From 0294112ee3135fbd15eaa70015af8283642dd970 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Sat, 4 Jul 2015 03:09:03 +0200 Subject: [PATCH 1/4] ACPI / PNP: Reserve ACPI resources at the fs_initcall_sync stage This effectively reverts the following three commits: 7bc10388ccdd ACPI / resources: free memory on error in add_region_before() 0f1b414d1907 ACPI / PNP: Avoid conflicting resource reservations b9a5e5e18fbf ACPI / init: Fix the ordering of acpi_reserve_resources() (commit b9a5e5e18fbf introduced regressions some of which, but not all, were addressed by commit 0f1b414d1907 and commit 7bc10388ccdd was a fixup on top of the latter) and causes ACPI fixed hardware resources to be reserved at the fs_initcall_sync stage of system initialization. The story is as follows. First, a boot regression was reported due to an apparent resource reservation ordering change after a commit that shouldn't lead to such changes. Investigation led to the conclusion that the problem happened because acpi_reserve_resources() was executed at the device_initcall() stage of system initialization which wasn't strictly ordered with respect to driver initialization (and with respect to the initialization of the pcieport driver in particular), so a random change causing the device initcalls to be run in a different order might break things. The response to that was to attempt to run acpi_reserve_resources() as soon as we knew that ACPI would be in use (commit b9a5e5e18fbf). However, that turned out to be too early, because it caused resource reservations made by the PNP system driver to fail on at least one system and that failure was addressed by commit 0f1b414d1907. That fix still turned out to be insufficient, though, because calling acpi_reserve_resources() before the fs_initcall stage of system initialization caused a boot regression to happen on the eCAFE EC-800-H20G/S netbook. That meant that we only could call acpi_reserve_resources() at the fs_initcall initialization stage or later, but then we might just as well call it after the PNP initalization in which case commit 0f1b414d1907 wouldn't be necessary any more. For this reason, the changes made by commit 0f1b414d1907 are reverted (along with a memory leak fixup on top of that commit), the changes made by commit b9a5e5e18fbf that went too far are reverted too and acpi_reserve_resources() is changed into fs_initcall_sync, which will cause it to be executed after the PNP subsystem initialization (which is an fs_initcall) and before device initcalls (including the pcieport driver initialization) which should avoid the initial issue. Link: https://bugzilla.kernel.org/show_bug.cgi?id=100581 Link: http://marc.info/?t=143092384600002&r=1&w=2 Link: https://bugzilla.kernel.org/show_bug.cgi?id=99831 Link: http://marc.info/?t=143389402600001&r=1&w=2 Fixes: b9a5e5e18fbf "ACPI / init: Fix the ordering of acpi_reserve_resources()" Reported-by: Roland Dreier Cc: All applicable Signed-off-by: Rafael J. Wysocki --- drivers/acpi/osl.c | 12 ++- drivers/acpi/resource.c | 162 ---------------------------------------- drivers/pnp/system.c | 35 +++------ include/linux/acpi.h | 10 --- 4 files changed, 18 insertions(+), 201 deletions(-) diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c index c262e4acd68d..3b8963f21b36 100644 --- a/drivers/acpi/osl.c +++ b/drivers/acpi/osl.c @@ -175,10 +175,14 @@ static void __init acpi_request_region (struct acpi_generic_address *gas, if (!addr || !length) return; - acpi_reserve_region(addr, length, gas->space_id, 0, desc); + /* Resources are never freed */ + if (gas->space_id == ACPI_ADR_SPACE_SYSTEM_IO) + request_region(addr, length, desc); + else if (gas->space_id == ACPI_ADR_SPACE_SYSTEM_MEMORY) + request_mem_region(addr, length, desc); } -static void __init acpi_reserve_resources(void) +static int __init acpi_reserve_resources(void) { acpi_request_region(&acpi_gbl_FADT.xpm1a_event_block, acpi_gbl_FADT.pm1_event_length, "ACPI PM1a_EVT_BLK"); @@ -207,7 +211,10 @@ static void __init acpi_reserve_resources(void) if (!(acpi_gbl_FADT.gpe1_block_length & 0x1)) acpi_request_region(&acpi_gbl_FADT.xgpe1_block, acpi_gbl_FADT.gpe1_block_length, "ACPI GPE1_BLK"); + + return 0; } +fs_initcall_sync(acpi_reserve_resources); void acpi_os_printf(const char *fmt, ...) { @@ -1862,7 +1869,6 @@ acpi_status __init acpi_os_initialize(void) acpi_status __init acpi_os_initialize1(void) { - acpi_reserve_resources(); kacpid_wq = alloc_workqueue("kacpid", 0, 1); kacpi_notify_wq = alloc_workqueue("kacpi_notify", 0, 1); kacpi_hotplug_wq = alloc_ordered_workqueue("kacpi_hotplug", 0); diff --git a/drivers/acpi/resource.c b/drivers/acpi/resource.c index 10561ce16ed1..8244f013f210 100644 --- a/drivers/acpi/resource.c +++ b/drivers/acpi/resource.c @@ -26,7 +26,6 @@ #include #include #include -#include #include #ifdef CONFIG_X86 @@ -622,164 +621,3 @@ int acpi_dev_filter_resource_type(struct acpi_resource *ares, return (type & types) ? 0 : 1; } EXPORT_SYMBOL_GPL(acpi_dev_filter_resource_type); - -struct reserved_region { - struct list_head node; - u64 start; - u64 end; -}; - -static LIST_HEAD(reserved_io_regions); -static LIST_HEAD(reserved_mem_regions); - -static int request_range(u64 start, u64 end, u8 space_id, unsigned long flags, - char *desc) -{ - unsigned int length = end - start + 1; - struct resource *res; - - res = space_id == ACPI_ADR_SPACE_SYSTEM_IO ? - request_region(start, length, desc) : - request_mem_region(start, length, desc); - if (!res) - return -EIO; - - res->flags &= ~flags; - return 0; -} - -static int add_region_before(u64 start, u64 end, u8 space_id, - unsigned long flags, char *desc, - struct list_head *head) -{ - struct reserved_region *reg; - int error; - - reg = kmalloc(sizeof(*reg), GFP_KERNEL); - if (!reg) - return -ENOMEM; - - error = request_range(start, end, space_id, flags, desc); - if (error) { - kfree(reg); - return error; - } - - reg->start = start; - reg->end = end; - list_add_tail(®->node, head); - return 0; -} - -/** - * acpi_reserve_region - Reserve an I/O or memory region as a system resource. - * @start: Starting address of the region. - * @length: Length of the region. - * @space_id: Identifier of address space to reserve the region from. - * @flags: Resource flags to clear for the region after requesting it. - * @desc: Region description (for messages). - * - * Reserve an I/O or memory region as a system resource to prevent others from - * using it. If the new region overlaps with one of the regions (in the given - * address space) already reserved by this routine, only the non-overlapping - * parts of it will be reserved. - * - * Returned is either 0 (success) or a negative error code indicating a resource - * reservation problem. It is the code of the first encountered error, but the - * routine doesn't abort until it has attempted to request all of the parts of - * the new region that don't overlap with other regions reserved previously. - * - * The resources requested by this routine are never released. - */ -int acpi_reserve_region(u64 start, unsigned int length, u8 space_id, - unsigned long flags, char *desc) -{ - struct list_head *regions; - struct reserved_region *reg; - u64 end = start + length - 1; - int ret = 0, error = 0; - - if (space_id == ACPI_ADR_SPACE_SYSTEM_IO) - regions = &reserved_io_regions; - else if (space_id == ACPI_ADR_SPACE_SYSTEM_MEMORY) - regions = &reserved_mem_regions; - else - return -EINVAL; - - if (list_empty(regions)) - return add_region_before(start, end, space_id, flags, desc, regions); - - list_for_each_entry(reg, regions, node) - if (reg->start == end + 1) { - /* The new region can be prepended to this one. */ - ret = request_range(start, end, space_id, flags, desc); - if (!ret) - reg->start = start; - - return ret; - } else if (reg->start > end) { - /* No overlap. Add the new region here and get out. */ - return add_region_before(start, end, space_id, flags, - desc, ®->node); - } else if (reg->end == start - 1) { - goto combine; - } else if (reg->end >= start) { - goto overlap; - } - - /* The new region goes after the last existing one. */ - return add_region_before(start, end, space_id, flags, desc, regions); - - overlap: - /* - * The new region overlaps an existing one. - * - * The head part of the new region immediately preceding the existing - * overlapping one can be combined with it right away. - */ - if (reg->start > start) { - error = request_range(start, reg->start - 1, space_id, flags, desc); - if (error) - ret = error; - else - reg->start = start; - } - - combine: - /* - * The new region is adjacent to an existing one. If it extends beyond - * that region all the way to the next one, it is possible to combine - * all three of them. - */ - while (reg->end < end) { - struct reserved_region *next = NULL; - u64 a = reg->end + 1, b = end; - - if (!list_is_last(®->node, regions)) { - next = list_next_entry(reg, node); - if (next->start <= end) - b = next->start - 1; - } - error = request_range(a, b, space_id, flags, desc); - if (!error) { - if (next && next->start == b + 1) { - reg->end = next->end; - list_del(&next->node); - kfree(next); - } else { - reg->end = end; - break; - } - } else if (next) { - if (!ret) - ret = error; - - reg = next; - } else { - break; - } - } - - return ret ? ret : error; -} -EXPORT_SYMBOL_GPL(acpi_reserve_region); diff --git a/drivers/pnp/system.c b/drivers/pnp/system.c index 515f33882ab8..49c1720df59a 100644 --- a/drivers/pnp/system.c +++ b/drivers/pnp/system.c @@ -7,7 +7,6 @@ * Bjorn Helgaas */ -#include #include #include #include @@ -23,41 +22,25 @@ static const struct pnp_device_id pnp_dev_table[] = { {"", 0} }; -#ifdef CONFIG_ACPI -static bool __reserve_range(u64 start, unsigned int length, bool io, char *desc) -{ - u8 space_id = io ? ACPI_ADR_SPACE_SYSTEM_IO : ACPI_ADR_SPACE_SYSTEM_MEMORY; - return !acpi_reserve_region(start, length, space_id, IORESOURCE_BUSY, desc); -} -#else -static bool __reserve_range(u64 start, unsigned int length, bool io, char *desc) -{ - struct resource *res; - - res = io ? request_region(start, length, desc) : - request_mem_region(start, length, desc); - if (res) { - res->flags &= ~IORESOURCE_BUSY; - return true; - } - return false; -} -#endif - static void reserve_range(struct pnp_dev *dev, struct resource *r, int port) { char *regionid; const char *pnpid = dev_name(&dev->dev); resource_size_t start = r->start, end = r->end; - bool reserved; + struct resource *res; regionid = kmalloc(16, GFP_KERNEL); if (!regionid) return; snprintf(regionid, 16, "pnp %s", pnpid); - reserved = __reserve_range(start, end - start + 1, !!port, regionid); - if (!reserved) + if (port) + res = request_region(start, end - start + 1, regionid); + else + res = request_mem_region(start, end - start + 1, regionid); + if (res) + res->flags &= ~IORESOURCE_BUSY; + else kfree(regionid); /* @@ -66,7 +49,7 @@ static void reserve_range(struct pnp_dev *dev, struct resource *r, int port) * have double reservations. */ dev_info(&dev->dev, "%pR %s reserved\n", r, - reserved ? "has been" : "could not be"); + res ? "has been" : "could not be"); } static void reserve_resources_of_dev(struct pnp_dev *dev) diff --git a/include/linux/acpi.h b/include/linux/acpi.h index c471dfc93b71..fedfd1bea3bf 100644 --- a/include/linux/acpi.h +++ b/include/linux/acpi.h @@ -309,9 +309,6 @@ int acpi_check_region(resource_size_t start, resource_size_t n, int acpi_resources_are_enforced(void); -int acpi_reserve_region(u64 start, unsigned int length, u8 space_id, - unsigned long flags, char *desc); - #ifdef CONFIG_HIBERNATION void __init acpi_no_s4_hw_signature(void); #endif @@ -507,13 +504,6 @@ static inline int acpi_check_region(resource_size_t start, resource_size_t n, return 0; } -static inline int acpi_reserve_region(u64 start, unsigned int length, - u8 space_id, unsigned long flags, - char *desc) -{ - return -ENXIO; -} - struct acpi_table_header; static inline int acpi_table_parse(char *id, int (*handler)(struct acpi_table_header *)) From d3e13ff3c1aa2403d9a5f371baac088daeb8f56d Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Tue, 7 Jul 2015 00:31:47 +0200 Subject: [PATCH 2/4] ACPI / LPSS: Fix up acpi_lpss_create_device() Fix a return value (which should be a negative error code) and a memory leak (the list allocated by acpi_dev_get_resources() needs to be freed on ioremap() errors too) in acpi_lpss_create_device() introduced by commit 4483d59e29fe 'ACPI / LPSS: check the result of ioremap()'. Fixes: 4483d59e29fe 'ACPI / LPSS: check the result of ioremap()' Reported-by: Dan Carpenter Cc: 4.0+ # 4.0+ Signed-off-by: Rafael J. Wysocki --- drivers/acpi/acpi_lpss.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/acpi/acpi_lpss.c b/drivers/acpi/acpi_lpss.c index 569ee090343f..46b58abb08c5 100644 --- a/drivers/acpi/acpi_lpss.c +++ b/drivers/acpi/acpi_lpss.c @@ -352,13 +352,16 @@ static int acpi_lpss_create_device(struct acpi_device *adev, pdata->mmio_size = resource_size(rentry->res); pdata->mmio_base = ioremap(rentry->res->start, pdata->mmio_size); - if (!pdata->mmio_base) - goto err_out; break; } acpi_dev_free_resource_list(&resource_list); + if (!pdata->mmio_base) { + ret = -ENOMEM; + goto err_out; + } + pdata->dev_desc = dev_desc; if (dev_desc->setup) From 93af5e93544328285a6f65f7d47bbea8979b28fb Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Fri, 26 Jun 2015 11:14:14 +0200 Subject: [PATCH 3/4] PM / Domains: Avoid infinite loops in attach/detach code If pm_genpd_{add,remove}_device() keeps on failing with -EAGAIN, we end up with an infinite loop in genpd_dev_pm_{at,de}tach(). This may happen due to a genpd.prepared_count imbalance. This is a bug elsewhere, but it will result in a system lock up, possibly during reboot of an otherwise functioning system. To avoid this, put a limit on the maximum number of loop iterations, using an exponential back-off mechanism. If the limit is reached, the operation will just fail. An error message is already printed. Signed-off-by: Geert Uytterhoeven Signed-off-by: Rafael J. Wysocki --- drivers/base/power/domain.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) diff --git a/drivers/base/power/domain.c b/drivers/base/power/domain.c index cdd547bd67df..0ee43c1056e0 100644 --- a/drivers/base/power/domain.c +++ b/drivers/base/power/domain.c @@ -6,6 +6,7 @@ * This file is released under the GPLv2. */ +#include #include #include #include @@ -19,6 +20,8 @@ #include #include +#define GENPD_RETRY_MAX_MS 250 /* Approximate */ + #define GENPD_DEV_CALLBACK(genpd, type, callback, dev) \ ({ \ type (*__routine)(struct device *__d); \ @@ -2131,6 +2134,7 @@ EXPORT_SYMBOL_GPL(of_genpd_get_from_provider); static void genpd_dev_pm_detach(struct device *dev, bool power_off) { struct generic_pm_domain *pd; + unsigned int i; int ret = 0; pd = pm_genpd_lookup_dev(dev); @@ -2139,10 +2143,12 @@ static void genpd_dev_pm_detach(struct device *dev, bool power_off) dev_dbg(dev, "removing from PM domain %s\n", pd->name); - while (1) { + for (i = 1; i < GENPD_RETRY_MAX_MS; i <<= 1) { ret = pm_genpd_remove_device(pd, dev); if (ret != -EAGAIN) break; + + mdelay(i); cond_resched(); } @@ -2183,6 +2189,7 @@ int genpd_dev_pm_attach(struct device *dev) { struct of_phandle_args pd_args; struct generic_pm_domain *pd; + unsigned int i; int ret; if (!dev->of_node) @@ -2218,10 +2225,12 @@ int genpd_dev_pm_attach(struct device *dev) dev_dbg(dev, "adding to PM domain %s\n", pd->name); - while (1) { + for (i = 1; i < GENPD_RETRY_MAX_MS; i <<= 1) { ret = pm_genpd_add_device(pd, dev); if (ret != -EAGAIN) break; + + mdelay(i); cond_resched(); } From b51f9b103f58db2c5c0a20d6cee26c8bc255d3ae Mon Sep 17 00:00:00 2001 From: Uwe Geuder Date: Mon, 29 Jun 2015 23:35:05 +0300 Subject: [PATCH 4/4] PM / hibernate: clarify resume documentation it was not the whole truth that kernel mode cannot be used with swap on LVM Signed-off-by: Uwe Geuder Acked-by: Pavel Machek Signed-off-by: Rafael J. Wysocki --- Documentation/power/swsusp.txt | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) diff --git a/Documentation/power/swsusp.txt b/Documentation/power/swsusp.txt index f732a8321e8a..8cc17ca71813 100644 --- a/Documentation/power/swsusp.txt +++ b/Documentation/power/swsusp.txt @@ -410,8 +410,17 @@ Documentation/usb/persist.txt. Q: Can I suspend-to-disk using a swap partition under LVM? -A: No. You can suspend successfully, but you'll not be able to -resume. uswsusp should be able to work with LVM. See suspend.sf.net. +A: Yes and No. You can suspend successfully, but the kernel will not be able +to resume on its own. You need an initramfs that can recognize the resume +situation, activate the logical volume containing the swap volume (but not +touch any filesystems!), and eventually call + +echo -n "$major:$minor" > /sys/power/resume + +where $major and $minor are the respective major and minor device numbers of +the swap volume. + +uswsusp works with LVM, too. See http://suspend.sourceforge.net/ Q: I upgraded the kernel from 2.6.15 to 2.6.16. Both kernels were compiled with the similar configuration files. Anyway I found that