From 114067b69e7b2c691faace0e33db2f04096f668d Mon Sep 17 00:00:00 2001 From: Namhyung Kim Date: Thu, 31 May 2012 14:43:27 +0900 Subject: perf tools: Check if callchain is corrupted We faced segmentation fault on perf top -G at very high sampling rate due to a corrupted callchain. While the root cause was not revealed (I failed to figure it out), this patch tries to protect us from the segfault on such cases. Reported-by: Arnaldo Carvalho de Melo Signed-off-by: Namhyung Kim Cc: Namhyung Kim Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Sunjin Yang Link: http://lkml.kernel.org/r/1338443007-24857-2-git-send-email-namhyung.kim@lge.com Signed-off-by: Arnaldo Carvalho de Melo --- include/linux/perf_event.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index f32578634d9d..1817d4015e5f 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -555,6 +555,8 @@ enum perf_event_type { PERF_RECORD_MAX, /* non-ABI */ }; +#define PERF_MAX_STACK_DEPTH 255 + enum perf_callchain_context { PERF_CONTEXT_HV = (__u64)-32, PERF_CONTEXT_KERNEL = (__u64)-128, @@ -609,8 +611,6 @@ struct perf_guest_info_callbacks { #include #include -#define PERF_MAX_STACK_DEPTH 255 - struct perf_callchain_entry { __u64 nr; __u64 ip[PERF_MAX_STACK_DEPTH]; -- cgit v1.2.3-71-gd317 From ae58d1e406986f31d1e88b32f5ac601506c196d8 Mon Sep 17 00:00:00 2001 From: Stephen Warren Date: Fri, 18 May 2012 09:29:34 -0600 Subject: i2c: Add generic I2C multiplexer using pinctrl API This is useful for SoCs whose I2C module's signals can be routed to different sets of pins at run-time, using the pinctrl API. +-----+ +-----+ | dev | | dev | +------------------------+ +-----+ +-----+ | SoC | | | | /----|------+--------+ | +---+ +------+ | child bus A, on first set of pins | |I2C|---|Pinmux| | | +---+ +------+ | child bus B, on second set of pins | \----|------+--------+--------+ | | | | | +------------------------+ +-----+ +-----+ +-----+ | dev | | dev | | dev | +-----+ +-----+ +-----+ Signed-off-by: Stephen Warren Acked-by: Linus Walleij Acked-by: Rob Herring Signed-off-by: Wolfram Sang --- .../devicetree/bindings/i2c/i2c-mux-pinctrl.txt | 93 +++++++ drivers/i2c/muxes/Kconfig | 12 + drivers/i2c/muxes/Makefile | 1 + drivers/i2c/muxes/i2c-mux-pinctrl.c | 279 +++++++++++++++++++++ include/linux/i2c-mux-pinctrl.h | 41 +++ 5 files changed, 426 insertions(+) create mode 100644 Documentation/devicetree/bindings/i2c/i2c-mux-pinctrl.txt create mode 100644 drivers/i2c/muxes/i2c-mux-pinctrl.c create mode 100644 include/linux/i2c-mux-pinctrl.h (limited to 'include/linux') diff --git a/Documentation/devicetree/bindings/i2c/i2c-mux-pinctrl.txt b/Documentation/devicetree/bindings/i2c/i2c-mux-pinctrl.txt new file mode 100644 index 000000000000..ae8af1694e95 --- /dev/null +++ b/Documentation/devicetree/bindings/i2c/i2c-mux-pinctrl.txt @@ -0,0 +1,93 @@ +Pinctrl-based I2C Bus Mux + +This binding describes an I2C bus multiplexer that uses pin multiplexing to +route the I2C signals, and represents the pin multiplexing configuration +using the pinctrl device tree bindings. + + +-----+ +-----+ + | dev | | dev | + +------------------------+ +-----+ +-----+ + | SoC | | | + | /----|------+--------+ + | +---+ +------+ | child bus A, on first set of pins + | |I2C|---|Pinmux| | + | +---+ +------+ | child bus B, on second set of pins + | \----|------+--------+--------+ + | | | | | + +------------------------+ +-----+ +-----+ +-----+ + | dev | | dev | | dev | + +-----+ +-----+ +-----+ + +Required properties: +- compatible: i2c-mux-pinctrl +- i2c-parent: The phandle of the I2C bus that this multiplexer's master-side + port is connected to. + +Also required are: + +* Standard pinctrl properties that specify the pin mux state for each child + bus. See ../pinctrl/pinctrl-bindings.txt. + +* Standard I2C mux properties. See mux.txt in this directory. + +* I2C child bus nodes. See mux.txt in this directory. + +For each named state defined in the pinctrl-names property, an I2C child bus +will be created. I2C child bus numbers are assigned based on the index into +the pinctrl-names property. + +The only exception is that no bus will be created for a state named "idle". If +such a state is defined, it must be the last entry in pinctrl-names. For +example: + + pinctrl-names = "ddc", "pta", "idle" -> ddc = bus 0, pta = bus 1 + pinctrl-names = "ddc", "idle", "pta" -> Invalid ("idle" not last) + pinctrl-names = "idle", "ddc", "pta" -> Invalid ("idle" not last) + +Whenever an access is made to a device on a child bus, the relevant pinctrl +state will be programmed into hardware. + +If an idle state is defined, whenever an access is not being made to a device +on a child bus, the idle pinctrl state will be programmed into hardware. + +If an idle state is not defined, the most recently used pinctrl state will be +left programmed into hardware whenever no access is being made of a device on +a child bus. + +Example: + + i2cmux { + compatible = "i2c-mux-pinctrl"; + #address-cells = <1>; + #size-cells = <0>; + + i2c-parent = <&i2c1>; + + pinctrl-names = "ddc", "pta", "idle"; + pinctrl-0 = <&state_i2cmux_ddc>; + pinctrl-1 = <&state_i2cmux_pta>; + pinctrl-2 = <&state_i2cmux_idle>; + + i2c@0 { + reg = <0>; + #address-cells = <1>; + #size-cells = <0>; + + eeprom { + compatible = "eeprom"; + reg = <0x50>; + }; + }; + + i2c@1 { + reg = <1>; + #address-cells = <1>; + #size-cells = <0>; + + eeprom { + compatible = "eeprom"; + reg = <0x50>; + }; + }; + }; + diff --git a/drivers/i2c/muxes/Kconfig b/drivers/i2c/muxes/Kconfig index beb2491db274..a0edd9854218 100644 --- a/drivers/i2c/muxes/Kconfig +++ b/drivers/i2c/muxes/Kconfig @@ -37,4 +37,16 @@ config I2C_MUX_PCA954x This driver can also be built as a module. If so, the module will be called i2c-mux-pca954x. +config I2C_MUX_PINCTRL + tristate "pinctrl-based I2C multiplexer" + depends on PINCTRL + help + If you say yes to this option, support will be included for an I2C + multiplexer that uses the pinctrl subsystem, i.e. pin multiplexing. + This is useful for SoCs whose I2C module's signals can be routed to + different sets of pins at run-time. + + This driver can also be built as a module. If so, the module will be + called pinctrl-i2cmux. + endmenu diff --git a/drivers/i2c/muxes/Makefile b/drivers/i2c/muxes/Makefile index 5826249b29ca..76da8692afff 100644 --- a/drivers/i2c/muxes/Makefile +++ b/drivers/i2c/muxes/Makefile @@ -4,5 +4,6 @@ obj-$(CONFIG_I2C_MUX_GPIO) += i2c-mux-gpio.o obj-$(CONFIG_I2C_MUX_PCA9541) += i2c-mux-pca9541.o obj-$(CONFIG_I2C_MUX_PCA954x) += i2c-mux-pca954x.o +obj-$(CONFIG_I2C_MUX_PINCTRL) += i2c-mux-pinctrl.o ccflags-$(CONFIG_I2C_DEBUG_BUS) := -DDEBUG diff --git a/drivers/i2c/muxes/i2c-mux-pinctrl.c b/drivers/i2c/muxes/i2c-mux-pinctrl.c new file mode 100644 index 000000000000..46a669763476 --- /dev/null +++ b/drivers/i2c/muxes/i2c-mux-pinctrl.c @@ -0,0 +1,279 @@ +/* + * I2C multiplexer using pinctrl API + * + * Copyright (c) 2012, NVIDIA CORPORATION. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +struct i2c_mux_pinctrl { + struct device *dev; + struct i2c_mux_pinctrl_platform_data *pdata; + struct pinctrl *pinctrl; + struct pinctrl_state **states; + struct pinctrl_state *state_idle; + struct i2c_adapter *parent; + struct i2c_adapter **busses; +}; + +static int i2c_mux_pinctrl_select(struct i2c_adapter *adap, void *data, + u32 chan) +{ + struct i2c_mux_pinctrl *mux = data; + + return pinctrl_select_state(mux->pinctrl, mux->states[chan]); +} + +static int i2c_mux_pinctrl_deselect(struct i2c_adapter *adap, void *data, + u32 chan) +{ + struct i2c_mux_pinctrl *mux = data; + + return pinctrl_select_state(mux->pinctrl, mux->state_idle); +} + +#ifdef CONFIG_OF +static int i2c_mux_pinctrl_parse_dt(struct i2c_mux_pinctrl *mux, + struct platform_device *pdev) +{ + struct device_node *np = pdev->dev.of_node; + int num_names, i, ret; + struct device_node *adapter_np; + struct i2c_adapter *adapter; + + if (!np) + return 0; + + mux->pdata = devm_kzalloc(&pdev->dev, sizeof(*mux->pdata), GFP_KERNEL); + if (!mux->pdata) { + dev_err(mux->dev, + "Cannot allocate i2c_mux_pinctrl_platform_data\n"); + return -ENOMEM; + } + + num_names = of_property_count_strings(np, "pinctrl-names"); + if (num_names < 0) { + dev_err(mux->dev, "Cannot parse pinctrl-names: %d\n", + num_names); + return num_names; + } + + mux->pdata->pinctrl_states = devm_kzalloc(&pdev->dev, + sizeof(*mux->pdata->pinctrl_states) * num_names, + GFP_KERNEL); + if (!mux->pdata->pinctrl_states) { + dev_err(mux->dev, "Cannot allocate pinctrl_states\n"); + return -ENOMEM; + } + + for (i = 0; i < num_names; i++) { + ret = of_property_read_string_index(np, "pinctrl-names", i, + &mux->pdata->pinctrl_states[mux->pdata->bus_count]); + if (ret < 0) { + dev_err(mux->dev, "Cannot parse pinctrl-names: %d\n", + ret); + return ret; + } + if (!strcmp(mux->pdata->pinctrl_states[mux->pdata->bus_count], + "idle")) { + if (i != num_names - 1) { + dev_err(mux->dev, "idle state must be last\n"); + return -EINVAL; + } + mux->pdata->pinctrl_state_idle = "idle"; + } else { + mux->pdata->bus_count++; + } + } + + adapter_np = of_parse_phandle(np, "i2c-parent", 0); + if (!adapter_np) { + dev_err(mux->dev, "Cannot parse i2c-parent\n"); + return -ENODEV; + } + adapter = of_find_i2c_adapter_by_node(adapter_np); + if (!adapter) { + dev_err(mux->dev, "Cannot find parent bus\n"); + return -ENODEV; + } + mux->pdata->parent_bus_num = i2c_adapter_id(adapter); + put_device(&adapter->dev); + + return 0; +} +#else +static inline int i2c_mux_pinctrl_parse_dt(struct i2c_mux_pinctrl *mux, + struct platform_device *pdev) +{ + return 0; +} +#endif + +static int __devinit i2c_mux_pinctrl_probe(struct platform_device *pdev) +{ + struct i2c_mux_pinctrl *mux; + int (*deselect)(struct i2c_adapter *, void *, u32); + int i, ret; + + mux = devm_kzalloc(&pdev->dev, sizeof(*mux), GFP_KERNEL); + if (!mux) { + dev_err(&pdev->dev, "Cannot allocate i2c_mux_pinctrl\n"); + ret = -ENOMEM; + goto err; + } + platform_set_drvdata(pdev, mux); + + mux->dev = &pdev->dev; + + mux->pdata = pdev->dev.platform_data; + if (!mux->pdata) { + ret = i2c_mux_pinctrl_parse_dt(mux, pdev); + if (ret < 0) + goto err; + } + if (!mux->pdata) { + dev_err(&pdev->dev, "Missing platform data\n"); + ret = -ENODEV; + goto err; + } + + mux->states = devm_kzalloc(&pdev->dev, + sizeof(*mux->states) * mux->pdata->bus_count, + GFP_KERNEL); + if (!mux->states) { + dev_err(&pdev->dev, "Cannot allocate states\n"); + ret = -ENOMEM; + goto err; + } + + mux->busses = devm_kzalloc(&pdev->dev, + sizeof(mux->busses) * mux->pdata->bus_count, + GFP_KERNEL); + if (!mux->states) { + dev_err(&pdev->dev, "Cannot allocate busses\n"); + ret = -ENOMEM; + goto err; + } + + mux->pinctrl = devm_pinctrl_get(&pdev->dev); + if (IS_ERR(mux->pinctrl)) { + ret = PTR_ERR(mux->pinctrl); + dev_err(&pdev->dev, "Cannot get pinctrl: %d\n", ret); + goto err; + } + for (i = 0; i < mux->pdata->bus_count; i++) { + mux->states[i] = pinctrl_lookup_state(mux->pinctrl, + mux->pdata->pinctrl_states[i]); + if (IS_ERR(mux->states[i])) { + ret = PTR_ERR(mux->states[i]); + dev_err(&pdev->dev, + "Cannot look up pinctrl state %s: %d\n", + mux->pdata->pinctrl_states[i], ret); + goto err; + } + } + if (mux->pdata->pinctrl_state_idle) { + mux->state_idle = pinctrl_lookup_state(mux->pinctrl, + mux->pdata->pinctrl_state_idle); + if (IS_ERR(mux->state_idle)) { + ret = PTR_ERR(mux->state_idle); + dev_err(&pdev->dev, + "Cannot look up pinctrl state %s: %d\n", + mux->pdata->pinctrl_state_idle, ret); + goto err; + } + + deselect = i2c_mux_pinctrl_deselect; + } else { + deselect = NULL; + } + + mux->parent = i2c_get_adapter(mux->pdata->parent_bus_num); + if (!mux->parent) { + dev_err(&pdev->dev, "Parent adapter (%d) not found\n", + mux->pdata->parent_bus_num); + ret = -ENODEV; + goto err; + } + + for (i = 0; i < mux->pdata->bus_count; i++) { + u32 bus = mux->pdata->base_bus_num ? + (mux->pdata->base_bus_num + i) : 0; + + mux->busses[i] = i2c_add_mux_adapter(mux->parent, &pdev->dev, + mux, bus, i, + i2c_mux_pinctrl_select, + deselect); + if (!mux->busses[i]) { + ret = -ENODEV; + dev_err(&pdev->dev, "Failed to add adapter %d\n", i); + goto err_del_adapter; + } + } + + return 0; + +err_del_adapter: + for (; i > 0; i--) + i2c_del_mux_adapter(mux->busses[i - 1]); + i2c_put_adapter(mux->parent); +err: + return ret; +} + +static int __devexit i2c_mux_pinctrl_remove(struct platform_device *pdev) +{ + struct i2c_mux_pinctrl *mux = platform_get_drvdata(pdev); + int i; + + for (i = 0; i < mux->pdata->bus_count; i++) + i2c_del_mux_adapter(mux->busses[i]); + + i2c_put_adapter(mux->parent); + + return 0; +} + +#ifdef CONFIG_OF +static const struct of_device_id i2c_mux_pinctrl_of_match[] __devinitconst = { + { .compatible = "i2c-mux-pinctrl", }, + {}, +}; +MODULE_DEVICE_TABLE(of, i2c_mux_pinctrl_of_match); +#endif + +static struct platform_driver i2c_mux_pinctrl_driver = { + .driver = { + .name = "i2c-mux-pinctrl", + .owner = THIS_MODULE, + .of_match_table = of_match_ptr(i2c_mux_pinctrl_of_match), + }, + .probe = i2c_mux_pinctrl_probe, + .remove = __devexit_p(i2c_mux_pinctrl_remove), +}; +module_platform_driver(i2c_mux_pinctrl_driver); + +MODULE_DESCRIPTION("pinctrl-based I2C multiplexer driver"); +MODULE_AUTHOR("Stephen Warren "); +MODULE_LICENSE("GPL v2"); +MODULE_ALIAS("platform:i2c-mux-pinctrl"); diff --git a/include/linux/i2c-mux-pinctrl.h b/include/linux/i2c-mux-pinctrl.h new file mode 100644 index 000000000000..a65c86429e84 --- /dev/null +++ b/include/linux/i2c-mux-pinctrl.h @@ -0,0 +1,41 @@ +/* + * i2c-mux-pinctrl platform data + * + * Copyright (c) 2012, NVIDIA CORPORATION. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +#ifndef _LINUX_I2C_MUX_PINCTRL_H +#define _LINUX_I2C_MUX_PINCTRL_H + +/** + * struct i2c_mux_pinctrl_platform_data - Platform data for i2c-mux-pinctrl + * @parent_bus_num: Parent I2C bus number + * @base_bus_num: Base I2C bus number for the child busses. 0 for dynamic. + * @bus_count: Number of child busses. Also the number of elements in + * @pinctrl_states + * @pinctrl_states: The names of the pinctrl state to select for each child bus + * @pinctrl_state_idle: The pinctrl state to select when no child bus is being + * accessed. If NULL, the most recently used pinctrl state will be left + * selected. + */ +struct i2c_mux_pinctrl_platform_data { + int parent_bus_num; + int base_bus_num; + int bus_count; + const char **pinctrl_states; + const char *pinctrl_state_idle; +}; + +#endif -- cgit v1.2.3-71-gd317 From 1549210fcc17e9ae20c09ac8cd4c48a8dfd431bd Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 5 Jun 2012 09:16:47 -0400 Subject: NFSv4: Fix an Oops in the open recovery code The open recovery code does not need to request a new value for the mdsthreshold, and so does not allocate a struct nfs4_threshold. The problem is that encode_getfattr_open() will still request an mdsthreshold, and so we end up Oopsing in decode_attr_mdsthreshold. This patch fixes encode_getfattr_open so that it doesn't request an mdsthreshold when the caller isn't asking for one. It also fixes decode_attr_mdsthreshold so that it errors if the server returns an mdsthreshold that we didn't ask for (instead of Oopsing). Signed-off-by: Trond Myklebust Cc: Andy Adamson --- fs/nfs/nfs4_fs.h | 2 +- fs/nfs/nfs4proc.c | 22 +++++++++++++++++++++- fs/nfs/nfs4xdr.c | 12 ++++++++---- include/linux/nfs_xdr.h | 1 + 4 files changed, 31 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index c6827f93ab57..cc5900ac61b5 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -295,7 +295,7 @@ is_ds_client(struct nfs_client *clp) extern const struct nfs4_minor_version_ops *nfs_v4_minor_ops[]; -extern const u32 nfs4_fattr_bitmap[2]; +extern const u32 nfs4_fattr_bitmap[3]; extern const u32 nfs4_statfs_bitmap[2]; extern const u32 nfs4_pathconf_bitmap[2]; extern const u32 nfs4_fsinfo_bitmap[3]; diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index d48dbefa0e71..ad1515521a6a 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -116,7 +116,7 @@ static int nfs4_map_errors(int err) /* * This is our standard bitmap for GETATTR requests. */ -const u32 nfs4_fattr_bitmap[2] = { +const u32 nfs4_fattr_bitmap[3] = { FATTR4_WORD0_TYPE | FATTR4_WORD0_CHANGE | FATTR4_WORD0_SIZE @@ -133,6 +133,24 @@ const u32 nfs4_fattr_bitmap[2] = { | FATTR4_WORD1_TIME_MODIFY }; +static const u32 nfs4_pnfs_open_bitmap[3] = { + FATTR4_WORD0_TYPE + | FATTR4_WORD0_CHANGE + | FATTR4_WORD0_SIZE + | FATTR4_WORD0_FSID + | FATTR4_WORD0_FILEID, + FATTR4_WORD1_MODE + | FATTR4_WORD1_NUMLINKS + | FATTR4_WORD1_OWNER + | FATTR4_WORD1_OWNER_GROUP + | FATTR4_WORD1_RAWDEV + | FATTR4_WORD1_SPACE_USED + | FATTR4_WORD1_TIME_ACCESS + | FATTR4_WORD1_TIME_METADATA + | FATTR4_WORD1_TIME_MODIFY, + FATTR4_WORD2_MDSTHRESHOLD +}; + const u32 nfs4_statfs_bitmap[2] = { FATTR4_WORD0_FILES_AVAIL | FATTR4_WORD0_FILES_FREE @@ -844,6 +862,7 @@ static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry, p->o_arg.name = &dentry->d_name; p->o_arg.server = server; p->o_arg.bitmask = server->attr_bitmask; + p->o_arg.open_bitmap = &nfs4_fattr_bitmap[0]; p->o_arg.claim = NFS4_OPEN_CLAIM_NULL; if (attrs != NULL && attrs->ia_valid != 0) { __be32 verf[2]; @@ -1820,6 +1839,7 @@ static int _nfs4_do_open(struct inode *dir, opendata->f_attr.mdsthreshold = pnfs_mdsthreshold_alloc(); if (!opendata->f_attr.mdsthreshold) goto err_opendata_put; + opendata->o_arg.open_bitmap = &nfs4_pnfs_open_bitmap[0]; } if (dentry->d_inode != NULL) opendata->state = nfs4_get_open_state(dentry->d_inode, sp); diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index ee4a74db95d0..9ca1428da9d9 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -1198,12 +1198,13 @@ static void encode_getfattr(struct xdr_stream *xdr, const u32* bitmask, struct c } static void encode_getfattr_open(struct xdr_stream *xdr, const u32 *bitmask, + const u32 *open_bitmap, struct compound_hdr *hdr) { encode_getattr_three(xdr, - bitmask[0] & nfs4_fattr_bitmap[0], - bitmask[1] & nfs4_fattr_bitmap[1], - bitmask[2] & FATTR4_WORD2_MDSTHRESHOLD, + bitmask[0] & open_bitmap[0], + bitmask[1] & open_bitmap[1], + bitmask[2] & open_bitmap[2], hdr); } @@ -2221,7 +2222,7 @@ static void nfs4_xdr_enc_open(struct rpc_rqst *req, struct xdr_stream *xdr, encode_putfh(xdr, args->fh, &hdr); encode_open(xdr, args, &hdr); encode_getfh(xdr, &hdr); - encode_getfattr_open(xdr, args->bitmask, &hdr); + encode_getfattr_open(xdr, args->bitmask, args->open_bitmap, &hdr); encode_nops(&hdr); } @@ -4360,6 +4361,9 @@ static int decode_attr_mdsthreshold(struct xdr_stream *xdr, if (unlikely(bitmap[2] & (FATTR4_WORD2_MDSTHRESHOLD - 1U))) return -EIO; if (likely(bitmap[2] & FATTR4_WORD2_MDSTHRESHOLD)) { + /* Did the server return an unrequested attribute? */ + if (unlikely(res == NULL)) + return -EREMOTEIO; p = xdr_inline_decode(xdr, 4); if (unlikely(!p)) goto out_overflow; diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index d1a7bf51c326..7519baef025b 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -348,6 +348,7 @@ struct nfs_openargs { const struct qstr * name; const struct nfs_server *server; /* Needed for ID mapping */ const u32 * bitmask; + const u32 * open_bitmap; __u32 claim; struct nfs4_sequence_args seq_args; }; -- cgit v1.2.3-71-gd317 From fffaee365fded09f9ebf2db19066065fa54323c3 Mon Sep 17 00:00:00 2001 From: Konstantin Khlebnikov Date: Tue, 5 Jun 2012 21:36:33 +0400 Subject: radix-tree: fix contiguous iterator MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This patch fixes bug in macro radix_tree_for_each_contig(). If radix_tree_next_slot() sees NULL in next slot it returns NULL, but following radix_tree_next_chunk() switches iterating into next chunk. As result iterating becomes non-contiguous and breaks vfs "splice" and all its users. Signed-off-by: Konstantin Khlebnikov Reported-and-bisected-by: Hans de Bruin Reported-and-bisected-by: Ondrej Zary Reported-bisected-and-tested-by: Toralf Förster Link: https://lkml.org/lkml/2012/6/5/64 Cc: stable # 3.4.x Signed-off-by: Linus Torvalds --- include/linux/radix-tree.h | 5 ++++- lib/radix-tree.c | 3 +++ 2 files changed, 7 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/radix-tree.h b/include/linux/radix-tree.h index 0d04cd69ab9b..ffc444c38b0a 100644 --- a/include/linux/radix-tree.h +++ b/include/linux/radix-tree.h @@ -368,8 +368,11 @@ radix_tree_next_slot(void **slot, struct radix_tree_iter *iter, unsigned flags) iter->index++; if (likely(*slot)) return slot; - if (flags & RADIX_TREE_ITER_CONTIG) + if (flags & RADIX_TREE_ITER_CONTIG) { + /* forbid switching to the next chunk */ + iter->next_index = 0; break; + } } } return NULL; diff --git a/lib/radix-tree.c b/lib/radix-tree.c index d7c878cc006c..e7964296fd50 100644 --- a/lib/radix-tree.c +++ b/lib/radix-tree.c @@ -686,6 +686,9 @@ void **radix_tree_next_chunk(struct radix_tree_root *root, * during iterating; it can be zero only at the beginning. * And we cannot overflow iter->next_index in a single step, * because RADIX_TREE_MAP_SHIFT < BITS_PER_LONG. + * + * This condition also used by radix_tree_next_slot() to stop + * contiguous iterating, and forbid swithing to the next chunk. */ index = iter->next_index; if (!index && iter->index) -- cgit v1.2.3-71-gd317 From 9bce008bae8b57bc7b007bcc2071d1247a527120 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 5 Jun 2012 18:32:03 -0400 Subject: NFS: Fix a commit bug The new commit code fails to copy the verifier into the wb_verf field of _all_ the nfs_page structures; it only copies it into the first entry. The consequence is that most requests end up failing to match in nfs_commit_release. Fix is to copy the verifier into the req->wb_verf field in nfs_write_completion. Signed-off-by: Trond Myklebust Cc: Fred Isaman --- fs/nfs/direct.c | 4 ++-- fs/nfs/write.c | 7 ++++--- include/linux/nfs_xdr.h | 2 ++ 3 files changed, 8 insertions(+), 5 deletions(-) (limited to 'include/linux') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index 23d170bc44f4..b5385a7efd56 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -710,12 +710,12 @@ static void nfs_direct_write_completion(struct nfs_pgio_header *hdr) if (dreq->flags == NFS_ODIRECT_RESCHED_WRITES) bit = NFS_IOHDR_NEED_RESCHED; else if (dreq->flags == 0) { - memcpy(&dreq->verf, &req->wb_verf, + memcpy(&dreq->verf, hdr->verf, sizeof(dreq->verf)); bit = NFS_IOHDR_NEED_COMMIT; dreq->flags = NFS_ODIRECT_DO_COMMIT; } else if (dreq->flags == NFS_ODIRECT_DO_COMMIT) { - if (memcmp(&dreq->verf, &req->wb_verf, sizeof(dreq->verf))) { + if (memcmp(&dreq->verf, hdr->verf, sizeof(dreq->verf))) { dreq->flags = NFS_ODIRECT_RESCHED_WRITES; bit = NFS_IOHDR_NEED_RESCHED; } else diff --git a/fs/nfs/write.c b/fs/nfs/write.c index e6fe3d69d14c..4d6861c0dc14 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -80,6 +80,7 @@ struct nfs_write_header *nfs_writehdr_alloc(void) INIT_LIST_HEAD(&hdr->rpc_list); spin_lock_init(&hdr->lock); atomic_set(&hdr->refcnt, 0); + hdr->verf = &p->verf; } return p; } @@ -619,6 +620,7 @@ static void nfs_write_completion(struct nfs_pgio_header *hdr) goto next; } if (test_bit(NFS_IOHDR_NEED_COMMIT, &hdr->flags)) { + memcpy(&req->wb_verf, hdr->verf, sizeof(req->wb_verf)); nfs_mark_request_commit(req, hdr->lseg, &cinfo); goto next; } @@ -1255,15 +1257,14 @@ static void nfs_writeback_release_common(void *calldata) struct nfs_write_data *data = calldata; struct nfs_pgio_header *hdr = data->header; int status = data->task.tk_status; - struct nfs_page *req = hdr->req; if ((status >= 0) && nfs_write_need_commit(data)) { spin_lock(&hdr->lock); if (test_bit(NFS_IOHDR_NEED_RESCHED, &hdr->flags)) ; /* Do nothing */ else if (!test_and_set_bit(NFS_IOHDR_NEED_COMMIT, &hdr->flags)) - memcpy(&req->wb_verf, &data->verf, sizeof(req->wb_verf)); - else if (memcmp(&req->wb_verf, &data->verf, sizeof(req->wb_verf))) + memcpy(hdr->verf, &data->verf, sizeof(*hdr->verf)); + else if (memcmp(hdr->verf, &data->verf, sizeof(*hdr->verf))) set_bit(NFS_IOHDR_NEED_RESCHED, &hdr->flags); spin_unlock(&hdr->lock); } diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 7519baef025b..8aadd90b808a 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1237,6 +1237,7 @@ struct nfs_pgio_header { struct list_head rpc_list; atomic_t refcnt; struct nfs_page *req; + struct nfs_writeverf *verf; struct pnfs_layout_segment *lseg; loff_t io_start; const struct rpc_call_ops *mds_ops; @@ -1274,6 +1275,7 @@ struct nfs_write_data { struct nfs_write_header { struct nfs_pgio_header header; struct nfs_write_data rpc_data; + struct nfs_writeverf verf; }; struct nfs_mds_commit_info { -- cgit v1.2.3-71-gd317 From c1174876874dcf8986806e4dad3d7d07af20b439 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 31 May 2012 14:47:33 +0200 Subject: sched: Fix domain iteration Weird topologies can lead to asymmetric domain setups. This needs further consideration since these setups are typically non-minimal too. For now, make it work by adding an extra mask selecting which CPUs are allowed to iterate up. The topology that triggered it is the one from David Rientjes: 10 20 20 30 20 10 20 20 20 20 10 20 30 20 20 10 resulting in boxes that wouldn't even boot. Reported-by: David Rientjes Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/n/tip-3p86l9cuaqnxz7uxsojmz5rm@git.kernel.org Signed-off-by: Ingo Molnar --- include/linux/sched.h | 11 +++++++++ kernel/sched/core.c | 64 ++++++++++++++++++++++++++++++++++++++++++++------- kernel/sched/fair.c | 5 ++-- kernel/sched/sched.h | 2 ++ 4 files changed, 72 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sched.h b/include/linux/sched.h index 6029d8c54476..ac321d753470 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -876,6 +876,8 @@ struct sched_group_power { * Number of busy cpus in this group. */ atomic_t nr_busy_cpus; + + unsigned long cpumask[0]; /* iteration mask */ }; struct sched_group { @@ -900,6 +902,15 @@ static inline struct cpumask *sched_group_cpus(struct sched_group *sg) return to_cpumask(sg->cpumask); } +/* + * cpumask masking which cpus in the group are allowed to iterate up the domain + * tree. + */ +static inline struct cpumask *sched_group_mask(struct sched_group *sg) +{ + return to_cpumask(sg->sgp->cpumask); +} + /** * group_first_cpu - Returns the first cpu in the cpumask of a sched_group. * @group: The group whose first cpu is to be returned. diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 6546083af3e0..781acb91a50a 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -5994,6 +5994,44 @@ struct sched_domain_topology_level { struct sd_data data; }; +/* + * Build an iteration mask that can exclude certain CPUs from the upwards + * domain traversal. + * + * Asymmetric node setups can result in situations where the domain tree is of + * unequal depth, make sure to skip domains that already cover the entire + * range. + * + * In that case build_sched_domains() will have terminated the iteration early + * and our sibling sd spans will be empty. Domains should always include the + * cpu they're built on, so check that. + * + */ +static void build_group_mask(struct sched_domain *sd, struct sched_group *sg) +{ + const struct cpumask *span = sched_domain_span(sd); + struct sd_data *sdd = sd->private; + struct sched_domain *sibling; + int i; + + for_each_cpu(i, span) { + sibling = *per_cpu_ptr(sdd->sd, i); + if (!cpumask_test_cpu(i, sched_domain_span(sibling))) + continue; + + cpumask_set_cpu(i, sched_group_mask(sg)); + } +} + +/* + * Return the canonical balance cpu for this group, this is the first cpu + * of this group that's also in the iteration mask. + */ +int group_balance_cpu(struct sched_group *sg) +{ + return cpumask_first_and(sched_group_cpus(sg), sched_group_mask(sg)); +} + static int build_overlap_sched_groups(struct sched_domain *sd, int cpu) { @@ -6012,6 +6050,12 @@ build_overlap_sched_groups(struct sched_domain *sd, int cpu) if (cpumask_test_cpu(i, covered)) continue; + child = *per_cpu_ptr(sdd->sd, i); + + /* See the comment near build_group_mask(). */ + if (!cpumask_test_cpu(i, sched_domain_span(child))) + continue; + sg = kzalloc_node(sizeof(struct sched_group) + cpumask_size(), GFP_KERNEL, cpu_to_node(cpu)); @@ -6019,8 +6063,6 @@ build_overlap_sched_groups(struct sched_domain *sd, int cpu) goto fail; sg_span = sched_group_cpus(sg); - - child = *per_cpu_ptr(sdd->sd, i); if (child->child) { child = child->child; cpumask_copy(sg_span, sched_domain_span(child)); @@ -6030,13 +6072,18 @@ build_overlap_sched_groups(struct sched_domain *sd, int cpu) cpumask_or(covered, covered, sg_span); sg->sgp = *per_cpu_ptr(sdd->sgp, i); - atomic_inc(&sg->sgp->ref); + if (atomic_inc_return(&sg->sgp->ref) == 1) + build_group_mask(sd, sg); + + /* + * Make sure the first group of this domain contains the + * canonical balance cpu. Otherwise the sched_domain iteration + * breaks. See update_sg_lb_stats(). + */ if ((!groups && cpumask_test_cpu(cpu, sg_span)) || - cpumask_first(sg_span) == cpu) { - WARN_ON_ONCE(!cpumask_test_cpu(cpu, sg_span)); + group_balance_cpu(sg) == cpu) groups = sg; - } if (!first) first = sg; @@ -6109,6 +6156,7 @@ build_sched_groups(struct sched_domain *sd, int cpu) cpumask_clear(sched_group_cpus(sg)); sg->sgp->power = 0; + cpumask_setall(sched_group_mask(sg)); for_each_cpu(j, span) { if (get_group(j, sdd, NULL) != group) @@ -6150,7 +6198,7 @@ static void init_sched_groups_power(int cpu, struct sched_domain *sd) sg = sg->next; } while (sg != sd->groups); - if (cpu != group_first_cpu(sg)) + if (cpu != group_balance_cpu(sg)) return; update_group_power(sd, cpu); @@ -6525,7 +6573,7 @@ static int __sdt_alloc(const struct cpumask *cpu_map) *per_cpu_ptr(sdd->sg, j) = sg; - sgp = kzalloc_node(sizeof(struct sched_group_power), + sgp = kzalloc_node(sizeof(struct sched_group_power) + cpumask_size(), GFP_KERNEL, cpu_to_node(j)); if (!sgp) return -ENOMEM; diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index b2a2d236f27b..54cbaa4e7b37 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -3652,7 +3652,7 @@ static inline void update_sg_lb_stats(struct lb_env *env, int i; if (local_group) - balance_cpu = group_first_cpu(group); + balance_cpu = group_balance_cpu(group); /* Tally up the load of all CPUs in the group */ max_cpu_load = 0; @@ -3667,7 +3667,8 @@ static inline void update_sg_lb_stats(struct lb_env *env, /* Bias balancing toward cpus of our domain */ if (local_group) { - if (idle_cpu(i) && !first_idle_cpu) { + if (idle_cpu(i) && !first_idle_cpu && + cpumask_test_cpu(i, sched_group_mask(group))) { first_idle_cpu = 1; balance_cpu = i; } diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index ba9dccfd24ce..6d52cea7f33d 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -526,6 +526,8 @@ static inline struct sched_domain *highest_flag_domain(int cpu, int flag) DECLARE_PER_CPU(struct sched_domain *, sd_llc); DECLARE_PER_CPU(int, sd_llc_id); +extern int group_balance_cpu(struct sched_group *sg); + #endif /* CONFIG_SMP */ #include "stats.h" -- cgit v1.2.3-71-gd317 From 0b0d9cf6ec7bab91977da2d71c09157f110f7c2e Mon Sep 17 00:00:00 2001 From: Arun Sharma Date: Fri, 20 Apr 2012 15:41:34 -0700 Subject: perf: Limit callchains to 127 Stack depth of 255 seems excessive, given that copy_from_user_nmi() could be slow. Signed-off-by: Arun Sharma Cc: Linus Torvalds Cc: linux-kernel@vger.kernel.org Signed-off-by: Peter Zijlstra Link: http://lkml.kernel.org/r/1334961696-19580-3-git-send-email-asharma@fb.com Signed-off-by: Ingo Molnar --- include/linux/perf_event.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 1817d4015e5f..45db49f64bb4 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -555,7 +555,7 @@ enum perf_event_type { PERF_RECORD_MAX, /* non-ABI */ }; -#define PERF_MAX_STACK_DEPTH 255 +#define PERF_MAX_STACK_DEPTH 127 enum perf_callchain_context { PERF_CONTEXT_HV = (__u64)-32, -- cgit v1.2.3-71-gd317 From aa9b16306e3243229580ff889cc59fd66bf77973 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Thu, 10 May 2012 16:41:44 -0700 Subject: rcu: Precompute RCU_FAST_NO_HZ timer offsets When a CPU is entering dyntick-idle mode, tick_nohz_stop_sched_tick() calls rcu_needs_cpu() see if RCU needs that CPU, and, if not, computes the next wakeup time based on the timer wheels. Only later, when actually entering the idle loop, rcu_prepare_for_idle() will be invoked. In some cases, rcu_prepare_for_idle() will post timers to wake the CPU back up. But all for naught: The next wakeup time for the CPU has already been computed, and posting a timer afterwards does not force that wakeup time to be recomputed. This means that rcu_prepare_for_idle()'s have no effect. This is not a problem on a busy system because something else will wake up the CPU soon enough. However, on lightly loaded systems, the CPU might stay asleep for a considerable length of time. If that CPU has a callback that the rest of the system is waiting on, the system might run very slowly or (in theory) even hang. This commit avoids this problem by having rcu_needs_cpu() give tick_nohz_stop_sched_tick() an estimate of when RCU will need the CPU to wake back up, which tick_nohz_stop_sched_tick() takes into account when programming the CPU's wakeup time. An alternative approach is for rcu_prepare_for_idle() to use hrtimers instead of normal timers, but timers are much more efficient than are hrtimers for frequently and repeatedly posting and cancelling a given timer, which is exactly what RCU_FAST_NO_HZ does. Reported-by: Pascal Chapperon Reported-by: Heiko Carstens Signed-off-by: Paul E. McKenney Signed-off-by: Paul E. McKenney Tested-by: Heiko Carstens Tested-by: Pascal Chapperon --- include/linux/rcutiny.h | 6 +++-- include/linux/rcutree.h | 2 +- kernel/rcutree_plugin.h | 66 +++++++++++++++++++++++++++++++----------------- kernel/time/tick-sched.c | 7 ++++- 4 files changed, 54 insertions(+), 27 deletions(-) (limited to 'include/linux') diff --git a/include/linux/rcutiny.h b/include/linux/rcutiny.h index adb5e5a38cae..854dc4c5c271 100644 --- a/include/linux/rcutiny.h +++ b/include/linux/rcutiny.h @@ -87,8 +87,9 @@ static inline void kfree_call_rcu(struct rcu_head *head, #ifdef CONFIG_TINY_RCU -static inline int rcu_needs_cpu(int cpu) +static inline int rcu_needs_cpu(int cpu, unsigned long *delta_jiffies) { + *delta_jiffies = ULONG_MAX; return 0; } @@ -96,8 +97,9 @@ static inline int rcu_needs_cpu(int cpu) int rcu_preempt_needs_cpu(void); -static inline int rcu_needs_cpu(int cpu) +static inline int rcu_needs_cpu(int cpu, unsigned long *delta_jiffies) { + *delta_jiffies = ULONG_MAX; return rcu_preempt_needs_cpu(); } diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h index 3c6083cde4fc..952b79339304 100644 --- a/include/linux/rcutree.h +++ b/include/linux/rcutree.h @@ -32,7 +32,7 @@ extern void rcu_init(void); extern void rcu_note_context_switch(int cpu); -extern int rcu_needs_cpu(int cpu); +extern int rcu_needs_cpu(int cpu, unsigned long *delta_jiffies); extern void rcu_cpu_stall_reset(void); /* diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h index 6bd9637d5d83..5271a020887e 100644 --- a/kernel/rcutree_plugin.h +++ b/kernel/rcutree_plugin.h @@ -1886,8 +1886,9 @@ static void __cpuinit rcu_prepare_kthreads(int cpu) * Because we not have RCU_FAST_NO_HZ, just check whether this CPU needs * any flavor of RCU. */ -int rcu_needs_cpu(int cpu) +int rcu_needs_cpu(int cpu, unsigned long *delta_jiffies) { + *delta_jiffies = ULONG_MAX; return rcu_cpu_has_callbacks(cpu); } @@ -1962,28 +1963,6 @@ static void rcu_idle_count_callbacks_posted(void) #define RCU_IDLE_GP_DELAY 6 /* Roughly one grace period. */ #define RCU_IDLE_LAZY_GP_DELAY (6 * HZ) /* Roughly six seconds. */ -/* - * Allow the CPU to enter dyntick-idle mode if either: (1) There are no - * callbacks on this CPU, (2) this CPU has not yet attempted to enter - * dyntick-idle mode, or (3) this CPU is in the process of attempting to - * enter dyntick-idle mode. Otherwise, if we have recently tried and failed - * to enter dyntick-idle mode, we refuse to try to enter it. After all, - * it is better to incur scheduling-clock interrupts than to spin - * continuously for the same time duration! - */ -int rcu_needs_cpu(int cpu) -{ - struct rcu_dynticks *rdtp = &per_cpu(rcu_dynticks, cpu); - - /* Flag a new idle sojourn to the idle-entry state machine. */ - rdtp->idle_first_pass = 1; - /* If no callbacks, RCU doesn't need the CPU. */ - if (!rcu_cpu_has_callbacks(cpu)) - return 0; - /* Otherwise, RCU needs the CPU only if it recently tried and failed. */ - return rdtp->dyntick_holdoff == jiffies; -} - /* * Does the specified flavor of RCU have non-lazy callbacks pending on * the specified CPU? Both RCU flavor and CPU are specified by the @@ -2026,6 +2005,47 @@ static bool rcu_cpu_has_nonlazy_callbacks(int cpu) rcu_preempt_cpu_has_nonlazy_callbacks(cpu); } +/* + * Allow the CPU to enter dyntick-idle mode if either: (1) There are no + * callbacks on this CPU, (2) this CPU has not yet attempted to enter + * dyntick-idle mode, or (3) this CPU is in the process of attempting to + * enter dyntick-idle mode. Otherwise, if we have recently tried and failed + * to enter dyntick-idle mode, we refuse to try to enter it. After all, + * it is better to incur scheduling-clock interrupts than to spin + * continuously for the same time duration! + * + * The delta_jiffies argument is used to store the time when RCU is + * going to need the CPU again if it still has callbacks. The reason + * for this is that rcu_prepare_for_idle() might need to post a timer, + * but if so, it will do so after tick_nohz_stop_sched_tick() has set + * the wakeup time for this CPU. This means that RCU's timer can be + * delayed until the wakeup time, which defeats the purpose of posting + * a timer. + */ +int rcu_needs_cpu(int cpu, unsigned long *delta_jiffies) +{ + struct rcu_dynticks *rdtp = &per_cpu(rcu_dynticks, cpu); + + /* Flag a new idle sojourn to the idle-entry state machine. */ + rdtp->idle_first_pass = 1; + /* If no callbacks, RCU doesn't need the CPU. */ + if (!rcu_cpu_has_callbacks(cpu)) { + *delta_jiffies = ULONG_MAX; + return 0; + } + if (rdtp->dyntick_holdoff == jiffies) { + /* RCU recently tried and failed, so don't try again. */ + *delta_jiffies = 1; + return 1; + } + /* Set up for the possibility that RCU will post a timer. */ + if (rcu_cpu_has_nonlazy_callbacks(cpu)) + *delta_jiffies = RCU_IDLE_GP_DELAY; + else + *delta_jiffies = RCU_IDLE_LAZY_GP_DELAY; + return 0; +} + /* * Handler for smp_call_function_single(). The only point of this * handler is to wake the CPU up, so the handler does only tracing. diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 6a3a5b9ff561..52f5ebbd443b 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -274,6 +274,7 @@ EXPORT_SYMBOL_GPL(get_cpu_iowait_time_us); static void tick_nohz_stop_sched_tick(struct tick_sched *ts) { unsigned long seq, last_jiffies, next_jiffies, delta_jiffies; + unsigned long rcu_delta_jiffies; ktime_t last_update, expires, now; struct clock_event_device *dev = __get_cpu_var(tick_cpu_device).evtdev; u64 time_delta; @@ -322,7 +323,7 @@ static void tick_nohz_stop_sched_tick(struct tick_sched *ts) time_delta = timekeeping_max_deferment(); } while (read_seqretry(&xtime_lock, seq)); - if (rcu_needs_cpu(cpu) || printk_needs_cpu(cpu) || + if (rcu_needs_cpu(cpu, &rcu_delta_jiffies) || printk_needs_cpu(cpu) || arch_needs_cpu(cpu)) { next_jiffies = last_jiffies + 1; delta_jiffies = 1; @@ -330,6 +331,10 @@ static void tick_nohz_stop_sched_tick(struct tick_sched *ts) /* Get the next timer wheel timer */ next_jiffies = get_next_timer_interrupt(last_jiffies); delta_jiffies = next_jiffies - last_jiffies; + if (rcu_delta_jiffies < delta_jiffies) { + next_jiffies = last_jiffies + rcu_delta_jiffies; + delta_jiffies = rcu_delta_jiffies; + } } /* * Do not stop the tick, if we are only one off -- cgit v1.2.3-71-gd317 From d1992b169d31f339dc5ea4e9f312567c8cf322a3 Mon Sep 17 00:00:00 2001 From: Hans Schillstrom Date: Thu, 17 May 2012 22:35:46 +0000 Subject: netfilter: xt_HMARK: fix endianness and provide consistent hashing This patch addresses two issues: a) Fix usage of u32 and __be32 that causes endianess warnings via sparse. b) Ensure consistent hashing in a cluster that is composed of big and little endian systems. Thus, we obtain the same hash mark in an heterogeneous cluster. Reported-by: Dan Carpenter Signed-off-by: Hans Schillstrom Signed-off-by: Pablo Neira Ayuso --- include/linux/netfilter/xt_HMARK.h | 5 +++ net/netfilter/xt_HMARK.c | 72 ++++++++++++++++++++++---------------- 2 files changed, 46 insertions(+), 31 deletions(-) (limited to 'include/linux') diff --git a/include/linux/netfilter/xt_HMARK.h b/include/linux/netfilter/xt_HMARK.h index abb1650940d2..826fc5807577 100644 --- a/include/linux/netfilter/xt_HMARK.h +++ b/include/linux/netfilter/xt_HMARK.h @@ -27,7 +27,12 @@ union hmark_ports { __u16 src; __u16 dst; } p16; + struct { + __be16 src; + __be16 dst; + } b16; __u32 v32; + __be32 b32; }; struct xt_hmark_info { diff --git a/net/netfilter/xt_HMARK.c b/net/netfilter/xt_HMARK.c index 0a96a43108ed..1686ca1b53a1 100644 --- a/net/netfilter/xt_HMARK.c +++ b/net/netfilter/xt_HMARK.c @@ -32,13 +32,13 @@ MODULE_ALIAS("ipt_HMARK"); MODULE_ALIAS("ip6t_HMARK"); struct hmark_tuple { - u32 src; - u32 dst; + __be32 src; + __be32 dst; union hmark_ports uports; - uint8_t proto; + u8 proto; }; -static inline u32 hmark_addr6_mask(const __u32 *addr32, const __u32 *mask) +static inline __be32 hmark_addr6_mask(const __be32 *addr32, const __be32 *mask) { return (addr32[0] & mask[0]) ^ (addr32[1] & mask[1]) ^ @@ -46,8 +46,8 @@ static inline u32 hmark_addr6_mask(const __u32 *addr32, const __u32 *mask) (addr32[3] & mask[3]); } -static inline u32 -hmark_addr_mask(int l3num, const __u32 *addr32, const __u32 *mask) +static inline __be32 +hmark_addr_mask(int l3num, const __be32 *addr32, const __be32 *mask) { switch (l3num) { case AF_INET: @@ -58,6 +58,22 @@ hmark_addr_mask(int l3num, const __u32 *addr32, const __u32 *mask) return 0; } +static inline void hmark_swap_ports(union hmark_ports *uports, + const struct xt_hmark_info *info) +{ + union hmark_ports hp; + u16 src, dst; + + hp.b32 = (uports->b32 & info->port_mask.b32) | info->port_set.b32; + src = ntohs(hp.b16.src); + dst = ntohs(hp.b16.dst); + + if (dst > src) + uports->v32 = (dst << 16) | src; + else + uports->v32 = (src << 16) | dst; +} + static int hmark_ct_set_htuple(const struct sk_buff *skb, struct hmark_tuple *t, const struct xt_hmark_info *info) @@ -74,22 +90,19 @@ hmark_ct_set_htuple(const struct sk_buff *skb, struct hmark_tuple *t, otuple = &ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple; rtuple = &ct->tuplehash[IP_CT_DIR_REPLY].tuple; - t->src = hmark_addr_mask(otuple->src.l3num, otuple->src.u3.all, - info->src_mask.all); - t->dst = hmark_addr_mask(otuple->src.l3num, rtuple->src.u3.all, - info->dst_mask.all); + t->src = hmark_addr_mask(otuple->src.l3num, otuple->src.u3.ip6, + info->src_mask.ip6); + t->dst = hmark_addr_mask(otuple->src.l3num, rtuple->src.u3.ip6, + info->dst_mask.ip6); if (info->flags & XT_HMARK_FLAG(XT_HMARK_METHOD_L3)) return 0; t->proto = nf_ct_protonum(ct); if (t->proto != IPPROTO_ICMP) { - t->uports.p16.src = otuple->src.u.all; - t->uports.p16.dst = rtuple->src.u.all; - t->uports.v32 = (t->uports.v32 & info->port_mask.v32) | - info->port_set.v32; - if (t->uports.p16.dst < t->uports.p16.src) - swap(t->uports.p16.dst, t->uports.p16.src); + t->uports.b16.src = otuple->src.u.all; + t->uports.b16.dst = rtuple->src.u.all; + hmark_swap_ports(&t->uports, info); } return 0; @@ -98,15 +111,19 @@ hmark_ct_set_htuple(const struct sk_buff *skb, struct hmark_tuple *t, #endif } +/* This hash function is endian independent, to ensure consistent hashing if + * the cluster is composed of big and little endian systems. */ static inline u32 hmark_hash(struct hmark_tuple *t, const struct xt_hmark_info *info) { u32 hash; + u32 src = ntohl(t->src); + u32 dst = ntohl(t->dst); - if (t->dst < t->src) - swap(t->src, t->dst); + if (dst < src) + swap(src, dst); - hash = jhash_3words(t->src, t->dst, t->uports.v32, info->hashrnd); + hash = jhash_3words(src, dst, t->uports.v32, info->hashrnd); hash = hash ^ (t->proto & info->proto_mask); return (((u64)hash * info->hmodulus) >> 32) + info->hoffset; @@ -126,11 +143,7 @@ hmark_set_tuple_ports(const struct sk_buff *skb, unsigned int nhoff, if (skb_copy_bits(skb, nhoff, &t->uports, sizeof(t->uports)) < 0) return; - t->uports.v32 = (t->uports.v32 & info->port_mask.v32) | - info->port_set.v32; - - if (t->uports.p16.dst < t->uports.p16.src) - swap(t->uports.p16.dst, t->uports.p16.src); + hmark_swap_ports(&t->uports, info); } #if IS_ENABLED(CONFIG_IP6_NF_IPTABLES) @@ -178,8 +191,8 @@ hmark_pkt_set_htuple_ipv6(const struct sk_buff *skb, struct hmark_tuple *t, return -1; } noicmp: - t->src = hmark_addr6_mask(ip6->saddr.s6_addr32, info->src_mask.all); - t->dst = hmark_addr6_mask(ip6->daddr.s6_addr32, info->dst_mask.all); + t->src = hmark_addr6_mask(ip6->saddr.s6_addr32, info->src_mask.ip6); + t->dst = hmark_addr6_mask(ip6->daddr.s6_addr32, info->dst_mask.ip6); if (info->flags & XT_HMARK_FLAG(XT_HMARK_METHOD_L3)) return 0; @@ -255,11 +268,8 @@ hmark_pkt_set_htuple_ipv4(const struct sk_buff *skb, struct hmark_tuple *t, } } - t->src = (__force u32) ip->saddr; - t->dst = (__force u32) ip->daddr; - - t->src &= info->src_mask.ip; - t->dst &= info->dst_mask.ip; + t->src = ip->saddr & info->src_mask.ip; + t->dst = ip->daddr & info->dst_mask.ip; if (info->flags & XT_HMARK_FLAG(XT_HMARK_METHOD_L3)) return 0; -- cgit v1.2.3-71-gd317 From bafb282df29c1524b1617019adebd6d0c3eb7a47 Mon Sep 17 00:00:00 2001 From: Konstantin Khlebnikov Date: Thu, 7 Jun 2012 14:21:11 -0700 Subject: c/r: prctl: update prctl_set_mm_exe_file() after mm->num_exe_file_vmas removal A fix for commit b32dfe377102 ("c/r: prctl: add ability to set new mm_struct::exe_file"). After removing mm->num_exe_file_vmas kernel keeps mm->exe_file until final mmput(), it never becomes NULL while task is alive. We can check for other mapped files in mm instead of checking mm->num_exe_file_vmas, and mark mm with flag MMF_EXE_FILE_CHANGED in order to forbid second changing of mm->exe_file. Signed-off-by: Konstantin Khlebnikov Reviewed-by: Cyrill Gorcunov Cc: Oleg Nesterov Cc: Matt Helsley Cc: Kees Cook Cc: KOSAKI Motohiro Cc: Tejun Heo Cc: Pavel Emelyanov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/sched.h | 1 + kernel/sys.c | 31 +++++++++++++++++++------------ 2 files changed, 20 insertions(+), 12 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sched.h b/include/linux/sched.h index 6029d8c54476..c688d4cc2e40 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -439,6 +439,7 @@ extern int get_dumpable(struct mm_struct *mm); /* leave room for more dump flags */ #define MMF_VM_MERGEABLE 16 /* KSM may merge identical pages */ #define MMF_VM_HUGEPAGE 17 /* set when VM_HUGEPAGE is set on vma */ +#define MMF_EXE_FILE_CHANGED 18 /* see prctl_set_mm_exe_file() */ #define MMF_INIT_MASK (MMF_DUMPABLE_MASK | MMF_DUMP_FILTER_MASK) diff --git a/kernel/sys.c b/kernel/sys.c index 9ff89cb9657a..54f20fdee93c 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -1796,17 +1796,11 @@ static bool vma_flags_mismatch(struct vm_area_struct *vma, static int prctl_set_mm_exe_file(struct mm_struct *mm, unsigned int fd) { + struct vm_area_struct *vma; struct file *exe_file; struct dentry *dentry; int err; - /* - * Setting new mm::exe_file is only allowed when no VM_EXECUTABLE vma's - * remain. So perform a quick test first. - */ - if (mm->num_exe_file_vmas) - return -EBUSY; - exe_file = fget(fd); if (!exe_file) return -EBADF; @@ -1827,17 +1821,30 @@ static int prctl_set_mm_exe_file(struct mm_struct *mm, unsigned int fd) if (err) goto exit; + down_write(&mm->mmap_sem); + + /* + * Forbid mm->exe_file change if there are mapped other files. + */ + err = -EBUSY; + for (vma = mm->mmap; vma; vma = vma->vm_next) { + if (vma->vm_file && !path_equal(&vma->vm_file->f_path, + &exe_file->f_path)) + goto exit_unlock; + } + /* * The symlink can be changed only once, just to disallow arbitrary * transitions malicious software might bring in. This means one * could make a snapshot over all processes running and monitor * /proc/pid/exe changes to notice unusual activity if needed. */ - down_write(&mm->mmap_sem); - if (likely(!mm->exe_file)) - set_mm_exe_file(mm, exe_file); - else - err = -EBUSY; + err = -EPERM; + if (test_and_set_bit(MMF_EXE_FILE_CHANGED, &mm->flags)) + goto exit_unlock; + + set_mm_exe_file(mm, exe_file); +exit_unlock: up_write(&mm->mmap_sem); exit: -- cgit v1.2.3-71-gd317 From 300f786b2683f8bb1ec0afb6e1851183a479c86d Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Thu, 7 Jun 2012 14:21:12 -0700 Subject: c/r: prctl: add ability to get clear_tid_address Zero is written at clear_tid_address when the process exits. This functionality is used by pthread_join(). We already have sys_set_tid_address() to change this address for the current task but there is no way to obtain it from user space. Without the ability to find this address and dump it we can't restore pthread'ed apps which call pthread_join() once they have been restored. This patch introduces the PR_GET_TID_ADDRESS prctl option which allows the current process to obtain own clear_tid_address. This feature is available iif CONFIG_CHECKPOINT_RESTORE is set. [akpm@linux-foundation.org: fix prctl numbering] Signed-off-by: Andrew Vagin Signed-off-by: Cyrill Gorcunov Cc: Pedro Alves Cc: Oleg Nesterov Cc: Pavel Emelyanov Cc: Tejun Heo Acked-by: Kees Cook Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/prctl.h | 10 ++++++---- kernel/sys.c | 13 +++++++++++++ 2 files changed, 19 insertions(+), 4 deletions(-) (limited to 'include/linux') diff --git a/include/linux/prctl.h b/include/linux/prctl.h index 711e0a30aacc..3988012255dc 100644 --- a/include/linux/prctl.h +++ b/include/linux/prctl.h @@ -127,8 +127,8 @@ #define PR_SET_PTRACER 0x59616d61 # define PR_SET_PTRACER_ANY ((unsigned long)-1) -#define PR_SET_CHILD_SUBREAPER 36 -#define PR_GET_CHILD_SUBREAPER 37 +#define PR_SET_CHILD_SUBREAPER 36 +#define PR_GET_CHILD_SUBREAPER 37 /* * If no_new_privs is set, then operations that grant new privileges (i.e. @@ -142,7 +142,9 @@ * asking selinux for a specific new context (e.g. with runcon) will result * in execve returning -EPERM. */ -#define PR_SET_NO_NEW_PRIVS 38 -#define PR_GET_NO_NEW_PRIVS 39 +#define PR_SET_NO_NEW_PRIVS 38 +#define PR_GET_NO_NEW_PRIVS 39 + +#define PR_GET_TID_ADDRESS 40 #endif /* _LINUX_PRCTL_H */ diff --git a/kernel/sys.c b/kernel/sys.c index 19a2c7139960..0ec1942ba7ea 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -1988,12 +1988,22 @@ out: up_read(&mm->mmap_sem); return error; } + +static int prctl_get_tid_address(struct task_struct *me, int __user **tid_addr) +{ + return put_user(me->clear_child_tid, tid_addr); +} + #else /* CONFIG_CHECKPOINT_RESTORE */ static int prctl_set_mm(int opt, unsigned long addr, unsigned long arg4, unsigned long arg5) { return -EINVAL; } +static int prctl_get_tid_address(struct task_struct *me, int __user **tid_addr) +{ + return -EINVAL; +} #endif SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, @@ -2131,6 +2141,9 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, else return -EINVAL; break; + case PR_GET_TID_ADDRESS: + error = prctl_get_tid_address(me, (int __user **)arg2); + break; default: return -EINVAL; } -- cgit v1.2.3-71-gd317 From ae82fdb1406ad41d68f07027fe31f2d35ba22a90 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Fri, 8 Jun 2012 14:58:13 +0930 Subject: module_param: stop double-calling parameters. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Commit 026cee0086fe1df4cf74691cf273062cc769617d "params: _initcall-like kernel parameters" set old-style module parameters to level 0. And we call those level 0 calls where we used to, early in start_kernel(). We also loop through the initcall levels and call the levelled module_params before the corresponding initcall. Unfortunately level 0 is early_init(), so we call the standard module_param calls twice. (Turns out most things don't care, but at least ubi.mtd does). Change the level to -1 for standard module_param calls. Reported-by: Benoît Thébaudeau Signed-off-by: Rusty Russell Cc: stable@kernel.org --- include/linux/moduleparam.h | 10 +++++----- init/main.c | 2 +- 2 files changed, 6 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/include/linux/moduleparam.h b/include/linux/moduleparam.h index 1b14d25162cb..d6a58065c09c 100644 --- a/include/linux/moduleparam.h +++ b/include/linux/moduleparam.h @@ -128,7 +128,7 @@ struct kparam_array * The ops can have NULL set or get functions. */ #define module_param_cb(name, ops, arg, perm) \ - __module_param_call(MODULE_PARAM_PREFIX, name, ops, arg, perm, 0) + __module_param_call(MODULE_PARAM_PREFIX, name, ops, arg, perm, -1) /** * _param_cb - general callback for a module/cmdline parameter @@ -192,7 +192,7 @@ struct kparam_array { (void *)set, (void *)get }; \ __module_param_call(MODULE_PARAM_PREFIX, \ name, &__param_ops_##name, arg, \ - (perm) + sizeof(__check_old_set_param(set))*0, 0) + (perm) + sizeof(__check_old_set_param(set))*0, -1) /* We don't get oldget: it's often a new-style param_get_uint, etc. */ static inline int @@ -272,7 +272,7 @@ static inline void __kernel_param_unlock(void) */ #define core_param(name, var, type, perm) \ param_check_##type(name, &(var)); \ - __module_param_call("", name, ¶m_ops_##type, &var, perm, 0) + __module_param_call("", name, ¶m_ops_##type, &var, perm, -1) #endif /* !MODULE */ /** @@ -290,7 +290,7 @@ static inline void __kernel_param_unlock(void) = { len, string }; \ __module_param_call(MODULE_PARAM_PREFIX, name, \ ¶m_ops_string, \ - .str = &__param_string_##name, perm, 0); \ + .str = &__param_string_##name, perm, -1); \ __MODULE_PARM_TYPE(name, "string") /** @@ -432,7 +432,7 @@ extern int param_set_bint(const char *val, const struct kernel_param *kp); __module_param_call(MODULE_PARAM_PREFIX, name, \ ¶m_array_ops, \ .arr = &__param_arr_##name, \ - perm, 0); \ + perm, -1); \ __MODULE_PARM_TYPE(name, "array of " #type) extern struct kernel_param_ops param_array_ops; diff --git a/init/main.c b/init/main.c index 1ca6b32c4828..37e12098eac1 100644 --- a/init/main.c +++ b/init/main.c @@ -508,7 +508,7 @@ asmlinkage void __init start_kernel(void) parse_early_param(); parse_args("Booting kernel", static_command_line, __start___param, __stop___param - __start___param, - 0, 0, &unknown_bootoption); + -1, -1, &unknown_bootoption); jump_label_init(); -- cgit v1.2.3-71-gd317 From c8e9cf7bb240049117d2fa64d1540476c289396d Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Thu, 7 Jun 2012 12:15:15 +0200 Subject: vga_switcheroo: Add a helper function to get the client state Add vga_switcheroo_get_client_state() to get the current state of the client. This is necessary to determine the proper initial state of audio clients in HD-audio driver. Acked-by: Dave Airlie Signed-off-by: Takashi Iwai --- drivers/gpu/vga/vga_switcheroo.c | 13 +++++++++++++ include/linux/vga_switcheroo.h | 7 +++++++ 2 files changed, 20 insertions(+) (limited to 'include/linux') diff --git a/drivers/gpu/vga/vga_switcheroo.c b/drivers/gpu/vga/vga_switcheroo.c index 38f9534ac513..eb4f64f0a565 100644 --- a/drivers/gpu/vga/vga_switcheroo.c +++ b/drivers/gpu/vga/vga_switcheroo.c @@ -190,6 +190,19 @@ find_active_client(struct list_head *head) return NULL; } +int vga_switcheroo_get_client_state(struct pci_dev *pdev) +{ + struct vga_switcheroo_client *client; + + client = find_client_from_pci(&vgasr_priv.clients, pdev); + if (!client) + return VGA_SWITCHEROO_NOT_FOUND; + if (!vgasr_priv.active) + return VGA_SWITCHEROO_INIT; + return client->pwr_state; +} +EXPORT_SYMBOL(vga_switcheroo_get_client_state); + void vga_switcheroo_unregister_client(struct pci_dev *pdev) { struct vga_switcheroo_client *client; diff --git a/include/linux/vga_switcheroo.h b/include/linux/vga_switcheroo.h index b455c7c212eb..b176342ca031 100644 --- a/include/linux/vga_switcheroo.h +++ b/include/linux/vga_switcheroo.h @@ -12,6 +12,9 @@ enum vga_switcheroo_state { VGA_SWITCHEROO_OFF, VGA_SWITCHEROO_ON, + /* below are referred only from vga_switcheroo_get_client_state() */ + VGA_SWITCHEROO_INIT, + VGA_SWITCHEROO_NOT_FOUND, }; enum vga_switcheroo_client_id { @@ -50,6 +53,8 @@ void vga_switcheroo_unregister_handler(void); int vga_switcheroo_process_delayed_switch(void); +int vga_switcheroo_get_client_state(struct pci_dev *dev); + #else static inline void vga_switcheroo_unregister_client(struct pci_dev *dev) {} @@ -62,5 +67,7 @@ static inline int vga_switcheroo_register_audio_client(struct pci_dev *pdev, int id, bool active) { return 0; } static inline void vga_switcheroo_unregister_handler(void) {} static inline int vga_switcheroo_process_delayed_switch(void) { return 0; } +static inline int vga_switcheroo_get_client_state(struct pci_dev *dev) { return VGA_SWITCHEROO_CLIENT_ON; } + #endif -- cgit v1.2.3-71-gd317 From 505cff00de9c303b95c204eb4544066e3e707911 Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Fri, 8 Jun 2012 12:49:17 +0200 Subject: vga_switcheroo: Fix error without CONFIG_VGA_SWITCHEROO Fix a typo that is built only when CONFIG_VGA_SWITCHEROO=n. Signed-off-by: Takashi Iwai --- include/linux/vga_switcheroo.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/vga_switcheroo.h b/include/linux/vga_switcheroo.h index b176342ca031..60da41fe9dc2 100644 --- a/include/linux/vga_switcheroo.h +++ b/include/linux/vga_switcheroo.h @@ -67,7 +67,7 @@ static inline int vga_switcheroo_register_audio_client(struct pci_dev *pdev, int id, bool active) { return 0; } static inline void vga_switcheroo_unregister_handler(void) {} static inline int vga_switcheroo_process_delayed_switch(void) { return 0; } -static inline int vga_switcheroo_get_client_state(struct pci_dev *dev) { return VGA_SWITCHEROO_CLIENT_ON; } +static inline int vga_switcheroo_get_client_state(struct pci_dev *dev) { return VGA_SWITCHEROO_ON; } #endif -- cgit v1.2.3-71-gd317 From 8876d6b5f81f4e242a6660da22bbd92f17a8d058 Mon Sep 17 00:00:00 2001 From: Paul Pluzhnikov Date: Sat, 9 Jun 2012 07:53:03 -0700 Subject: net: Make linux/tcp.h C++ friendly (trivial) I originally sent this patch to , but Jiri Kosina did not feel that this is fully appropriate for the trivial tree. Using linux/tcp.h from C++ results in: cat t.cc #include int main() { } g++ -c t.cc In file included from t.cc:1: /usr/include/linux/tcp.h:72: error: '__u32 __fswab32(__u32)' cannot appear in a constant-expression /usr/include/linux/tcp.h:72: error: a function call cannot appear in a constant-expression ... Attached trivial patch fixes this problem. Tested: - the t.cc above compiles with g++ and - the following program generates the same output before/after the patch: #include #include int main () { #define P(a) printf("%s: %08x\n", #a, (int)a) P(TCP_FLAG_CWR); P(TCP_FLAG_ECE); P(TCP_FLAG_URG); P(TCP_FLAG_ACK); P(TCP_FLAG_PSH); P(TCP_FLAG_RST); P(TCP_FLAG_SYN); P(TCP_FLAG_FIN); P(TCP_RESERVED_BITS); P(TCP_DATA_OFFSET); #undef P return 0; } Signed-off-by: Paul Pluzhnikov Signed-off-by: David S. Miller --- include/linux/tcp.h | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/include/linux/tcp.h b/include/linux/tcp.h index 4c5b63283377..5f359dbfcdce 100644 --- a/include/linux/tcp.h +++ b/include/linux/tcp.h @@ -69,16 +69,16 @@ union tcp_word_hdr { #define tcp_flag_word(tp) ( ((union tcp_word_hdr *)(tp))->words [3]) enum { - TCP_FLAG_CWR = __cpu_to_be32(0x00800000), - TCP_FLAG_ECE = __cpu_to_be32(0x00400000), - TCP_FLAG_URG = __cpu_to_be32(0x00200000), - TCP_FLAG_ACK = __cpu_to_be32(0x00100000), - TCP_FLAG_PSH = __cpu_to_be32(0x00080000), - TCP_FLAG_RST = __cpu_to_be32(0x00040000), - TCP_FLAG_SYN = __cpu_to_be32(0x00020000), - TCP_FLAG_FIN = __cpu_to_be32(0x00010000), - TCP_RESERVED_BITS = __cpu_to_be32(0x0F000000), - TCP_DATA_OFFSET = __cpu_to_be32(0xF0000000) + TCP_FLAG_CWR = __constant_cpu_to_be32(0x00800000), + TCP_FLAG_ECE = __constant_cpu_to_be32(0x00400000), + TCP_FLAG_URG = __constant_cpu_to_be32(0x00200000), + TCP_FLAG_ACK = __constant_cpu_to_be32(0x00100000), + TCP_FLAG_PSH = __constant_cpu_to_be32(0x00080000), + TCP_FLAG_RST = __constant_cpu_to_be32(0x00040000), + TCP_FLAG_SYN = __constant_cpu_to_be32(0x00020000), + TCP_FLAG_FIN = __constant_cpu_to_be32(0x00010000), + TCP_RESERVED_BITS = __constant_cpu_to_be32(0x0F000000), + TCP_DATA_OFFSET = __constant_cpu_to_be32(0xF0000000) }; /* -- cgit v1.2.3-71-gd317 From c2fb8a3fa25513de8fedb38509b1f15a5bbee47b Mon Sep 17 00:00:00 2001 From: Alan Stern Date: Wed, 13 Jun 2012 11:20:19 -0400 Subject: USB: add NO_D3_DURING_SLEEP flag and revert 151b61284776be2 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This patch (as1558) fixes a problem affecting several ASUS computers: The machine crashes or corrupts memory when going into suspend if the ehci-hcd driver is bound to any controllers. Users have been forced to unbind or unload ehci-hcd before putting their systems to sleep. After extensive testing, it was determined that the machines don't like going into suspend when any EHCI controllers are in the PCI D3 power state. Presumably this is a firmware bug, but there's nothing we can do about it except to avoid putting the controllers in D3 during system sleep. The patch adds a new flag to indicate whether the problem is present, and avoids changing the controller's power state if the flag is set. Runtime suspend is unaffected; this matters only for system suspend. However as a side effect, the controller will not respond to remote wakeup requests while the system is asleep. Hence USB wakeup is not functional -- but of course, this is already true in the current state of affairs. A similar patch has already been applied as commit 151b61284776be2d6f02d48c23c3625678960b97 (USB: EHCI: fix crash during suspend on ASUS computers). The patch supersedes that one and reverts it. There are two differences: The old patch added the flag at the USB level; this patch adds it at the PCI level. The old patch applied to all chipsets with the same vendor, subsystem vendor, and product IDs; this patch makes an exception for a known-good system (based on DMI information). Signed-off-by: Alan Stern Tested-by: Dâniel Fraga Tested-by: Andrey Rahmatullin Tested-by: Steven Rostedt Cc: stable Reviewed-by: Rafael J. Wysocki Signed-off-by: Greg Kroah-Hartman --- drivers/pci/pci.c | 5 +++++ drivers/pci/quirks.c | 26 ++++++++++++++++++++++++++ drivers/usb/core/hcd-pci.c | 9 --------- drivers/usb/host/ehci-pci.c | 8 -------- include/linux/pci.h | 2 ++ include/linux/usb/hcd.h | 2 -- 6 files changed, 33 insertions(+), 19 deletions(-) (limited to 'include/linux') diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 447e83472c01..77cb54a65cde 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -1744,6 +1744,11 @@ int pci_prepare_to_sleep(struct pci_dev *dev) if (target_state == PCI_POWER_ERROR) return -EIO; + /* Some devices mustn't be in D3 during system sleep */ + if (target_state == PCI_D3hot && + (dev->dev_flags & PCI_DEV_FLAGS_NO_D3_DURING_SLEEP)) + return 0; + pci_enable_wake(dev, target_state, device_may_wakeup(&dev->dev)); error = pci_set_power_state(dev, target_state); diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c index 2a7521677541..194b243a2817 100644 --- a/drivers/pci/quirks.c +++ b/drivers/pci/quirks.c @@ -2929,6 +2929,32 @@ static void __devinit disable_igfx_irq(struct pci_dev *dev) DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL, 0x0102, disable_igfx_irq); DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL, 0x010a, disable_igfx_irq); +/* + * The Intel 6 Series/C200 Series chipset's EHCI controllers on many + * ASUS motherboards will cause memory corruption or a system crash + * if they are in D3 while the system is put into S3 sleep. + */ +static void __devinit asus_ehci_no_d3(struct pci_dev *dev) +{ + const char *sys_info; + static const char good_Asus_board[] = "P8Z68-V"; + + if (dev->dev_flags & PCI_DEV_FLAGS_NO_D3_DURING_SLEEP) + return; + if (dev->subsystem_vendor != PCI_VENDOR_ID_ASUSTEK) + return; + sys_info = dmi_get_system_info(DMI_BOARD_NAME); + if (sys_info && memcmp(sys_info, good_Asus_board, + sizeof(good_Asus_board) - 1) == 0) + return; + + dev_info(&dev->dev, "broken D3 during system sleep on ASUS\n"); + dev->dev_flags |= PCI_DEV_FLAGS_NO_D3_DURING_SLEEP; + device_set_wakeup_capable(&dev->dev, false); +} +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL, 0x1c26, asus_ehci_no_d3); +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL, 0x1c2d, asus_ehci_no_d3); + static void pci_do_fixups(struct pci_dev *dev, struct pci_fixup *f, struct pci_fixup *end) { diff --git a/drivers/usb/core/hcd-pci.c b/drivers/usb/core/hcd-pci.c index 57ed9e400c06..622b4a48e732 100644 --- a/drivers/usb/core/hcd-pci.c +++ b/drivers/usb/core/hcd-pci.c @@ -493,15 +493,6 @@ static int hcd_pci_suspend_noirq(struct device *dev) pci_save_state(pci_dev); - /* - * Some systems crash if an EHCI controller is in D3 during - * a sleep transition. We have to leave such controllers in D0. - */ - if (hcd->broken_pci_sleep) { - dev_dbg(dev, "Staying in PCI D0\n"); - return retval; - } - /* If the root hub is dead rather than suspended, disallow remote * wakeup. usb_hc_died() should ensure that both hosts are marked as * dying, so we only need to check the primary roothub. diff --git a/drivers/usb/host/ehci-pci.c b/drivers/usb/host/ehci-pci.c index bc94d7bf072d..123481793a47 100644 --- a/drivers/usb/host/ehci-pci.c +++ b/drivers/usb/host/ehci-pci.c @@ -144,14 +144,6 @@ static int ehci_pci_setup(struct usb_hcd *hcd) hcd->has_tt = 1; tdi_reset(ehci); } - if (pdev->subsystem_vendor == PCI_VENDOR_ID_ASUSTEK) { - /* EHCI #1 or #2 on 6 Series/C200 Series chipset */ - if (pdev->device == 0x1c26 || pdev->device == 0x1c2d) { - ehci_info(ehci, "broken D3 during system sleep on ASUS\n"); - hcd->broken_pci_sleep = 1; - device_set_wakeup_capable(&pdev->dev, false); - } - } break; case PCI_VENDOR_ID_TDI: if (pdev->device == PCI_DEVICE_ID_TDI_EHCI) { diff --git a/include/linux/pci.h b/include/linux/pci.h index d8c379dba6ad..fefb4e19bf6a 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -176,6 +176,8 @@ enum pci_dev_flags { PCI_DEV_FLAGS_NO_D3 = (__force pci_dev_flags_t) 2, /* Provide indication device is assigned by a Virtual Machine Manager */ PCI_DEV_FLAGS_ASSIGNED = (__force pci_dev_flags_t) 4, + /* Device causes system crash if in D3 during S3 sleep */ + PCI_DEV_FLAGS_NO_D3_DURING_SLEEP = (__force pci_dev_flags_t) 8, }; enum pci_irq_reroute_variant { diff --git a/include/linux/usb/hcd.h b/include/linux/usb/hcd.h index 7f855d50cdf5..49b3ac29726a 100644 --- a/include/linux/usb/hcd.h +++ b/include/linux/usb/hcd.h @@ -126,8 +126,6 @@ struct usb_hcd { unsigned wireless:1; /* Wireless USB HCD */ unsigned authorized_default:1; unsigned has_tt:1; /* Integrated TT in root hub */ - unsigned broken_pci_sleep:1; /* Don't put the - controller in PCI-D3 for system sleep */ unsigned int irq; /* irq allocated */ void __iomem *regs; /* device memory/io */ -- cgit v1.2.3-71-gd317 From 9b15b817f3d62409290fd56fe3cbb076a931bb0a Mon Sep 17 00:00:00 2001 From: Hugh Dickins Date: Fri, 15 Jun 2012 17:55:50 -0700 Subject: swap: fix shmem swapping when more than 8 areas Minchan Kim reports that when a system has many swap areas, and tmpfs swaps out to the ninth or more, shmem_getpage_gfp()'s attempts to read back the page cannot locate it, and the read fails with -ENOMEM. Whoops. Yes, I blindly followed read_swap_header()'s pte_to_swp_entry( swp_entry_to_pte()) technique for determining maximum usable swap offset, without stopping to realize that that actually depends upon the pte swap encoding shifting swap offset to the higher bits and truncating it there. Whereas our radix_tree swap encoding leaves offset in the lower bits: it's swap "type" (that is, index of swap area) that was truncated. Fix it by reducing the SWP_TYPE_SHIFT() in swapops.h, and removing the broken radix_to_swp_entry(swp_to_radix_entry()) from read_swap_header(). This does not reduce the usable size of a swap area any further, it leaves it as claimed when making the original commit: no change from 3.0 on x86_64, nor on i386 without PAE; but 3.0's 512GB is reduced to 128GB per swapfile on i386 with PAE. It's not a change I would have risked five years ago, but with x86_64 supported for ten years, I believe it's appropriate now. Hmm, and what if some architecture implements its swap pte with offset encoded below type? That would equally break the maximum usable swap offset check. Happily, they all follow the same tradition of encoding offset above type, but I'll prepare a check on that for next. Reported-and-Reviewed-and-Tested-by: Minchan Kim Signed-off-by: Hugh Dickins Cc: stable@vger.kernel.org [3.1, 3.2, 3.3, 3.4] Signed-off-by: Linus Torvalds --- include/linux/swapops.h | 8 +++++--- mm/swapfile.c | 12 ++++-------- 2 files changed, 9 insertions(+), 11 deletions(-) (limited to 'include/linux') diff --git a/include/linux/swapops.h b/include/linux/swapops.h index 792d16d9cbc7..47ead515c811 100644 --- a/include/linux/swapops.h +++ b/include/linux/swapops.h @@ -9,13 +9,15 @@ * get good packing density in that tree, so the index should be dense in * the low-order bits. * - * We arrange the `type' and `offset' fields so that `type' is at the five + * We arrange the `type' and `offset' fields so that `type' is at the seven * high-order bits of the swp_entry_t and `offset' is right-aligned in the - * remaining bits. + * remaining bits. Although `type' itself needs only five bits, we allow for + * shmem/tmpfs to shift it all up a further two bits: see swp_to_radix_entry(). * * swp_entry_t's are *never* stored anywhere in their arch-dependent format. */ -#define SWP_TYPE_SHIFT(e) (sizeof(e.val) * 8 - MAX_SWAPFILES_SHIFT) +#define SWP_TYPE_SHIFT(e) ((sizeof(e.val) * 8) - \ + (MAX_SWAPFILES_SHIFT + RADIX_TREE_EXCEPTIONAL_SHIFT)) #define SWP_OFFSET_MASK(e) ((1UL << SWP_TYPE_SHIFT(e)) - 1) /* diff --git a/mm/swapfile.c b/mm/swapfile.c index de5bc51c4a66..71373d03fcee 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c @@ -1916,24 +1916,20 @@ static unsigned long read_swap_header(struct swap_info_struct *p, /* * Find out how many pages are allowed for a single swap - * device. There are three limiting factors: 1) the number + * device. There are two limiting factors: 1) the number * of bits for the swap offset in the swp_entry_t type, and * 2) the number of bits in the swap pte as defined by the - * the different architectures, and 3) the number of free bits - * in an exceptional radix_tree entry. In order to find the + * different architectures. In order to find the * largest possible bit mask, a swap entry with swap type 0 * and swap offset ~0UL is created, encoded to a swap pte, * decoded to a swp_entry_t again, and finally the swap * offset is extracted. This will mask all the bits from * the initial ~0UL mask that can't be encoded in either * the swp_entry_t or the architecture definition of a - * swap pte. Then the same is done for a radix_tree entry. + * swap pte. */ maxpages = swp_offset(pte_to_swp_entry( - swp_entry_to_pte(swp_entry(0, ~0UL)))); - maxpages = swp_offset(radix_to_swp_entry( - swp_to_radix_entry(swp_entry(0, maxpages)))) + 1; - + swp_entry_to_pte(swp_entry(0, ~0UL)))) + 1; if (maxpages > swap_header->info.last_page) { maxpages = swap_header->info.last_page + 1; /* p->max is an unsigned int: don't overflow it */ -- cgit v1.2.3-71-gd317