * sizeof(int)) { printk(KERN_WARNING EFIKA_PLATFORM_NAME - ": Can't get bus-range for %s\n", pcictrl->full_name); + ": Can't get bus-range for %pOF\n", pcictrl); goto out_put; } @@ -109,14 +109,14 @@ static void __init efika_pcisetup(void) else printk(KERN_INFO EFIKA_PLATFORM_NAME ": PCI buses %d..%d", bus_range[0], bus_range[1]); - printk(" controlled by %s\n", pcictrl->full_name); + printk(" controlled by %pOF\n", pcictrl); printk("\n"); hose = pcibios_alloc_controller(pcictrl); if (!hose) { printk(KERN_WARNING EFIKA_PLATFORM_NAME - ": Can't allocate PCI controller structure for %s\n", - pcictrl->full_name); + ": Can't allocate PCI controller structure for %pOF\n", + pcictrl); goto out_put; } diff --git a/arch/powerpc/platforms/52xx/media5200.c b/arch/powerpc/platforms/52xx/media5200.c index a3227040cc86..1fcab233d2f2 100644 --- a/arch/powerpc/platforms/52xx/media5200.c +++ b/arch/powerpc/platforms/52xx/media5200.c @@ -156,7 +156,7 @@ static void __init media5200_init_irq(void) fpga_np = of_find_compatible_node(NULL, NULL, "fsl,media5200-fpga"); if (!fpga_np) goto out; - pr_debug("%s: found fpga node: %s\n", __func__, fpga_np->full_name); + pr_debug("%s: found fpga node: %pOF\n", __func__, fpga_np); media5200_irq.regs = of_iomap(fpga_np, 0); if (!media5200_irq.regs) diff --git a/arch/powerpc/platforms/52xx/mpc52xx_gpt.c b/arch/powerpc/platforms/52xx/mpc52xx_gpt.c index 22645a7c6b8a..9e974b1e1697 100644 --- a/arch/powerpc/platforms/52xx/mpc52xx_gpt.c +++ b/arch/powerpc/platforms/52xx/mpc52xx_gpt.c @@ -226,7 +226,7 @@ static int mpc52xx_gpt_irq_xlate(struct irq_domain *h, struct device_node *ct, dev_dbg(gpt->dev, "%s: flags=%i\n", __func__, intspec[0]); if ((intsize < 1) || (intspec[0] > 3)) { - dev_err(gpt->dev, "bad irq specifier in %s\n", ct->full_name); + dev_err(gpt->dev, "bad irq specifier in %pOF\n", ct); return -EINVAL; } @@ -331,7 +331,7 @@ mpc52xx_gpt_gpio_setup(struct mpc52xx_gpt_priv *gpt, struct device_node *node) if (!of_find_property(node, "gpio-controller", NULL)) return; - gpt->gc.label = kstrdup(node->full_name, GFP_KERNEL); + gpt->gc.label = kasprintf(GFP_KERNEL, "%pOF", node); if (!gpt->gc.label) { dev_err(gpt->dev, "out of memory\n"); return; diff --git a/arch/powerpc/platforms/52xx/mpc52xx_pci.c b/arch/powerpc/platforms/52xx/mpc52xx_pci.c index 00282c2b0cae..af0f79995214 100644 --- a/arch/powerpc/platforms/52xx/mpc52xx_pci.c +++ b/arch/powerpc/platforms/52xx/mpc52xx_pci.c @@ -369,19 +369,19 @@ mpc52xx_add_bridge(struct device_node *node) const int *bus_range; struct resource rsrc; - pr_debug("Adding MPC52xx PCI host bridge %s\n", node->full_name); + pr_debug("Adding MPC52xx PCI host bridge %pOF\n", node); pci_add_flags(PCI_REASSIGN_ALL_BUS); if (of_address_to_resource(node, 0, &rsrc) != 0) { - printk(KERN_ERR "Can't get %s resources\n", node->full_name); + printk(KERN_ERR "Can't get %pOF resources\n", node); return -EINVAL; } bus_range = of_get_property(node, "bus-range", &len); if (bus_range == NULL || len < 2 * sizeof(int)) { - printk(KERN_WARNING "Can't get %s bus-range, assume bus 0\n", - node->full_name); + printk(KERN_WARNING "Can't get %pOF bus-range, assume bus 0\n", + node); bus_range = NULL; } diff --git a/arch/powerpc/platforms/83xx/mcu_mpc8349emitx.c b/arch/powerpc/platforms/83xx/mcu_mpc8349emitx.c index 63c5ab6489c9..96bb55ca61d3 100644 --- a/arch/powerpc/platforms/83xx/mcu_mpc8349emitx.c +++ b/arch/powerpc/platforms/83xx/mcu_mpc8349emitx.c @@ -128,7 +128,7 @@ static int mcu_gpiochip_add(struct mcu *mcu) return -ENODEV; gc->owner = THIS_MODULE; - gc->label = np->full_name; + gc->label = kasprintf(GFP_KERNEL, "%pOF", np); gc->can_sleep = 1; gc->ngpio = MCU_NUM_GPIO; gc->base = -1; @@ -141,6 +141,7 @@ static int mcu_gpiochip_add(struct mcu *mcu) static int mcu_gpiochip_remove(struct mcu *mcu) { + kfree(mcu->gc.label); gpiochip_remove(&mcu->gc); return 0; } diff --git a/arch/powerpc/platforms/83xx/mpc832x_rdb.c b/arch/powerpc/platforms/83xx/mpc832x_rdb.c index 763ffca9628d..a4539c5accb0 100644 --- a/arch/powerpc/platforms/83xx/mpc832x_rdb.c +++ b/arch/powerpc/platforms/83xx/mpc832x_rdb.c @@ -113,7 +113,7 @@ static int __init of_fsl_spi_probe(char *type, char *compatible, u32 sysclk, unreg: platform_device_del(pdev); err: - pr_err("%s: registration failed\n", np->full_name); + pr_err("%pOF: registration failed\n", np); next: i++; } diff --git a/arch/powerpc/platforms/85xx/p1022_ds.c b/arch/powerpc/platforms/85xx/p1022_ds.c index 0908abd7e36f..9fb57f78cdbe 100644 --- a/arch/powerpc/platforms/85xx/p1022_ds.c +++ b/arch/powerpc/platforms/85xx/p1022_ds.c @@ -508,8 +508,8 @@ static void __init p1022_ds_setup_arch(void) * allocate one static local variable for each * call to this function. */ - pr_info("p1022ds: disabling %s node", - np2->full_name); + pr_info("p1022ds: disabling %pOF node", + np2); of_update_property(np2, &nor_status); of_node_put(np2); } @@ -524,8 +524,8 @@ static void __init p1022_ds_setup_arch(void) .length = sizeof("disabled"), }; - pr_info("p1022ds: disabling %s node", - np2->full_name); + pr_info("p1022ds: disabling %pOF node", + np2); of_update_property(np2, &nand_status); of_node_put(np2); } diff --git a/arch/powerpc/platforms/85xx/xes_mpc85xx.c b/arch/powerpc/platforms/85xx/xes_mpc85xx.c index cd6ce845f398..77e618dce4a8 100644 --- a/arch/powerpc/platforms/85xx/xes_mpc85xx.c +++ b/arch/powerpc/platforms/85xx/xes_mpc85xx.c @@ -100,8 +100,8 @@ static void xes_mpc85xx_fixups(void) err = of_address_to_resource(np, 0, &r[0]); if (err) { printk(KERN_WARNING "xes_mpc85xx: Could not get " - "resource for device tree node '%s'", - np->full_name); + "resource for device tree node '%pOF'", + np); continue; } diff --git a/arch/powerpc/platforms/amigaone/setup.c b/arch/powerpc/platforms/amigaone/setup.c index 45cb9821173c..b9d466cc2b8a 100644 --- a/arch/powerpc/platforms/amigaone/setup.c +++ b/arch/powerpc/platforms/amigaone/setup.c @@ -40,7 +40,7 @@ static int __init amigaone_add_bridge(struct device_node *dev) const int *bus_range; struct pci_controller *hose; - printk(KERN_INFO "Adding PCI host bridge %s\n", dev->full_name); + printk(KERN_INFO "Adding PCI host bridge %pOF\n", dev); cfg_addr = of_get_address(dev, 0, NULL, NULL); cfg_data = of_get_address(dev, 1, NULL, NULL); @@ -49,8 +49,8 @@ static int __init amigaone_add_bridge(struct device_node *dev) bus_range = of_get_property(dev, "bus-range", &len); if ((bus_range == NULL) || (len < 2 * sizeof(int))) - printk(KERN_WARNING "Can't get bus-range for %s, assume" - " bus 0\n", dev->full_name); + printk(KERN_WARNING "Can't get bus-range for %pOF, assume" + " bus 0\n", dev); hose = pcibios_alloc_controller(dev); if (hose == NULL) diff --git a/arch/powerpc/platforms/cell/axon_msi.c b/arch/powerpc/platforms/cell/axon_msi.c index 8d3ae2cc52bf..6ea3f248b155 100644 --- a/arch/powerpc/platforms/cell/axon_msi.c +++ b/arch/powerpc/platforms/cell/axon_msi.c @@ -187,8 +187,8 @@ static struct axon_msic *find_msi_translator(struct pci_dev *dev) irq_domain = irq_find_host(dn); if (!irq_domain) { - dev_dbg(&dev->dev, "axon_msi: no irq_domain found for node %s\n", - dn->full_name); + dev_dbg(&dev->dev, "axon_msi: no irq_domain found for node %pOF\n", + dn); goto out_error; } @@ -326,8 +326,8 @@ static void axon_msi_shutdown(struct platform_device *device) struct axon_msic *msic = dev_get_drvdata(&device->dev); u32 tmp; - pr_devel("axon_msi: disabling %s\n", - irq_domain_get_of_node(msic->irq_domain)->full_name); + pr_devel("axon_msi: disabling %pOF\n", + irq_domain_get_of_node(msic->irq_domain)); tmp = dcr_read(msic->dcr_host, MSIC_CTRL_REG); tmp &= ~MSIC_CTRL_ENABLE & ~MSIC_CTRL_IRQ_ENABLE; msic_dcr_write(msic, MSIC_CTRL_REG, tmp); @@ -340,12 +340,12 @@ static int axon_msi_probe(struct platform_device *device) unsigned int virq; int dcr_base, dcr_len; - pr_devel("axon_msi: setting up dn %s\n", dn->full_name); + pr_devel("axon_msi: setting up dn %pOF\n", dn); msic = kzalloc(sizeof(struct axon_msic), GFP_KERNEL); if (!msic) { - printk(KERN_ERR "axon_msi: couldn't allocate msic for %s\n", - dn->full_name); + printk(KERN_ERR "axon_msi: couldn't allocate msic for %pOF\n", + dn); goto out; } @@ -354,30 +354,30 @@ static int axon_msi_probe(struct platform_device *device) if (dcr_base == 0 || dcr_len == 0) { printk(KERN_ERR - "axon_msi: couldn't parse dcr properties on %s\n", - dn->full_name); + "axon_msi: couldn't parse dcr properties on %pOF\n", + dn); goto out_free_msic; } msic->dcr_host = dcr_map(dn, dcr_base, dcr_len); if (!DCR_MAP_OK(msic->dcr_host)) { - printk(KERN_ERR "axon_msi: dcr_map failed for %s\n", - dn->full_name); + printk(KERN_ERR "axon_msi: dcr_map failed for %pOF\n", + dn); goto out_free_msic; } msic->fifo_virt = dma_alloc_coherent(&device->dev, MSIC_FIFO_SIZE_BYTES, &msic->fifo_phys, GFP_KERNEL); if (!msic->fifo_virt) { - printk(KERN_ERR "axon_msi: couldn't allocate fifo for %s\n", - dn->full_name); + printk(KERN_ERR "axon_msi: couldn't allocate fifo for %pOF\n", + dn); goto out_free_msic; } virq = irq_of_parse_and_map(dn, 0); if (!virq) { - printk(KERN_ERR "axon_msi: irq parse and map failed for %s\n", - dn->full_name); + printk(KERN_ERR "axon_msi: irq parse and map failed for %pOF\n", + dn); goto out_free_fifo; } memset(msic->fifo_virt, 0xff, MSIC_FIFO_SIZE_BYTES); @@ -385,8 +385,8 @@ static int axon_msi_probe(struct platform_device *device) /* We rely on being able to stash a virq in a u16, so limit irqs to < 65536 */ msic->irq_domain = irq_domain_add_nomap(dn, 65536, &msic_host_ops, msic); if (!msic->irq_domain) { - printk(KERN_ERR "axon_msi: couldn't allocate irq_domain for %s\n", - dn->full_name); + printk(KERN_ERR "axon_msi: couldn't allocate irq_domain for %pOF\n", + dn); goto out_free_fifo; } @@ -412,7 +412,7 @@ static int axon_msi_probe(struct platform_device *device) axon_msi_debug_setup(dn, msic); - printk(KERN_DEBUG "axon_msi: setup MSIC on %s\n", dn->full_name); + printk(KERN_DEBUG "axon_msi: setup MSIC on %pOF\n", dn); return 0; diff --git a/arch/powerpc/platforms/cell/interrupt.c b/arch/powerpc/platforms/cell/interrupt.c index 871d38479a25..6fc85e29dc08 100644 --- a/arch/powerpc/platforms/cell/interrupt.c +++ b/arch/powerpc/platforms/cell/interrupt.c @@ -303,8 +303,8 @@ static void __init init_one_iic(unsigned int hw_cpu, unsigned long addr, iic->node = of_node_get(node); out_be64(&iic->regs->prio, 0); - printk(KERN_INFO "IIC for CPU %d target id 0x%x : %s\n", - hw_cpu, iic->target_id, node->full_name); + printk(KERN_INFO "IIC for CPU %d target id 0x%x : %pOF\n", + hw_cpu, iic->target_id, node); } static int __init setup_iic(void) diff --git a/arch/powerpc/platforms/cell/iommu.c b/arch/powerpc/platforms/cell/iommu.c index 29d4f96ed33e..4b91ad08eefd 100644 --- a/arch/powerpc/platforms/cell/iommu.c +++ b/arch/powerpc/platforms/cell/iommu.c @@ -278,8 +278,8 @@ static int cell_iommu_find_ioc(int nid, unsigned long *base) if (of_node_to_nid(np) != nid) continue; if (of_address_to_resource(np, 0, &r)) { - printk(KERN_ERR "iommu: can't get address for %s\n", - np->full_name); + printk(KERN_ERR "iommu: can't get address for %pOF\n", + np); continue; } *base = r.start; @@ -458,8 +458,8 @@ static inline u32 cell_iommu_get_ioid(struct device_node *np) ioid = of_get_property(np, "ioid", NULL); if (ioid == NULL) { - printk(KERN_WARNING "iommu: missing ioid for %s using 0\n", - np->full_name); + printk(KERN_WARNING "iommu: missing ioid for %pOF using 0\n", + np); return 0; } @@ -559,8 +559,8 @@ static struct iommu_table *cell_get_iommu_table(struct device *dev) */ iommu = cell_iommu_for_node(dev_to_node(dev)); if (iommu == NULL || list_empty(&iommu->windows)) { - dev_err(dev, "iommu: missing iommu for %s (node %d)\n", - of_node_full_name(dev->of_node), dev_to_node(dev)); + dev_err(dev, "iommu: missing iommu for %pOF (node %d)\n", + dev->of_node, dev_to_node(dev)); return NULL; } window = list_entry(iommu->windows.next, struct iommu_window, list); @@ -720,12 +720,12 @@ static struct cbe_iommu * __init cell_iommu_alloc(struct device_node *np) /* Get node ID */ nid = of_node_to_nid(np); if (nid < 0) { - printk(KERN_ERR "iommu: failed to get node for %s\n", - np->full_name); + printk(KERN_ERR "iommu: failed to get node for %pOF\n", + np); return NULL; } - pr_debug("iommu: setting up iommu for node %d (%s)\n", - nid, np->full_name); + pr_debug("iommu: setting up iommu for node %d (%pOF)\n", + nid, np); /* XXX todo: If we can have multiple windows on the same IOMMU, which * isn't the case today, we probably want here to check whether the @@ -736,8 +736,8 @@ static struct cbe_iommu * __init cell_iommu_alloc(struct device_node *np) */ if (cbe_nr_iommus >= NR_IOMMUS) { - printk(KERN_ERR "iommu: too many IOMMUs detected ! (%s)\n", - np->full_name); + printk(KERN_ERR "iommu: too many IOMMUs detected ! (%pOF)\n", + np); return NULL; } diff --git a/arch/powerpc/platforms/cell/ras.c b/arch/powerpc/platforms/cell/ras.c index 460ab392f0e7..2f704afe9af3 100644 --- a/arch/powerpc/platforms/cell/ras.c +++ b/arch/powerpc/platforms/cell/ras.c @@ -196,8 +196,8 @@ static int __init cbe_ptcal_enable(void) for_each_node_by_type(np, "cpu") { const u32 *nid = of_get_property(np, "node-id", NULL); if (!nid) { - printk(KERN_ERR "%s: node %s is missing node-id?\n", - __func__, np->full_name); + printk(KERN_ERR "%s: node %pOF is missing node-id?\n", + __func__, np); continue; } cbe_ptcal_enable_on_node(*nid, order); diff --git a/arch/powerpc/platforms/cell/spider-pci.c b/arch/powerpc/platforms/cell/spider-pci.c index f1f7878893f3..d1e61e273e64 100644 --- a/arch/powerpc/platforms/cell/spider-pci.c +++ b/arch/powerpc/platforms/cell/spider-pci.c @@ -130,8 +130,8 @@ int __init spiderpci_iowa_init(struct iowa_bus *bus, void *data) struct resource r; unsigned long offset = (unsigned long)data; - pr_debug("SPIDERPCI-IOWA:Bus initialize for spider(%s)\n", - np->full_name); + pr_debug("SPIDERPCI-IOWA:Bus initialize for spider(%pOF)\n", + np); priv = kzalloc(sizeof(struct spiderpci_iowa_private), GFP_KERNEL); if (!priv) { diff --git a/arch/powerpc/platforms/cell/spider-pic.c b/arch/powerpc/platforms/cell/spider-pic.c index ff924af00e78..aa44bfc46467 100644 --- a/arch/powerpc/platforms/cell/spider-pic.c +++ b/arch/powerpc/platforms/cell/spider-pic.c @@ -323,8 +323,8 @@ static void __init spider_init_one(struct device_node *of_node, int chip, irq_set_handler_data(virq, pic); irq_set_chained_handler(virq, spider_irq_cascade); - printk(KERN_INFO "spider_pic: node %d, addr: 0x%lx %s\n", - pic->node_id, addr, of_node->full_name); + printk(KERN_INFO "spider_pic: node %d, addr: 0x%lx %pOF\n", + pic->node_id, addr, of_node); /* Enable the interrupt detection enable bit. Do this last! */ out_be32(pic->regs + TIR_DEN, in_be32(pic->regs + TIR_DEN) | 0x1); diff --git a/arch/powerpc/platforms/cell/spu_manage.c b/arch/powerpc/platforms/cell/spu_manage.c index 672d310dcf14..f636ee22b203 100644 --- a/arch/powerpc/platforms/cell/spu_manage.c +++ b/arch/powerpc/platforms/cell/spu_manage.c @@ -191,8 +191,8 @@ static int __init spu_map_interrupts(struct spu *spu, struct device_node *np) goto err; } ret = -EINVAL; - pr_debug(" irq %d no 0x%x on %s\n", i, oirq.args[0], - oirq.np->full_name); + pr_debug(" irq %d no 0x%x on %pOF\n", i, oirq.args[0], + oirq.np); spu->irqs[i] = irq_create_of_mapping(&oirq); if (!spu->irqs[i]) { pr_debug("spu_new: failed to map it !\n"); @@ -243,32 +243,32 @@ static int __init spu_map_device(struct spu *spu) ret = spu_map_resource(spu, 0, (void __iomem**)&spu->local_store, &spu->local_store_phys); if (ret) { - pr_debug("spu_new: failed to map %s resource 0\n", - np->full_name); + pr_debug("spu_new: failed to map %pOF resource 0\n", + np); goto out; } ret = spu_map_resource(spu, 1, (void __iomem**)&spu->problem, &spu->problem_phys); if (ret) { - pr_debug("spu_new: failed to map %s resource 1\n", - np->full_name); + pr_debug("spu_new: failed to map %pOF resource 1\n", + np); goto out_unmap; } ret = spu_map_resource(spu, 2, (void __iomem**)&spu->priv2, NULL); if (ret) { - pr_debug("spu_new: failed to map %s resource 2\n", - np->full_name); + pr_debug("spu_new: failed to map %pOF resource 2\n", + np); goto out_unmap; } if (!firmware_has_feature(FW_FEATURE_LPAR)) ret = spu_map_resource(spu, 3, (void __iomem**)&spu->priv1, NULL); if (ret) { - pr_debug("spu_new: failed to map %s resource 3\n", - np->full_name); + pr_debug("spu_new: failed to map %pOF resource 3\n", + np); goto out_unmap; } - pr_debug("spu_new: %s maps:\n", np->full_name); + pr_debug("spu_new: %pOF maps:\n", np); pr_debug(" local store : 0x%016lx -> 0x%p\n", spu->local_store_phys, spu->local_store); pr_debug(" problem state : 0x%016lx -> 0x%p\n", @@ -316,8 +316,8 @@ static int __init of_create_spu(struct spu *spu, void *data) spu->node = of_node_to_nid(spe); if (spu->node >= MAX_NUMNODES) { - printk(KERN_WARNING "SPE %s on node %d ignored," - " node number too big\n", spe->full_name, spu->node); + printk(KERN_WARNING "SPE %pOF on node %d ignored," + " node number too big\n", spe, spu->node); printk(KERN_WARNING "Check if CONFIG_NUMA is enabled.\n"); ret = -ENODEV; goto out; diff --git a/arch/powerpc/platforms/chrp/pci.c b/arch/powerpc/platforms/chrp/pci.c index 1b87e198faa7..27264794f5c0 100644 --- a/arch/powerpc/platforms/chrp/pci.c +++ b/arch/powerpc/platforms/chrp/pci.c @@ -235,14 +235,14 @@ chrp_find_bridges(void) ++index; /* The GG2 bridge on the LongTrail doesn't have an address */ if (of_address_to_resource(dev, 0, &r) && !is_longtrail) { - printk(KERN_WARNING "Can't use %s: no address\n", - dev->full_name); + printk(KERN_WARNING "Can't use %pOF: no address\n", + dev); continue; } bus_range = of_get_property(dev, "bus-range", &len); if (bus_range == NULL || len < 2 * sizeof(int)) { - printk(KERN_WARNING "Can't get bus-range for %s\n", - dev->full_name); + printk(KERN_WARNING "Can't get bus-range for %pOF\n", + dev); continue; } if (bus_range[1] == bus_range[0]) @@ -250,15 +250,15 @@ chrp_find_bridges(void) else printk(KERN_INFO "PCI buses %d..%d", bus_range[0], bus_range[1]); - printk(" controlled by %s", dev->full_name); + printk(" controlled by %pOF", dev); if (!is_longtrail) printk(" at %llx", (unsigned long long)r.start); printk("\n"); hose = pcibios_alloc_controller(dev); if (!hose) { - printk("Can't allocate PCI controller structure for %s\n", - dev->full_name); + printk("Can't allocate PCI controller structure for %pOF\n", + dev); continue; } hose->first_busno = hose->self_busno = bus_range[0]; @@ -297,8 +297,8 @@ chrp_find_bridges(void) } } } else { - printk("No methods for %s (model %s), using RTAS\n", - dev->full_name, model); + printk("No methods for %pOF (model %s), using RTAS\n", + dev, model); hose->ops = &rtas_pci_ops; } diff --git a/arch/powerpc/platforms/embedded6xx/linkstation.c b/arch/powerpc/platforms/embedded6xx/linkstation.c index f29cf29b11f8..f514d5d28cd4 100644 --- a/arch/powerpc/platforms/embedded6xx/linkstation.c +++ b/arch/powerpc/platforms/embedded6xx/linkstation.c @@ -41,12 +41,12 @@ static int __init linkstation_add_bridge(struct device_node *dev) struct pci_controller *hose; const int *bus_range; - printk("Adding PCI host bridge %s\n", dev->full_name); + printk("Adding PCI host bridge %pOF\n", dev); bus_range = of_get_property(dev, "bus-range", &len); if (bus_range == NULL || len < 2 * sizeof(int)) - printk(KERN_WARNING "Can't get bus-range for %s, assume" - " bus 0\n", dev->full_name); + printk(KERN_WARNING "Can't get bus-range for %pOF, assume" + " bus 0\n", dev); hose = pcibios_alloc_controller(dev); if (hose == NULL) diff --git a/arch/powerpc/platforms/embedded6xx/mvme5100.c b/arch/powerpc/platforms/embedded6xx/mvme5100.c index 8e3590941960..273dfa3f0252 100644 --- a/arch/powerpc/platforms/embedded6xx/mvme5100.c +++ b/arch/powerpc/platforms/embedded6xx/mvme5100.c @@ -115,7 +115,7 @@ static int __init mvme5100_add_bridge(struct device_node *dev) struct pci_controller *hose; unsigned short devid; - pr_info("Adding PCI host bridge %s\n", dev->full_name); + pr_info("Adding PCI host bridge %pOF\n", dev); bus_range = of_get_property(dev, "bus-range", &len); diff --git a/arch/powerpc/platforms/embedded6xx/storcenter.c b/arch/powerpc/platforms/embedded6xx/storcenter.c index 471a50bcd074..ed1914dd34bb 100644 --- a/arch/powerpc/platforms/embedded6xx/storcenter.c +++ b/arch/powerpc/platforms/embedded6xx/storcenter.c @@ -44,7 +44,7 @@ static int __init storcenter_add_bridge(struct device_node *dev) struct pci_controller *hose; const int *bus_range; - printk("Adding PCI host bridge %s\n", dev->full_name); + printk("Adding PCI host bridge %pOF\n", dev); hose = pcibios_alloc_controller(dev); if (hose == NULL) diff --git a/arch/powerpc/platforms/maple/pci.c b/arch/powerpc/platforms/maple/pci.c index 69794d9389c2..e3821379e86f 100644 --- a/arch/powerpc/platforms/maple/pci.c +++ b/arch/powerpc/platforms/maple/pci.c @@ -73,8 +73,8 @@ static void __init fixup_bus_range(struct device_node *bridge) /* Lookup the "bus-range" property for the hose */ prop = of_find_property(bridge, "bus-range", &len); if (prop == NULL || prop->value == NULL || len < 2 * sizeof(int)) { - printk(KERN_WARNING "Can't get bus-range for %s\n", - bridge->full_name); + printk(KERN_WARNING "Can't get bus-range for %pOF\n", + bridge); return; } bus_range = prop->value; @@ -498,12 +498,12 @@ static int __init maple_add_bridge(struct device_node *dev) const int *bus_range; int primary = 1; - DBG("Adding PCI host bridge %s\n", dev->full_name); + DBG("Adding PCI host bridge %pOF\n", dev); bus_range = of_get_property(dev, "bus-range", &len); if (bus_range == NULL || len < 2 * sizeof(int)) { - printk(KERN_WARNING "Can't get bus-range for %s, assume bus 0\n", - dev->full_name); + printk(KERN_WARNING "Can't get bus-range for %pOF, assume bus 0\n", + dev); } hose = pcibios_alloc_controller(dev); diff --git a/arch/powerpc/platforms/pasemi/pci.c b/arch/powerpc/platforms/pasemi/pci.c index 10c4e8fc6ea9..5ff6108f19e9 100644 --- a/arch/powerpc/platforms/pasemi/pci.c +++ b/arch/powerpc/platforms/pasemi/pci.c @@ -193,7 +193,7 @@ static int __init pas_add_bridge(struct device_node *dev) { struct pci_controller *hose; - pr_debug("Adding PCI host bridge %s\n", dev->full_name); + pr_debug("Adding PCI host bridge %pOF\n", dev); hose = pcibios_alloc_controller(dev); if (!hose) diff --git a/arch/powerpc/platforms/powermac/feature.c b/arch/powerpc/platforms/powermac/feature.c index 1e02328c3f2d..9e3f39d36e88 100644 --- a/arch/powerpc/platforms/powermac/feature.c +++ b/arch/powerpc/platforms/powermac/feature.c @@ -2658,25 +2658,25 @@ static void __init probe_one_macio(const char *name, const char *compat, int typ if (i >= MAX_MACIO_CHIPS) { printk(KERN_ERR "pmac_feature: Please increase MAX_MACIO_CHIPS !\n"); - printk(KERN_ERR "pmac_feature: %s skipped\n", node->full_name); + printk(KERN_ERR "pmac_feature: %pOF skipped\n", node); return; } addrp = of_get_pci_address(node, 0, &size, NULL); if (addrp == NULL) { - printk(KERN_ERR "pmac_feature: %s: can't find base !\n", - node->full_name); + printk(KERN_ERR "pmac_feature: %pOF: can't find base !\n", + node); return; } addr = of_translate_address(node, addrp); if (addr == 0) { - printk(KERN_ERR "pmac_feature: %s, can't translate base !\n", - node->full_name); + printk(KERN_ERR "pmac_feature: %pOF, can't translate base !\n", + node); return; } base = ioremap(addr, (unsigned long)size); if (!base) { - printk(KERN_ERR "pmac_feature: %s, can't map mac-io chip !\n", - node->full_name); + printk(KERN_ERR "pmac_feature: %pOF, can't map mac-io chip !\n", + node); return; } if (type == macio_keylargo || type == macio_keylargo2) { diff --git a/arch/powerpc/platforms/powermac/low_i2c.c b/arch/powerpc/platforms/powermac/low_i2c.c index f627c9fd7b48..70183eb3d5c8 100644 --- a/arch/powerpc/platforms/powermac/low_i2c.c +++ b/arch/powerpc/platforms/powermac/low_i2c.c @@ -494,8 +494,8 @@ static struct pmac_i2c_host_kw *__init kw_i2c_host_init(struct device_node *np) host = kzalloc(sizeof(struct pmac_i2c_host_kw), GFP_KERNEL); if (host == NULL) { - printk(KERN_ERR "low_i2c: Can't allocate host for %s\n", - np->full_name); + printk(KERN_ERR "low_i2c: Can't allocate host for %pOF\n", + np); return NULL; } @@ -505,8 +505,8 @@ static struct pmac_i2c_host_kw *__init kw_i2c_host_init(struct device_node *np) */ addrp = of_get_property(np, "AAPL,address", NULL); if (addrp == NULL) { - printk(KERN_ERR "low_i2c: Can't find address for %s\n", - np->full_name); + printk(KERN_ERR "low_i2c: Can't find address for %pOF\n", + np); kfree(host); return NULL; } @@ -538,13 +538,13 @@ static struct pmac_i2c_host_kw *__init kw_i2c_host_init(struct device_node *np) host->irq = irq_of_parse_and_map(np, 0); if (!host->irq) printk(KERN_WARNING - "low_i2c: Failed to map interrupt for %s\n", - np->full_name); + "low_i2c: Failed to map interrupt for %pOF\n", + np); host->base = ioremap((*addrp), 0x1000); if (host->base == NULL) { - printk(KERN_ERR "low_i2c: Can't map registers for %s\n", - np->full_name); + printk(KERN_ERR "low_i2c: Can't map registers for %pOF\n", + np); kfree(host); return NULL; } @@ -560,8 +560,8 @@ static struct pmac_i2c_host_kw *__init kw_i2c_host_init(struct device_node *np) "keywest i2c", host)) host->irq = 0; - printk(KERN_INFO "KeyWest i2c @0x%08x irq %d %s\n", - *addrp, host->irq, np->full_name); + printk(KERN_INFO "KeyWest i2c @0x%08x irq %d %pOF\n", + *addrp, host->irq, np); return host; } @@ -798,7 +798,7 @@ static void __init pmu_i2c_probe(void) if (busnode == NULL) return; - printk(KERN_INFO "PMU i2c %s\n", busnode->full_name); + printk(KERN_INFO "PMU i2c %pOF\n", busnode); /* * We add bus 1 and 2 only for now, bus 0 is "special" @@ -913,7 +913,7 @@ static void __init smu_i2c_probe(void) if (controller == NULL) return; - printk(KERN_INFO "SMU i2c %s\n", controller->full_name); + printk(KERN_INFO "SMU i2c %pOF\n", controller); /* Look for childs, note that they might not be of the right * type as older device trees mix i2c busses and other things @@ -945,8 +945,8 @@ static void __init smu_i2c_probe(void) bus->flags = 0; list_add(&bus->link, &pmac_i2c_busses); - printk(KERN_INFO " channel %x bus %s\n", - bus->channel, busnode->full_name); + printk(KERN_INFO " channel %x bus %pOF\n", + bus->channel, busnode); } } @@ -1129,7 +1129,7 @@ int pmac_i2c_setmode(struct pmac_i2c_bus *bus, int mode) */ if (mode < pmac_i2c_mode_dumb || mode > pmac_i2c_mode_combined) { printk(KERN_ERR "low_i2c: Invalid mode %d requested on" - " bus %s !\n", mode, bus->busnode->full_name); + " bus %pOF !\n", mode, bus->busnode); return -EINVAL; } bus->mode = mode; @@ -1146,8 +1146,8 @@ int pmac_i2c_xfer(struct pmac_i2c_bus *bus, u8 addrdir, int subsize, WARN_ON(!bus->opened); DBG("xfer() chan=%d, addrdir=0x%x, mode=%d, subsize=%d, subaddr=0x%x," - " %d bytes, bus %s\n", bus->channel, addrdir, bus->mode, subsize, - subaddr, len, bus->busnode->full_name); + " %d bytes, bus %pOF\n", bus->channel, addrdir, bus->mode, subsize, + subaddr, len, bus->busnode); rc = bus->xfer(bus, addrdir, subsize, subaddr, data, len); @@ -1241,13 +1241,13 @@ static void* pmac_i2c_do_begin(struct pmf_function *func, struct pmf_args *args) bus = pmac_i2c_find_bus(func->node); if (bus == NULL) { - printk(KERN_ERR "low_i2c: Can't find bus for %s (pfunc)\n", - func->node->full_name); + printk(KERN_ERR "low_i2c: Can't find bus for %pOF (pfunc)\n", + func->node); return NULL; } if (pmac_i2c_open(bus, 0)) { - printk(KERN_ERR "low_i2c: Can't open i2c bus for %s (pfunc)\n", - func->node->full_name); + printk(KERN_ERR "low_i2c: Can't open i2c bus for %pOF (pfunc)\n", + func->node); return NULL; } @@ -1417,7 +1417,7 @@ static struct pmf_handlers pmac_i2c_pfunc_handlers = { static void __init pmac_i2c_dev_create(struct device_node *np, int quirks) { - DBG("dev_create(%s)\n", np->full_name); + DBG("dev_create(%pOF)\n", np); pmf_register_driver(np, &pmac_i2c_pfunc_handlers, (void *)(long)quirks); @@ -1425,20 +1425,20 @@ static void __init pmac_i2c_dev_create(struct device_node *np, int quirks) static void __init pmac_i2c_dev_init(struct device_node *np, int quirks) { - DBG("dev_create(%s)\n", np->full_name); + DBG("dev_create(%pOF)\n", np); pmf_do_functions(np, NULL, 0, PMF_FLAGS_ON_INIT, NULL); } static void pmac_i2c_dev_suspend(struct device_node *np, int quirks) { - DBG("dev_suspend(%s)\n", np->full_name); + DBG("dev_suspend(%pOF)\n", np); pmf_do_functions(np, NULL, 0, PMF_FLAGS_ON_SLEEP, NULL); } static void pmac_i2c_dev_resume(struct device_node *np, int quirks) { - DBG("dev_resume(%s)\n", np->full_name); + DBG("dev_resume(%pOF)\n", np); pmf_do_functions(np, NULL, 0, PMF_FLAGS_ON_WAKE, NULL); } diff --git a/arch/powerpc/platforms/powermac/pci.c b/arch/powerpc/platforms/powermac/pci.c index 6e06c3be2e9a..0b8174a79993 100644 --- a/arch/powerpc/platforms/powermac/pci.c +++ b/arch/powerpc/platforms/powermac/pci.c @@ -783,7 +783,7 @@ static int __init pmac_add_bridge(struct device_node *dev) const int *bus_range; int primary = 1, has_address = 0; - DBG("Adding PCI host bridge %s\n", dev->full_name); + DBG("Adding PCI host bridge %pOF\n", dev); /* Fetch host bridge registers address */ has_address = (of_address_to_resource(dev, 0, &rsrc) == 0); @@ -791,8 +791,8 @@ static int __init pmac_add_bridge(struct device_node *dev) /* Get bus range if any */ bus_range = of_get_property(dev, "bus-range", &len); if (bus_range == NULL || len < 2 * sizeof(int)) { - printk(KERN_WARNING "Can't get bus-range for %s, assume" - " bus 0\n", dev->full_name); + printk(KERN_WARNING "Can't get bus-range for %pOF, assume" + " bus 0\n", dev); } hose = pcibios_alloc_controller(dev); diff --git a/arch/powerpc/platforms/powermac/pfunc_base.c b/arch/powerpc/platforms/powermac/pfunc_base.c index 459138ed4571..860159d46ab8 100644 --- a/arch/powerpc/platforms/powermac/pfunc_base.c +++ b/arch/powerpc/platforms/powermac/pfunc_base.c @@ -54,8 +54,8 @@ static int macio_do_gpio_write(PMF_STD_ARGS, u8 value, u8 mask) raw_spin_lock_irqsave(&feature_lock, flags); tmp = readb(addr); tmp = (tmp & ~mask) | (value & mask); - DBG("Do write 0x%02x to GPIO %s (%p)\n", - tmp, func->node->full_name, addr); + DBG("Do write 0x%02x to GPIO %pOF (%p)\n", + tmp, func->node, addr); writeb(tmp, addr); raw_spin_unlock_irqrestore(&feature_lock, flags); @@ -107,8 +107,8 @@ static void macio_gpio_init_one(struct macio_chip *macio) if (gparent == NULL) return; - DBG("Installing GPIO functions for macio %s\n", - macio->of_node->full_name); + DBG("Installing GPIO functions for macio %pOF\n", + macio->of_node); /* * Ok, got one, we dont need anything special to track them down, so @@ -129,8 +129,8 @@ static void macio_gpio_init_one(struct macio_chip *macio) pmf_register_driver(gp, &macio_gpio_handlers, (void *)offset); } - DBG("Calling initial GPIO functions for macio %s\n", - macio->of_node->full_name); + DBG("Calling initial GPIO functions for macio %pOF\n", + macio->of_node); /* And now we run all the init ones */ for (gp = NULL; (gp = of_get_next_child(gparent, gp)) != NULL;) @@ -267,8 +267,8 @@ static struct pmf_handlers macio_mmio_handlers = { static void macio_mmio_init_one(struct macio_chip *macio) { - DBG("Installing MMIO functions for macio %s\n", - macio->of_node->full_name); + DBG("Installing MMIO functions for macio %pOF\n", + macio->of_node); pmf_register_driver(macio->of_node, &macio_mmio_handlers, macio); } @@ -298,8 +298,8 @@ static void uninorth_install_pfunc(void) { struct device_node *np; - DBG("Installing functions for UniN %s\n", - uninorth_node->full_name); + DBG("Installing functions for UniN %pOF\n", + uninorth_node); /* * Install handlers for the bridge itself @@ -317,8 +317,8 @@ static void uninorth_install_pfunc(void) break; } if (unin_hwclock) { - DBG("Installing functions for UniN clock %s\n", - unin_hwclock->full_name); + DBG("Installing functions for UniN clock %pOF\n", + unin_hwclock); pmf_register_driver(unin_hwclock, &unin_mmio_handlers, NULL); pmf_do_functions(unin_hwclock, NULL, 0, PMF_FLAGS_ON_INIT, NULL); diff --git a/arch/powerpc/platforms/powermac/pfunc_core.c b/arch/powerpc/platforms/powermac/pfunc_core.c index 695e8c4d4224..df3c93bef228 100644 --- a/arch/powerpc/platforms/powermac/pfunc_core.c +++ b/arch/powerpc/platforms/powermac/pfunc_core.c @@ -708,7 +708,7 @@ int pmf_register_driver(struct device_node *np, if (handlers == NULL) return -EINVAL; - DBG("pmf: registering driver for node %s\n", np->full_name); + DBG("pmf: registering driver for node %pOF\n", np); spin_lock_irqsave(&pmf_lock, flags); dev = pmf_find_device(np); @@ -781,7 +781,7 @@ void pmf_unregister_driver(struct device_node *np) struct pmf_device *dev; unsigned long flags; - DBG("pmf: unregistering driver for node %s\n", np->full_name); + DBG("pmf: unregistering driver for node %pOF\n", np); spin_lock_irqsave(&pmf_lock, flags); dev = pmf_find_device(np); @@ -940,7 +940,7 @@ int pmf_call_one(struct pmf_function *func, struct pmf_args *args) void *instdata = NULL; int rc = 0; - DBG(" ** pmf_call_one(%s/%s) **\n", dev->node->full_name, func->name); + DBG(" ** pmf_call_one(%pOF/%s) **\n", dev->node, func->name); if (dev->handlers->begin) instdata = dev->handlers->begin(func, args); diff --git a/arch/powerpc/platforms/powermac/pic.c b/arch/powerpc/platforms/powermac/pic.c index f5f9ad7c3398..5e0719b27294 100644 --- a/arch/powerpc/platforms/powermac/pic.c +++ b/arch/powerpc/platforms/powermac/pic.c @@ -364,8 +364,8 @@ static void __init pmac_pic_probe_oldstyle(void) (addr + 0x10); of_node_put(master); - printk(KERN_INFO "irq: Found primary Apple PIC %s for %d irqs\n", - master->full_name, max_real_irqs); + printk(KERN_INFO "irq: Found primary Apple PIC %pOF for %d irqs\n", + master, max_real_irqs); /* Map interrupts of cascaded controller */ if (slave && !of_address_to_resource(slave, 0, &r)) { @@ -378,8 +378,8 @@ static void __init pmac_pic_probe_oldstyle(void) (addr + 0x10); pmac_irq_cascade = irq_of_parse_and_map(slave, 0); - printk(KERN_INFO "irq: Found slave Apple PIC %s for %d irqs" - " cascade: %d\n", slave->full_name, + printk(KERN_INFO "irq: Found slave Apple PIC %pOF for %d irqs" + " cascade: %d\n", slave, max_irqs - max_real_irqs, pmac_irq_cascade); } of_node_put(slave); diff --git a/arch/powerpc/platforms/powermac/setup.c b/arch/powerpc/platforms/powermac/setup.c index 6b4e9d181126..ab668cb72263 100644 --- a/arch/powerpc/platforms/powermac/setup.c +++ b/arch/powerpc/platforms/powermac/setup.c @@ -556,7 +556,7 @@ static int __init check_pmac_serial_console(void) pr_debug(" can't find stdout package %s !\n", name); return -ENODEV; } - pr_debug("stdout is %s\n", prom_stdout->full_name); + pr_debug("stdout is %pOF\n", prom_stdout); name = of_get_property(prom_stdout, "name", NULL); if (!name) { diff --git a/arch/powerpc/platforms/powernv/opal-async.c b/arch/powerpc/platforms/powernv/opal-async.c index 83bebeec0fea..cf33769a7b72 100644 --- a/arch/powerpc/platforms/powernv/opal-async.c +++ b/arch/powerpc/platforms/powernv/opal-async.c @@ -171,8 +171,8 @@ int __init opal_async_comp_init(void) async = of_get_property(opal_node, "opal-msg-async-num", NULL); if (!async) { - pr_err("%s: %s has no opal-msg-async-num\n", - __func__, opal_node->full_name); + pr_err("%s: %pOF has no opal-msg-async-num\n", + __func__, opal_node); err = -ENOENT; goto out_opal_node; } diff --git a/arch/powerpc/platforms/powernv/opal-xscom.c b/arch/powerpc/platforms/powernv/opal-xscom.c index 28651fb25417..81c0a943dea9 100644 --- a/arch/powerpc/platforms/powernv/opal-xscom.c +++ b/arch/powerpc/platforms/powernv/opal-xscom.c @@ -36,14 +36,14 @@ static scom_map_t opal_scom_map(struct device_node *dev, u64 reg, u64 count) const __be32 *gcid; if (!of_get_property(dev, "scom-controller", NULL)) { - pr_err("%s: device %s is not a SCOM controller\n", - __func__, dev->full_name); + pr_err("%s: device %pOF is not a SCOM controller\n", + __func__, dev); return SCOM_MAP_INVALID; } gcid = of_get_property(dev, "ibm,chip-id", NULL); if (!gcid) { - pr_err("%s: device %s has no ibm,chip-id\n", - __func__, dev->full_name); + pr_err("%s: device %pOF has no ibm,chip-id\n", + __func__, dev); return SCOM_MAP_INVALID; } m = kmalloc(sizeof(struct opal_scom_map), GFP_KERNEL); diff --git a/arch/powerpc/platforms/powernv/pci-ioda.c b/arch/powerpc/platforms/powernv/pci-ioda.c index 48de308224d6..57f9e55f4352 100644 --- a/arch/powerpc/platforms/powernv/pci-ioda.c +++ b/arch/powerpc/platforms/powernv/pci-ioda.c @@ -444,8 +444,8 @@ static void __init pnv_ioda_parse_m64_window(struct pnv_phb *phb) r = of_get_property(dn, "ibm,opal-m64-window", NULL); if (!r) { - pr_info(" No <ibm,opal-m64-window> on %s\n", - dn->full_name); + pr_info(" No <ibm,opal-m64-window> on %pOF\n", + dn); return; } @@ -3796,8 +3796,7 @@ static void __init pnv_pci_init_ioda_phb(struct device_node *np, if (!of_device_is_available(np)) return; - pr_info("Initializing %s PHB (%s)\n", - pnv_phb_names[ioda_type], of_node_full_name(np)); + pr_info("Initializing %s PHB (%pOF)\n", pnv_phb_names[ioda_type], np); prop64 = of_get_property(np, "ibm,opal-phbid", NULL); if (!prop64) { @@ -3812,8 +3811,8 @@ static void __init pnv_pci_init_ioda_phb(struct device_node *np, /* Allocate PCI controller */ phb->hose = hose = pcibios_alloc_controller(np); if (!phb->hose) { - pr_err(" Can't allocate PCI controller for %s\n", - np->full_name); + pr_err(" Can't allocate PCI controller for %pOF\n", + np); memblock_free(__pa(phb), sizeof(struct pnv_phb)); return; } @@ -3824,7 +3823,7 @@ static void __init pnv_pci_init_ioda_phb(struct device_node *np, hose->first_busno = be32_to_cpu(prop32[0]); hose->last_busno = be32_to_cpu(prop32[1]); } else { - pr_warn(" Broken <bus-range> on %s\n", np->full_name); + pr_warn(" Broken <bus-range> on %pOF\n", np); hose->first_busno = 0; hose->last_busno = 0xff; } @@ -4045,7 +4044,7 @@ void __init pnv_pci_init_ioda_hub(struct device_node *np) const __be64 *prop64; u64 hub_id; - pr_info("Probing IODA IO-Hub %s\n", np->full_name); + pr_info("Probing IODA IO-Hub %pOF\n", np); prop64 = of_get_property(np, "ibm,opal-hubid", NULL); if (!prop64) { diff --git a/arch/powerpc/platforms/powernv/rng.c b/arch/powerpc/platforms/powernv/rng.c index c5ce3a8bd4c9..718f50ed22f1 100644 --- a/arch/powerpc/platforms/powernv/rng.c +++ b/arch/powerpc/platforms/powernv/rng.c @@ -125,7 +125,7 @@ static __init void rng_init_per_cpu(struct powernv_rng *rng, chip_id = of_get_ibm_chip_id(dn); if (chip_id == -1) - pr_warn("No ibm,chip-id found for %s.\n", dn->full_name); + pr_warn("No ibm,chip-id found for %pOF.\n", dn); for_each_possible_cpu(cpu) { if (per_cpu(powernv_rng, cpu) == NULL || @@ -178,8 +178,8 @@ static __init int rng_init(void) for_each_compatible_node(dn, NULL, "ibm,power-rng") { rc = rng_create(dn); if (rc) { - pr_err("Failed creating rng for %s (%d).\n", - dn->full_name, rc); + pr_err("Failed creating rng for %pOF (%d).\n", + dn, rc); continue; } diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c index 39187696ee74..80b84c9c8509 100644 --- a/arch/powerpc/platforms/pseries/dlpar.c +++ b/arch/powerpc/platforms/pseries/dlpar.c @@ -264,8 +264,7 @@ int dlpar_attach_node(struct device_node *dn) rc = of_attach_node(dn); if (rc) { - printk(KERN_ERR "Failed to add device node %s\n", - dn->full_name); + printk(KERN_ERR "Failed to add device node %pOF\n", dn); return rc; } diff --git a/arch/powerpc/platforms/pseries/event_sources.c b/arch/powerpc/platforms/pseries/event_sources.c index 32187dc76730..6eeb0d4bab61 100644 --- a/arch/powerpc/platforms/pseries/event_sources.c +++ b/arch/powerpc/platforms/pseries/event_sources.c @@ -36,8 +36,8 @@ void request_event_sources_irqs(struct device_node *np, virqs[count] = irq_create_of_mapping(&oirq); if (!virqs[count]) { pr_err("event-sources: Unable to allocate " - "interrupt number for %s\n", - np->full_name); + "interrupt number for %pOF\n", + np); WARN_ON(1); } else { count++; @@ -48,7 +48,7 @@ void request_event_sources_irqs(struct device_node *np, for (i = 0; i < count; i++) { if (request_irq(virqs[i], handler, 0, name, NULL)) { pr_err("event-sources: Unable to request interrupt " - "%d for %s\n", virqs[i], np->full_name); + "%d for %pOF\n", virqs[i], np); WARN_ON(1); return; } diff --git a/arch/powerpc/platforms/pseries/hotplug-cpu.c b/arch/powerpc/platforms/pseries/hotplug-cpu.c index 6afd1efd3633..0a93093fbcef 100644 --- a/arch/powerpc/platforms/pseries/hotplug-cpu.c +++ b/arch/powerpc/platforms/pseries/hotplug-cpu.c @@ -264,8 +264,8 @@ static int pseries_add_processor(struct device_node *np) /* If we get here, it most likely means that NR_CPUS is * less than the partition's max processors setting. */ - printk(KERN_ERR "Cannot add cpu %s; this system configuration" - " supports %d logical cpus.\n", np->full_name, + printk(KERN_ERR "Cannot add cpu %pOF; this system configuration" + " supports %d logical cpus.\n", np, num_possible_cpus()); goto out_unlock; } diff --git a/arch/powerpc/platforms/pseries/ibmebus.c b/arch/powerpc/platforms/pseries/ibmebus.c index 52146b1356d2..408a86044133 100644 --- a/arch/powerpc/platforms/pseries/ibmebus.c +++ b/arch/powerpc/platforms/pseries/ibmebus.c @@ -150,8 +150,7 @@ static const struct dma_map_ops ibmebus_dma_ops = { static int ibmebus_match_path(struct device *dev, void *data) { struct device_node *dn = to_platform_device(dev)->dev.of_node; - return (dn->full_name && - (strcasecmp((char *)data, dn->full_name) == 0)); + return (of_find_node_by_path(data) == dn); } static int ibmebus_match_node(struct device *dev, void *data) @@ -395,7 +394,7 @@ static ssize_t devspec_show(struct device *dev, struct platform_device *ofdev; ofdev = to_platform_device(dev); - return sprintf(buf, "%s\n", ofdev->dev.of_node->full_name); + return sprintf(buf, "%pOF\n", ofdev->dev.of_node); } static DEVICE_ATTR_RO(devspec); diff --git a/arch/powerpc/platforms/pseries/iommu.c b/arch/powerpc/platforms/pseries/iommu.c index 8374adee27e3..7c181467d0ad 100644 --- a/arch/powerpc/platforms/pseries/iommu.c +++ b/arch/powerpc/platforms/pseries/iommu.c @@ -511,8 +511,8 @@ static void iommu_table_setparms(struct pci_controller *phb, basep = of_get_property(node, "linux,tce-base", NULL); sizep = of_get_property(node, "linux,tce-size", NULL); if (basep == NULL || sizep == NULL) { - printk(KERN_ERR "PCI_DMA: iommu_table_setparms: %s has " - "missing tce entries !\n", dn->full_name); + printk(KERN_ERR "PCI_DMA: iommu_table_setparms: %pOF has " + "missing tce entries !\n", dn); return; } @@ -587,7 +587,7 @@ static void pci_dma_bus_setup_pSeries(struct pci_bus *bus) dn = pci_bus_to_OF_node(bus); - pr_debug("pci_dma_bus_setup_pSeries: setting up bus %s\n", dn->full_name); + pr_debug("pci_dma_bus_setup_pSeries: setting up bus %pOF\n", dn); if (bus->self) { /* This is not a root bus, any setup will be done for the @@ -701,8 +701,8 @@ static void pci_dma_bus_setup_pSeriesLP(struct pci_bus *bus) dn = pci_bus_to_OF_node(bus); - pr_debug("pci_dma_bus_setup_pSeriesLP: setting up bus %s\n", - dn->full_name); + pr_debug("pci_dma_bus_setup_pSeriesLP: setting up bus %pOF\n", + dn); /* Find nearest ibm,dma-window, walking up the device tree */ for (pdn = dn; pdn != NULL; pdn = pdn->parent) { @@ -718,8 +718,8 @@ static void pci_dma_bus_setup_pSeriesLP(struct pci_bus *bus) ppci = PCI_DN(pdn); - pr_debug(" parent is %s, iommu_table: 0x%p\n", - pdn->full_name, ppci->table_group); + pr_debug(" parent is %pOF, iommu_table: 0x%p\n", + pdn, ppci->table_group); if (!ppci->table_group) { ppci->table_group = iommu_pseries_alloc_group(ppci->phb->node); @@ -817,28 +817,28 @@ static void remove_ddw(struct device_node *np, bool remove_prop) ret = tce_clearrange_multi_pSeriesLP(0, 1ULL << (be32_to_cpu(dwp->window_shift) - PAGE_SHIFT), dwp); if (ret) - pr_warning("%s failed to clear tces in window.\n", - np->full_name); + pr_warning("%pOF failed to clear tces in window.\n", + np); else - pr_debug("%s successfully cleared tces in window.\n", - np->full_name); + pr_debug("%pOF successfully cleared tces in window.\n", + np); ret = rtas_call(ddw_avail[2], 1, 1, NULL, liobn); if (ret) - pr_warning("%s: failed to remove direct window: rtas returned " + pr_warning("%pOF: failed to remove direct window: rtas returned " "%d to ibm,remove-pe-dma-window(%x) %llx\n", - np->full_name, ret, ddw_avail[2], liobn); + np, ret, ddw_avail[2], liobn); else - pr_debug("%s: successfully removed direct window: rtas returned " + pr_debug("%pOF: successfully removed direct window: rtas returned " "%d to ibm,remove-pe-dma-window(%x) %llx\n", - np->full_name, ret, ddw_avail[2], liobn); + np, ret, ddw_avail[2], liobn); delprop: if (remove_prop) ret = of_remove_property(np, win64); if (ret) - pr_warning("%s: failed to remove direct window property: %d\n", - np->full_name, ret); + pr_warning("%pOF: failed to remove direct window property: %d\n", + np, ret); } static u64 find_existing_ddw(struct device_node *pdn) @@ -1004,7 +1004,7 @@ static u64 enable_ddw(struct pci_dev *dev, struct device_node *pdn) * list. */ list_for_each_entry(fpdn, &failed_ddw_pdn_list, list) { - if (!strcmp(fpdn->pdn->full_name, pdn->full_name)) + if (fpdn->pdn == pdn) goto out_unlock; } @@ -1087,8 +1087,8 @@ static u64 enable_ddw(struct pci_dev *dev, struct device_node *pdn) ddwprop->tce_shift = cpu_to_be32(page_shift); ddwprop->window_shift = cpu_to_be32(len); - dev_dbg(&dev->dev, "created tce table LIOBN 0x%x for %s\n", - create.liobn, dn->full_name); + dev_dbg(&dev->dev, "created tce table LIOBN 0x%x for %pOF\n", + create.liobn, dn); window = kzalloc(sizeof(*window), GFP_KERNEL); if (!window) @@ -1097,15 +1097,15 @@ static u64 enable_ddw(struct pci_dev *dev, struct device_node *pdn) ret = walk_system_ram_range(0, memblock_end_of_DRAM() >> PAGE_SHIFT, win64->value, tce_setrange_multi_pSeriesLP_walk); if (ret) { - dev_info(&dev->dev, "failed to map direct window for %s: %d\n", - dn->full_name, ret); + dev_info(&dev->dev, "failed to map direct window for %pOF: %d\n", + dn, ret); goto out_free_window; } ret = of_add_property(pdn, win64); if (ret) { - dev_err(&dev->dev, "unable to add dma window property for %s: %d", - pdn->full_name, ret); + dev_err(&dev->dev, "unable to add dma window property for %pOF: %d", + pdn, ret); goto out_free_window; } @@ -1158,7 +1158,7 @@ static void pci_dma_dev_setup_pSeriesLP(struct pci_dev *dev) * already allocated. */ dn = pci_device_to_OF_node(dev); - pr_debug(" node is %s\n", dn->full_name); + pr_debug(" node is %pOF\n", dn); for (pdn = dn; pdn && PCI_DN(pdn) && !PCI_DN(pdn)->table_group; pdn = pdn->parent) { @@ -1169,11 +1169,11 @@ static void pci_dma_dev_setup_pSeriesLP(struct pci_dev *dev) if (!pdn || !PCI_DN(pdn)) { printk(KERN_WARNING "pci_dma_dev_setup_pSeriesLP: " - "no DMA window found for pci dev=%s dn=%s\n", - pci_name(dev), of_node_full_name(dn)); + "no DMA window found for pci dev=%s dn=%pOF\n", + pci_name(dev), dn); return; } - pr_debug(" parent is %s\n", pdn->full_name); + pr_debug(" parent is %pOF\n", pdn); pci = PCI_DN(pdn); if (!pci->table_group) { @@ -1213,7 +1213,7 @@ static int dma_set_mask_pSeriesLP(struct device *dev, u64 dma_mask) /* only attempt to use a new window if 64-bit DMA is requested */ if (!disable_ddw && dma_mask == DMA_BIT_MASK(64)) { dn = pci_device_to_OF_node(pdev); - dev_dbg(dev, "node is %s\n", dn->full_name); + dev_dbg(dev, "node is %pOF\n", dn); /* * the device tree might contain the dma-window properties diff --git a/arch/powerpc/platforms/pseries/msi.c b/arch/powerpc/platforms/pseries/msi.c index 326ef0dd6038..353912e700eb 100644 --- a/arch/powerpc/platforms/pseries/msi.c +++ b/arch/powerpc/platforms/pseries/msi.c @@ -144,7 +144,7 @@ static int check_req(struct pci_dev *pdev, int nvec, char *prop_name) p = of_get_property(dn, prop_name, NULL); if (!p) { - pr_debug("rtas_msi: No %s on %s\n", prop_name, dn->full_name); + pr_debug("rtas_msi: No %s on %pOF\n", prop_name, dn); return -ENOENT; } @@ -182,8 +182,8 @@ static struct device_node *find_pe_total_msi(struct pci_dev *dev, int *total) while (dn) { p = of_get_property(dn, "ibm,pe-total-#msi", NULL); if (p) { - pr_debug("rtas_msi: found prop on dn %s\n", - dn->full_name); + pr_debug("rtas_msi: found prop on dn %pOF\n", + dn); *total = be32_to_cpup(p); return dn; } @@ -222,7 +222,7 @@ static struct device_node *find_pe_dn(struct pci_dev *dev, int *total) /* Hardcode of 8 for old firmwares */ *total = 8; - pr_debug("rtas_msi: using PE dn %s\n", dn->full_name); + pr_debug("rtas_msi: using PE dn %pOF\n", dn); return dn; } @@ -242,7 +242,7 @@ static void *count_non_bridge_devices(struct device_node *dn, void *data) const __be32 *p; u32 class; - pr_debug("rtas_msi: counting %s\n", dn->full_name); + pr_debug("rtas_msi: counting %pOF\n", dn); p = of_get_property(dn, "class-code", NULL); class = p ? be32_to_cpup(p) : 0; @@ -300,7 +300,7 @@ static int msi_quota_for_device(struct pci_dev *dev, int request) goto out; } - pr_debug("rtas_msi: found PE %s\n", pe_dn->full_name); + pr_debug("rtas_msi: found PE %pOF\n", pe_dn); memset(&counts, 0, sizeof(struct msi_counts)); diff --git a/arch/powerpc/platforms/pseries/pci_dlpar.c b/arch/powerpc/platforms/pseries/pci_dlpar.c index 547fd13e4f8e..561917fa54a8 100644 --- a/arch/powerpc/platforms/pseries/pci_dlpar.c +++ b/arch/powerpc/platforms/pseries/pci_dlpar.c @@ -38,7 +38,7 @@ struct pci_controller *init_phb_dynamic(struct device_node *dn) { struct pci_controller *phb; - pr_debug("PCI: Initializing new hotplug PHB %s\n", dn->full_name); + pr_debug("PCI: Initializing new hotplug PHB %pOF\n", dn); phb = pcibios_alloc_controller(dn); if (!phb) diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c index 8a47f168476b..aa5ca74316fa 100644 --- a/arch/powerpc/platforms/pseries/vio.c +++ b/arch/powerpc/platforms/pseries/vio.c @@ -1362,9 +1362,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) else if (!strcmp(parent_node->full_name, "/vdevice")) family = VDEVICE; else { - pr_warn("%s: parent(%s) of %s not recognized.\n", + pr_warn("%s: parent(%pOF) of %s not recognized.\n", __func__, - parent_node->full_name, + parent_node, of_node_name); of_node_put(parent_node); return NULL; @@ -1555,7 +1555,7 @@ static ssize_t devspec_show(struct device *dev, { struct device_node *of_node = dev->of_node; - return sprintf(buf, "%s\n", of_node_full_name(of_node)); + return sprintf(buf, "%pOF\n", of_node); } static DEVICE_ATTR_RO(devspec); diff --git a/arch/powerpc/sysdev/axonram.c b/arch/powerpc/sysdev/axonram.c index 2799706106c6..072296eba3e7 100644 --- a/arch/powerpc/sysdev/axonram.c +++ b/arch/powerpc/sysdev/axonram.c @@ -188,8 +188,8 @@ static int axon_ram_probe(struct platform_device *device) axon_ram_bank_id++; - dev_info(&device->dev, "Found memory controller on %s\n", - device->dev.of_node->full_name); + dev_info(&device->dev, "Found memory controller on %pOF\n", + device->dev.of_node); bank = kzalloc(sizeof(struct axon_ram_bank), GFP_KERNEL); if (bank == NULL) { diff --git a/arch/powerpc/sysdev/dcr.c b/arch/powerpc/sysdev/dcr.c index 121e26fffd50..d72eda568b7d 100644 --- a/arch/powerpc/sysdev/dcr.c +++ b/arch/powerpc/sysdev/dcr.c @@ -195,8 +195,8 @@ dcr_host_mmio_t dcr_map_mmio(struct device_node *dev, dcr_host_mmio_t ret = { .token = NULL, .stride = 0, .base = dcr_n }; u64 addr; - pr_debug("dcr_map(%s, 0x%x, 0x%x)\n", - dev->full_name, dcr_n, dcr_c); + pr_debug("dcr_map(%pOF, 0x%x, 0x%x)\n", + dev, dcr_n, dcr_c); addr = of_translate_dcr_address(dev, dcr_n, &ret.stride); pr_debug("translates to addr: 0x%llx, stride: 0x%x\n", diff --git a/arch/powerpc/sysdev/fsl_85xx_cache_sram.c b/arch/powerpc/sysdev/fsl_85xx_cache_sram.c index 37a69097e022..00ccf3e4fcb4 100644 --- a/arch/powerpc/sysdev/fsl_85xx_cache_sram.c +++ b/arch/powerpc/sysdev/fsl_85xx_cache_sram.c @@ -101,8 +101,8 @@ int __init instantiate_cache_sram(struct platform_device *dev, if (!request_mem_region(cache_sram->base_phys, cache_sram->size, "fsl_85xx_cache_sram")) { - dev_err(&dev->dev, "%s: request memory failed\n", - dev->dev.of_node->full_name); + dev_err(&dev->dev, "%pOF: request memory failed\n", + dev->dev.of_node); ret = -ENXIO; goto out_free; } @@ -110,16 +110,16 @@ int __init instantiate_cache_sram(struct platform_device *dev, cache_sram->base_virt = ioremap_prot(cache_sram->base_phys, cache_sram->size, _PAGE_COHERENT | PAGE_KERNEL); if (!cache_sram->base_virt) { - dev_err(&dev->dev, "%s: ioremap_prot failed\n", - dev->dev.of_node->full_name); + dev_err(&dev->dev, "%pOF: ioremap_prot failed\n", + dev->dev.of_node); ret = -ENOMEM; goto out_release; } cache_sram->rh = rh_create(sizeof(unsigned int)); if (IS_ERR(cache_sram->rh)) { - dev_err(&dev->dev, "%s: Unable to create remote heap\n", - dev->dev.of_node->full_name); + dev_err(&dev->dev, "%pOF: Unable to create remote heap\n", + dev->dev.of_node); ret = PTR_ERR(cache_sram->rh); goto out_unmap; } diff --git a/arch/powerpc/sysdev/fsl_gtm.c b/arch/powerpc/sysdev/fsl_gtm.c index a6f0b96ce2c9..d902306f4718 100644 --- a/arch/powerpc/sysdev/fsl_gtm.c +++ b/arch/powerpc/sysdev/fsl_gtm.c @@ -388,8 +388,8 @@ static int __init fsl_gtm_init(void) gtm = kzalloc(sizeof(*gtm), GFP_KERNEL); if (!gtm) { - pr_err("%s: unable to allocate memory\n", - np->full_name); + pr_err("%pOF: unable to allocate memory\n", + np); continue; } @@ -397,7 +397,7 @@ static int __init fsl_gtm_init(void) clock = of_get_property(np, "clock-frequency", &size); if (!clock || size != sizeof(*clock)) { - pr_err("%s: no clock-frequency\n", np->full_name); + pr_err("%pOF: no clock-frequency\n", np); goto err; } gtm->clock = *clock; @@ -407,8 +407,8 @@ static int __init fsl_gtm_init(void) irq = irq_of_parse_and_map(np, i); if (!irq) { - pr_err("%s: not enough interrupts specified\n", - np->full_name); + pr_err("%pOF: not enough interrupts specified\n", + np); goto err; } gtm->timers[i].irq = irq; @@ -417,8 +417,8 @@ static int __init fsl_gtm_init(void) gtm->regs = of_iomap(np, 0); if (!gtm->regs) { - pr_err("%s: unable to iomap registers\n", - np->full_name); + pr_err("%pOF: unable to iomap registers\n", + np); goto err; } diff --git a/arch/powerpc/sysdev/fsl_msi.c b/arch/powerpc/sysdev/fsl_msi.c index 8a244828782e..44cbf4c12ea1 100644 --- a/arch/powerpc/sysdev/fsl_msi.c +++ b/arch/powerpc/sysdev/fsl_msi.c @@ -214,8 +214,8 @@ static int fsl_setup_msi_irqs(struct pci_dev *pdev, int nvec, int type) phandle = np->phandle; else { dev_err(&pdev->dev, - "node %s has an invalid fsl,msi phandle %u\n", - hose->dn->full_name, np->phandle); + "node %pOF has an invalid fsl,msi phandle %u\n", + hose->dn, np->phandle); return -EINVAL; } } @@ -438,16 +438,16 @@ static int fsl_of_msi_probe(struct platform_device *dev) if ((features->fsl_pic_ip & FSL_PIC_IP_MASK) != FSL_PIC_IP_VMPIC) { err = of_address_to_resource(dev->dev.of_node, 0, &res); if (err) { - dev_err(&dev->dev, "invalid resource for node %s\n", - dev->dev.of_node->full_name); + dev_err(&dev->dev, "invalid resource for node %pOF\n", + dev->dev.of_node); goto error_out; } msi->msi_regs = ioremap(res.start, resource_size(&res)); if (!msi->msi_regs) { err = -ENOMEM; - dev_err(&dev->dev, "could not map node %s\n", - dev->dev.of_node->full_name); + dev_err(&dev->dev, "could not map node %pOF\n", + dev->dev.of_node); goto error_out; } msi->msiir_offset = @@ -522,8 +522,8 @@ static int fsl_of_msi_probe(struct platform_device *dev) for (irq_index = 0, i = 0; i < len / (2 * sizeof(u32)); i++) { if (p[i * 2] % IRQS_PER_MSI_REG || p[i * 2 + 1] % IRQS_PER_MSI_REG) { - pr_warn("%s: %s: msi available range of %u at %u is not IRQ-aligned\n", - __func__, dev->dev.of_node->full_name, + pr_warn("%s: %pOF: msi available range of %u at %u is not IRQ-aligned\n", + __func__, dev->dev.of_node, p[i * 2 + 1], p[i * 2]); err = -EINVAL; goto error_out; diff --git a/arch/powerpc/sysdev/fsl_pci.c b/arch/powerpc/sysdev/fsl_pci.c index d3a597456b6e..22d98057f773 100644 --- a/arch/powerpc/sysdev/fsl_pci.c +++ b/arch/powerpc/sysdev/fsl_pci.c @@ -202,7 +202,6 @@ static void setup_pci_atmu(struct pci_controller *hose) u32 pcicsrbar = 0, pcicsrbar_sz; u32 piwar = PIWAR_EN | PIWAR_PF | PIWAR_TGI_LOCAL | PIWAR_READ_SNOOP | PIWAR_WRITE_SNOOP; - const char *name = hose->dn->full_name; const u64 *reg; int len; bool setup_inbound; @@ -290,12 +289,12 @@ static void setup_pci_atmu(struct pci_controller *hose) paddr_lo -= offset; if (paddr_hi == paddr_lo) { - pr_err("%s: No outbound window space\n", name); + pr_err("%pOF: No outbound window space\n", hose->dn); return; } if (paddr_lo == 0) { - pr_err("%s: No space for inbound window\n", name); + pr_err("%pOF: No space for inbound window\n", hose->dn); return; } @@ -313,7 +312,7 @@ static void setup_pci_atmu(struct pci_controller *hose) paddr_lo = min(paddr_lo, (u64)pcicsrbar); - pr_info("%s: PCICSRBAR @ 0x%x\n", name, pcicsrbar); + pr_info("%pOF: PCICSRBAR @ 0x%x\n", hose->dn, pcicsrbar); /* Setup inbound mem window */ mem = memblock_end_of_DRAM(); @@ -336,12 +335,12 @@ static void setup_pci_atmu(struct pci_controller *hose) u64 address = be64_to_cpup(reg); if ((address >= mem) && (address < (mem + PAGE_SIZE))) { - pr_info("%s: extending DDR ATMU to cover MSIIR", name); + pr_info("%pOF: extending DDR ATMU to cover MSIIR", hose->dn); mem += PAGE_SIZE; } else { /* TODO: Create a new ATMU for MSIIR */ - pr_warn("%s: msi-address-64 address of %llx is " - "unsupported\n", name, address); + pr_warn("%pOF: msi-address-64 address of %llx is " + "unsupported\n", hose->dn, address); } } @@ -354,8 +353,8 @@ static void setup_pci_atmu(struct pci_controller *hose) if ((1ull << mem_log) != mem) { mem_log++; if ((1ull << mem_log) > mem) - pr_info("%s: Setting PCI inbound window " - "greater than memory size\n", name); + pr_info("%pOF: Setting PCI inbound window " + "greater than memory size\n", hose->dn); } piwar |= ((mem_log - 1) & PIWAR_SZ_MASK); @@ -402,7 +401,7 @@ static void setup_pci_atmu(struct pci_controller *hose) */ ppc_md.dma_set_mask = fsl_pci_dma_set_mask; - pr_info("%s: Setup 64-bit PCI DMA window\n", name); + pr_info("%pOF: Setup 64-bit PCI DMA window\n", hose->dn); } } else { u64 paddr = 0; @@ -443,18 +442,18 @@ static void setup_pci_atmu(struct pci_controller *hose) #ifdef CONFIG_SWIOTLB ppc_swiotlb_enable = 1; #else - pr_err("%s: ERROR: Memory size exceeds PCI ATMU ability to " + pr_err("%pOF: ERROR: Memory size exceeds PCI ATMU ability to " "map - enable CONFIG_SWIOTLB to avoid dma errors.\n", - name); + hose->dn); #endif /* adjusting outbound windows could reclaim space in mem map */ if (paddr_hi < 0xffffffffull) - pr_warning("%s: WARNING: Outbound window cfg leaves " + pr_warning("%pOF: WARNING: Outbound window cfg leaves " "gaps in memory map. Adjusting the memory map " "could reduce unnecessary bounce buffering.\n", - name); + hose->dn); - pr_info("%s: DMA window size is 0x%llx\n", name, + pr_info("%pOF: DMA window size is 0x%llx\n", hose->dn, (u64)hose->dma_window_size); } } @@ -532,11 +531,11 @@ int fsl_add_bridge(struct platform_device *pdev, int is_primary) dev = pdev->dev.of_node; if (!of_device_is_available(dev)) { - pr_warning("%s: disabled\n", dev->full_name); + pr_warning("%pOF: disabled\n", dev); return -ENODEV; } - pr_debug("Adding PCI host bridge %s\n", dev->full_name); + pr_debug("Adding PCI host bridge %pOF\n", dev); /* Fetch host bridge registers address */ if (of_address_to_resource(dev, 0, &rsrc)) { @@ -547,8 +546,8 @@ int fsl_add_bridge(struct platform_device *pdev, int is_primary) /* Get bus range if any */ bus_range = of_get_property(dev, "bus-range", &len); if (bus_range == NULL || len < 2 * sizeof(int)) - printk(KERN_WARNING "Can't get bus-range for %s, assume" - " bus 0\n", dev->full_name); + printk(KERN_WARNING "Can't get bus-range for %pOF, assume" + " bus 0\n", dev); pci_add_flags(PCI_REASSIGN_ALL_BUS); hose = pcibios_alloc_controller(dev); @@ -809,11 +808,11 @@ int __init mpc83xx_add_bridge(struct device_node *dev) is_mpc83xx_pci = 1; if (!of_device_is_available(dev)) { - pr_warning("%s: disabled by the firmware.\n", - dev->full_name); + pr_warning("%pOF: disabled by the firmware.\n", + dev); return -ENODEV; } - pr_debug("Adding PCI host bridge %s\n", dev->full_name); + pr_debug("Adding PCI host bridge %pOF\n", dev); /* Fetch host bridge registers address */ if (of_address_to_resource(dev, 0, &rsrc_reg)) { @@ -848,8 +847,8 @@ int __init mpc83xx_add_bridge(struct device_node *dev) /* Get bus range if any */ bus_range = of_get_property(dev, "bus-range", &len); if (bus_range == NULL || len < 2 * sizeof(int)) { - printk(KERN_WARNING "Can't get bus-range for %s, assume" - " bus 0\n", dev->full_name); + printk(KERN_WARNING "Can't get bus-range for %pOF, assume" + " bus 0\n", dev); } pci_add_flags(PCI_REASSIGN_ALL_BUS); diff --git a/arch/powerpc/sysdev/fsl_rio.c b/arch/powerpc/sysdev/fsl_rio.c index 1c41c51f22cb..9234be1e66f5 100644 --- a/arch/powerpc/sysdev/fsl_rio.c +++ b/arch/powerpc/sysdev/fsl_rio.c @@ -450,12 +450,12 @@ int fsl_rio_setup(struct platform_device *dev) rc = of_address_to_resource(dev->dev.of_node, 0, ®s); if (rc) { - dev_err(&dev->dev, "Can't get %s property 'reg'\n", - dev->dev.of_node->full_name); + dev_err(&dev->dev, "Can't get %pOF property 'reg'\n", + dev->dev.of_node); return -EFAULT; } - dev_info(&dev->dev, "Of-device full name %s\n", - dev->dev.of_node->full_name); + dev_info(&dev->dev, "Of-device full name %pOF\n", + dev->dev.of_node); dev_info(&dev->dev, "Regs: %pR\n", ®s); rio_regs_win = ioremap(regs.start, resource_size(®s)); @@ -494,8 +494,8 @@ int fsl_rio_setup(struct platform_device *dev) } rc = of_address_to_resource(rmu_node, 0, &rmu_regs); if (rc) { - dev_err(&dev->dev, "Can't get %s property 'reg'\n", - rmu_node->full_name); + dev_err(&dev->dev, "Can't get %pOF property 'reg'\n", + rmu_node); goto err_rmu; } rmu_regs_win = ioremap(rmu_regs.start, resource_size(&rmu_regs)); @@ -529,8 +529,8 @@ int fsl_rio_setup(struct platform_device *dev) aw = of_n_addr_cells(np); dt_range = of_get_property(np, "reg", &rlen); if (!dt_range) { - pr_err("%s: unable to find 'reg' property\n", - np->full_name); + pr_err("%pOF: unable to find 'reg' property\n", + np); rc = -ENOMEM; goto err_pw; } @@ -557,8 +557,8 @@ int fsl_rio_setup(struct platform_device *dev) aw = of_n_addr_cells(np); dt_range = of_get_property(np, "reg", &rlen); if (!dt_range) { - pr_err("%s: unable to find 'reg' property\n", - np->full_name); + pr_err("%pOF: unable to find 'reg' property\n", + np); rc = -ENOMEM; goto err; } @@ -569,15 +569,15 @@ int fsl_rio_setup(struct platform_device *dev) for_each_child_of_node(dev->dev.of_node, np) { port_index = of_get_property(np, "cell-index", NULL); if (!port_index) { - dev_err(&dev->dev, "Can't get %s property 'cell-index'\n", - np->full_name); + dev_err(&dev->dev, "Can't get %pOF property 'cell-index'\n", + np); continue; } dt_range = of_get_property(np, "ranges", &rlen); if (!dt_range) { - dev_err(&dev->dev, "Can't get %s property 'ranges'\n", - np->full_name); + dev_err(&dev->dev, "Can't get %pOF property 'ranges'\n", + np); continue; } @@ -598,8 +598,8 @@ int fsl_rio_setup(struct platform_device *dev) range_start = of_read_number(dt_range + aw, paw); range_size = of_read_number(dt_range + aw + paw, sw); - dev_info(&dev->dev, "%s: LAW start 0x%016llx, size 0x%016llx.\n", - np->full_name, range_start, range_size); + dev_info(&dev->dev, "%pOF: LAW start 0x%016llx, size 0x%016llx.\n", + np, range_start, range_size); port = kzalloc(sizeof(struct rio_mport), GFP_KERNEL); if (!port) @@ -757,8 +757,8 @@ int fsl_rio_setup(struct platform_device *dev) */ static int fsl_of_rio_rpn_probe(struct platform_device *dev) { - printk(KERN_INFO "Setting up RapidIO peer-to-peer network %s\n", - dev->dev.of_node->full_name); + printk(KERN_INFO "Setting up RapidIO peer-to-peer network %pOF\n", + dev->dev.of_node); return fsl_rio_setup(dev); }; diff --git a/arch/powerpc/sysdev/fsl_rmu.c b/arch/powerpc/sysdev/fsl_rmu.c index c1826de4e749..ab7a74c75be8 100644 --- a/arch/powerpc/sysdev/fsl_rmu.c +++ b/arch/powerpc/sysdev/fsl_rmu.c @@ -1074,8 +1074,8 @@ int fsl_rio_setup_rmu(struct rio_mport *mport, struct device_node *node) priv = mport->priv; if (!node) { - dev_warn(priv->dev, "Can't get %s property 'fsl,rmu'\n", - priv->dev->of_node->full_name); + dev_warn(priv->dev, "Can't get %pOF property 'fsl,rmu'\n", + priv->dev->of_node); return -EINVAL; } @@ -1086,8 +1086,8 @@ int fsl_rio_setup_rmu(struct rio_mport *mport, struct device_node *node) aw = of_n_addr_cells(node); msg_addr = of_get_property(node, "reg", &mlen); if (!msg_addr) { - pr_err("%s: unable to find 'reg' property of message-unit\n", - node->full_name); + pr_err("%pOF: unable to find 'reg' property of message-unit\n", + node); kfree(rmu); return -ENOMEM; } @@ -1098,8 +1098,8 @@ int fsl_rio_setup_rmu(struct rio_mport *mport, struct device_node *node) rmu->txirq = irq_of_parse_and_map(node, 0); rmu->rxirq = irq_of_parse_and_map(node, 1); - printk(KERN_INFO "%s: txirq: %d, rxirq %d\n", - node->full_name, rmu->txirq, rmu->rxirq); + printk(KERN_INFO "%pOF: txirq: %d, rxirq %d\n", + node, rmu->txirq, rmu->rxirq); priv->rmm_handle = rmu; diff --git a/arch/powerpc/sysdev/mpic.c b/arch/powerpc/sysdev/mpic.c index b9aac951a90f..ead3e2549ebf 100644 --- a/arch/powerpc/sysdev/mpic.c +++ b/arch/powerpc/sysdev/mpic.c @@ -1650,8 +1650,8 @@ void __init mpic_init(struct mpic *mpic) if (mpic->flags & MPIC_SECONDARY) { int virq = irq_of_parse_and_map(mpic->node, 0); if (virq) { - printk(KERN_INFO "%s: hooking up to IRQ %d\n", - mpic->node->full_name, virq); + printk(KERN_INFO "%pOF: hooking up to IRQ %d\n", + mpic->node, virq); irq_set_handler_data(virq, mpic); irq_set_chained_handler(virq, &mpic_cascade); } diff --git a/arch/powerpc/sysdev/mpic_msgr.c b/arch/powerpc/sysdev/mpic_msgr.c index db2286be5d9a..eb69a5186243 100644 --- a/arch/powerpc/sysdev/mpic_msgr.c +++ b/arch/powerpc/sysdev/mpic_msgr.c @@ -192,7 +192,7 @@ static int mpic_msgr_probe(struct platform_device *dev) return -ENOMEM; } } - dev_info(&dev->dev, "Of-device full name %s\n", np->full_name); + dev_info(&dev->dev, "Of-device full name %pOF\n", np); /* IO map the message register block. */ of_address_to_resource(np, 0, &rsrc); diff --git a/arch/powerpc/sysdev/mpic_msi.c b/arch/powerpc/sysdev/mpic_msi.c index 1d48a5385905..9ed860aee9c3 100644 --- a/arch/powerpc/sysdev/mpic_msi.c +++ b/arch/powerpc/sysdev/mpic_msi.c @@ -60,7 +60,7 @@ static int mpic_msi_reserve_u3_hwirqs(struct mpic *mpic) np = NULL; while ((np = of_find_all_nodes(np))) { - pr_debug("mpic: mapping hwirqs for %s\n", np->full_name); + pr_debug("mpic: mapping hwirqs for %pOF\n", np); index = 0; while (of_irq_parse_one(np, index++, &oirq) == 0) { diff --git a/arch/powerpc/sysdev/mpic_timer.c b/arch/powerpc/sysdev/mpic_timer.c index 9d9b06217f8b..a418579591be 100644 --- a/arch/powerpc/sysdev/mpic_timer.c +++ b/arch/powerpc/sysdev/mpic_timer.c @@ -466,8 +466,7 @@ static int timer_group_get_irq(struct device_node *np, p = of_get_property(np, "fsl,available-ranges", &len); if (p && len % (2 * sizeof(u32)) != 0) { - pr_err("%s: malformed available-ranges property.\n", - np->full_name); + pr_err("%pOF: malformed available-ranges property.\n", np); return -EINVAL; } @@ -484,8 +483,7 @@ static int timer_group_get_irq(struct device_node *np, for (j = 0; j < count; j++) { irq = irq_of_parse_and_map(np, irq_index); if (!irq) { - pr_err("%s: irq parse and map failed.\n", - np->full_name); + pr_err("%pOF: irq parse and map failed.\n", np); return -EINVAL; } @@ -508,8 +506,7 @@ static void timer_group_init(struct device_node *np) priv = kzalloc(sizeof(struct timer_group_priv), GFP_KERNEL); if (!priv) { - pr_err("%s: cannot allocate memory for group.\n", - np->full_name); + pr_err("%pOF: cannot allocate memory for group.\n", np); return; } @@ -518,29 +515,27 @@ static void timer_group_init(struct device_node *np) priv->regs = of_iomap(np, i++); if (!priv->regs) { - pr_err("%s: cannot ioremap timer register address.\n", - np->full_name); + pr_err("%pOF: cannot ioremap timer register address.\n", np); goto out; } if (priv->flags & FSL_GLOBAL_TIMER) { priv->group_tcr = of_iomap(np, i++); if (!priv->group_tcr) { - pr_err("%s: cannot ioremap tcr address.\n", - np->full_name); + pr_err("%pOF: cannot ioremap tcr address.\n", np); goto out; } } ret = timer_group_get_freq(np, priv); if (ret < 0) { - pr_err("%s: cannot get timer frequency.\n", np->full_name); + pr_err("%pOF: cannot get timer frequency.\n", np); goto out; } ret = timer_group_get_irq(np, priv); if (ret < 0) { - pr_err("%s: cannot get timer irqs.\n", np->full_name); + pr_err("%pOF: cannot get timer irqs.\n", np); goto out; } diff --git a/arch/powerpc/sysdev/msi_bitmap.c b/arch/powerpc/sysdev/msi_bitmap.c index 5ebd3f018295..c4dae27172b3 100644 --- a/arch/powerpc/sysdev/msi_bitmap.c +++ b/arch/powerpc/sysdev/msi_bitmap.c @@ -86,13 +86,13 @@ int msi_bitmap_reserve_dt_hwirqs(struct msi_bitmap *bmp) p = of_get_property(bmp->of_node, "msi-available-ranges", &len); if (!p) { pr_debug("msi_bitmap: no msi-available-ranges property " \ - "found on %s\n", bmp->of_node->full_name); + "found on %pOF\n", bmp->of_node); return 1; } if (len % (2 * sizeof(u32)) != 0) { printk(KERN_WARNING "msi_bitmap: Malformed msi-available-ranges" - " property on %s\n", bmp->of_node->full_name); + " property on %pOF\n", bmp->of_node); return -EINVAL; } diff --git a/arch/powerpc/sysdev/mv64x60_dev.c b/arch/powerpc/sysdev/mv64x60_dev.c index 026bbc3b2c47..185a67e742a6 100644 --- a/arch/powerpc/sysdev/mv64x60_dev.c +++ b/arch/powerpc/sysdev/mv64x60_dev.c @@ -452,8 +452,8 @@ static int __init mv64x60_device_setup(void) err = mv64x60_mpsc_device_setup(np, id++); if (err) printk(KERN_ERR "Failed to initialize MV64x60 " - "serial device %s: error %d.\n", - np->full_name, err); + "serial device %pOF: error %d.\n", + np, err); } id = 0; @@ -463,8 +463,8 @@ static int __init mv64x60_device_setup(void) if (IS_ERR(pdev)) { err = PTR_ERR(pdev); printk(KERN_ERR "Failed to initialize MV64x60 " - "network block %s: error %d.\n", - np->full_name, err); + "network block %pOF: error %d.\n", + np, err); continue; } for_each_child_of_node(np, np2) { @@ -474,9 +474,9 @@ static int __init mv64x60_device_setup(void) err = mv64x60_eth_device_setup(np2, id2++, pdev); if (err) printk(KERN_ERR "Failed to initialize " - "MV64x60 network device %s: " + "MV64x60 network device %pOF: " "error %d.\n", - np2->full_name, err); + np2, err); } } @@ -485,8 +485,8 @@ static int __init mv64x60_device_setup(void) err = mv64x60_i2c_device_setup(np, id++); if (err) printk(KERN_ERR "Failed to initialize MV64x60 I2C " - "bus %s: error %d.\n", - np->full_name, err); + "bus %pOF: error %d.\n", + np, err); } /* support up to one watchdog timer */ @@ -494,8 +494,8 @@ static int __init mv64x60_device_setup(void) if (np) { if ((err = mv64x60_wdt_device_setup(np, id))) printk(KERN_ERR "Failed to initialize MV64x60 " - "Watchdog %s: error %d.\n", - np->full_name, err); + "Watchdog %pOF: error %d.\n", + np, err); of_node_put(np); } diff --git a/arch/powerpc/sysdev/mv64x60_pci.c b/arch/powerpc/sysdev/mv64x60_pci.c index 330d56613c5a..5261750dacc4 100644 --- a/arch/powerpc/sysdev/mv64x60_pci.c +++ b/arch/powerpc/sysdev/mv64x60_pci.c @@ -136,8 +136,8 @@ static int __init mv64x60_add_bridge(struct device_node *dev) /* Get bus range if any */ bus_range = of_get_property(dev, "bus-range", &len); if (bus_range == NULL || len < 2 * sizeof(int)) - printk(KERN_WARNING "Can't get bus-range for %s, assume" - " bus 0\n", dev->full_name); + printk(KERN_WARNING "Can't get bus-range for %pOF, assume" + " bus 0\n", dev); hose = pcibios_alloc_controller(dev); if (!hose) diff --git a/arch/powerpc/sysdev/of_rtc.c b/arch/powerpc/sysdev/of_rtc.c index 6f54b54b1328..153fdac4720f 100644 --- a/arch/powerpc/sysdev/of_rtc.c +++ b/arch/powerpc/sysdev/of_rtc.c @@ -38,21 +38,21 @@ void __init of_instantiate_rtc(void) res = kmalloc(sizeof(*res), GFP_KERNEL); if (!res) { printk(KERN_ERR "OF RTC: Out of memory " - "allocating resource structure for %s\n", - node->full_name); + "allocating resource structure for %pOF\n", + node); continue; } err = of_address_to_resource(node, 0, res); if (err) { printk(KERN_ERR "OF RTC: Error " - "translating resources for %s\n", - node->full_name); + "translating resources for %pOF\n", + node); continue; } - printk(KERN_INFO "OF_RTC: %s is a %s @ 0x%llx-0x%llx\n", - node->full_name, plat_name, + printk(KERN_INFO "OF_RTC: %pOF is a %s @ 0x%llx-0x%llx\n", + node, plat_name, (unsigned long long)res->start, (unsigned long long)res->end); platform_device_register_simple(plat_name, -1, res, 1); diff --git a/arch/powerpc/sysdev/scom.c b/arch/powerpc/sysdev/scom.c index 76ea32c1b664..0f6fd5d04d33 100644 --- a/arch/powerpc/sysdev/scom.c +++ b/arch/powerpc/sysdev/scom.c @@ -194,12 +194,13 @@ static int scom_debug_init_one(struct dentry *root, struct device_node *dn, ent->dn = of_node_get(dn); snprintf(ent->name, 16, "%08x", i); - ent->path.data = (void*) dn->full_name; - ent->path.size = strlen(dn->full_name); + ent->path.data = (void*)kasprintf(GFP_KERNEL, "%pOF", dn); + ent->path.size = strlen((char *)ent->path.data); dir = debugfs_create_dir(ent->name, root); if (!dir) { of_node_put(dn); + kfree(ent->path.data); kfree(ent); return -1; } diff --git a/arch/powerpc/sysdev/simple_gpio.c b/arch/powerpc/sysdev/simple_gpio.c index 6afddae2fb47..f02d4576138c 100644 --- a/arch/powerpc/sysdev/simple_gpio.c +++ b/arch/powerpc/sysdev/simple_gpio.c @@ -142,7 +142,6 @@ void __init simple_gpiochip_init(const char *compatible) } continue; err: - pr_err("%s: registration failed, status %d\n", - np->full_name, ret); + pr_err("%pOF: registration failed, status %d\n", np, ret); } } diff --git a/arch/powerpc/sysdev/tsi108_pci.c b/arch/powerpc/sysdev/tsi108_pci.c index 5692dd569b9b..28ff1f53cefc 100644 --- a/arch/powerpc/sysdev/tsi108_pci.c +++ b/arch/powerpc/sysdev/tsi108_pci.c @@ -213,8 +213,8 @@ int __init tsi108_setup_pci(struct device_node *dev, u32 cfg_phys, int primary) /* Get bus range if any */ bus_range = of_get_property(dev, "bus-range", &len); if (bus_range == NULL || len < 2 * sizeof(int)) { - printk(KERN_WARNING "Can't get bus-range for %s, assume" - " bus 0\n", dev->full_name); + printk(KERN_WARNING "Can't get bus-range for %pOF, assume" + " bus 0\n", dev); } hose = pcibios_alloc_controller(dev); diff --git a/arch/powerpc/sysdev/xive/native.c b/arch/powerpc/sysdev/xive/native.c index 1dbf782c9239..edc2e81b7280 100644 --- a/arch/powerpc/sysdev/xive/native.c +++ b/arch/powerpc/sysdev/xive/native.c @@ -551,7 +551,7 @@ bool __init xive_native_init(void) pr_devel("not found !\n"); return false; } - pr_devel("Found %s\n", np->full_name); + pr_devel("Found %pOF\n", np); /* Resource 1 is HV window */ if (of_address_to_resource(np, 1, &r)) { -- 2.11.0 ^ permalink raw reply related [flat|nested] 53+ messages in thread
* Re: [v2,1/5] powerpc: Convert to using %pOF instead of full_name @ 2017-08-24 12:37 ` Michael Ellerman 0 siblings, 0 replies; 53+ messages in thread From: Michael Ellerman @ 2017-08-24 12:37 UTC (permalink / raw) To: Rob Herring, devicetree, linuxppc-dev Cc: Arnd Bergmann, linux-kernel, Scott Wood, David Laight, Paul Mackerras, Anatolij Gustschin, Frank Rowand On Mon, 2017-08-21 at 15:16:47 UTC, Rob Herring wrote: > Now that we have a custom printf format specifier, convert users of > full_name to use %pOF instead. This is preparation to remove storing > of the full path string for each node. > > Signed-off-by: Rob Herring <robh@kernel.org> > Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> > Cc: Paul Mackerras <paulus@samba.org> > Cc: Michael Ellerman <mpe@ellerman.id.au> > Cc: Anatolij Gustschin <agust@denx.de> > Cc: Scott Wood <oss@buserror.net> > Cc: Kumar Gala <galak@kernel.crashing.org> > Cc: Arnd Bergmann <arnd@arndb.de> > Cc: linuxppc-dev@lists.ozlabs.org > Reviewed-by: Tyrel Datwyler <tyreld@linux.vnet.ibm.com> Applied to powerpc next, thanks. https://git.kernel.org/powerpc/c/b7c670d673d1186e9a6aafaad36aac I forgot to add in the change log I took the chance to reformat some of the affected printks, hopefully that doesn't muck you up when you're merging/rebasing on top of this. cheers ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [v2,1/5] powerpc: Convert to using %pOF instead of full_name @ 2017-08-24 12:37 ` Michael Ellerman 0 siblings, 0 replies; 53+ messages in thread From: Michael Ellerman @ 2017-08-24 12:37 UTC (permalink / raw) To: Rob Herring, devicetree-u79uwXL29TY76Z2rM5mHXA, linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ Cc: Arnd Bergmann, linux-kernel-u79uwXL29TY76Z2rM5mHXA, Scott Wood, David Laight, Paul Mackerras, Anatolij Gustschin, Frank Rowand On Mon, 2017-08-21 at 15:16:47 UTC, Rob Herring wrote: > Now that we have a custom printf format specifier, convert users of > full_name to use %pOF instead. This is preparation to remove storing > of the full path string for each node. > > Signed-off-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> > Cc: Benjamin Herrenschmidt <benh-XVmvHMARGAS8U2dJNN8I7kB+6BGkLq7r@public.gmane.org> > Cc: Paul Mackerras <paulus-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org> > Cc: Michael Ellerman <mpe-Gsx/Oe8HsFggBc27wqDAHg@public.gmane.org> > Cc: Anatolij Gustschin <agust-ynQEQJNshbs@public.gmane.org> > Cc: Scott Wood <oss-fOR+EgIDQEHk1uMJSBkQmQ@public.gmane.org> > Cc: Kumar Gala <galak-XVmvHMARGAS8U2dJNN8I7kB+6BGkLq7r@public.gmane.org> > Cc: Arnd Bergmann <arnd-r2nGTMty4D4@public.gmane.org> > Cc: linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ@public.gmane.org > Reviewed-by: Tyrel Datwyler <tyreld-23VcF4HTsmIX0ybBhKVfKdBPR1lH4CV8@public.gmane.org> Applied to powerpc next, thanks. https://git.kernel.org/powerpc/c/b7c670d673d1186e9a6aafaad36aac I forgot to add in the change log I took the chance to reformat some of the affected printks, hopefully that doesn't muck you up when you're merging/rebasing on top of this. cheers -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* [PATCH v2 2/5] powerpc: pseries: vio: match parent nodes with of_find_node_by_path 2017-08-21 15:16 ` Rob Herring (?) (?) @ 2017-08-21 15:16 ` Rob Herring 2017-08-22 5:12 ` Michael Ellerman -1 siblings, 1 reply; 53+ messages in thread From: Rob Herring @ 2017-08-21 15:16 UTC (permalink / raw) To: devicetree, Michael Ellerman, linuxppc-dev Cc: linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight In preparation to remove the full path from device_node.full_name, use of_find_node_by_path instead of open coding with strcmp. Signed-off-by: Rob Herring <robh@kernel.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: linuxppc-dev@lists.ozlabs.org --- v2: - rebased to linux-next and removed spurious change fro patch 1. arch/powerpc/platforms/pseries/vio.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c index aa5ca74316fa..5754572deb23 100644 --- a/arch/powerpc/platforms/pseries/vio.c +++ b/arch/powerpc/platforms/pseries/vio.c @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) */ parent_node = of_get_parent(of_node); if (parent_node) { - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) + if (parent_node == of_find_node_by_path("/ibm,platform-facilities")) family = PFO; - else if (!strcmp(parent_node->full_name, "/vdevice")) + else if (parent_node == of_find_node_by_path("/vdevice")) family = VDEVICE; else { pr_warn("%s: parent(%pOF) of %s not recognized.\n", -- 2.11.0 ^ permalink raw reply related [flat|nested] 53+ messages in thread
* Re: [PATCH v2 2/5] powerpc: pseries: vio: match parent nodes with of_find_node_by_path @ 2017-08-22 5:12 ` Michael Ellerman 0 siblings, 0 replies; 53+ messages in thread From: Michael Ellerman @ 2017-08-22 5:12 UTC (permalink / raw) To: Rob Herring, devicetree, linuxppc-dev Cc: linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight Rob Herring <robh@kernel.org> writes: > In preparation to remove the full path from device_node.full_name, use > of_find_node_by_path instead of open coding with strcmp. > > Signed-off-by: Rob Herring <robh@kernel.org> > Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> > Cc: Paul Mackerras <paulus@samba.org> > Cc: Michael Ellerman <mpe@ellerman.id.au> > Cc: linuxppc-dev@lists.ozlabs.org > --- > v2: > - rebased to linux-next and removed spurious change fro patch 1. > > arch/powerpc/platforms/pseries/vio.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c > index aa5ca74316fa..5754572deb23 100644 > --- a/arch/powerpc/platforms/pseries/vio.c > +++ b/arch/powerpc/platforms/pseries/vio.c > @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) > */ > parent_node = of_get_parent(of_node); > if (parent_node) { > - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) > + if (parent_node == of_find_node_by_path("/ibm,platform-facilities")) > family = PFO; > - else if (!strcmp(parent_node->full_name, "/vdevice")) > + else if (parent_node == of_find_node_by_path("/vdevice")) > family = VDEVICE; This leaks references to the looked up nodes. Both these nodes are defined in PAPR (our hypervisor spec), and both of them must have a device_type, either "ibm,platform-facilities" or "vdevice". Looking at the commit that added the code I don't see any particular reason it used the comparison against full_name, rather than using the device_type. So I'm inclined to do this instead: diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c index 8a47f168476b..f26f906e6021 100644 --- a/arch/powerpc/platforms/pseries/vio.c +++ b/arch/powerpc/platforms/pseries/vio.c @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) */ parent_node = of_get_parent(of_node); if (parent_node) { - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) + if (!strcmp(parent_node->type, "ibm,platform-facilities")) family = PFO; - else if (!strcmp(parent_node->full_name, "/vdevice")) + else if (!strcmp(parent_node->type, "vdevice")) family = VDEVICE; else { pr_warn("%s: parent(%s) of %s not recognized.\n", I've checked both Qemu and kvmtool add the device_type, and I'm fairly confident that PowerVM does too. Anyway I'll test it on all the machines I can find. cheers ^ permalink raw reply related [flat|nested] 53+ messages in thread
* Re: [PATCH v2 2/5] powerpc: pseries: vio: match parent nodes with of_find_node_by_path @ 2017-08-22 5:12 ` Michael Ellerman 0 siblings, 0 replies; 53+ messages in thread From: Michael Ellerman @ 2017-08-22 5:12 UTC (permalink / raw) To: Rob Herring, devicetree-u79uwXL29TY76Z2rM5mHXA, linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ Cc: linux-kernel-u79uwXL29TY76Z2rM5mHXA, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> writes: > In preparation to remove the full path from device_node.full_name, use > of_find_node_by_path instead of open coding with strcmp. > > Signed-off-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> > Cc: Benjamin Herrenschmidt <benh-XVmvHMARGAS8U2dJNN8I7kB+6BGkLq7r@public.gmane.org> > Cc: Paul Mackerras <paulus-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org> > Cc: Michael Ellerman <mpe-Gsx/Oe8HsFggBc27wqDAHg@public.gmane.org> > Cc: linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ@public.gmane.org > --- > v2: > - rebased to linux-next and removed spurious change fro patch 1. > > arch/powerpc/platforms/pseries/vio.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c > index aa5ca74316fa..5754572deb23 100644 > --- a/arch/powerpc/platforms/pseries/vio.c > +++ b/arch/powerpc/platforms/pseries/vio.c > @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) > */ > parent_node = of_get_parent(of_node); > if (parent_node) { > - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) > + if (parent_node == of_find_node_by_path("/ibm,platform-facilities")) > family = PFO; > - else if (!strcmp(parent_node->full_name, "/vdevice")) > + else if (parent_node == of_find_node_by_path("/vdevice")) > family = VDEVICE; This leaks references to the looked up nodes. Both these nodes are defined in PAPR (our hypervisor spec), and both of them must have a device_type, either "ibm,platform-facilities" or "vdevice". Looking at the commit that added the code I don't see any particular reason it used the comparison against full_name, rather than using the device_type. So I'm inclined to do this instead: diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c index 8a47f168476b..f26f906e6021 100644 --- a/arch/powerpc/platforms/pseries/vio.c +++ b/arch/powerpc/platforms/pseries/vio.c @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) */ parent_node = of_get_parent(of_node); if (parent_node) { - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) + if (!strcmp(parent_node->type, "ibm,platform-facilities")) family = PFO; - else if (!strcmp(parent_node->full_name, "/vdevice")) + else if (!strcmp(parent_node->type, "vdevice")) family = VDEVICE; else { pr_warn("%s: parent(%s) of %s not recognized.\n", I've checked both Qemu and kvmtool add the device_type, and I'm fairly confident that PowerVM does too. Anyway I'll test it on all the machines I can find. cheers -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply related [flat|nested] 53+ messages in thread
* Re: [PATCH v2 2/5] powerpc: pseries: vio: match parent nodes with of_find_node_by_path @ 2017-08-24 18:08 ` Rob Herring 0 siblings, 0 replies; 53+ messages in thread From: Rob Herring @ 2017-08-24 18:08 UTC (permalink / raw) To: Michael Ellerman Cc: devicetree, linuxppc-dev, linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight On Tue, Aug 22, 2017 at 12:12 AM, Michael Ellerman <mpe@ellerman.id.au> wrote: > Rob Herring <robh@kernel.org> writes: > >> In preparation to remove the full path from device_node.full_name, use >> of_find_node_by_path instead of open coding with strcmp. >> >> Signed-off-by: Rob Herring <robh@kernel.org> >> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> >> Cc: Paul Mackerras <paulus@samba.org> >> Cc: Michael Ellerman <mpe@ellerman.id.au> >> Cc: linuxppc-dev@lists.ozlabs.org >> --- >> v2: >> - rebased to linux-next and removed spurious change fro patch 1. >> >> arch/powerpc/platforms/pseries/vio.c | 4 ++-- >> 1 file changed, 2 insertions(+), 2 deletions(-) >> >> diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c >> index aa5ca74316fa..5754572deb23 100644 >> --- a/arch/powerpc/platforms/pseries/vio.c >> +++ b/arch/powerpc/platforms/pseries/vio.c >> @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) >> */ >> parent_node = of_get_parent(of_node); >> if (parent_node) { >> - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) >> + if (parent_node == of_find_node_by_path("/ibm,platform-facilities")) >> family = PFO; >> - else if (!strcmp(parent_node->full_name, "/vdevice")) >> + else if (parent_node == of_find_node_by_path("/vdevice")) >> family = VDEVICE; > > This leaks references to the looked up nodes. > > Both these nodes are defined in PAPR (our hypervisor spec), and both of > them must have a device_type, either "ibm,platform-facilities" or > "vdevice". > > Looking at the commit that added the code I don't see any particular > reason it used the comparison against full_name, rather than using the > device_type. > > So I'm inclined to do this instead: > > diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c > index 8a47f168476b..f26f906e6021 100644 > --- a/arch/powerpc/platforms/pseries/vio.c > +++ b/arch/powerpc/platforms/pseries/vio.c > @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) > */ > parent_node = of_get_parent(of_node); > if (parent_node) { > - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) > + if (!strcmp(parent_node->type, "ibm,platform-facilities")) > family = PFO; > - else if (!strcmp(parent_node->full_name, "/vdevice")) > + else if (!strcmp(parent_node->type, "vdevice")) > family = VDEVICE; > else { > pr_warn("%s: parent(%s) of %s not recognized.\n", > > > I've checked both Qemu and kvmtool add the device_type, and I'm fairly > confident that PowerVM does too. Anyway I'll test it on all the machines > I can find. Okay, do you want me to respin the patch or will you update it with the above change? Rob ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 2/5] powerpc: pseries: vio: match parent nodes with of_find_node_by_path @ 2017-08-24 18:08 ` Rob Herring 0 siblings, 0 replies; 53+ messages in thread From: Rob Herring @ 2017-08-24 18:08 UTC (permalink / raw) To: Michael Ellerman Cc: devicetree-u79uwXL29TY76Z2rM5mHXA, linuxppc-dev, linux-kernel-u79uwXL29TY76Z2rM5mHXA, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight On Tue, Aug 22, 2017 at 12:12 AM, Michael Ellerman <mpe-Gsx/Oe8HsFggBc27wqDAHg@public.gmane.org> wrote: > Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> writes: > >> In preparation to remove the full path from device_node.full_name, use >> of_find_node_by_path instead of open coding with strcmp. >> >> Signed-off-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> >> Cc: Benjamin Herrenschmidt <benh-XVmvHMARGAS8U2dJNN8I7kB+6BGkLq7r@public.gmane.org> >> Cc: Paul Mackerras <paulus-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org> >> Cc: Michael Ellerman <mpe-Gsx/Oe8HsFggBc27wqDAHg@public.gmane.org> >> Cc: linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ@public.gmane.org >> --- >> v2: >> - rebased to linux-next and removed spurious change fro patch 1. >> >> arch/powerpc/platforms/pseries/vio.c | 4 ++-- >> 1 file changed, 2 insertions(+), 2 deletions(-) >> >> diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c >> index aa5ca74316fa..5754572deb23 100644 >> --- a/arch/powerpc/platforms/pseries/vio.c >> +++ b/arch/powerpc/platforms/pseries/vio.c >> @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) >> */ >> parent_node = of_get_parent(of_node); >> if (parent_node) { >> - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) >> + if (parent_node == of_find_node_by_path("/ibm,platform-facilities")) >> family = PFO; >> - else if (!strcmp(parent_node->full_name, "/vdevice")) >> + else if (parent_node == of_find_node_by_path("/vdevice")) >> family = VDEVICE; > > This leaks references to the looked up nodes. > > Both these nodes are defined in PAPR (our hypervisor spec), and both of > them must have a device_type, either "ibm,platform-facilities" or > "vdevice". > > Looking at the commit that added the code I don't see any particular > reason it used the comparison against full_name, rather than using the > device_type. > > So I'm inclined to do this instead: > > diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c > index 8a47f168476b..f26f906e6021 100644 > --- a/arch/powerpc/platforms/pseries/vio.c > +++ b/arch/powerpc/platforms/pseries/vio.c > @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) > */ > parent_node = of_get_parent(of_node); > if (parent_node) { > - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) > + if (!strcmp(parent_node->type, "ibm,platform-facilities")) > family = PFO; > - else if (!strcmp(parent_node->full_name, "/vdevice")) > + else if (!strcmp(parent_node->type, "vdevice")) > family = VDEVICE; > else { > pr_warn("%s: parent(%s) of %s not recognized.\n", > > > I've checked both Qemu and kvmtool add the device_type, and I'm fairly > confident that PowerVM does too. Anyway I'll test it on all the machines > I can find. Okay, do you want me to respin the patch or will you update it with the above change? Rob -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 2/5] powerpc: pseries: vio: match parent nodes with of_find_node_by_path @ 2017-08-25 0:10 ` Michael Ellerman 0 siblings, 0 replies; 53+ messages in thread From: Michael Ellerman @ 2017-08-25 0:10 UTC (permalink / raw) To: Rob Herring Cc: devicetree, linuxppc-dev, linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight Rob Herring <robh@kernel.org> writes: > On Tue, Aug 22, 2017 at 12:12 AM, Michael Ellerman <mpe@ellerman.id.au> wrote: >> Rob Herring <robh@kernel.org> writes: >> >>> In preparation to remove the full path from device_node.full_name, use >>> of_find_node_by_path instead of open coding with strcmp. >>> >>> Signed-off-by: Rob Herring <robh@kernel.org> >>> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> >>> Cc: Paul Mackerras <paulus@samba.org> >>> Cc: Michael Ellerman <mpe@ellerman.id.au> >>> Cc: linuxppc-dev@lists.ozlabs.org >>> --- >>> v2: >>> - rebased to linux-next and removed spurious change fro patch 1. >>> >>> arch/powerpc/platforms/pseries/vio.c | 4 ++-- >>> 1 file changed, 2 insertions(+), 2 deletions(-) >>> >>> diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c >>> index aa5ca74316fa..5754572deb23 100644 >>> --- a/arch/powerpc/platforms/pseries/vio.c >>> +++ b/arch/powerpc/platforms/pseries/vio.c >>> @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) >>> */ >>> parent_node = of_get_parent(of_node); >>> if (parent_node) { >>> - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) >>> + if (parent_node == of_find_node_by_path("/ibm,platform-facilities")) >>> family = PFO; >>> - else if (!strcmp(parent_node->full_name, "/vdevice")) >>> + else if (parent_node == of_find_node_by_path("/vdevice")) >>> family = VDEVICE; >> >> This leaks references to the looked up nodes. >> >> Both these nodes are defined in PAPR (our hypervisor spec), and both of >> them must have a device_type, either "ibm,platform-facilities" or >> "vdevice". >> >> Looking at the commit that added the code I don't see any particular >> reason it used the comparison against full_name, rather than using the >> device_type. >> >> So I'm inclined to do this instead: >> >> diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c >> index 8a47f168476b..f26f906e6021 100644 >> --- a/arch/powerpc/platforms/pseries/vio.c >> +++ b/arch/powerpc/platforms/pseries/vio.c >> @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) >> */ >> parent_node = of_get_parent(of_node); >> if (parent_node) { >> - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) >> + if (!strcmp(parent_node->type, "ibm,platform-facilities")) >> family = PFO; >> - else if (!strcmp(parent_node->full_name, "/vdevice")) >> + else if (!strcmp(parent_node->type, "vdevice")) >> family = VDEVICE; >> else { >> pr_warn("%s: parent(%s) of %s not recognized.\n", >> >> >> I've checked both Qemu and kvmtool add the device_type, and I'm fairly >> confident that PowerVM does too. Anyway I'll test it on all the machines >> I can find. > > Okay, do you want me to respin the patch or will you update it with I merged it. Should be in next today. cheers ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 2/5] powerpc: pseries: vio: match parent nodes with of_find_node_by_path @ 2017-08-25 0:10 ` Michael Ellerman 0 siblings, 0 replies; 53+ messages in thread From: Michael Ellerman @ 2017-08-25 0:10 UTC (permalink / raw) To: Rob Herring Cc: devicetree@vger.kernel.org, linuxppc-dev, linux-kernel@vger.kernel.org, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> writes: > On Tue, Aug 22, 2017 at 12:12 AM, Michael Ellerman <mpe-Gsx/Oe8HsFggBc27wqDAHg@public.gmane.org> wrote: >> Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> writes: >> >>> In preparation to remove the full path from device_node.full_name, use >>> of_find_node_by_path instead of open coding with strcmp. >>> >>> Signed-off-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> >>> Cc: Benjamin Herrenschmidt <benh-XVmvHMARGAS8U2dJNN8I7kB+6BGkLq7r@public.gmane.org> >>> Cc: Paul Mackerras <paulus-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org> >>> Cc: Michael Ellerman <mpe-Gsx/Oe8HsFggBc27wqDAHg@public.gmane.org> >>> Cc: linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ@public.gmane.org >>> --- >>> v2: >>> - rebased to linux-next and removed spurious change fro patch 1. >>> >>> arch/powerpc/platforms/pseries/vio.c | 4 ++-- >>> 1 file changed, 2 insertions(+), 2 deletions(-) >>> >>> diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c >>> index aa5ca74316fa..5754572deb23 100644 >>> --- a/arch/powerpc/platforms/pseries/vio.c >>> +++ b/arch/powerpc/platforms/pseries/vio.c >>> @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) >>> */ >>> parent_node = of_get_parent(of_node); >>> if (parent_node) { >>> - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) >>> + if (parent_node == of_find_node_by_path("/ibm,platform-facilities")) >>> family = PFO; >>> - else if (!strcmp(parent_node->full_name, "/vdevice")) >>> + else if (parent_node == of_find_node_by_path("/vdevice")) >>> family = VDEVICE; >> >> This leaks references to the looked up nodes. >> >> Both these nodes are defined in PAPR (our hypervisor spec), and both of >> them must have a device_type, either "ibm,platform-facilities" or >> "vdevice". >> >> Looking at the commit that added the code I don't see any particular >> reason it used the comparison against full_name, rather than using the >> device_type. >> >> So I'm inclined to do this instead: >> >> diff --git a/arch/powerpc/platforms/pseries/vio.c b/arch/powerpc/platforms/pseries/vio.c >> index 8a47f168476b..f26f906e6021 100644 >> --- a/arch/powerpc/platforms/pseries/vio.c >> +++ b/arch/powerpc/platforms/pseries/vio.c >> @@ -1357,9 +1357,9 @@ struct vio_dev *vio_register_device_node(struct device_node *of_node) >> */ >> parent_node = of_get_parent(of_node); >> if (parent_node) { >> - if (!strcmp(parent_node->full_name, "/ibm,platform-facilities")) >> + if (!strcmp(parent_node->type, "ibm,platform-facilities")) >> family = PFO; >> - else if (!strcmp(parent_node->full_name, "/vdevice")) >> + else if (!strcmp(parent_node->type, "vdevice")) >> family = VDEVICE; >> else { >> pr_warn("%s: parent(%s) of %s not recognized.\n", >> >> >> I've checked both Qemu and kvmtool add the device_type, and I'm fairly >> confident that PowerVM does too. Anyway I'll test it on all the machines >> I can find. > > Okay, do you want me to respin the patch or will you update it with I merged it. Should be in next today. cheers -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* [PATCH v2 3/5] powerpc: pseries: remove dlpar_attach_node dependency on full path 2017-08-21 15:16 ` Rob Herring ` (2 preceding siblings ...) (?) @ 2017-08-21 15:16 ` Rob Herring 2017-08-24 12:37 ` [v2, " Michael Ellerman -1 siblings, 1 reply; 53+ messages in thread From: Rob Herring @ 2017-08-21 15:16 UTC (permalink / raw) To: devicetree, Michael Ellerman, linuxppc-dev Cc: linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight In preparation to stop storing the full node path in full_name, remove the dependency on full_name from dlpar_attach_node(). Callers of dlpar_attach_node() already have the parent device_node, so just pass the parent node into dlpar_attach_node instead of the path. This avoids doing a lookup of the parent node by the path. Signed-off-by: Rob Herring <robh@kernel.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: linuxppc-dev@lists.ozlabs.org --- v2: - rebased to linux-next arch/powerpc/platforms/pseries/dlpar.c | 6 ++---- arch/powerpc/platforms/pseries/hotplug-cpu.c | 2 +- arch/powerpc/platforms/pseries/mobility.c | 2 +- arch/powerpc/platforms/pseries/pseries.h | 2 +- 4 files changed, 5 insertions(+), 7 deletions(-) diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c index 80b84c9c8509..783f36364690 100644 --- a/arch/powerpc/platforms/pseries/dlpar.c +++ b/arch/powerpc/platforms/pseries/dlpar.c @@ -254,13 +254,11 @@ struct device_node *dlpar_configure_connector(__be32 drc_index, return first_dn; } -int dlpar_attach_node(struct device_node *dn) +int dlpar_attach_node(struct device_node *dn, struct device_node *parent) { int rc; - dn->parent = pseries_of_derive_parent(dn->full_name); - if (IS_ERR(dn->parent)) - return PTR_ERR(dn->parent); + dn->parent = parent; rc = of_attach_node(dn); if (rc) { diff --git a/arch/powerpc/platforms/pseries/hotplug-cpu.c b/arch/powerpc/platforms/pseries/hotplug-cpu.c index 0a93093fbcef..b357f1ae0b0a 100644 --- a/arch/powerpc/platforms/pseries/hotplug-cpu.c +++ b/arch/powerpc/platforms/pseries/hotplug-cpu.c @@ -463,7 +463,7 @@ static ssize_t dlpar_cpu_add(u32 drc_index) return -EINVAL; } - rc = dlpar_attach_node(dn); + rc = dlpar_attach_node(dn, parent); if (rc) { saved_rc = rc; pr_warn("Failed to attach node %s, rc: %d, drc index: %x\n", diff --git a/arch/powerpc/platforms/pseries/mobility.c b/arch/powerpc/platforms/pseries/mobility.c index 2da4851eff99..210ce632d63e 100644 --- a/arch/powerpc/platforms/pseries/mobility.c +++ b/arch/powerpc/platforms/pseries/mobility.c @@ -229,7 +229,7 @@ static int add_dt_node(__be32 parent_phandle, __be32 drc_index) if (!dn) return -ENOENT; - rc = dlpar_attach_node(dn); + rc = dlpar_attach_node(dn, parent_dn); if (rc) dlpar_free_cc_nodes(dn); diff --git a/arch/powerpc/platforms/pseries/pseries.h b/arch/powerpc/platforms/pseries/pseries.h index 1361a9db534b..4470a3194311 100644 --- a/arch/powerpc/platforms/pseries/pseries.h +++ b/arch/powerpc/platforms/pseries/pseries.h @@ -46,7 +46,7 @@ extern void dlpar_free_cc_nodes(struct device_node *); extern void dlpar_free_cc_property(struct property *); extern struct device_node *dlpar_configure_connector(__be32, struct device_node *); -extern int dlpar_attach_node(struct device_node *); +extern int dlpar_attach_node(struct device_node *, struct device_node *); extern int dlpar_detach_node(struct device_node *); extern int dlpar_acquire_drc(u32 drc_index); extern int dlpar_release_drc(u32 drc_index); -- 2.11.0 ^ permalink raw reply related [flat|nested] 53+ messages in thread
* Re: [v2, 3/5] powerpc: pseries: remove dlpar_attach_node dependency on full path 2017-08-21 15:16 ` [PATCH v2 3/5] powerpc: pseries: remove dlpar_attach_node dependency on full path Rob Herring @ 2017-08-24 12:37 ` Michael Ellerman 0 siblings, 0 replies; 53+ messages in thread From: Michael Ellerman @ 2017-08-24 12:37 UTC (permalink / raw) To: Rob Herring, devicetree, linuxppc-dev Cc: David Laight, Frank Rowand, linux-kernel, Paul Mackerras On Mon, 2017-08-21 at 15:16:49 UTC, Rob Herring wrote: > In preparation to stop storing the full node path in full_name, remove the > dependency on full_name from dlpar_attach_node(). Callers of > dlpar_attach_node() already have the parent device_node, so just pass the > parent node into dlpar_attach_node instead of the path. This avoids doing > a lookup of the parent node by the path. > > Signed-off-by: Rob Herring <robh@kernel.org> > Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> > Cc: Paul Mackerras <paulus@samba.org> > Cc: Michael Ellerman <mpe@ellerman.id.au> > Cc: linuxppc-dev@lists.ozlabs.org Applied to powerpc next, thanks. https://git.kernel.org/powerpc/c/215ee763f8cb9a2912d411f96f6f67 cheers ^ permalink raw reply [flat|nested] 53+ messages in thread
* [PATCH v2 4/5] powerpc: pseries: only store the device node basename in full_name 2017-08-21 15:16 ` Rob Herring ` (3 preceding siblings ...) (?) @ 2017-08-21 15:16 ` Rob Herring 2017-10-03 9:26 ` Michael Ellerman -1 siblings, 1 reply; 53+ messages in thread From: Rob Herring @ 2017-08-21 15:16 UTC (permalink / raw) To: devicetree, Michael Ellerman, linuxppc-dev Cc: linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight With dependencies on full_name containing the entire device node path removed, stop storing the full_name in nodes created by dlpar_configure_connector() and pSeries_reconfig_add_node(). Signed-off-by: Rob Herring <robh@kernel.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: linuxppc-dev@lists.ozlabs.org --- v2: - Rework dlpar_parse_cc_node() to a single allocation and strcpy instead of kasprintf arch/powerpc/platforms/pseries/dlpar.c | 30 +++++++----------------------- arch/powerpc/platforms/pseries/reconfig.c | 2 +- 2 files changed, 8 insertions(+), 24 deletions(-) diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c index 783f36364690..31a0615aeea1 100644 --- a/arch/powerpc/platforms/pseries/dlpar.c +++ b/arch/powerpc/platforms/pseries/dlpar.c @@ -75,28 +75,17 @@ static struct property *dlpar_parse_cc_property(struct cc_workarea *ccwa) return prop; } -static struct device_node *dlpar_parse_cc_node(struct cc_workarea *ccwa, - const char *path) +static struct device_node *(dlpar_parse_cc_node)(struct cc_workarea *ccwa) { struct device_node *dn; - char *name; - - /* If parent node path is "/" advance path to NULL terminator to - * prevent double leading slashs in full_name. - */ - if (!path[1]) - path++; + const char *name = (const char *)ccwa + be32_to_cpu(ccwa->name_offset); - dn = kzalloc(sizeof(*dn), GFP_KERNEL); + dn = kzalloc(sizeof(*dn) + strlen(name) + 1, GFP_KERNEL); if (!dn) return NULL; - name = (char *)ccwa + be32_to_cpu(ccwa->name_offset); - dn->full_name = kasprintf(GFP_KERNEL, "%s/%s", path, name); - if (!dn->full_name) { - kfree(dn); - return NULL; - } + dn->full_name = (char *)(dn + 1); + strcpy((char *)dn->full_name, name); of_node_set_flag(dn, OF_DYNAMIC); of_node_init(dn); @@ -148,7 +137,6 @@ struct device_node *dlpar_configure_connector(__be32 drc_index, struct property *last_property = NULL; struct cc_workarea *ccwa; char *data_buf; - const char *parent_path = parent->full_name; int cc_token; int rc = -1; @@ -182,7 +170,7 @@ struct device_node *dlpar_configure_connector(__be32 drc_index, break; case NEXT_SIBLING: - dn = dlpar_parse_cc_node(ccwa, parent_path); + dn = dlpar_parse_cc_node(ccwa); if (!dn) goto cc_error; @@ -192,10 +180,7 @@ struct device_node *dlpar_configure_connector(__be32 drc_index, break; case NEXT_CHILD: - if (first_dn) - parent_path = last_dn->full_name; - - dn = dlpar_parse_cc_node(ccwa, parent_path); + dn = dlpar_parse_cc_node(ccwa); if (!dn) goto cc_error; @@ -226,7 +211,6 @@ struct device_node *dlpar_configure_connector(__be32 drc_index, case PREV_PARENT: last_dn = last_dn->parent; - parent_path = last_dn->parent->full_name; break; case CALL_AGAIN: diff --git a/arch/powerpc/platforms/pseries/reconfig.c b/arch/powerpc/platforms/pseries/reconfig.c index 296c188fd5ca..f24d8159c9e1 100644 --- a/arch/powerpc/platforms/pseries/reconfig.c +++ b/arch/powerpc/platforms/pseries/reconfig.c @@ -33,7 +33,7 @@ static int pSeries_reconfig_add_node(const char *path, struct property *proplist if (!np) goto out_err; - np->full_name = kstrdup(path, GFP_KERNEL); + np->full_name = kstrdup(kbasename(path), GFP_KERNEL); if (!np->full_name) goto out_err; -- 2.11.0 ^ permalink raw reply related [flat|nested] 53+ messages in thread
* Re: [PATCH v2 4/5] powerpc: pseries: only store the device node basename in full_name 2017-08-21 15:16 ` [PATCH v2 4/5] powerpc: pseries: only store the device node basename in full_name Rob Herring @ 2017-10-03 9:26 ` Michael Ellerman 0 siblings, 0 replies; 53+ messages in thread From: Michael Ellerman @ 2017-10-03 9:26 UTC (permalink / raw) To: Rob Herring, devicetree, linuxppc-dev Cc: linux-kernel, David Laight, Paul Mackerras, tyreld, nfont, Frank Rowand Hi Rob, Unfortunately this one has a bug, which only showed up after some stress testing. Rob Herring <robh@kernel.org> writes: > With dependencies on full_name containing the entire device node path > removed, stop storing the full_name in nodes created by > dlpar_configure_connector() and pSeries_reconfig_add_node(). > > Signed-off-by: Rob Herring <robh@kernel.org> > Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> > Cc: Paul Mackerras <paulus@samba.org> > Cc: Michael Ellerman <mpe@ellerman.id.au> > Cc: linuxppc-dev@lists.ozlabs.org > --- > v2: > - Rework dlpar_parse_cc_node() to a single allocation and strcpy instead > of kasprintf > > arch/powerpc/platforms/pseries/dlpar.c | 30 +++++++----------------------- > arch/powerpc/platforms/pseries/reconfig.c | 2 +- > 2 files changed, 8 insertions(+), 24 deletions(-) > > diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c > index 783f36364690..31a0615aeea1 100644 > --- a/arch/powerpc/platforms/pseries/dlpar.c > +++ b/arch/powerpc/platforms/pseries/dlpar.c > @@ -75,28 +75,17 @@ static struct property *dlpar_parse_cc_property(struct cc_workarea *ccwa) > return prop; > } > > -static struct device_node *dlpar_parse_cc_node(struct cc_workarea *ccwa, > - const char *path) > +static struct device_node *(dlpar_parse_cc_node)(struct cc_workarea *ccwa) > { > struct device_node *dn; > - char *name; > - > - /* If parent node path is "/" advance path to NULL terminator to > - * prevent double leading slashs in full_name. > - */ > - if (!path[1]) > - path++; > + const char *name = (const char *)ccwa + be32_to_cpu(ccwa->name_offset); > > - dn = kzalloc(sizeof(*dn), GFP_KERNEL); > + dn = kzalloc(sizeof(*dn) + strlen(name) + 1, GFP_KERNEL); > if (!dn) > return NULL; > > - name = (char *)ccwa + be32_to_cpu(ccwa->name_offset); > - dn->full_name = kasprintf(GFP_KERNEL, "%s/%s", path, name); > - if (!dn->full_name) { > - kfree(dn); > - return NULL; > - } > + dn->full_name = (char *)(dn + 1); > + strcpy((char *)dn->full_name, name); Allocating the full_name on the tail of the struct this way breaks the call to kfree(node->full_name) in of_node_release(). eg: [ 322.543581] Freeing node->full_name c0000007f4ed4090 [ 322.543583] ============================================================================= [ 322.543586] BUG kmalloc-192 (Tainted: G B ): Invalid object pointer 0xc0000007f4ed4090 [ 322.543588] ----------------------------------------------------------------------------- [ 322.543592] INFO: Slab 0xf000000001fd3b40 objects=292 used=72 fp=0xc0000007f4ed41a8 flags=0x13ffff800000101 [ 322.543595] CPU: 40 PID: 5 Comm: kworker/u192:0 Tainted: G B 4.14.0-rc2-gcc-6.3.1-next-20170929-dirty #429 [ 322.543600] Workqueue: pseries hotplug workque pseries_hp_work_fn [ 322.543602] Call Trace: [ 322.543605] [c0000003f769f660] [c000000000a6b710] dump_stack+0xb0/0xf0 (unreliable) [ 322.543609] [c0000003f769f6a0] [c00000000032cd5c] slab_err+0x9c/0xc0 [ 322.543612] [c0000003f769f790] [c00000000033440c] free_debug_processing+0x19c/0x490 [ 322.543615] [c0000003f769f860] [c0000000003349fc] __slab_free+0x2fc/0x4b0 [ 322.543619] [c0000003f769f960] [c0000000008f2484] of_node_release+0xe4/0x1a0 [ 322.543622] [c0000003f769fa00] [c000000000a71c04] kobject_put+0xd4/0x160 [ 322.543625] [c0000003f769fa80] [c0000000008f10c4] of_node_put+0x34/0x50 [ 322.543628] [c0000003f769fab0] [c0000000000c0248] dlpar_cpu_remove_by_index+0x108/0x130 [ 322.543631] [c0000003f769fb40] [c0000000000c166c] dlpar_cpu+0x27c/0x510 [ 322.543634] [c0000003f769fbf0] [c0000000000bb2d0] handle_dlpar_errorlog+0xc0/0x160 [ 322.543638] [c0000003f769fc60] [c0000000000bb3ac] pseries_hp_work_fn+0x3c/0xa0 [ 322.543641] [c0000003f769fc90] [c00000000012200c] process_one_work+0x2bc/0x560 [ 322.543645] [c0000003f769fd20] [c000000000122348] worker_thread+0x98/0x5d0 [ 322.543648] [c0000003f769fdc0] [c00000000012b4e4] kthread+0x164/0x1b0 [ 322.543651] [c0000003f769fe30] [c00000000000bae0] ret_from_kernel_thread+0x5c/0x7c The obvious fix is just to allocate it separately as before, eg ~=: diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c index 267b01e30ac0..8617d7e28bf7 100644 --- a/arch/powerpc/platforms/pseries/dlpar.c +++ b/arch/powerpc/platforms/pseries/dlpar.c @@ -80,11 +80,16 @@ static struct device_node *dlpar_parse_cc_node(struct cc_workarea *ccwa) struct device_node *dn; const char *name = (const char *)ccwa + be32_to_cpu(ccwa->name_offset); - dn = kzalloc(sizeof(*dn) + strlen(name) + 1, GFP_KERNEL); + dn = kzalloc(sizeof(*dn), GFP_KERNEL); if (!dn) return NULL; - dn->full_name = (char *)(dn + 1); + dn->full_name = kzalloc(strlen(name) + 1, GFP_KERNEL); + if (!dn->full_name) { + kfree(dn); + return NULL; + } + strcpy((char *)dn->full_name, name); of_node_set_flag(dn, OF_DYNAMIC); cheers ^ permalink raw reply related [flat|nested] 53+ messages in thread
* Re: [PATCH v2 4/5] powerpc: pseries: only store the device node basename in full_name @ 2017-10-03 9:26 ` Michael Ellerman 0 siblings, 0 replies; 53+ messages in thread From: Michael Ellerman @ 2017-10-03 9:26 UTC (permalink / raw) To: Rob Herring, devicetree, linuxppc-dev Cc: linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight, nfont Hi Rob, Unfortunately this one has a bug, which only showed up after some stress testing. Rob Herring <robh@kernel.org> writes: > With dependencies on full_name containing the entire device node path > removed, stop storing the full_name in nodes created by > dlpar_configure_connector() and pSeries_reconfig_add_node(). > > Signed-off-by: Rob Herring <robh@kernel.org> > Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> > Cc: Paul Mackerras <paulus@samba.org> > Cc: Michael Ellerman <mpe@ellerman.id.au> > Cc: linuxppc-dev@lists.ozlabs.org > --- > v2: > - Rework dlpar_parse_cc_node() to a single allocation and strcpy instead > of kasprintf > > arch/powerpc/platforms/pseries/dlpar.c | 30 +++++++----------------------- > arch/powerpc/platforms/pseries/reconfig.c | 2 +- > 2 files changed, 8 insertions(+), 24 deletions(-) > > diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c > index 783f36364690..31a0615aeea1 100644 > --- a/arch/powerpc/platforms/pseries/dlpar.c > +++ b/arch/powerpc/platforms/pseries/dlpar.c > @@ -75,28 +75,17 @@ static struct property *dlpar_parse_cc_property(struct cc_workarea *ccwa) > return prop; > } > > -static struct device_node *dlpar_parse_cc_node(struct cc_workarea *ccwa, > - const char *path) > +static struct device_node *(dlpar_parse_cc_node)(struct cc_workarea *ccwa) > { > struct device_node *dn; > - char *name; > - > - /* If parent node path is "/" advance path to NULL terminator to > - * prevent double leading slashs in full_name. > - */ > - if (!path[1]) > - path++; > + const char *name = (const char *)ccwa + be32_to_cpu(ccwa->name_offset); > > - dn = kzalloc(sizeof(*dn), GFP_KERNEL); > + dn = kzalloc(sizeof(*dn) + strlen(name) + 1, GFP_KERNEL); > if (!dn) > return NULL; > > - name = (char *)ccwa + be32_to_cpu(ccwa->name_offset); > - dn->full_name = kasprintf(GFP_KERNEL, "%s/%s", path, name); > - if (!dn->full_name) { > - kfree(dn); > - return NULL; > - } > + dn->full_name = (char *)(dn + 1); > + strcpy((char *)dn->full_name, name); Allocating the full_name on the tail of the struct this way breaks the call to kfree(node->full_name) in of_node_release(). eg: [ 322.543581] Freeing node->full_name c0000007f4ed4090 [ 322.543583] ============================================================================= [ 322.543586] BUG kmalloc-192 (Tainted: G B ): Invalid object pointer 0xc0000007f4ed4090 [ 322.543588] ----------------------------------------------------------------------------- [ 322.543592] INFO: Slab 0xf000000001fd3b40 objects=292 used=72 fp=0xc0000007f4ed41a8 flags=0x13ffff800000101 [ 322.543595] CPU: 40 PID: 5 Comm: kworker/u192:0 Tainted: G B 4.14.0-rc2-gcc-6.3.1-next-20170929-dirty #429 [ 322.543600] Workqueue: pseries hotplug workque pseries_hp_work_fn [ 322.543602] Call Trace: [ 322.543605] [c0000003f769f660] [c000000000a6b710] dump_stack+0xb0/0xf0 (unreliable) [ 322.543609] [c0000003f769f6a0] [c00000000032cd5c] slab_err+0x9c/0xc0 [ 322.543612] [c0000003f769f790] [c00000000033440c] free_debug_processing+0x19c/0x490 [ 322.543615] [c0000003f769f860] [c0000000003349fc] __slab_free+0x2fc/0x4b0 [ 322.543619] [c0000003f769f960] [c0000000008f2484] of_node_release+0xe4/0x1a0 [ 322.543622] [c0000003f769fa00] [c000000000a71c04] kobject_put+0xd4/0x160 [ 322.543625] [c0000003f769fa80] [c0000000008f10c4] of_node_put+0x34/0x50 [ 322.543628] [c0000003f769fab0] [c0000000000c0248] dlpar_cpu_remove_by_index+0x108/0x130 [ 322.543631] [c0000003f769fb40] [c0000000000c166c] dlpar_cpu+0x27c/0x510 [ 322.543634] [c0000003f769fbf0] [c0000000000bb2d0] handle_dlpar_errorlog+0xc0/0x160 [ 322.543638] [c0000003f769fc60] [c0000000000bb3ac] pseries_hp_work_fn+0x3c/0xa0 [ 322.543641] [c0000003f769fc90] [c00000000012200c] process_one_work+0x2bc/0x560 [ 322.543645] [c0000003f769fd20] [c000000000122348] worker_thread+0x98/0x5d0 [ 322.543648] [c0000003f769fdc0] [c00000000012b4e4] kthread+0x164/0x1b0 [ 322.543651] [c0000003f769fe30] [c00000000000bae0] ret_from_kernel_thread+0x5c/0x7c The obvious fix is just to allocate it separately as before, eg ~=: diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c index 267b01e30ac0..8617d7e28bf7 100644 --- a/arch/powerpc/platforms/pseries/dlpar.c +++ b/arch/powerpc/platforms/pseries/dlpar.c @@ -80,11 +80,16 @@ static struct device_node *dlpar_parse_cc_node(struct cc_workarea *ccwa) struct device_node *dn; const char *name = (const char *)ccwa + be32_to_cpu(ccwa->name_offset); - dn = kzalloc(sizeof(*dn) + strlen(name) + 1, GFP_KERNEL); + dn = kzalloc(sizeof(*dn), GFP_KERNEL); if (!dn) return NULL; - dn->full_name = (char *)(dn + 1); + dn->full_name = kzalloc(strlen(name) + 1, GFP_KERNEL); + if (!dn->full_name) { + kfree(dn); + return NULL; + } + strcpy((char *)dn->full_name, name); of_node_set_flag(dn, OF_DYNAMIC); cheers ^ permalink raw reply related [flat|nested] 53+ messages in thread
* Re: [PATCH v2 4/5] powerpc: pseries: only store the device node basename in full_name 2017-10-03 9:26 ` Michael Ellerman (?) @ 2017-10-03 18:44 ` Rob Herring 2017-10-04 12:37 ` Michael Ellerman -1 siblings, 1 reply; 53+ messages in thread From: Rob Herring @ 2017-10-03 18:44 UTC (permalink / raw) To: Michael Ellerman Cc: devicetree, linuxppc-dev, linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight, nfont, tyreld On Tue, Oct 3, 2017 at 4:26 AM, Michael Ellerman <mpe@ellerman.id.au> wrote: > Hi Rob, > > Unfortunately this one has a bug, which only showed up after some stress > testing. > > Rob Herring <robh@kernel.org> writes: >> With dependencies on full_name containing the entire device node path >> removed, stop storing the full_name in nodes created by >> dlpar_configure_connector() and pSeries_reconfig_add_node(). >> >> Signed-off-by: Rob Herring <robh@kernel.org> >> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> >> Cc: Paul Mackerras <paulus@samba.org> >> Cc: Michael Ellerman <mpe@ellerman.id.au> >> Cc: linuxppc-dev@lists.ozlabs.org >> --- >> v2: >> - Rework dlpar_parse_cc_node() to a single allocation and strcpy instead >> of kasprintf >> >> arch/powerpc/platforms/pseries/dlpar.c | 30 +++++++----------------------- >> arch/powerpc/platforms/pseries/reconfig.c | 2 +- >> 2 files changed, 8 insertions(+), 24 deletions(-) >> >> diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c >> index 783f36364690..31a0615aeea1 100644 >> --- a/arch/powerpc/platforms/pseries/dlpar.c >> +++ b/arch/powerpc/platforms/pseries/dlpar.c >> @@ -75,28 +75,17 @@ static struct property *dlpar_parse_cc_property(struct cc_workarea *ccwa) >> return prop; >> } >> >> -static struct device_node *dlpar_parse_cc_node(struct cc_workarea *ccwa, >> - const char *path) >> +static struct device_node *(dlpar_parse_cc_node)(struct cc_workarea *ccwa) >> { >> struct device_node *dn; >> - char *name; >> - >> - /* If parent node path is "/" advance path to NULL terminator to >> - * prevent double leading slashs in full_name. >> - */ >> - if (!path[1]) >> - path++; >> + const char *name = (const char *)ccwa + be32_to_cpu(ccwa->name_offset); >> >> - dn = kzalloc(sizeof(*dn), GFP_KERNEL); >> + dn = kzalloc(sizeof(*dn) + strlen(name) + 1, GFP_KERNEL); >> if (!dn) >> return NULL; >> >> - name = (char *)ccwa + be32_to_cpu(ccwa->name_offset); >> - dn->full_name = kasprintf(GFP_KERNEL, "%s/%s", path, name); >> - if (!dn->full_name) { >> - kfree(dn); >> - return NULL; >> - } >> + dn->full_name = (char *)(dn + 1); >> + strcpy((char *)dn->full_name, name); > > Allocating the full_name on the tail of the struct this way breaks the > call to kfree(node->full_name) in of_node_release(). > > eg: > > > [ 322.543581] Freeing node->full_name c0000007f4ed4090 > [ 322.543583] ============================================================================= > [ 322.543586] BUG kmalloc-192 (Tainted: G B ): Invalid object pointer 0xc0000007f4ed4090 > [ 322.543588] ----------------------------------------------------------------------------- > > [ 322.543592] INFO: Slab 0xf000000001fd3b40 objects=292 used=72 fp=0xc0000007f4ed41a8 flags=0x13ffff800000101 > [ 322.543595] CPU: 40 PID: 5 Comm: kworker/u192:0 Tainted: G B 4.14.0-rc2-gcc-6.3.1-next-20170929-dirty #429 > [ 322.543600] Workqueue: pseries hotplug workque pseries_hp_work_fn > [ 322.543602] Call Trace: > [ 322.543605] [c0000003f769f660] [c000000000a6b710] dump_stack+0xb0/0xf0 (unreliable) > [ 322.543609] [c0000003f769f6a0] [c00000000032cd5c] slab_err+0x9c/0xc0 > [ 322.543612] [c0000003f769f790] [c00000000033440c] free_debug_processing+0x19c/0x490 > [ 322.543615] [c0000003f769f860] [c0000000003349fc] __slab_free+0x2fc/0x4b0 > [ 322.543619] [c0000003f769f960] [c0000000008f2484] of_node_release+0xe4/0x1a0 > [ 322.543622] [c0000003f769fa00] [c000000000a71c04] kobject_put+0xd4/0x160 > [ 322.543625] [c0000003f769fa80] [c0000000008f10c4] of_node_put+0x34/0x50 > [ 322.543628] [c0000003f769fab0] [c0000000000c0248] dlpar_cpu_remove_by_index+0x108/0x130 > [ 322.543631] [c0000003f769fb40] [c0000000000c166c] dlpar_cpu+0x27c/0x510 > [ 322.543634] [c0000003f769fbf0] [c0000000000bb2d0] handle_dlpar_errorlog+0xc0/0x160 > [ 322.543638] [c0000003f769fc60] [c0000000000bb3ac] pseries_hp_work_fn+0x3c/0xa0 > [ 322.543641] [c0000003f769fc90] [c00000000012200c] process_one_work+0x2bc/0x560 > [ 322.543645] [c0000003f769fd20] [c000000000122348] worker_thread+0x98/0x5d0 > [ 322.543648] [c0000003f769fdc0] [c00000000012b4e4] kthread+0x164/0x1b0 > [ 322.543651] [c0000003f769fe30] [c00000000000bae0] ret_from_kernel_thread+0x5c/0x7c > > > The obvious fix is just to allocate it separately as before, eg ~=: Yes, I'll go back to doing 2 allocs like v1, but using kstrdup as was also pointed out. Rob ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 4/5] powerpc: pseries: only store the device node basename in full_name @ 2017-10-04 12:37 ` Michael Ellerman 0 siblings, 0 replies; 53+ messages in thread From: Michael Ellerman @ 2017-10-04 12:37 UTC (permalink / raw) To: Rob Herring Cc: devicetree, linuxppc-dev, linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight, nfont, tyreld Rob Herring <robh@kernel.org> writes: > On Tue, Oct 3, 2017 at 4:26 AM, Michael Ellerman <mpe@ellerman.id.au> wrote: ... >> >> The obvious fix is just to allocate it separately as before, eg ~=: > > Yes, I'll go back to doing 2 allocs like v1, but using kstrdup as was > also pointed out. Yeah that would be better still I guess. cheers ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 4/5] powerpc: pseries: only store the device node basename in full_name @ 2017-10-04 12:37 ` Michael Ellerman 0 siblings, 0 replies; 53+ messages in thread From: Michael Ellerman @ 2017-10-04 12:37 UTC (permalink / raw) To: Rob Herring Cc: devicetree@vger.kernel.org, linuxppc-dev, linux-kernel@vger.kernel.org, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight, nfont@linux.vnet.ibm.com, tyreld@linux.vnet.ibm.com Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> writes: > On Tue, Oct 3, 2017 at 4:26 AM, Michael Ellerman <mpe-Gsx/Oe8HsFggBc27wqDAHg@public.gmane.org> wrote: ... >> >> The obvious fix is just to allocate it separately as before, eg ~=: > > Yes, I'll go back to doing 2 allocs like v1, but using kstrdup as was > also pointed out. Yeah that would be better still I guess. cheers -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* [PATCH v2 5/5] of/fdt: only store the device node basename in full_name 2017-08-21 15:16 ` Rob Herring ` (4 preceding siblings ...) (?) @ 2017-08-21 15:16 ` Rob Herring 2017-10-17 21:32 ` Alan Tull 2017-11-28 13:13 ` Geert Uytterhoeven -1 siblings, 2 replies; 53+ messages in thread From: Rob Herring @ 2017-08-21 15:16 UTC (permalink / raw) To: devicetree, Michael Ellerman, linuxppc-dev Cc: linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight With dependencies on a statically allocated full path name converted to use %pOF format specifier, we can store just the basename of node, and the unflattening of the FDT can be simplified. This commit will affect the remaining users of full_name. After analyzing these users, the remaining cases should only change some print messages. The main users of full_name are providing a name for struct resource. The resource names shouldn't be important other than providing /proc/iomem names. We no longer distinguish between pre and post 0x10 dtb formats as either a full path or basename will work. However, less than 0x10 formats have been broken since the conversion to use libfdt (and no one has cared). The conversion of the unflattening code to be non-recursive also broke pre 0x10 formats as the populate_node function would return 0 in that case. Signed-off-by: Rob Herring <robh@kernel.org> --- v2: - rebase to linux-next drivers/of/fdt.c | 69 +++++++++----------------------------------------------- 1 file changed, 11 insertions(+), 58 deletions(-) diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c index ce30c9a588a4..27c535af0be8 100644 --- a/drivers/of/fdt.c +++ b/drivers/of/fdt.c @@ -266,74 +266,32 @@ static void populate_properties(const void *blob, *pprev = NULL; } -static unsigned int populate_node(const void *blob, - int offset, - void **mem, - struct device_node *dad, - unsigned int fpsize, - struct device_node **pnp, - bool dryrun) +static bool populate_node(const void *blob, + int offset, + void **mem, + struct device_node *dad, + struct device_node **pnp, + bool dryrun) { struct device_node *np; const char *pathp; unsigned int l, allocl; - int new_format = 0; pathp = fdt_get_name(blob, offset, &l); if (!pathp) { *pnp = NULL; - return 0; + return false; } allocl = ++l; - /* version 0x10 has a more compact unit name here instead of the full - * path. we accumulate the full path size using "fpsize", we'll rebuild - * it later. We detect this because the first character of the name is - * not '/'. - */ - if ((*pathp) != '/') { - new_format = 1; - if (fpsize == 0) { - /* root node: special case. fpsize accounts for path - * plus terminating zero. root node only has '/', so - * fpsize should be 2, but we want to avoid the first - * level nodes to have two '/' so we use fpsize 1 here - */ - fpsize = 1; - allocl = 2; - l = 1; - pathp = ""; - } else { - /* account for '/' and path size minus terminal 0 - * already in 'l' - */ - fpsize += l; - allocl = fpsize; - } - } - np = unflatten_dt_alloc(mem, sizeof(struct device_node) + allocl, __alignof__(struct device_node)); if (!dryrun) { char *fn; of_node_init(np); np->full_name = fn = ((char *)np) + sizeof(*np); - if (new_format) { - /* rebuild full path for new format */ - if (dad && dad->parent) { - strcpy(fn, dad->full_name); -#ifdef DEBUG - if ((strlen(fn) + l + 1) != allocl) { - pr_debug("%s: p: %d, l: %d, a: %d\n", - pathp, (int)strlen(fn), - l, allocl); - } -#endif - fn += strlen(fn); - } - *(fn++) = '/'; - } + memcpy(fn, pathp, l); if (dad != NULL) { @@ -355,7 +313,7 @@ static unsigned int populate_node(const void *blob, } *pnp = np; - return fpsize; + return true; } static void reverse_nodes(struct device_node *parent) @@ -399,7 +357,6 @@ static int unflatten_dt_nodes(const void *blob, struct device_node *root; int offset = 0, depth = 0, initial_depth = 0; #define FDT_MAX_DEPTH 64 - unsigned int fpsizes[FDT_MAX_DEPTH]; struct device_node *nps[FDT_MAX_DEPTH]; void *base = mem; bool dryrun = !base; @@ -418,7 +375,6 @@ static int unflatten_dt_nodes(const void *blob, depth = initial_depth = 1; root = dad; - fpsizes[depth] = dad ? strlen(of_node_full_name(dad)) : 0; nps[depth] = dad; for (offset = 0; @@ -427,11 +383,8 @@ static int unflatten_dt_nodes(const void *blob, if (WARN_ON_ONCE(depth >= FDT_MAX_DEPTH)) continue; - fpsizes[depth+1] = populate_node(blob, offset, &mem, - nps[depth], - fpsizes[depth], - &nps[depth+1], dryrun); - if (!fpsizes[depth+1]) + if (!populate_node(blob, offset, &mem, nps[depth], + &nps[depth+1], dryrun)) return mem - base; if (!dryrun && nodepp && !*nodepp) -- 2.11.0 ^ permalink raw reply related [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-17 21:32 ` Alan Tull 0 siblings, 0 replies; 53+ messages in thread From: Alan Tull @ 2017-10-17 21:32 UTC (permalink / raw) To: Rob Herring Cc: devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: Hi Rob, > With dependencies on a statically allocated full path name converted to > use %pOF format specifier, we can store just the basename of node, and > the unflattening of the FDT can be simplified. > > This commit will affect the remaining users of full_name. After > analyzing these users, the remaining cases should only change some print > messages. The main users of full_name are providing a name for struct > resource. The resource names shouldn't be important other than providing > /proc/iomem names. > > We no longer distinguish between pre and post 0x10 dtb formats as either > a full path or basename will work. However, less than 0x10 formats have > been broken since the conversion to use libfdt (and no one has cared). > The conversion of the unflattening code to be non-recursive also broke > pre 0x10 formats as the populate_node function would return 0 in that > case. > > Signed-off-by: Rob Herring <robh@kernel.org> > --- > v2: > - rebase to linux-next > > drivers/of/fdt.c | 69 +++++++++----------------------------------------------- > 1 file changed, 11 insertions(+), 58 deletions(-) I've just updated to the latest next branch and am finding problems applying overlays. Reverting this commit alleviates things. The errors I get are: [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 [ 88.513447] OF: overlay: apply failed '/__symbols__' [ 88.518423] create_overlay: Failed to create overlay (err=-12) My branch also includes Pantelis' patch that turns on symbols [1]. Alan [1] https://github.com/pantoniou/linux-beagle-track-mainline/commit/486c87d450f2895a766c6097328d0c6538ec7a31 ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-17 21:32 ` Alan Tull 0 siblings, 0 replies; 53+ messages in thread From: Alan Tull @ 2017-10-17 21:32 UTC (permalink / raw) To: Rob Herring Cc: devicetree-u79uwXL29TY76Z2rM5mHXA, Michael Ellerman, linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ, linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: Hi Rob, > With dependencies on a statically allocated full path name converted to > use %pOF format specifier, we can store just the basename of node, and > the unflattening of the FDT can be simplified. > > This commit will affect the remaining users of full_name. After > analyzing these users, the remaining cases should only change some print > messages. The main users of full_name are providing a name for struct > resource. The resource names shouldn't be important other than providing > /proc/iomem names. > > We no longer distinguish between pre and post 0x10 dtb formats as either > a full path or basename will work. However, less than 0x10 formats have > been broken since the conversion to use libfdt (and no one has cared). > The conversion of the unflattening code to be non-recursive also broke > pre 0x10 formats as the populate_node function would return 0 in that > case. > > Signed-off-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> > --- > v2: > - rebase to linux-next > > drivers/of/fdt.c | 69 +++++++++----------------------------------------------- > 1 file changed, 11 insertions(+), 58 deletions(-) I've just updated to the latest next branch and am finding problems applying overlays. Reverting this commit alleviates things. The errors I get are: [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 [ 88.513447] OF: overlay: apply failed '/__symbols__' [ 88.518423] create_overlay: Failed to create overlay (err=-12) My branch also includes Pantelis' patch that turns on symbols [1]. Alan [1] https://github.com/pantoniou/linux-beagle-track-mainline/commit/486c87d450f2895a766c6097328d0c6538ec7a31 -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name 2017-10-17 21:32 ` Alan Tull (?) @ 2017-10-17 21:46 ` Rob Herring 2017-10-17 23:51 ` Frank Rowand -1 siblings, 1 reply; 53+ messages in thread From: Rob Herring @ 2017-10-17 21:46 UTC (permalink / raw) To: Alan Tull Cc: devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: > On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: > > Hi Rob, > >> With dependencies on a statically allocated full path name converted to >> use %pOF format specifier, we can store just the basename of node, and >> the unflattening of the FDT can be simplified. >> >> This commit will affect the remaining users of full_name. After >> analyzing these users, the remaining cases should only change some print >> messages. The main users of full_name are providing a name for struct >> resource. The resource names shouldn't be important other than providing >> /proc/iomem names. >> >> We no longer distinguish between pre and post 0x10 dtb formats as either >> a full path or basename will work. However, less than 0x10 formats have >> been broken since the conversion to use libfdt (and no one has cared). >> The conversion of the unflattening code to be non-recursive also broke >> pre 0x10 formats as the populate_node function would return 0 in that >> case. >> >> Signed-off-by: Rob Herring <robh@kernel.org> >> --- >> v2: >> - rebase to linux-next >> >> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >> 1 file changed, 11 insertions(+), 58 deletions(-) > > I've just updated to the latest next branch and am finding problems > applying overlays. Reverting this commit alleviates things. The > errors I get are: > > [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 > [ 88.513447] OF: overlay: apply failed '/__symbols__' > [ 88.518423] create_overlay: Failed to create overlay (err=-12) Frank's series with overlay updates should fix this. Rob ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-17 23:51 ` Frank Rowand 0 siblings, 0 replies; 53+ messages in thread From: Frank Rowand @ 2017-10-17 23:51 UTC (permalink / raw) To: Rob Herring, Alan Tull Cc: devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight On 10/17/17 14:46, Rob Herring wrote: > On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: >> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: >> >> Hi Rob, >> >>> With dependencies on a statically allocated full path name converted to >>> use %pOF format specifier, we can store just the basename of node, and >>> the unflattening of the FDT can be simplified. >>> >>> This commit will affect the remaining users of full_name. After >>> analyzing these users, the remaining cases should only change some print >>> messages. The main users of full_name are providing a name for struct >>> resource. The resource names shouldn't be important other than providing >>> /proc/iomem names. >>> >>> We no longer distinguish between pre and post 0x10 dtb formats as either >>> a full path or basename will work. However, less than 0x10 formats have >>> been broken since the conversion to use libfdt (and no one has cared). >>> The conversion of the unflattening code to be non-recursive also broke >>> pre 0x10 formats as the populate_node function would return 0 in that >>> case. >>> >>> Signed-off-by: Rob Herring <robh@kernel.org> >>> --- >>> v2: >>> - rebase to linux-next >>> >>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >>> 1 file changed, 11 insertions(+), 58 deletions(-) >> >> I've just updated to the latest next branch and am finding problems >> applying overlays. Reverting this commit alleviates things. The >> errors I get are: >> >> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >> [ 88.513447] OF: overlay: apply failed '/__symbols__' >> [ 88.518423] create_overlay: Failed to create overlay (err=-12) > > Frank's series with overlay updates should fix this. Yes, it does: [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name -Frank ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-17 23:51 ` Frank Rowand 0 siblings, 0 replies; 53+ messages in thread From: Frank Rowand @ 2017-10-17 23:51 UTC (permalink / raw) To: Rob Herring, Alan Tull Cc: devicetree-u79uwXL29TY76Z2rM5mHXA, Michael Ellerman, linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight On 10/17/17 14:46, Rob Herring wrote: > On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >> >> Hi Rob, >> >>> With dependencies on a statically allocated full path name converted to >>> use %pOF format specifier, we can store just the basename of node, and >>> the unflattening of the FDT can be simplified. >>> >>> This commit will affect the remaining users of full_name. After >>> analyzing these users, the remaining cases should only change some print >>> messages. The main users of full_name are providing a name for struct >>> resource. The resource names shouldn't be important other than providing >>> /proc/iomem names. >>> >>> We no longer distinguish between pre and post 0x10 dtb formats as either >>> a full path or basename will work. However, less than 0x10 formats have >>> been broken since the conversion to use libfdt (and no one has cared). >>> The conversion of the unflattening code to be non-recursive also broke >>> pre 0x10 formats as the populate_node function would return 0 in that >>> case. >>> >>> Signed-off-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> >>> --- >>> v2: >>> - rebase to linux-next >>> >>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >>> 1 file changed, 11 insertions(+), 58 deletions(-) >> >> I've just updated to the latest next branch and am finding problems >> applying overlays. Reverting this commit alleviates things. The >> errors I get are: >> >> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >> [ 88.513447] OF: overlay: apply failed '/__symbols__' >> [ 88.518423] create_overlay: Failed to create overlay (err=-12) > > Frank's series with overlay updates should fix this. Yes, it does: [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name -Frank -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-18 15:12 ` Alan Tull 0 siblings, 0 replies; 53+ messages in thread From: Alan Tull @ 2017-10-18 15:12 UTC (permalink / raw) To: Frank Rowand, Pantelis Antoniou Cc: Rob Herring, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list@gmail.com> wrote: > On 10/17/17 14:46, Rob Herring wrote: >> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: >>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: >>> >>> Hi Rob, >>> >>>> With dependencies on a statically allocated full path name converted to >>>> use %pOF format specifier, we can store just the basename of node, and >>>> the unflattening of the FDT can be simplified. >>>> >>>> This commit will affect the remaining users of full_name. After >>>> analyzing these users, the remaining cases should only change some print >>>> messages. The main users of full_name are providing a name for struct >>>> resource. The resource names shouldn't be important other than providing >>>> /proc/iomem names. >>>> >>>> We no longer distinguish between pre and post 0x10 dtb formats as either >>>> a full path or basename will work. However, less than 0x10 formats have >>>> been broken since the conversion to use libfdt (and no one has cared). >>>> The conversion of the unflattening code to be non-recursive also broke >>>> pre 0x10 formats as the populate_node function would return 0 in that >>>> case. >>>> >>>> Signed-off-by: Rob Herring <robh@kernel.org> >>>> --- >>>> v2: >>>> - rebase to linux-next >>>> >>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >>>> 1 file changed, 11 insertions(+), 58 deletions(-) >>> >>> I've just updated to the latest next branch and am finding problems >>> applying overlays. Reverting this commit alleviates things. The >>> errors I get are: >>> >>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >> >> Frank's series with overlay updates should fix this. > > Yes, it does: > > [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name Thanks for the fast response. I fetched the dt/next branch to test this but there are sufficient changes that Pantelis' "OF: DT-Overlay configfs interface (v7)" is broken now. I've been adding that downstream since 4.4. We're using it as an interface for applying overlays to program FPGAs. If we fix it again, is there any chance that can go upstream now? Alan ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-18 15:12 ` Alan Tull 0 siblings, 0 replies; 53+ messages in thread From: Alan Tull @ 2017-10-18 15:12 UTC (permalink / raw) To: Frank Rowand, Pantelis Antoniou Cc: Rob Herring, devicetree-u79uwXL29TY76Z2rM5mHXA, Michael Ellerman, linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga-u79uwXL29TY76Z2rM5mHXA On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote: > On 10/17/17 14:46, Rob Herring wrote: >> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >>> >>> Hi Rob, >>> >>>> With dependencies on a statically allocated full path name converted to >>>> use %pOF format specifier, we can store just the basename of node, and >>>> the unflattening of the FDT can be simplified. >>>> >>>> This commit will affect the remaining users of full_name. After >>>> analyzing these users, the remaining cases should only change some print >>>> messages. The main users of full_name are providing a name for struct >>>> resource. The resource names shouldn't be important other than providing >>>> /proc/iomem names. >>>> >>>> We no longer distinguish between pre and post 0x10 dtb formats as either >>>> a full path or basename will work. However, less than 0x10 formats have >>>> been broken since the conversion to use libfdt (and no one has cared). >>>> The conversion of the unflattening code to be non-recursive also broke >>>> pre 0x10 formats as the populate_node function would return 0 in that >>>> case. >>>> >>>> Signed-off-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> >>>> --- >>>> v2: >>>> - rebase to linux-next >>>> >>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >>>> 1 file changed, 11 insertions(+), 58 deletions(-) >>> >>> I've just updated to the latest next branch and am finding problems >>> applying overlays. Reverting this commit alleviates things. The >>> errors I get are: >>> >>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >> >> Frank's series with overlay updates should fix this. > > Yes, it does: > > [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name Thanks for the fast response. I fetched the dt/next branch to test this but there are sufficient changes that Pantelis' "OF: DT-Overlay configfs interface (v7)" is broken now. I've been adding that downstream since 4.4. We're using it as an interface for applying overlays to program FPGAs. If we fix it again, is there any chance that can go upstream now? Alan -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name 2017-10-18 15:12 ` Alan Tull (?) @ 2017-10-18 15:44 ` Rob Herring 2017-10-18 15:53 ` Pantelis Antoniou -1 siblings, 1 reply; 53+ messages in thread From: Rob Herring @ 2017-10-18 15:44 UTC (permalink / raw) To: Alan Tull Cc: Frank Rowand, Pantelis Antoniou, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull@kernel.org> wrote: > On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list@gmail.com> wrote: >> On 10/17/17 14:46, Rob Herring wrote: >>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: >>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: >>>> >>>> Hi Rob, >>>> >>>>> With dependencies on a statically allocated full path name converted to >>>>> use %pOF format specifier, we can store just the basename of node, and >>>>> the unflattening of the FDT can be simplified. >>>>> >>>>> This commit will affect the remaining users of full_name. After >>>>> analyzing these users, the remaining cases should only change some print >>>>> messages. The main users of full_name are providing a name for struct >>>>> resource. The resource names shouldn't be important other than providing >>>>> /proc/iomem names. >>>>> >>>>> We no longer distinguish between pre and post 0x10 dtb formats as either >>>>> a full path or basename will work. However, less than 0x10 formats have >>>>> been broken since the conversion to use libfdt (and no one has cared). >>>>> The conversion of the unflattening code to be non-recursive also broke >>>>> pre 0x10 formats as the populate_node function would return 0 in that >>>>> case. >>>>> >>>>> Signed-off-by: Rob Herring <robh@kernel.org> >>>>> --- >>>>> v2: >>>>> - rebase to linux-next >>>>> >>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >>>> >>>> I've just updated to the latest next branch and am finding problems >>>> applying overlays. Reverting this commit alleviates things. The >>>> errors I get are: >>>> >>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >>> >>> Frank's series with overlay updates should fix this. >> >> Yes, it does: >> >> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name > > Thanks for the fast response. I fetched the dt/next branch to test > this but there are sufficient changes that Pantelis' "OF: DT-Overlay > configfs interface (v7)" is broken now. I've been adding that > downstream since 4.4. We're using it as an interface for applying > overlays to program FPGAs. If we fix it again, is there any chance > that can go upstream now? With a drive-by posting once every few years, no. The issue remains that the kernel is not really setup to deal with any random property or node to be changed at any point in run-time. I think there needs to be some restrictions around what the overlays can touch. We can't have it be wide open and then lock things down later and break users. One example of what you could do is you can only add sub-trees to whitelisted nodes. That's probably acceptable for your usecase. Rob ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name 2017-10-18 15:44 ` Rob Herring @ 2017-10-18 15:53 ` Pantelis Antoniou 0 siblings, 0 replies; 53+ messages in thread From: Pantelis Antoniou @ 2017-10-18 15:53 UTC (permalink / raw) To: Rob Herring Cc: Alan Tull, Frank Rowand, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: > On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull@kernel.org> wrote: > > On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list@gmail.com> wrote: > >> On 10/17/17 14:46, Rob Herring wrote: > >>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: > >>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: > >>>> > >>>> Hi Rob, > >>>> > >>>>> With dependencies on a statically allocated full path name converted to > >>>>> use %pOF format specifier, we can store just the basename of node, and > >>>>> the unflattening of the FDT can be simplified. > >>>>> > >>>>> This commit will affect the remaining users of full_name. After > >>>>> analyzing these users, the remaining cases should only change some print > >>>>> messages. The main users of full_name are providing a name for struct > >>>>> resource. The resource names shouldn't be important other than providing > >>>>> /proc/iomem names. > >>>>> > >>>>> We no longer distinguish between pre and post 0x10 dtb formats as either > >>>>> a full path or basename will work. However, less than 0x10 formats have > >>>>> been broken since the conversion to use libfdt (and no one has cared). > >>>>> The conversion of the unflattening code to be non-recursive also broke > >>>>> pre 0x10 formats as the populate_node function would return 0 in that > >>>>> case. > >>>>> > >>>>> Signed-off-by: Rob Herring <robh@kernel.org> > >>>>> --- > >>>>> v2: > >>>>> - rebase to linux-next > >>>>> > >>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- > >>>>> 1 file changed, 11 insertions(+), 58 deletions(-) > >>>> > >>>> I've just updated to the latest next branch and am finding problems > >>>> applying overlays. Reverting this commit alleviates things. The > >>>> errors I get are: > >>>> > >>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 > >>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' > >>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) > >>> > >>> Frank's series with overlay updates should fix this. > >> > >> Yes, it does: > >> > >> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name > > > > Thanks for the fast response. I fetched the dt/next branch to test > > this but there are sufficient changes that Pantelis' "OF: DT-Overlay > > configfs interface (v7)" is broken now. I've been adding that > > downstream since 4.4. We're using it as an interface for applying > > overlays to program FPGAs. If we fix it again, is there any chance > > that can go upstream now? > > With a drive-by posting once every few years, no. > I take offense to that. There's nothing changed in the patch for years. Reposting the same patch without changes would achieve nothing. > The issue remains that the kernel is not really setup to deal with any > random property or node to be changed at any point in run-time. I > think there needs to be some restrictions around what the overlays can > touch. We can't have it be wide open and then lock things down later > and break users. One example of what you could do is you can only add > sub-trees to whitelisted nodes. That's probably acceptable for your > usecase. > Defining what can and what cannot be changed is not as trivial as a list of white-listed nodes. In some cases there is a whole node hierarchy being inserted (like in a FPGA). In others, it's merely changing a status property to "okay" and a few device parameters. The real issue is that the kernel has no way to verify that a given device tree, either at boot time or at overlay application time, is correct. When the tree is wrong at boot-time you'll hang (if you're lucky). If the tree is wrong at run-time you'll get some into some unidentified funky state. Finally what is, and what is not 'correct' is not for the kernel to decide arbitrarily, it's a matter of policy, different for each use-case. > Rob Regards -- Pantelis ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-18 15:53 ` Pantelis Antoniou 0 siblings, 0 replies; 53+ messages in thread From: Pantelis Antoniou @ 2017-10-18 15:53 UTC (permalink / raw) To: Rob Herring Cc: Alan Tull, Frank Rowand, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: > On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull@kernel.org> wrote: > > On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list@gmail.com> wrote: > >> On 10/17/17 14:46, Rob Herring wrote: > >>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: > >>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: > >>>> > >>>> Hi Rob, > >>>> > >>>>> With dependencies on a statically allocated full path name converted to > >>>>> use %pOF format specifier, we can store just the basename of node, and > >>>>> the unflattening of the FDT can be simplified. > >>>>> > >>>>> This commit will affect the remaining users of full_name. After > >>>>> analyzing these users, the remaining cases should only change some print > >>>>> messages. The main users of full_name are providing a name for struct > >>>>> resource. The resource names shouldn't be important other than providing > >>>>> /proc/iomem names. > >>>>> > >>>>> We no longer distinguish between pre and post 0x10 dtb formats as either > >>>>> a full path or basename will work. However, less than 0x10 formats have > >>>>> been broken since the conversion to use libfdt (and no one has cared). > >>>>> The conversion of the unflattening code to be non-recursive also broke > >>>>> pre 0x10 formats as the populate_node function would return 0 in that > >>>>> case. > >>>>> > >>>>> Signed-off-by: Rob Herring <robh@kernel.org> > >>>>> --- > >>>>> v2: > >>>>> - rebase to linux-next > >>>>> > >>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- > >>>>> 1 file changed, 11 insertions(+), 58 deletions(-) > >>>> > >>>> I've just updated to the latest next branch and am finding problems > >>>> applying overlays. Reverting this commit alleviates things. The > >>>> errors I get are: > >>>> > >>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 > >>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' > >>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) > >>> > >>> Frank's series with overlay updates should fix this. > >> > >> Yes, it does: > >> > >> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name > > > > Thanks for the fast response. I fetched the dt/next branch to test > > this but there are sufficient changes that Pantelis' "OF: DT-Overlay > > configfs interface (v7)" is broken now. I've been adding that > > downstream since 4.4. We're using it as an interface for applying > > overlays to program FPGAs. If we fix it again, is there any chance > > that can go upstream now? > > With a drive-by posting once every few years, no. > I take offense to that. There's nothing changed in the patch for years. Reposting the same patch without changes would achieve nothing. > The issue remains that the kernel is not really setup to deal with any > random property or node to be changed at any point in run-time. I > think there needs to be some restrictions around what the overlays can > touch. We can't have it be wide open and then lock things down later > and break users. One example of what you could do is you can only add > sub-trees to whitelisted nodes. That's probably acceptable for your > usecase. > Defining what can and what cannot be changed is not as trivial as a list of white-listed nodes. In some cases there is a whole node hierarchy being inserted (like in a FPGA). In others, it's merely changing a status property to "okay" and a few device parameters. The real issue is that the kernel has no way to verify that a given device tree, either at boot time or at overlay application time, is correct. When the tree is wrong at boot-time you'll hang (if you're lucky). If the tree is wrong at run-time you'll get some into some unidentified funky state. Finally what is, and what is not 'correct' is not for the kernel to decide arbitrarily, it's a matter of policy, different for each use-case. > Rob Regards -- Pantelis ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name 2017-10-18 15:53 ` Pantelis Antoniou (?) @ 2017-10-18 18:30 ` Rob Herring 2017-10-18 21:46 ` Frank Rowand 2017-10-19 8:51 ` Pantelis Antoniou -1 siblings, 2 replies; 53+ messages in thread From: Rob Herring @ 2017-10-18 18:30 UTC (permalink / raw) To: Pantelis Antoniou Cc: Alan Tull, Frank Rowand, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou <pantelis.antoniou@konsulko.com> wrote: > On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: >> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull@kernel.org> wrote: >> > On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list@gmail.com> wrote: >> >> On 10/17/17 14:46, Rob Herring wrote: >> >>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: >> >>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: >> >>>> >> >>>> Hi Rob, >> >>>> >> >>>>> With dependencies on a statically allocated full path name converted to >> >>>>> use %pOF format specifier, we can store just the basename of node, and >> >>>>> the unflattening of the FDT can be simplified. >> >>>>> >> >>>>> This commit will affect the remaining users of full_name. After >> >>>>> analyzing these users, the remaining cases should only change some print >> >>>>> messages. The main users of full_name are providing a name for struct >> >>>>> resource. The resource names shouldn't be important other than providing >> >>>>> /proc/iomem names. >> >>>>> >> >>>>> We no longer distinguish between pre and post 0x10 dtb formats as either >> >>>>> a full path or basename will work. However, less than 0x10 formats have >> >>>>> been broken since the conversion to use libfdt (and no one has cared). >> >>>>> The conversion of the unflattening code to be non-recursive also broke >> >>>>> pre 0x10 formats as the populate_node function would return 0 in that >> >>>>> case. >> >>>>> >> >>>>> Signed-off-by: Rob Herring <robh@kernel.org> >> >>>>> --- >> >>>>> v2: >> >>>>> - rebase to linux-next >> >>>>> >> >>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >> >>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >> >>>> >> >>>> I've just updated to the latest next branch and am finding problems >> >>>> applying overlays. Reverting this commit alleviates things. The >> >>>> errors I get are: >> >>>> >> >>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >> >>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >> >>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >> >>> >> >>> Frank's series with overlay updates should fix this. >> >> >> >> Yes, it does: >> >> >> >> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name >> > >> > Thanks for the fast response. I fetched the dt/next branch to test >> > this but there are sufficient changes that Pantelis' "OF: DT-Overlay >> > configfs interface (v7)" is broken now. I've been adding that >> > downstream since 4.4. We're using it as an interface for applying >> > overlays to program FPGAs. If we fix it again, is there any chance >> > that can go upstream now? >> >> With a drive-by posting once every few years, no. >> > > I take offense to that. There's nothing changed in the patch for years. > Reposting the same patch without changes would achieve nothing. Are you still expecting review comments on it or something? Furthermore, If something is posted infrequently, then I'm not inclined to comment or care if the next posting is going to be after I forget what I previously said (which is not very long). I'm just saying, don't expect to forward port, post and it will be accepted. Below is minimally one of the issues that needs to be addressed. >> The issue remains that the kernel is not really setup to deal with any >> random property or node to be changed at any point in run-time. I >> think there needs to be some restrictions around what the overlays can >> touch. We can't have it be wide open and then lock things down later >> and break users. One example of what you could do is you can only add >> sub-trees to whitelisted nodes. That's probably acceptable for your >> usecase. >> > > Defining what can and what cannot be changed is not as trivial as a > list of white-listed nodes. No, but we have to start somewhere and we are not starting with any change allowed anywhere at anytime. If that is what people want, then they are going to get to maintain that out of tree. > In some cases there is a whole node hierarchy being inserted (like in > a FPGA). Yes, so you'd have a target fpga region. That sounds fine to me. Maybe its not a static whitelist, but drivers have to register target nodes/paths. > In others, it's merely changing a status property to "okay" and > a few device parameters. That seems fine too. Disabled nodes could be allowed. But what if you add/change properties on a node that is not disabled? Once a node is enabled, who is responsible for registering the device? What about changing a node from enabled to disabled? The kernel would need to handle that or not allow it. > The real issue is that the kernel has no way to verify that a given > device tree, either at boot time or at overlay application time, is > correct. > > When the tree is wrong at boot-time you'll hang (if you're lucky). > If the tree is wrong at run-time you'll get some into some unidentified > funky state. Or have some security hole or a mechanism for userspace to crash the system. > Finally what is, and what is not 'correct' is not for the kernel to > decide arbitrarily, it's a matter of policy, different for each > use-case. It is if the kernel will break doing so. Rob ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name 2017-10-18 18:30 ` Rob Herring @ 2017-10-18 21:46 ` Frank Rowand 2017-10-19 8:41 ` Pantelis Antoniou 2017-10-19 8:51 ` Pantelis Antoniou 1 sibling, 1 reply; 53+ messages in thread From: Frank Rowand @ 2017-10-18 21:46 UTC (permalink / raw) To: Rob Herring, Pantelis Antoniou Cc: Alan Tull, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga On 10/18/17 11:30, Rob Herring wrote: > On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou > <pantelis.antoniou@konsulko.com> wrote: >> On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: >>> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull@kernel.org> wrote: >>>> On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list@gmail.com> wrote: >>>>> On 10/17/17 14:46, Rob Herring wrote: >>>>>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: >>>>>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: >>>>>>> >>>>>>> Hi Rob, >>>>>>> >>>>>>>> With dependencies on a statically allocated full path name converted to >>>>>>>> use %pOF format specifier, we can store just the basename of node, and >>>>>>>> the unflattening of the FDT can be simplified. >>>>>>>> >>>>>>>> This commit will affect the remaining users of full_name. After >>>>>>>> analyzing these users, the remaining cases should only change some print >>>>>>>> messages. The main users of full_name are providing a name for struct >>>>>>>> resource. The resource names shouldn't be important other than providing >>>>>>>> /proc/iomem names. >>>>>>>> >>>>>>>> We no longer distinguish between pre and post 0x10 dtb formats as either >>>>>>>> a full path or basename will work. However, less than 0x10 formats have >>>>>>>> been broken since the conversion to use libfdt (and no one has cared). >>>>>>>> The conversion of the unflattening code to be non-recursive also broke >>>>>>>> pre 0x10 formats as the populate_node function would return 0 in that >>>>>>>> case. >>>>>>>> >>>>>>>> Signed-off-by: Rob Herring <robh@kernel.org> >>>>>>>> --- >>>>>>>> v2: >>>>>>>> - rebase to linux-next >>>>>>>> >>>>>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >>>>>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >>>>>>> >>>>>>> I've just updated to the latest next branch and am finding problems >>>>>>> applying overlays. Reverting this commit alleviates things. The >>>>>>> errors I get are: >>>>>>> >>>>>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >>>>>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >>>>>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >>>>>> >>>>>> Frank's series with overlay updates should fix this. >>>>> >>>>> Yes, it does: >>>>> >>>>> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name >>>> >>>> Thanks for the fast response. I fetched the dt/next branch to test >>>> this but there are sufficient changes that Pantelis' "OF: DT-Overlay >>>> configfs interface (v7)" is broken now. I've been adding that >>>> downstream since 4.4. We're using it as an interface for applying >>>> overlays to program FPGAs. If we fix it again, is there any chance >>>> that can go upstream now? >>> >>> With a drive-by posting once every few years, no. >>> >> >> I take offense to that. There's nothing changed in the patch for years. >> Reposting the same patch without changes would achieve nothing. > > Are you still expecting review comments on it or something? > Furthermore, If something is posted infrequently, then I'm not > inclined to comment or care if the next posting is going to be after I > forget what I previously said (which is not very long). > > I'm just saying, don't expect to forward port, post and it will be > accepted. Below is minimally one of the issues that needs to be > addressed. > >>> The issue remains that the kernel is not really setup to deal with any >>> random property or node to be changed at any point in run-time. I >>> think there needs to be some restrictions around what the overlays can >>> touch. We can't have it be wide open and then lock things down later >>> and break users. That paragraph is key to any discussion of accepting code to apply overlays. Solving that issue has been stated to be a gating factor for such code from the beginning of overlay development. >>> One example of what you could do is you can only add >>> sub-trees to whitelisted nodes. That's probably acceptable for your >>> usecase. >>> >> >> Defining what can and what cannot be changed is not as trivial as a >> list of white-listed nodes. > > No, but we have to start somewhere and we are not starting with any > change allowed anywhere at anytime. If that is what people want, then > they are going to get to maintain that out of tree. > >> In some cases there is a whole node hierarchy being inserted (like in >> a FPGA). > > Yes, so you'd have a target fpga region. That sounds fine to me. Maybe > its not a static whitelist, but drivers have to register target > nodes/paths. > >> In others, it's merely changing a status property to "okay" and >> a few device parameters. > > That seems fine too. Disabled nodes could be allowed. But what if you > add/change properties on a node that is not disabled? Once a node is > enabled, who is responsible for registering the device? > > What about changing a node from enabled to disabled? The kernel would > need to handle that or not allow it. > >> The real issue is that the kernel has no way to verify that a given >> device tree, either at boot time or at overlay application time, is >> correct. >> >> When the tree is wrong at boot-time you'll hang (if you're lucky). >> If the tree is wrong at run-time you'll get some into some unidentified >> funky state. > > Or have some security hole or a mechanism for userspace to crash the system. > >> Finally what is, and what is not 'correct' is not for the kernel to >> decide arbitrarily, it's a matter of policy, different for each >> use-case. > > It is if the kernel will break doing so. > > Rob > ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-19 8:41 ` Pantelis Antoniou 0 siblings, 0 replies; 53+ messages in thread From: Pantelis Antoniou @ 2017-10-19 8:41 UTC (permalink / raw) To: Frank Rowand Cc: Rob Herring, Alan Tull, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga Hi Frank, > On Oct 19, 2017, at 00:46 , Frank Rowand <frowand.list@gmail.com> wrote: > > On 10/18/17 11:30, Rob Herring wrote: >> On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou >> <pantelis.antoniou@konsulko.com> wrote: >>> On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: >>>> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull@kernel.org> wrote: >>>>> On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list@gmail.com> wrote: >>>>>> On 10/17/17 14:46, Rob Herring wrote: >>>>>>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: >>>>>>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: >>>>>>>> >>>>>>>> Hi Rob, >>>>>>>> >>>>>>>>> With dependencies on a statically allocated full path name converted to >>>>>>>>> use %pOF format specifier, we can store just the basename of node, and >>>>>>>>> the unflattening of the FDT can be simplified. >>>>>>>>> >>>>>>>>> This commit will affect the remaining users of full_name. After >>>>>>>>> analyzing these users, the remaining cases should only change some print >>>>>>>>> messages. The main users of full_name are providing a name for struct >>>>>>>>> resource. The resource names shouldn't be important other than providing >>>>>>>>> /proc/iomem names. >>>>>>>>> >>>>>>>>> We no longer distinguish between pre and post 0x10 dtb formats as either >>>>>>>>> a full path or basename will work. However, less than 0x10 formats have >>>>>>>>> been broken since the conversion to use libfdt (and no one has cared). >>>>>>>>> The conversion of the unflattening code to be non-recursive also broke >>>>>>>>> pre 0x10 formats as the populate_node function would return 0 in that >>>>>>>>> case. >>>>>>>>> >>>>>>>>> Signed-off-by: Rob Herring <robh@kernel.org> >>>>>>>>> --- >>>>>>>>> v2: >>>>>>>>> - rebase to linux-next >>>>>>>>> >>>>>>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >>>>>>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >>>>>>>> >>>>>>>> I've just updated to the latest next branch and am finding problems >>>>>>>> applying overlays. Reverting this commit alleviates things. The >>>>>>>> errors I get are: >>>>>>>> >>>>>>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >>>>>>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >>>>>>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >>>>>>> >>>>>>> Frank's series with overlay updates should fix this. >>>>>> >>>>>> Yes, it does: >>>>>> >>>>>> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name >>>>> >>>>> Thanks for the fast response. I fetched the dt/next branch to test >>>>> this but there are sufficient changes that Pantelis' "OF: DT-Overlay >>>>> configfs interface (v7)" is broken now. I've been adding that >>>>> downstream since 4.4. We're using it as an interface for applying >>>>> overlays to program FPGAs. If we fix it again, is there any chance >>>>> that can go upstream now? >>>> >>>> With a drive-by posting once every few years, no. >>>> >>> >>> I take offense to that. There's nothing changed in the patch for years. >>> Reposting the same patch without changes would achieve nothing. >> >> Are you still expecting review comments on it or something? >> Furthermore, If something is posted infrequently, then I'm not >> inclined to comment or care if the next posting is going to be after I >> forget what I previously said (which is not very long). >> >> I'm just saying, don't expect to forward port, post and it will be >> accepted. Below is minimally one of the issues that needs to be >> addressed. >> > > >>>> The issue remains that the kernel is not really setup to deal with any >>>> random property or node to be changed at any point in run-time. I >>>> think there needs to be some restrictions around what the overlays can >>>> touch. We can't have it be wide open and then lock things down later >>>> and break users. > > That paragraph is key to any discussion of accepting code to apply overlays. > Solving that issue has been stated to be a gating factor for such code from > the beginning of overlay development. > Overlays are not only used only for cases where you have external expansion boards, or FPGAs where every change is contained under a few designated nodes, so that’s why I’m pushing for a in-kernel validator that’s more flexible than a single whitelist. An eBPF validator would handle a whitelist trivially easy, and would be flexible enough for any other more complicated use case. > >>>> One example of what you could do is you can only add >>>> sub-trees to whitelisted nodes. That's probably acceptable for your >>>> usecase. >>>> >>> >>> Defining what can and what cannot be changed is not as trivial as a >>> list of white-listed nodes. >> >> No, but we have to start somewhere and we are not starting with any >> change allowed anywhere at anytime. If that is what people want, then >> they are going to get to maintain that out of tree. >> >>> In some cases there is a whole node hierarchy being inserted (like in >>> a FPGA). >> >> Yes, so you'd have a target fpga region. That sounds fine to me. Maybe >> its not a static whitelist, but drivers have to register target >> nodes/paths. >> >>> In others, it's merely changing a status property to "okay" and >>> a few device parameters. >> >> That seems fine too. Disabled nodes could be allowed. But what if you >> add/change properties on a node that is not disabled? Once a node is >> enabled, who is responsible for registering the device? >> >> What about changing a node from enabled to disabled? The kernel would >> need to handle that or not allow it. >> >>> The real issue is that the kernel has no way to verify that a given >>> device tree, either at boot time or at overlay application time, is >>> correct. >>> >>> When the tree is wrong at boot-time you'll hang (if you're lucky). >>> If the tree is wrong at run-time you'll get some into some unidentified >>> funky state. >> >> Or have some security hole or a mechanism for userspace to crash the system. >> >>> Finally what is, and what is not 'correct' is not for the kernel to >>> decide arbitrarily, it's a matter of policy, different for each >>> use-case. >> >> It is if the kernel will break doing so. >> >> Rob >> > ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-19 8:41 ` Pantelis Antoniou 0 siblings, 0 replies; 53+ messages in thread From: Pantelis Antoniou @ 2017-10-19 8:41 UTC (permalink / raw) To: Frank Rowand Cc: Rob Herring, Alan Tull, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga Hi Frank, > On Oct 19, 2017, at 00:46 , Frank Rowand <frowand.list@gmail.com> = wrote: >=20 > On 10/18/17 11:30, Rob Herring wrote: >> On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou >> <pantelis.antoniou@konsulko.com> wrote: >>> On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: >>>> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull@kernel.org> = wrote: >>>>> On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand = <frowand.list@gmail.com> wrote: >>>>>> On 10/17/17 14:46, Rob Herring wrote: >>>>>>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> = wrote: >>>>>>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> = wrote: >>>>>>>>=20 >>>>>>>> Hi Rob, >>>>>>>>=20 >>>>>>>>> With dependencies on a statically allocated full path name = converted to >>>>>>>>> use %pOF format specifier, we can store just the basename of = node, and >>>>>>>>> the unflattening of the FDT can be simplified. >>>>>>>>>=20 >>>>>>>>> This commit will affect the remaining users of full_name. = After >>>>>>>>> analyzing these users, the remaining cases should only change = some print >>>>>>>>> messages. The main users of full_name are providing a name for = struct >>>>>>>>> resource. The resource names shouldn't be important other than = providing >>>>>>>>> /proc/iomem names. >>>>>>>>>=20 >>>>>>>>> We no longer distinguish between pre and post 0x10 dtb formats = as either >>>>>>>>> a full path or basename will work. However, less than 0x10 = formats have >>>>>>>>> been broken since the conversion to use libfdt (and no one has = cared). >>>>>>>>> The conversion of the unflattening code to be non-recursive = also broke >>>>>>>>> pre 0x10 formats as the populate_node function would return 0 = in that >>>>>>>>> case. >>>>>>>>>=20 >>>>>>>>> Signed-off-by: Rob Herring <robh@kernel.org> >>>>>>>>> --- >>>>>>>>> v2: >>>>>>>>> - rebase to linux-next >>>>>>>>>=20 >>>>>>>>> drivers/of/fdt.c | 69 = +++++++++----------------------------------------------- >>>>>>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >>>>>>>>=20 >>>>>>>> I've just updated to the latest next branch and am finding = problems >>>>>>>> applying overlays. Reverting this commit alleviates things. = The >>>>>>>> errors I get are: >>>>>>>>=20 >>>>>>>> [ 88.498704] OF: overlay: Failed to apply prop = @/__symbols__/clk_0 >>>>>>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >>>>>>>> [ 88.518423] create_overlay: Failed to create overlay = (err=3D-12) >>>>>>>=20 >>>>>>> Frank's series with overlay updates should fix this. >>>>>>=20 >>>>>> Yes, it does: >>>>>>=20 >>>>>> [PATCH v3 11/12] of: overlay: remove a dependency on device node = full_name >>>>>=20 >>>>> Thanks for the fast response. I fetched the dt/next branch to = test >>>>> this but there are sufficient changes that Pantelis' "OF: = DT-Overlay >>>>> configfs interface (v7)" is broken now. I've been adding that >>>>> downstream since 4.4. We're using it as an interface for applying >>>>> overlays to program FPGAs. If we fix it again, is there any = chance >>>>> that can go upstream now? >>>>=20 >>>> With a drive-by posting once every few years, no. >>>>=20 >>>=20 >>> I take offense to that. There's nothing changed in the patch for = years. >>> Reposting the same patch without changes would achieve nothing. >>=20 >> Are you still expecting review comments on it or something? >> Furthermore, If something is posted infrequently, then I'm not >> inclined to comment or care if the next posting is going to be after = I >> forget what I previously said (which is not very long). >>=20 >> I'm just saying, don't expect to forward port, post and it will be >> accepted. Below is minimally one of the issues that needs to be >> addressed. >>=20 >=20 >=20 >>>> The issue remains that the kernel is not really setup to deal with = any >>>> random property or node to be changed at any point in run-time. I >>>> think there needs to be some restrictions around what the overlays = can >>>> touch. We can't have it be wide open and then lock things down = later >>>> and break users. >=20 > That paragraph is key to any discussion of accepting code to apply = overlays. > Solving that issue has been stated to be a gating factor for such code = from > the beginning of overlay development. >=20 Overlays are not only used only for cases where you have external = expansion boards, or FPGAs where every change is contained under a few designated nodes, so = that=E2=80=99s why I=E2=80=99m pushing for a in-kernel validator that=E2=80=99s more flexible than a = single whitelist. An eBPF validator would handle a whitelist trivially easy, and would be = flexible enough for any other more complicated use case. >=20 >>>> One example of what you could do is you can only add >>>> sub-trees to whitelisted nodes. That's probably acceptable for your >>>> usecase. >>>>=20 >>>=20 >>> Defining what can and what cannot be changed is not as trivial as a >>> list of white-listed nodes. >>=20 >> No, but we have to start somewhere and we are not starting with any >> change allowed anywhere at anytime. If that is what people want, then >> they are going to get to maintain that out of tree. >>=20 >>> In some cases there is a whole node hierarchy being inserted (like = in >>> a FPGA). >>=20 >> Yes, so you'd have a target fpga region. That sounds fine to me. = Maybe >> its not a static whitelist, but drivers have to register target >> nodes/paths. >>=20 >>> In others, it's merely changing a status property to "okay" and >>> a few device parameters. >>=20 >> That seems fine too. Disabled nodes could be allowed. But what if you >> add/change properties on a node that is not disabled? Once a node is >> enabled, who is responsible for registering the device? >>=20 >> What about changing a node from enabled to disabled? The kernel would >> need to handle that or not allow it. >>=20 >>> The real issue is that the kernel has no way to verify that a given >>> device tree, either at boot time or at overlay application time, is >>> correct. >>>=20 >>> When the tree is wrong at boot-time you'll hang (if you're lucky). >>> If the tree is wrong at run-time you'll get some into some = unidentified >>> funky state. >>=20 >> Or have some security hole or a mechanism for userspace to crash the = system. >>=20 >>> Finally what is, and what is not 'correct' is not for the kernel to >>> decide arbitrarily, it's a matter of policy, different for each >>> use-case. >>=20 >> It is if the kernel will break doing so. >>=20 >> Rob >>=20 >=20 ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-19 8:41 ` Pantelis Antoniou 0 siblings, 0 replies; 53+ messages in thread From: Pantelis Antoniou @ 2017-10-19 8:41 UTC (permalink / raw) To: Frank Rowand Cc: Rob Herring, Alan Tull, devicetree-u79uwXL29TY76Z2rM5mHXA, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga-u79uwXL29TY76Z2rM5mHXA Hi Frank, > On Oct 19, 2017, at 00:46 , Frank Rowand <frowand.list-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote: > > On 10/18/17 11:30, Rob Herring wrote: >> On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou >> <pantelis.antoniou-OWPKS81ov/FWk0Htik3J/w@public.gmane.org> wrote: >>> On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: >>>> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >>>>> On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote: >>>>>> On 10/17/17 14:46, Rob Herring wrote: >>>>>>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >>>>>>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >>>>>>>> >>>>>>>> Hi Rob, >>>>>>>> >>>>>>>>> With dependencies on a statically allocated full path name converted to >>>>>>>>> use %pOF format specifier, we can store just the basename of node, and >>>>>>>>> the unflattening of the FDT can be simplified. >>>>>>>>> >>>>>>>>> This commit will affect the remaining users of full_name. After >>>>>>>>> analyzing these users, the remaining cases should only change some print >>>>>>>>> messages. The main users of full_name are providing a name for struct >>>>>>>>> resource. The resource names shouldn't be important other than providing >>>>>>>>> /proc/iomem names. >>>>>>>>> >>>>>>>>> We no longer distinguish between pre and post 0x10 dtb formats as either >>>>>>>>> a full path or basename will work. However, less than 0x10 formats have >>>>>>>>> been broken since the conversion to use libfdt (and no one has cared). >>>>>>>>> The conversion of the unflattening code to be non-recursive also broke >>>>>>>>> pre 0x10 formats as the populate_node function would return 0 in that >>>>>>>>> case. >>>>>>>>> >>>>>>>>> Signed-off-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> >>>>>>>>> --- >>>>>>>>> v2: >>>>>>>>> - rebase to linux-next >>>>>>>>> >>>>>>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >>>>>>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >>>>>>>> >>>>>>>> I've just updated to the latest next branch and am finding problems >>>>>>>> applying overlays. Reverting this commit alleviates things. The >>>>>>>> errors I get are: >>>>>>>> >>>>>>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >>>>>>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >>>>>>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >>>>>>> >>>>>>> Frank's series with overlay updates should fix this. >>>>>> >>>>>> Yes, it does: >>>>>> >>>>>> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name >>>>> >>>>> Thanks for the fast response. I fetched the dt/next branch to test >>>>> this but there are sufficient changes that Pantelis' "OF: DT-Overlay >>>>> configfs interface (v7)" is broken now. I've been adding that >>>>> downstream since 4.4. We're using it as an interface for applying >>>>> overlays to program FPGAs. If we fix it again, is there any chance >>>>> that can go upstream now? >>>> >>>> With a drive-by posting once every few years, no. >>>> >>> >>> I take offense to that. There's nothing changed in the patch for years. >>> Reposting the same patch without changes would achieve nothing. >> >> Are you still expecting review comments on it or something? >> Furthermore, If something is posted infrequently, then I'm not >> inclined to comment or care if the next posting is going to be after I >> forget what I previously said (which is not very long). >> >> I'm just saying, don't expect to forward port, post and it will be >> accepted. Below is minimally one of the issues that needs to be >> addressed. >> > > >>>> The issue remains that the kernel is not really setup to deal with any >>>> random property or node to be changed at any point in run-time. I >>>> think there needs to be some restrictions around what the overlays can >>>> touch. We can't have it be wide open and then lock things down later >>>> and break users. > > That paragraph is key to any discussion of accepting code to apply overlays. > Solving that issue has been stated to be a gating factor for such code from > the beginning of overlay development. > Overlays are not only used only for cases where you have external expansion boards, or FPGAs where every change is contained under a few designated nodes, so that’s why I’m pushing for a in-kernel validator that’s more flexible than a single whitelist. An eBPF validator would handle a whitelist trivially easy, and would be flexible enough for any other more complicated use case. > >>>> One example of what you could do is you can only add >>>> sub-trees to whitelisted nodes. That's probably acceptable for your >>>> usecase. >>>> >>> >>> Defining what can and what cannot be changed is not as trivial as a >>> list of white-listed nodes. >> >> No, but we have to start somewhere and we are not starting with any >> change allowed anywhere at anytime. If that is what people want, then >> they are going to get to maintain that out of tree. >> >>> In some cases there is a whole node hierarchy being inserted (like in >>> a FPGA). >> >> Yes, so you'd have a target fpga region. That sounds fine to me. Maybe >> its not a static whitelist, but drivers have to register target >> nodes/paths. >> >>> In others, it's merely changing a status property to "okay" and >>> a few device parameters. >> >> That seems fine too. Disabled nodes could be allowed. But what if you >> add/change properties on a node that is not disabled? Once a node is >> enabled, who is responsible for registering the device? >> >> What about changing a node from enabled to disabled? The kernel would >> need to handle that or not allow it. >> >>> The real issue is that the kernel has no way to verify that a given >>> device tree, either at boot time or at overlay application time, is >>> correct. >>> >>> When the tree is wrong at boot-time you'll hang (if you're lucky). >>> If the tree is wrong at run-time you'll get some into some unidentified >>> funky state. >> >> Or have some security hole or a mechanism for userspace to crash the system. >> >>> Finally what is, and what is not 'correct' is not for the kernel to >>> decide arbitrarily, it's a matter of policy, different for each >>> use-case. >> >> It is if the kernel will break doing so. >> >> Rob >> > -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-19 8:51 ` Pantelis Antoniou 0 siblings, 0 replies; 53+ messages in thread From: Pantelis Antoniou @ 2017-10-19 8:51 UTC (permalink / raw) To: Rob Herring Cc: Alan Tull, Frank Rowand, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga Hi Rob, > On Oct 18, 2017, at 21:30 , Rob Herring <robh@kernel.org> wrote: > > On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou > <pantelis.antoniou@konsulko.com> wrote: >> On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: >>> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull@kernel.org> wrote: >>>> On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list@gmail.com> wrote: >>>>> On 10/17/17 14:46, Rob Herring wrote: >>>>>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: >>>>>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: >>>>>>> >>>>>>> Hi Rob, >>>>>>> >>>>>>>> With dependencies on a statically allocated full path name converted to >>>>>>>> use %pOF format specifier, we can store just the basename of node, and >>>>>>>> the unflattening of the FDT can be simplified. >>>>>>>> >>>>>>>> This commit will affect the remaining users of full_name. After >>>>>>>> analyzing these users, the remaining cases should only change some print >>>>>>>> messages. The main users of full_name are providing a name for struct >>>>>>>> resource. The resource names shouldn't be important other than providing >>>>>>>> /proc/iomem names. >>>>>>>> >>>>>>>> We no longer distinguish between pre and post 0x10 dtb formats as either >>>>>>>> a full path or basename will work. However, less than 0x10 formats have >>>>>>>> been broken since the conversion to use libfdt (and no one has cared). >>>>>>>> The conversion of the unflattening code to be non-recursive also broke >>>>>>>> pre 0x10 formats as the populate_node function would return 0 in that >>>>>>>> case. >>>>>>>> >>>>>>>> Signed-off-by: Rob Herring <robh@kernel.org> >>>>>>>> --- >>>>>>>> v2: >>>>>>>> - rebase to linux-next >>>>>>>> >>>>>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >>>>>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >>>>>>> >>>>>>> I've just updated to the latest next branch and am finding problems >>>>>>> applying overlays. Reverting this commit alleviates things. The >>>>>>> errors I get are: >>>>>>> >>>>>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >>>>>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >>>>>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >>>>>> >>>>>> Frank's series with overlay updates should fix this. >>>>> >>>>> Yes, it does: >>>>> >>>>> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name >>>> >>>> Thanks for the fast response. I fetched the dt/next branch to test >>>> this but there are sufficient changes that Pantelis' "OF: DT-Overlay >>>> configfs interface (v7)" is broken now. I've been adding that >>>> downstream since 4.4. We're using it as an interface for applying >>>> overlays to program FPGAs. If we fix it again, is there any chance >>>> that can go upstream now? >>> >>> With a drive-by posting once every few years, no. >>> >> >> I take offense to that. There's nothing changed in the patch for years. >> Reposting the same patch without changes would achieve nothing. > > Are you still expecting review comments on it or something? > Furthermore, If something is posted infrequently, then I'm not > inclined to comment or care if the next posting is going to be after I > forget what I previously said (which is not very long). > > I'm just saying, don't expect to forward port, post and it will be > accepted. Below is minimally one of the issues that needs to be > addressed. > >>> The issue remains that the kernel is not really setup to deal with any >>> random property or node to be changed at any point in run-time. I >>> think there needs to be some restrictions around what the overlays can >>> touch. We can't have it be wide open and then lock things down later >>> and break users. One example of what you could do is you can only add >>> sub-trees to whitelisted nodes. That's probably acceptable for your >>> usecase. >>> >> >> Defining what can and what cannot be changed is not as trivial as a >> list of white-listed nodes. > > No, but we have to start somewhere and we are not starting with any > change allowed anywhere at anytime. If that is what people want, then > they are going to get to maintain that out of tree. > I am still not sold on this ‘dangerous’ idea. No-one is crazy enough to suggest overlays to be loadable by an unprivileged user. It’s exactly the same ‘danger’ as loading a kernel module, while is sure capable of much greater mischief. >> In some cases there is a whole node hierarchy being inserted (like in >> a FPGA). > > Yes, so you'd have a target fpga region. That sounds fine to me. Maybe > its not a static whitelist, but drivers have to register target > nodes/paths. > >> In others, it's merely changing a status property to "okay" and >> a few device parameters. > > That seems fine too. Disabled nodes could be allowed. But what if you > add/change properties on a node that is not disabled? Once a node is > enabled, who is responsible for registering the device? > > What about changing a node from enabled to disabled? The kernel would > need to handle that or not allow it. > So it seems a simple whitelist won’t cut it. We’re already talking about special casing for this or that property. My argument is that this kind of validation is not part of the core-device tree, but instead is a policy decision different for each board. >> The real issue is that the kernel has no way to verify that a given >> device tree, either at boot time or at overlay application time, is >> correct. >> >> When the tree is wrong at boot-time you'll hang (if you're lucky). >> If the tree is wrong at run-time you'll get some into some unidentified >> funky state. > > Or have some security hole or a mechanism for userspace to crash the system. > User-space as in regular users should never have enough privileges to apply an overlay, same as in loading a kernel module. >> Finally what is, and what is not 'correct' is not for the kernel to >> decide arbitrarily, it's a matter of policy, different for each >> use-case. > > It is if the kernel will break doing so. > I still haven’t seen a real example of the kernel breaking. I have seen a lot of cases where the kernel is crashing due to the device removal path being broken, but those are kernel bugs to fix, not something to use to hold back functionality that people want to use. > Rob Regards — Pantelis ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-19 8:51 ` Pantelis Antoniou 0 siblings, 0 replies; 53+ messages in thread From: Pantelis Antoniou @ 2017-10-19 8:51 UTC (permalink / raw) To: Rob Herring Cc: Alan Tull, Frank Rowand, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga Hi Rob, > On Oct 18, 2017, at 21:30 , Rob Herring <robh@kernel.org> wrote: >=20 > On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou > <pantelis.antoniou@konsulko.com> wrote: >> On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: >>> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull@kernel.org> = wrote: >>>> On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand = <frowand.list@gmail.com> wrote: >>>>> On 10/17/17 14:46, Rob Herring wrote: >>>>>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> = wrote: >>>>>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> = wrote: >>>>>>>=20 >>>>>>> Hi Rob, >>>>>>>=20 >>>>>>>> With dependencies on a statically allocated full path name = converted to >>>>>>>> use %pOF format specifier, we can store just the basename of = node, and >>>>>>>> the unflattening of the FDT can be simplified. >>>>>>>>=20 >>>>>>>> This commit will affect the remaining users of full_name. After >>>>>>>> analyzing these users, the remaining cases should only change = some print >>>>>>>> messages. The main users of full_name are providing a name for = struct >>>>>>>> resource. The resource names shouldn't be important other than = providing >>>>>>>> /proc/iomem names. >>>>>>>>=20 >>>>>>>> We no longer distinguish between pre and post 0x10 dtb formats = as either >>>>>>>> a full path or basename will work. However, less than 0x10 = formats have >>>>>>>> been broken since the conversion to use libfdt (and no one has = cared). >>>>>>>> The conversion of the unflattening code to be non-recursive = also broke >>>>>>>> pre 0x10 formats as the populate_node function would return 0 = in that >>>>>>>> case. >>>>>>>>=20 >>>>>>>> Signed-off-by: Rob Herring <robh@kernel.org> >>>>>>>> --- >>>>>>>> v2: >>>>>>>> - rebase to linux-next >>>>>>>>=20 >>>>>>>> drivers/of/fdt.c | 69 = +++++++++----------------------------------------------- >>>>>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >>>>>>>=20 >>>>>>> I've just updated to the latest next branch and am finding = problems >>>>>>> applying overlays. Reverting this commit alleviates things. = The >>>>>>> errors I get are: >>>>>>>=20 >>>>>>> [ 88.498704] OF: overlay: Failed to apply prop = @/__symbols__/clk_0 >>>>>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >>>>>>> [ 88.518423] create_overlay: Failed to create overlay = (err=3D-12) >>>>>>=20 >>>>>> Frank's series with overlay updates should fix this. >>>>>=20 >>>>> Yes, it does: >>>>>=20 >>>>> [PATCH v3 11/12] of: overlay: remove a dependency on device node = full_name >>>>=20 >>>> Thanks for the fast response. I fetched the dt/next branch to test >>>> this but there are sufficient changes that Pantelis' "OF: = DT-Overlay >>>> configfs interface (v7)" is broken now. I've been adding that >>>> downstream since 4.4. We're using it as an interface for applying >>>> overlays to program FPGAs. If we fix it again, is there any chance >>>> that can go upstream now? >>>=20 >>> With a drive-by posting once every few years, no. >>>=20 >>=20 >> I take offense to that. There's nothing changed in the patch for = years. >> Reposting the same patch without changes would achieve nothing. >=20 > Are you still expecting review comments on it or something? > Furthermore, If something is posted infrequently, then I'm not > inclined to comment or care if the next posting is going to be after I > forget what I previously said (which is not very long). >=20 > I'm just saying, don't expect to forward port, post and it will be > accepted. Below is minimally one of the issues that needs to be > addressed. >=20 >>> The issue remains that the kernel is not really setup to deal with = any >>> random property or node to be changed at any point in run-time. I >>> think there needs to be some restrictions around what the overlays = can >>> touch. We can't have it be wide open and then lock things down later >>> and break users. One example of what you could do is you can only = add >>> sub-trees to whitelisted nodes. That's probably acceptable for your >>> usecase. >>>=20 >>=20 >> Defining what can and what cannot be changed is not as trivial as a >> list of white-listed nodes. >=20 > No, but we have to start somewhere and we are not starting with any > change allowed anywhere at anytime. If that is what people want, then > they are going to get to maintain that out of tree. >=20 I am still not sold on this =E2=80=98dangerous=E2=80=99 idea. No-one is = crazy enough to suggest overlays to be loadable by an unprivileged user. It=E2=80=99s = exactly the same =E2=80=98danger=E2=80=99 as loading a kernel module, while is sure = capable of much greater mischief. >> In some cases there is a whole node hierarchy being inserted (like in >> a FPGA). >=20 > Yes, so you'd have a target fpga region. That sounds fine to me. Maybe > its not a static whitelist, but drivers have to register target > nodes/paths. >=20 >> In others, it's merely changing a status property to "okay" and >> a few device parameters. >=20 > That seems fine too. Disabled nodes could be allowed. But what if you > add/change properties on a node that is not disabled? Once a node is > enabled, who is responsible for registering the device? >=20 > What about changing a node from enabled to disabled? The kernel would > need to handle that or not allow it. >=20 So it seems a simple whitelist won=E2=80=99t cut it. We=E2=80=99re = already talking about special casing for this or that property. My argument is that this kind of validation is not part of the = core-device tree, but instead is a policy decision different for each board. =20 >> The real issue is that the kernel has no way to verify that a given >> device tree, either at boot time or at overlay application time, is >> correct. >>=20 >> When the tree is wrong at boot-time you'll hang (if you're lucky). >> If the tree is wrong at run-time you'll get some into some = unidentified >> funky state. >=20 > Or have some security hole or a mechanism for userspace to crash the = system. >=20 User-space as in regular users should never have enough privileges to = apply an overlay, same as in loading a kernel module. >> Finally what is, and what is not 'correct' is not for the kernel to >> decide arbitrarily, it's a matter of policy, different for each >> use-case. >=20 > It is if the kernel will break doing so. >=20 I still haven=E2=80=99t seen a real example of the kernel breaking. I have seen a lot of cases where the kernel is crashing due to the = device removal path being broken, but those are kernel bugs to fix, not = something to use to hold back functionality that people want to use. > Rob Regards =E2=80=94 Pantelis ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-19 8:51 ` Pantelis Antoniou 0 siblings, 0 replies; 53+ messages in thread From: Pantelis Antoniou @ 2017-10-19 8:51 UTC (permalink / raw) To: Rob Herring Cc: Alan Tull, Frank Rowand, devicetree-u79uwXL29TY76Z2rM5mHXA, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga-u79uwXL29TY76Z2rM5mHXA Hi Rob, > On Oct 18, 2017, at 21:30 , Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: > > On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou > <pantelis.antoniou-OWPKS81ov/FWk0Htik3J/w@public.gmane.org> wrote: >> On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: >>> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >>>> On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote: >>>>> On 10/17/17 14:46, Rob Herring wrote: >>>>>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >>>>>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >>>>>>> >>>>>>> Hi Rob, >>>>>>> >>>>>>>> With dependencies on a statically allocated full path name converted to >>>>>>>> use %pOF format specifier, we can store just the basename of node, and >>>>>>>> the unflattening of the FDT can be simplified. >>>>>>>> >>>>>>>> This commit will affect the remaining users of full_name. After >>>>>>>> analyzing these users, the remaining cases should only change some print >>>>>>>> messages. The main users of full_name are providing a name for struct >>>>>>>> resource. The resource names shouldn't be important other than providing >>>>>>>> /proc/iomem names. >>>>>>>> >>>>>>>> We no longer distinguish between pre and post 0x10 dtb formats as either >>>>>>>> a full path or basename will work. However, less than 0x10 formats have >>>>>>>> been broken since the conversion to use libfdt (and no one has cared). >>>>>>>> The conversion of the unflattening code to be non-recursive also broke >>>>>>>> pre 0x10 formats as the populate_node function would return 0 in that >>>>>>>> case. >>>>>>>> >>>>>>>> Signed-off-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> >>>>>>>> --- >>>>>>>> v2: >>>>>>>> - rebase to linux-next >>>>>>>> >>>>>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >>>>>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >>>>>>> >>>>>>> I've just updated to the latest next branch and am finding problems >>>>>>> applying overlays. Reverting this commit alleviates things. The >>>>>>> errors I get are: >>>>>>> >>>>>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >>>>>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >>>>>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >>>>>> >>>>>> Frank's series with overlay updates should fix this. >>>>> >>>>> Yes, it does: >>>>> >>>>> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name >>>> >>>> Thanks for the fast response. I fetched the dt/next branch to test >>>> this but there are sufficient changes that Pantelis' "OF: DT-Overlay >>>> configfs interface (v7)" is broken now. I've been adding that >>>> downstream since 4.4. We're using it as an interface for applying >>>> overlays to program FPGAs. If we fix it again, is there any chance >>>> that can go upstream now? >>> >>> With a drive-by posting once every few years, no. >>> >> >> I take offense to that. There's nothing changed in the patch for years. >> Reposting the same patch without changes would achieve nothing. > > Are you still expecting review comments on it or something? > Furthermore, If something is posted infrequently, then I'm not > inclined to comment or care if the next posting is going to be after I > forget what I previously said (which is not very long). > > I'm just saying, don't expect to forward port, post and it will be > accepted. Below is minimally one of the issues that needs to be > addressed. > >>> The issue remains that the kernel is not really setup to deal with any >>> random property or node to be changed at any point in run-time. I >>> think there needs to be some restrictions around what the overlays can >>> touch. We can't have it be wide open and then lock things down later >>> and break users. One example of what you could do is you can only add >>> sub-trees to whitelisted nodes. That's probably acceptable for your >>> usecase. >>> >> >> Defining what can and what cannot be changed is not as trivial as a >> list of white-listed nodes. > > No, but we have to start somewhere and we are not starting with any > change allowed anywhere at anytime. If that is what people want, then > they are going to get to maintain that out of tree. > I am still not sold on this ‘dangerous’ idea. No-one is crazy enough to suggest overlays to be loadable by an unprivileged user. It’s exactly the same ‘danger’ as loading a kernel module, while is sure capable of much greater mischief. >> In some cases there is a whole node hierarchy being inserted (like in >> a FPGA). > > Yes, so you'd have a target fpga region. That sounds fine to me. Maybe > its not a static whitelist, but drivers have to register target > nodes/paths. > >> In others, it's merely changing a status property to "okay" and >> a few device parameters. > > That seems fine too. Disabled nodes could be allowed. But what if you > add/change properties on a node that is not disabled? Once a node is > enabled, who is responsible for registering the device? > > What about changing a node from enabled to disabled? The kernel would > need to handle that or not allow it. > So it seems a simple whitelist won’t cut it. We’re already talking about special casing for this or that property. My argument is that this kind of validation is not part of the core-device tree, but instead is a policy decision different for each board. >> The real issue is that the kernel has no way to verify that a given >> device tree, either at boot time or at overlay application time, is >> correct. >> >> When the tree is wrong at boot-time you'll hang (if you're lucky). >> If the tree is wrong at run-time you'll get some into some unidentified >> funky state. > > Or have some security hole or a mechanism for userspace to crash the system. > User-space as in regular users should never have enough privileges to apply an overlay, same as in loading a kernel module. >> Finally what is, and what is not 'correct' is not for the kernel to >> decide arbitrarily, it's a matter of policy, different for each >> use-case. > > It is if the kernel will break doing so. > I still haven’t seen a real example of the kernel breaking. I have seen a lot of cases where the kernel is crashing due to the device removal path being broken, but those are kernel bugs to fix, not something to use to hold back functionality that people want to use. > Rob Regards — Pantelis -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name 2017-10-19 8:51 ` Pantelis Antoniou (?) (?) @ 2017-10-19 20:06 ` Moritz Fischer 2017-10-19 21:46 ` Frank Rowand -1 siblings, 1 reply; 53+ messages in thread From: Moritz Fischer @ 2017-10-19 20:06 UTC (permalink / raw) To: Pantelis Antoniou Cc: Rob Herring, Alan Tull, Frank Rowand, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga [-- Attachment #1: Type: text/plain, Size: 7797 bytes --] On Thu, Oct 19, 2017 at 11:51:40AM +0300, Pantelis Antoniou wrote: > Hi Rob, > > > On Oct 18, 2017, at 21:30 , Rob Herring <robh@kernel.org> wrote: > > > > On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou > > <pantelis.antoniou@konsulko.com> wrote: > >> On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: > >>> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull@kernel.org> wrote: > >>>> On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list@gmail.com> wrote: > >>>>> On 10/17/17 14:46, Rob Herring wrote: > >>>>>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: > >>>>>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: > >>>>>>> > >>>>>>> Hi Rob, > >>>>>>> > >>>>>>>> With dependencies on a statically allocated full path name converted to > >>>>>>>> use %pOF format specifier, we can store just the basename of node, and > >>>>>>>> the unflattening of the FDT can be simplified. > >>>>>>>> > >>>>>>>> This commit will affect the remaining users of full_name. After > >>>>>>>> analyzing these users, the remaining cases should only change some print > >>>>>>>> messages. The main users of full_name are providing a name for struct > >>>>>>>> resource. The resource names shouldn't be important other than providing > >>>>>>>> /proc/iomem names. > >>>>>>>> > >>>>>>>> We no longer distinguish between pre and post 0x10 dtb formats as either > >>>>>>>> a full path or basename will work. However, less than 0x10 formats have > >>>>>>>> been broken since the conversion to use libfdt (and no one has cared). > >>>>>>>> The conversion of the unflattening code to be non-recursive also broke > >>>>>>>> pre 0x10 formats as the populate_node function would return 0 in that > >>>>>>>> case. > >>>>>>>> > >>>>>>>> Signed-off-by: Rob Herring <robh@kernel.org> > >>>>>>>> --- > >>>>>>>> v2: > >>>>>>>> - rebase to linux-next > >>>>>>>> > >>>>>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- > >>>>>>>> 1 file changed, 11 insertions(+), 58 deletions(-) > >>>>>>> > >>>>>>> I've just updated to the latest next branch and am finding problems > >>>>>>> applying overlays. Reverting this commit alleviates things. The > >>>>>>> errors I get are: > >>>>>>> > >>>>>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 > >>>>>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' > >>>>>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) > >>>>>> > >>>>>> Frank's series with overlay updates should fix this. > >>>>> > >>>>> Yes, it does: > >>>>> > >>>>> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name > >>>> > >>>> Thanks for the fast response. I fetched the dt/next branch to test > >>>> this but there are sufficient changes that Pantelis' "OF: DT-Overlay > >>>> configfs interface (v7)" is broken now. I've been adding that > >>>> downstream since 4.4. We're using it as an interface for applying > >>>> overlays to program FPGAs. If we fix it again, is there any chance > >>>> that can go upstream now? > >>> > >>> With a drive-by posting once every few years, no. > >>> > >> > >> I take offense to that. There's nothing changed in the patch for years. > >> Reposting the same patch without changes would achieve nothing. > > > > Are you still expecting review comments on it or something? > > Furthermore, If something is posted infrequently, then I'm not > > inclined to comment or care if the next posting is going to be after I > > forget what I previously said (which is not very long). > > > > I'm just saying, don't expect to forward port, post and it will be > > accepted. Below is minimally one of the issues that needs to be > > addressed. > > > >>> The issue remains that the kernel is not really setup to deal with any > >>> random property or node to be changed at any point in run-time. I > >>> think there needs to be some restrictions around what the overlays can > >>> touch. We can't have it be wide open and then lock things down later > >>> and break users. One example of what you could do is you can only add > >>> sub-trees to whitelisted nodes. That's probably acceptable for your > >>> usecase. > >>> > >> > >> Defining what can and what cannot be changed is not as trivial as a > >> list of white-listed nodes. > > > > No, but we have to start somewhere and we are not starting with any > > change allowed anywhere at anytime. If that is what people want, then > > they are going to get to maintain that out of tree. > > > > I am still not sold on this ‘dangerous’ idea. No-one is crazy enough to > suggest overlays to be loadable by an unprivileged user. It’s exactly the > same ‘danger’ as loading a kernel module, while is sure capable of much > greater mischief. Agreed. > > >> In some cases there is a whole node hierarchy being inserted (like in > >> a FPGA). > > > > Yes, so you'd have a target fpga region. That sounds fine to me. Maybe > > its not a static whitelist, but drivers have to register target > > nodes/paths. > > > >> In others, it's merely changing a status property to "okay" and > >> a few device parameters. > > > > That seems fine too. Disabled nodes could be allowed. But what if you > > add/change properties on a node that is not disabled? Once a node is > > enabled, who is responsible for registering the device? > > > > What about changing a node from enabled to disabled? The kernel would > > need to handle that or not allow it. > > > > So it seems a simple whitelist won’t cut it. We’re already talking about > special casing for this or that property. > > My argument is that this kind of validation is not part of the core-device tree, > but instead is a policy decision different for each board. > > >> The real issue is that the kernel has no way to verify that a given > >> device tree, either at boot time or at overlay application time, is > >> correct. > >> > >> When the tree is wrong at boot-time you'll hang (if you're lucky). > >> If the tree is wrong at run-time you'll get some into some unidentified > >> funky state. > > > > Or have some security hole or a mechanism for userspace to crash the system. > > > > User-space as in regular users should never have enough privileges to apply an > overlay, same as in loading a kernel module. > > >> Finally what is, and what is not 'correct' is not for the kernel to > >> decide arbitrarily, it's a matter of policy, different for each > >> use-case. > > > > It is if the kernel will break doing so. > > > > I still haven’t seen a real example of the kernel breaking. > > I have seen a lot of cases where the kernel is crashing due to the device > removal path being broken, but those are kernel bugs to fix, not something > to use to hold back functionality that people want to use. We also have plenty of code that is just not aware of overlays, and assumes certain parts of the tree to stay static. I ran into that issue when I tried to add thermal zones via an overlay, I've been investigating how to fix the thermal framework to work with overlays since then and have some partially working code. Currently the thermal framework parses the thermal-zones node at boot, and assumes this stays static. This breaks with overlays. I agree we eventually need to fix the parts that break when we use overlays. > > > Rob > > Regards > > — Pantelis > > -- > To unsubscribe from this list: send the line "unsubscribe linux-fpga" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html Cheers, Moritz [-- Attachment #2: signature.asc --] [-- Type: application/pgp-signature, Size: 488 bytes --] ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name 2017-10-19 20:06 ` Moritz Fischer @ 2017-10-19 21:46 ` Frank Rowand 2017-10-20 8:06 ` Pantelis Antoniou 0 siblings, 1 reply; 53+ messages in thread From: Frank Rowand @ 2017-10-19 21:46 UTC (permalink / raw) To: Moritz Fischer, Pantelis Antoniou Cc: Rob Herring, Alan Tull, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga On 10/19/17 13:06, Moritz Fischer wrote: < snip > > We also have plenty of code that is just not aware of overlays, and > assumes certain parts of the tree to stay static. I would state that somewhat differently. :-) There is very little code that is aware of overlays, and most code assumes the device tree does not change after early boot. This is one of the areas where the creation of overlays needs to be done with care. > I ran into that issue when I tried to add thermal zones via an overlay, > I've been investigating how to fix the thermal framework to work with > overlays since then and have some partially working code. > Currently the thermal framework parses the thermal-zones node at boot, > and assumes this stays static. This breaks with overlays. > > I agree we eventually need to fix the parts that break when we use > overlays. ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-20 8:06 ` Pantelis Antoniou 0 siblings, 0 replies; 53+ messages in thread From: Pantelis Antoniou @ 2017-10-20 8:06 UTC (permalink / raw) To: Frank Rowand Cc: Moritz Fischer, Rob Herring, Alan Tull, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga Hi Frank, > On Oct 20, 2017, at 00:46 , Frank Rowand <frowand.list@gmail.com> wrote: > > On 10/19/17 13:06, Moritz Fischer wrote: > > < snip > > >> We also have plenty of code that is just not aware of overlays, and >> assumes certain parts of the tree to stay static. > > I would state that somewhat differently. :-) There is very little > code that is aware of overlays, and most code assumes the device tree > does not change after early boot. > > This is one of the areas where the creation of overlays needs to be > done with care. > Correct. But this is not breaking the kernel. In general we have the following case where we load overlays (at least well formed overlays that are not doing stupid things). 1. Activation of a new device. Usually this works since is something that’s normally done at boot. 2. Deactivation of a device. This might break because the removal paths of platform device especially are not well tested (or never executed for that matter). 3. Modification of properties in an already activated device. If the device driver has not installed a device tree modification hook (as in almost 99% of the devices) it will do absolutely nothing, since the device tree is parsed only at probe time. I can argue that for these cases we could have a catch-all hook that displays a message that nothing happened. 4. Modification of some part of the tree that’s not part of a device driver, perhaps the aliases or chosen node. This may potentially be harmful or harmless depending on what has been modified. For instance modifying an already existing alias might cause internal inconsistency about device naming, while adding a new aliases should be harmless. This is a matter of policy per board, whether to allow or not. Are there other cases that are potentially more harmful? > >> I ran into that issue when I tried to add thermal zones via an overlay, >> I've been investigating how to fix the thermal framework to work with >> overlays since then and have some partially working code. >> Currently the thermal framework parses the thermal-zones node at boot, >> and assumes this stays static. This breaks with overlays. >> >> I agree we eventually need to fix the parts that break when we use >> overlays. > Regards — Pantelis ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-20 8:06 ` Pantelis Antoniou 0 siblings, 0 replies; 53+ messages in thread From: Pantelis Antoniou @ 2017-10-20 8:06 UTC (permalink / raw) To: Frank Rowand Cc: Moritz Fischer, Rob Herring, Alan Tull, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga Hi Frank, > On Oct 20, 2017, at 00:46 , Frank Rowand <frowand.list@gmail.com> = wrote: >=20 > On 10/19/17 13:06, Moritz Fischer wrote: >=20 > < snip > >=20 >> We also have plenty of code that is just not aware of overlays, and >> assumes certain parts of the tree to stay static. >=20 > I would state that somewhat differently. :-) There is very little > code that is aware of overlays, and most code assumes the device tree > does not change after early boot. >=20 > This is one of the areas where the creation of overlays needs to be > done with care. >=20 Correct. But this is not breaking the kernel. In general we have the following case where we load overlays (at least well formed overlays that are not doing stupid things). 1. Activation of a new device. Usually this works since is something = that=E2=80=99s normally done at boot. 2. Deactivation of a device. This might break because the removal paths of platform device especially are not well tested (or never executed for = that matter). 3. Modification of properties in an already activated device. If the = device driver has not installed a device tree modification hook (as in almost 99% of = the devices) it will do absolutely nothing, since the device tree is parsed only at = probe time. I can argue that for these cases we could have a catch-all hook that = displays a message that nothing happened. 4. Modification of some part of the tree that=E2=80=99s not part of a = device driver, perhaps the aliases or chosen node. This may potentially be harmful or harmless = depending on what has been modified. For instance modifying an already existing alias = might cause internal inconsistency about device naming, while adding a new aliases = should be harmless. This is a matter of policy per board, whether to allow or not. Are there other cases that are potentially more harmful? >=20 >> I ran into that issue when I tried to add thermal zones via an = overlay, >> I've been investigating how to fix the thermal framework to work with >> overlays since then and have some partially working code. >> Currently the thermal framework parses the thermal-zones node at = boot, >> and assumes this stays static. This breaks with overlays. >>=20 >> I agree we eventually need to fix the parts that break when we use >> overlays. >=20 Regards =E2=80=94 Pantelis ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-20 8:06 ` Pantelis Antoniou 0 siblings, 0 replies; 53+ messages in thread From: Pantelis Antoniou @ 2017-10-20 8:06 UTC (permalink / raw) To: Frank Rowand Cc: Moritz Fischer, Rob Herring, Alan Tull, devicetree-u79uwXL29TY76Z2rM5mHXA, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga-u79uwXL29TY76Z2rM5mHXA Hi Frank, > On Oct 20, 2017, at 00:46 , Frank Rowand <frowand.list-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote: > > On 10/19/17 13:06, Moritz Fischer wrote: > > < snip > > >> We also have plenty of code that is just not aware of overlays, and >> assumes certain parts of the tree to stay static. > > I would state that somewhat differently. :-) There is very little > code that is aware of overlays, and most code assumes the device tree > does not change after early boot. > > This is one of the areas where the creation of overlays needs to be > done with care. > Correct. But this is not breaking the kernel. In general we have the following case where we load overlays (at least well formed overlays that are not doing stupid things). 1. Activation of a new device. Usually this works since is something that’s normally done at boot. 2. Deactivation of a device. This might break because the removal paths of platform device especially are not well tested (or never executed for that matter). 3. Modification of properties in an already activated device. If the device driver has not installed a device tree modification hook (as in almost 99% of the devices) it will do absolutely nothing, since the device tree is parsed only at probe time. I can argue that for these cases we could have a catch-all hook that displays a message that nothing happened. 4. Modification of some part of the tree that’s not part of a device driver, perhaps the aliases or chosen node. This may potentially be harmful or harmless depending on what has been modified. For instance modifying an already existing alias might cause internal inconsistency about device naming, while adding a new aliases should be harmless. This is a matter of policy per board, whether to allow or not. Are there other cases that are potentially more harmful? > >> I ran into that issue when I tried to add thermal zones via an overlay, >> I've been investigating how to fix the thermal framework to work with >> overlays since then and have some partially working code. >> Currently the thermal framework parses the thermal-zones node at boot, >> and assumes this stays static. This breaks with overlays. >> >> I agree we eventually need to fix the parts that break when we use >> overlays. > Regards — Pantelis -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name 2017-10-20 8:06 ` Pantelis Antoniou (?) (?) @ 2017-10-20 14:59 ` Rob Herring -1 siblings, 0 replies; 53+ messages in thread From: Rob Herring @ 2017-10-20 14:59 UTC (permalink / raw) To: Pantelis Antoniou Cc: Frank Rowand, Moritz Fischer, Alan Tull, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga On Fri, Oct 20, 2017 at 3:06 AM, Pantelis Antoniou <pantelis.antoniou@konsulko.com> wrote: > Hi Frank, > >> On Oct 20, 2017, at 00:46 , Frank Rowand <frowand.list@gmail.com> wrote: >> >> On 10/19/17 13:06, Moritz Fischer wrote: >> >> < snip > >> >>> We also have plenty of code that is just not aware of overlays, and >>> assumes certain parts of the tree to stay static. >> >> I would state that somewhat differently. :-) There is very little >> code that is aware of overlays, and most code assumes the device tree >> does not change after early boot. >> >> This is one of the areas where the creation of overlays needs to be >> done with care. >> > > Correct. But this is not breaking the kernel. > > In general we have the following case where we load overlays (at least > well formed overlays that are not doing stupid things). > > 1. Activation of a new device. Usually this works since is something that’s > normally done at boot. Depends where in the tree it is. > 2. Deactivation of a device. This might break because the removal paths > of platform device especially are not well tested (or never executed for that > matter). It's more than just the removal path of a driver. What if you try to remove a node that's a dependency of others. Generally, hot unplug of GPIOs, clocks, interrupt controllers, etc. don't work. Plus we already have a way to disable devices. Unbind the driver. > 3. Modification of properties in an already activated device. If the device driver > has not installed a device tree modification hook (as in almost 99% of the devices) > it will do absolutely nothing, since the device tree is parsed only at probe time. > I can argue that for these cases we could have a catch-all hook that displays a > message that nothing happened. We don't know if a driver has kept pointers to property values or not. That's why we don't ever delete properties. If we are modifying or removing a property, then we just add new ones. A rogue program repeatedly modifying a property and allocating memory might be a cause for concern. > 4. Modification of some part of the tree that’s not part of a device driver, perhaps > the aliases or chosen node. This may potentially be harmful or harmless depending on > what has been modified. For instance modifying an already existing alias might cause > internal inconsistency about device naming, while adding a new aliases should be harmless. > This is a matter of policy per board, whether to allow or not. > > Are there other cases that are potentially more harmful? > >> >>> I ran into that issue when I tried to add thermal zones via an overlay, >>> I've been investigating how to fix the thermal framework to work with >>> overlays since then and have some partially working code. >>> Currently the thermal framework parses the thermal-zones node at boot, >>> and assumes this stays static. This breaks with overlays. >>> >>> I agree we eventually need to fix the parts that break when we use >>> overlays. >> > > Regards > > — Pantelis > ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name 2017-10-18 15:53 ` Pantelis Antoniou @ 2017-10-18 18:39 ` Alan Tull -1 siblings, 0 replies; 53+ messages in thread From: Alan Tull @ 2017-10-18 18:39 UTC (permalink / raw) To: Pantelis Antoniou Cc: Rob Herring, Frank Rowand, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga, Yves Vandervennet On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou <pantelis.antoniou@konsulko.com> wrote: > On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: >> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull@kernel.org> wrote: >> > On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list@gmail.com> wrote: >> >> On 10/17/17 14:46, Rob Herring wrote: >> >>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: >> >>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: >> >>>> >> >>>> Hi Rob, >> >>>> >> >>>>> With dependencies on a statically allocated full path name converted to >> >>>>> use %pOF format specifier, we can store just the basename of node, and >> >>>>> the unflattening of the FDT can be simplified. >> >>>>> >> >>>>> This commit will affect the remaining users of full_name. After >> >>>>> analyzing these users, the remaining cases should only change some print >> >>>>> messages. The main users of full_name are providing a name for struct >> >>>>> resource. The resource names shouldn't be important other than providing >> >>>>> /proc/iomem names. >> >>>>> >> >>>>> We no longer distinguish between pre and post 0x10 dtb formats as either >> >>>>> a full path or basename will work. However, less than 0x10 formats have >> >>>>> been broken since the conversion to use libfdt (and no one has cared). >> >>>>> The conversion of the unflattening code to be non-recursive also broke >> >>>>> pre 0x10 formats as the populate_node function would return 0 in that >> >>>>> case. >> >>>>> >> >>>>> Signed-off-by: Rob Herring <robh@kernel.org> >> >>>>> --- >> >>>>> v2: >> >>>>> - rebase to linux-next >> >>>>> >> >>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >> >>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >> >>>> >> >>>> I've just updated to the latest next branch and am finding problems >> >>>> applying overlays. Reverting this commit alleviates things. The >> >>>> errors I get are: >> >>>> >> >>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >> >>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >> >>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >> >>> >> >>> Frank's series with overlay updates should fix this. >> >> >> >> Yes, it does: >> >> >> >> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name >> > >> > Thanks for the fast response. I fetched the dt/next branch to test >> > this but there are sufficient changes that Pantelis' "OF: DT-Overlay >> > configfs interface (v7)" is broken now. I've been adding that >> > downstream since 4.4. We're using it as an interface for applying >> > overlays to program FPGAs. If we fix it again, is there any chance >> > that can go upstream now? >> >> With a drive-by posting once every few years, no. >> > I take offense to that. There's nothing changed in the patch for years. > Reposting the same patch without changes would achieve nothing. > >> The issue remains that the kernel is not really setup to deal with any >> random property or node to be changed at any point in run-time. Yeah, I'm not super surprised :) I have some whitelist ideas below. >> I >> think there needs to be some restrictions around what the overlays can >> touch. We can't have it be wide open and then lock things down later >> and break users. One example of what you could do is you can only add >> sub-trees to whitelisted nodes. That's probably acceptable for your >> usecase. I can take a look at making OF_OVERLAY_PRE_APPLY and OF_OVERLAY_PRE_REMOVE notifiers mandatory if that's interesting. The behavior would be: If an overlay is applied, there's got to be some handler in the kernel that verifies that it is acceptable. In my case, the handler for FPGA regions would look at the overlay and see it only added stuff under a FPGA region. And we would change the default to be: if there is no handler, reject the overlay. >> > > Defining what can and what cannot be changed is not as trivial as a > list of white-listed nodes. > > In some cases there is a whole node hierarchy being inserted (like in > a FPGA). For FPGA, the insertion points are always FPGA regions. > In others, it's merely changing a status property to "okay" and > a few device parameters. > > The real issue is that the kernel has no way to verify that a given > device tree, either at boot time or at overlay application time, is > correct. > > When the tree is wrong at boot-time you'll hang (if you're lucky). > If the tree is wrong at run-time you'll get some into some unidentified > funky state. > > Finally what is, and what is not 'correct' is not for the kernel to > decide arbitrarily, it's a matter of policy, different for each > use-case. > >> Rob > > Regards > > -- Pantelis Alan Tull > > ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-18 18:39 ` Alan Tull 0 siblings, 0 replies; 53+ messages in thread From: Alan Tull @ 2017-10-18 18:39 UTC (permalink / raw) To: Pantelis Antoniou Cc: Rob Herring, Frank Rowand, devicetree-u79uwXL29TY76Z2rM5mHXA, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga-u79uwXL29TY76Z2rM5mHXA, Yves Vandervennet On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou <pantelis.antoniou-OWPKS81ov/FWk0Htik3J/w@public.gmane.org> wrote: > On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: >> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >> > On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote: >> >> On 10/17/17 14:46, Rob Herring wrote: >> >>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >> >>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >> >>>> >> >>>> Hi Rob, >> >>>> >> >>>>> With dependencies on a statically allocated full path name converted to >> >>>>> use %pOF format specifier, we can store just the basename of node, and >> >>>>> the unflattening of the FDT can be simplified. >> >>>>> >> >>>>> This commit will affect the remaining users of full_name. After >> >>>>> analyzing these users, the remaining cases should only change some print >> >>>>> messages. The main users of full_name are providing a name for struct >> >>>>> resource. The resource names shouldn't be important other than providing >> >>>>> /proc/iomem names. >> >>>>> >> >>>>> We no longer distinguish between pre and post 0x10 dtb formats as either >> >>>>> a full path or basename will work. However, less than 0x10 formats have >> >>>>> been broken since the conversion to use libfdt (and no one has cared). >> >>>>> The conversion of the unflattening code to be non-recursive also broke >> >>>>> pre 0x10 formats as the populate_node function would return 0 in that >> >>>>> case. >> >>>>> >> >>>>> Signed-off-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> >> >>>>> --- >> >>>>> v2: >> >>>>> - rebase to linux-next >> >>>>> >> >>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >> >>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >> >>>> >> >>>> I've just updated to the latest next branch and am finding problems >> >>>> applying overlays. Reverting this commit alleviates things. The >> >>>> errors I get are: >> >>>> >> >>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >> >>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >> >>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >> >>> >> >>> Frank's series with overlay updates should fix this. >> >> >> >> Yes, it does: >> >> >> >> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name >> > >> > Thanks for the fast response. I fetched the dt/next branch to test >> > this but there are sufficient changes that Pantelis' "OF: DT-Overlay >> > configfs interface (v7)" is broken now. I've been adding that >> > downstream since 4.4. We're using it as an interface for applying >> > overlays to program FPGAs. If we fix it again, is there any chance >> > that can go upstream now? >> >> With a drive-by posting once every few years, no. >> > I take offense to that. There's nothing changed in the patch for years. > Reposting the same patch without changes would achieve nothing. > >> The issue remains that the kernel is not really setup to deal with any >> random property or node to be changed at any point in run-time. Yeah, I'm not super surprised :) I have some whitelist ideas below. >> I >> think there needs to be some restrictions around what the overlays can >> touch. We can't have it be wide open and then lock things down later >> and break users. One example of what you could do is you can only add >> sub-trees to whitelisted nodes. That's probably acceptable for your >> usecase. I can take a look at making OF_OVERLAY_PRE_APPLY and OF_OVERLAY_PRE_REMOVE notifiers mandatory if that's interesting. The behavior would be: If an overlay is applied, there's got to be some handler in the kernel that verifies that it is acceptable. In my case, the handler for FPGA regions would look at the overlay and see it only added stuff under a FPGA region. And we would change the default to be: if there is no handler, reject the overlay. >> > > Defining what can and what cannot be changed is not as trivial as a > list of white-listed nodes. > > In some cases there is a whole node hierarchy being inserted (like in > a FPGA). For FPGA, the insertion points are always FPGA regions. > In others, it's merely changing a status property to "okay" and > a few device parameters. > > The real issue is that the kernel has no way to verify that a given > device tree, either at boot time or at overlay application time, is > correct. > > When the tree is wrong at boot-time you'll hang (if you're lucky). > If the tree is wrong at run-time you'll get some into some unidentified > funky state. > > Finally what is, and what is not 'correct' is not for the kernel to > decide arbitrarily, it's a matter of policy, different for each > use-case. > >> Rob > > Regards > > -- Pantelis Alan Tull > > -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-18 21:40 ` Frank Rowand 0 siblings, 0 replies; 53+ messages in thread From: Frank Rowand @ 2017-10-18 21:40 UTC (permalink / raw) To: Alan Tull, Pantelis Antoniou Cc: Rob Herring, devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga, Yves Vandervennet On 10/18/17 11:39, Alan Tull wrote: > On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou > <pantelis.antoniou@konsulko.com> wrote: >> On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: >>> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull@kernel.org> wrote: >>>> On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list@gmail.com> wrote: >>>>> On 10/17/17 14:46, Rob Herring wrote: >>>>>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull@kernel.org> wrote: >>>>>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh@kernel.org> wrote: >>>>>>> >>>>>>> Hi Rob, >>>>>>> >>>>>>>> With dependencies on a statically allocated full path name converted to >>>>>>>> use %pOF format specifier, we can store just the basename of node, and >>>>>>>> the unflattening of the FDT can be simplified. >>>>>>>> >>>>>>>> This commit will affect the remaining users of full_name. After >>>>>>>> analyzing these users, the remaining cases should only change some print >>>>>>>> messages. The main users of full_name are providing a name for struct >>>>>>>> resource. The resource names shouldn't be important other than providing >>>>>>>> /proc/iomem names. >>>>>>>> >>>>>>>> We no longer distinguish between pre and post 0x10 dtb formats as either >>>>>>>> a full path or basename will work. However, less than 0x10 formats have >>>>>>>> been broken since the conversion to use libfdt (and no one has cared). >>>>>>>> The conversion of the unflattening code to be non-recursive also broke >>>>>>>> pre 0x10 formats as the populate_node function would return 0 in that >>>>>>>> case. >>>>>>>> >>>>>>>> Signed-off-by: Rob Herring <robh@kernel.org> >>>>>>>> --- >>>>>>>> v2: >>>>>>>> - rebase to linux-next >>>>>>>> >>>>>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >>>>>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >>>>>>> >>>>>>> I've just updated to the latest next branch and am finding problems >>>>>>> applying overlays. Reverting this commit alleviates things. The >>>>>>> errors I get are: >>>>>>> >>>>>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >>>>>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >>>>>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >>>>>> >>>>>> Frank's series with overlay updates should fix this. >>>>> >>>>> Yes, it does: >>>>> >>>>> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name >>>> >>>> Thanks for the fast response. I fetched the dt/next branch to test >>>> this but there are sufficient changes that Pantelis' "OF: DT-Overlay >>>> configfs interface (v7)" is broken now. I've been adding that >>>> downstream since 4.4. We're using it as an interface for applying >>>> overlays to program FPGAs. If we fix it again, is there any chance >>>> that can go upstream now? >>> >>> With a drive-by posting once every few years, no. >>> >> I take offense to that. There's nothing changed in the patch for years. >> Reposting the same patch without changes would achieve nothing. >> >>> The issue remains that the kernel is not really setup to deal with any >>> random property or node to be changed at any point in run-time. > > Yeah, I'm not super surprised :) I have some whitelist ideas below. > >>> I >>> think there needs to be some restrictions around what the overlays can >>> touch. We can't have it be wide open and then lock things down later >>> and break users. One example of what you could do is you can only add >>> sub-trees to whitelisted nodes. That's probably acceptable for your >>> usecase. > > I can take a look at making OF_OVERLAY_PRE_APPLY and > OF_OVERLAY_PRE_REMOVE notifiers mandatory if that's interesting. The > behavior would be: If an overlay is applied, there's got to be some > handler in the kernel that verifies that it is acceptable. In my > case, the handler for FPGA regions would look at the overlay and see > it only added stuff under a FPGA region. > > And we would change the default to be: if there is no handler, reject > the overlay. I think the responsibility belongs to the device tree core code. It should be centralized and consistent. In the fpga case, I think the connector paradigm should be adequate. The connector describes what is available to the fpga and provides access to those things. The fpga overlay does not reach outside the connector to touch anything else in the device tree. >> Defining what can and what cannot be changed is not as trivial as a >> list of white-listed nodes. >> >> In some cases there is a whole node hierarchy being inserted (like in >> a FPGA). > > For FPGA, the insertion points are always FPGA regions. > >> In others, it's merely changing a status property to "okay" and >> a few device parameters. >> >> The real issue is that the kernel has no way to verify that a given >> device tree, either at boot time or at overlay application time, is >> correct. >> >> When the tree is wrong at boot-time you'll hang (if you're lucky). >> If the tree is wrong at run-time you'll get some into some unidentified >> funky state. >> >> Finally what is, and what is not 'correct' is not for the kernel to >> decide arbitrarily, it's a matter of policy, different for each >> use-case. >> >>> Rob >> >> Regards >> >> -- Pantelis > > Alan Tull > >> >> > ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-10-18 21:40 ` Frank Rowand 0 siblings, 0 replies; 53+ messages in thread From: Frank Rowand @ 2017-10-18 21:40 UTC (permalink / raw) To: Alan Tull, Pantelis Antoniou Cc: Rob Herring, devicetree-u79uwXL29TY76Z2rM5mHXA, Michael Ellerman, linuxppc-dev, linux-kernel, Benjamin Herrenschmidt, Paul Mackerras, David Laight, linux-fpga-u79uwXL29TY76Z2rM5mHXA, Yves Vandervennet On 10/18/17 11:39, Alan Tull wrote: > On Wed, Oct 18, 2017 at 10:53 AM, Pantelis Antoniou > <pantelis.antoniou-OWPKS81ov/FWk0Htik3J/w@public.gmane.org> wrote: >> On Wed, 2017-10-18 at 10:44 -0500, Rob Herring wrote: >>> On Wed, Oct 18, 2017 at 10:12 AM, Alan Tull <atull-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >>>> On Tue, Oct 17, 2017 at 6:51 PM, Frank Rowand <frowand.list-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote: >>>>> On 10/17/17 14:46, Rob Herring wrote: >>>>>> On Tue, Oct 17, 2017 at 4:32 PM, Alan Tull <atull-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >>>>>>> On Mon, Aug 21, 2017 at 10:16 AM, Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >>>>>>> >>>>>>> Hi Rob, >>>>>>> >>>>>>>> With dependencies on a statically allocated full path name converted to >>>>>>>> use %pOF format specifier, we can store just the basename of node, and >>>>>>>> the unflattening of the FDT can be simplified. >>>>>>>> >>>>>>>> This commit will affect the remaining users of full_name. After >>>>>>>> analyzing these users, the remaining cases should only change some print >>>>>>>> messages. The main users of full_name are providing a name for struct >>>>>>>> resource. The resource names shouldn't be important other than providing >>>>>>>> /proc/iomem names. >>>>>>>> >>>>>>>> We no longer distinguish between pre and post 0x10 dtb formats as either >>>>>>>> a full path or basename will work. However, less than 0x10 formats have >>>>>>>> been broken since the conversion to use libfdt (and no one has cared). >>>>>>>> The conversion of the unflattening code to be non-recursive also broke >>>>>>>> pre 0x10 formats as the populate_node function would return 0 in that >>>>>>>> case. >>>>>>>> >>>>>>>> Signed-off-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> >>>>>>>> --- >>>>>>>> v2: >>>>>>>> - rebase to linux-next >>>>>>>> >>>>>>>> drivers/of/fdt.c | 69 +++++++++----------------------------------------------- >>>>>>>> 1 file changed, 11 insertions(+), 58 deletions(-) >>>>>>> >>>>>>> I've just updated to the latest next branch and am finding problems >>>>>>> applying overlays. Reverting this commit alleviates things. The >>>>>>> errors I get are: >>>>>>> >>>>>>> [ 88.498704] OF: overlay: Failed to apply prop @/__symbols__/clk_0 >>>>>>> [ 88.513447] OF: overlay: apply failed '/__symbols__' >>>>>>> [ 88.518423] create_overlay: Failed to create overlay (err=-12) >>>>>> >>>>>> Frank's series with overlay updates should fix this. >>>>> >>>>> Yes, it does: >>>>> >>>>> [PATCH v3 11/12] of: overlay: remove a dependency on device node full_name >>>> >>>> Thanks for the fast response. I fetched the dt/next branch to test >>>> this but there are sufficient changes that Pantelis' "OF: DT-Overlay >>>> configfs interface (v7)" is broken now. I've been adding that >>>> downstream since 4.4. We're using it as an interface for applying >>>> overlays to program FPGAs. If we fix it again, is there any chance >>>> that can go upstream now? >>> >>> With a drive-by posting once every few years, no. >>> >> I take offense to that. There's nothing changed in the patch for years. >> Reposting the same patch without changes would achieve nothing. >> >>> The issue remains that the kernel is not really setup to deal with any >>> random property or node to be changed at any point in run-time. > > Yeah, I'm not super surprised :) I have some whitelist ideas below. > >>> I >>> think there needs to be some restrictions around what the overlays can >>> touch. We can't have it be wide open and then lock things down later >>> and break users. One example of what you could do is you can only add >>> sub-trees to whitelisted nodes. That's probably acceptable for your >>> usecase. > > I can take a look at making OF_OVERLAY_PRE_APPLY and > OF_OVERLAY_PRE_REMOVE notifiers mandatory if that's interesting. The > behavior would be: If an overlay is applied, there's got to be some > handler in the kernel that verifies that it is acceptable. In my > case, the handler for FPGA regions would look at the overlay and see > it only added stuff under a FPGA region. > > And we would change the default to be: if there is no handler, reject > the overlay. I think the responsibility belongs to the device tree core code. It should be centralized and consistent. In the fpga case, I think the connector paradigm should be adequate. The connector describes what is available to the fpga and provides access to those things. The fpga overlay does not reach outside the connector to touch anything else in the device tree. >> Defining what can and what cannot be changed is not as trivial as a >> list of white-listed nodes. >> >> In some cases there is a whole node hierarchy being inserted (like in >> a FPGA). > > For FPGA, the insertion points are always FPGA regions. > >> In others, it's merely changing a status property to "okay" and >> a few device parameters. >> >> The real issue is that the kernel has no way to verify that a given >> device tree, either at boot time or at overlay application time, is >> correct. >> >> When the tree is wrong at boot-time you'll hang (if you're lucky). >> If the tree is wrong at run-time you'll get some into some unidentified >> funky state. >> >> Finally what is, and what is not 'correct' is not for the kernel to >> decide arbitrarily, it's a matter of policy, different for each >> use-case. >> >>> Rob >> >> Regards >> >> -- Pantelis > > Alan Tull > >> >> > -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-11-28 13:13 ` Geert Uytterhoeven 0 siblings, 0 replies; 53+ messages in thread From: Geert Uytterhoeven @ 2017-11-28 13:13 UTC (permalink / raw) To: Rob Herring Cc: devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight Hi Rob, On Mon, Aug 21, 2017 at 5:16 PM, Rob Herring <robh@kernel.org> wrote: > With dependencies on a statically allocated full path name converted to > use %pOF format specifier, we can store just the basename of node, and > the unflattening of the FDT can be simplified. > > This commit will affect the remaining users of full_name. After > analyzing these users, the remaining cases should only change some print > messages. The main users of full_name are providing a name for struct > resource. The resource names shouldn't be important other than providing > /proc/iomem names. I guess the plan is to get rid in a subsequent step of all calls to kbasename() on a full name, which is now futile? Thanks! Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-11-28 13:13 ` Geert Uytterhoeven 0 siblings, 0 replies; 53+ messages in thread From: Geert Uytterhoeven @ 2017-11-28 13:13 UTC (permalink / raw) To: Rob Herring Cc: devicetree-u79uwXL29TY76Z2rM5mHXA, Michael Ellerman, linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ, linux-kernel-u79uwXL29TY76Z2rM5mHXA, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight Hi Rob, On Mon, Aug 21, 2017 at 5:16 PM, Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: > With dependencies on a statically allocated full path name converted to > use %pOF format specifier, we can store just the basename of node, and > the unflattening of the FDT can be simplified. > > This commit will affect the remaining users of full_name. After > analyzing these users, the remaining cases should only change some print > messages. The main users of full_name are providing a name for struct > resource. The resource names shouldn't be important other than providing > /proc/iomem names. I guess the plan is to get rid in a subsequent step of all calls to kbasename() on a full name, which is now futile? Thanks! Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert-Td1EMuHUCqxL1ZNQvxDV9g@public.gmane.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-11-28 13:26 ` Rob Herring 0 siblings, 0 replies; 53+ messages in thread From: Rob Herring @ 2017-11-28 13:26 UTC (permalink / raw) To: Geert Uytterhoeven Cc: devicetree, Michael Ellerman, linuxppc-dev, linux-kernel, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight On Tue, Nov 28, 2017 at 7:13 AM, Geert Uytterhoeven <geert@linux-m68k.org> wrote: > Hi Rob, > > On Mon, Aug 21, 2017 at 5:16 PM, Rob Herring <robh@kernel.org> wrote: >> With dependencies on a statically allocated full path name converted to >> use %pOF format specifier, we can store just the basename of node, and >> the unflattening of the FDT can be simplified. >> >> This commit will affect the remaining users of full_name. After >> analyzing these users, the remaining cases should only change some print >> messages. The main users of full_name are providing a name for struct >> resource. The resource names shouldn't be important other than providing >> /proc/iomem names. > > I guess the plan is to get rid in a subsequent step of all calls to kbasename() > on a full name, which is now futile? No. Sparc (PDT) is still the full path and I don't plan to change that. Rob ^ permalink raw reply [flat|nested] 53+ messages in thread
* Re: [PATCH v2 5/5] of/fdt: only store the device node basename in full_name @ 2017-11-28 13:26 ` Rob Herring 0 siblings, 0 replies; 53+ messages in thread From: Rob Herring @ 2017-11-28 13:26 UTC (permalink / raw) To: Geert Uytterhoeven Cc: devicetree-u79uwXL29TY76Z2rM5mHXA, Michael Ellerman, linuxppc-dev-uLR06cmDAlY/bJ5BZ2RsiQ, linux-kernel-u79uwXL29TY76Z2rM5mHXA, Frank Rowand, Benjamin Herrenschmidt, Paul Mackerras, David Laight On Tue, Nov 28, 2017 at 7:13 AM, Geert Uytterhoeven <geert-Td1EMuHUCqxL1ZNQvxDV9g@public.gmane.org> wrote: > Hi Rob, > > On Mon, Aug 21, 2017 at 5:16 PM, Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> wrote: >> With dependencies on a statically allocated full path name converted to >> use %pOF format specifier, we can store just the basename of node, and >> the unflattening of the FDT can be simplified. >> >> This commit will affect the remaining users of full_name. After >> analyzing these users, the remaining cases should only change some print >> messages. The main users of full_name are providing a name for struct >> resource. The resource names shouldn't be important other than providing >> /proc/iomem names. > > I guess the plan is to get rid in a subsequent step of all calls to kbasename() > on a full name, which is now futile? No. Sparc (PDT) is still the full path and I don't plan to change that. Rob -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 53+ messages in thread
end of thread, other threads:[~2017-11-28 13:26 UTC | newest] Thread overview: 53+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2017-08-21 15:16 [PATCH v2 0/5] Removing full paths from DT full_name Rob Herring 2017-08-21 15:16 ` Rob Herring 2017-08-21 15:16 ` [PATCH v2 1/5] powerpc: Convert to using %pOF instead of full_name Rob Herring 2017-08-24 12:37 ` [v2,1/5] " Michael Ellerman 2017-08-24 12:37 ` Michael Ellerman 2017-08-21 15:16 ` [PATCH v2 2/5] powerpc: pseries: vio: match parent nodes with of_find_node_by_path Rob Herring 2017-08-22 5:12 ` Michael Ellerman 2017-08-22 5:12 ` Michael Ellerman 2017-08-24 18:08 ` Rob Herring 2017-08-24 18:08 ` Rob Herring 2017-08-25 0:10 ` Michael Ellerman 2017-08-25 0:10 ` Michael Ellerman 2017-08-21 15:16 ` [PATCH v2 3/5] powerpc: pseries: remove dlpar_attach_node dependency on full path Rob Herring 2017-08-24 12:37 ` [v2, " Michael Ellerman 2017-08-21 15:16 ` [PATCH v2 4/5] powerpc: pseries: only store the device node basename in full_name Rob Herring 2017-10-03 9:26 ` Michael Ellerman 2017-10-03 9:26 ` Michael Ellerman 2017-10-03 18:44 ` Rob Herring 2017-10-04 12:37 ` Michael Ellerman 2017-10-04 12:37 ` Michael Ellerman 2017-08-21 15:16 ` [PATCH v2 5/5] of/fdt: " Rob Herring 2017-10-17 21:32 ` Alan Tull 2017-10-17 21:32 ` Alan Tull 2017-10-17 21:46 ` Rob Herring 2017-10-17 23:51 ` Frank Rowand 2017-10-17 23:51 ` Frank Rowand 2017-10-18 15:12 ` Alan Tull 2017-10-18 15:12 ` Alan Tull 2017-10-18 15:44 ` Rob Herring 2017-10-18 15:53 ` Pantelis Antoniou 2017-10-18 15:53 ` Pantelis Antoniou 2017-10-18 18:30 ` Rob Herring 2017-10-18 21:46 ` Frank Rowand 2017-10-19 8:41 ` Pantelis Antoniou 2017-10-19 8:41 ` Pantelis Antoniou 2017-10-19 8:41 ` Pantelis Antoniou 2017-10-19 8:51 ` Pantelis Antoniou 2017-10-19 8:51 ` Pantelis Antoniou 2017-10-19 8:51 ` Pantelis Antoniou 2017-10-19 20:06 ` Moritz Fischer 2017-10-19 21:46 ` Frank Rowand 2017-10-20 8:06 ` Pantelis Antoniou 2017-10-20 8:06 ` Pantelis Antoniou 2017-10-20 8:06 ` Pantelis Antoniou 2017-10-20 14:59 ` Rob Herring 2017-10-18 18:39 ` Alan Tull 2017-10-18 18:39 ` Alan Tull 2017-10-18 21:40 ` Frank Rowand 2017-10-18 21:40 ` Frank Rowand 2017-11-28 13:13 ` Geert Uytterhoeven 2017-11-28 13:13 ` Geert Uytterhoeven 2017-11-28 13:26 ` Rob Herring 2017-11-28 13:26 ` Rob Herring
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.