Merge block/IO core bits from Jens Axboe:
 "This is a bit bigger on the core side than usual, but that is purely
  because we decided to hold off on parts of Tejun's submission on 3.4
  to give it a bit more time to simmer.  As a consequence, it's seen a
  long cycle in for-next.
  It contains:
   - Bug fix from Dan, wrong locking type.
   - Relax splice gifting restriction from Eric.
   - A ton of updates from Tejun, primarily for blkcg.  This improves
     the code a lot, making the API nicer and cleaner, and also includes
     fixes for how we handle and tie policies and re-activate on
     switches.  The changes also include generic bug fixes.
   - A simple fix from Vivek, along with a fix for doing proper delayed
     allocation of the blkcg stats."
Fix up annoying conflict just due to different merge resolution in
Documentation/feature-removal-schedule.txt
* 'for-3.5/core' of git://git.kernel.dk/linux-block: (92 commits)
  blkcg: tg_stats_alloc_lock is an irq lock
  vmsplice: relax alignement requirements for SPLICE_F_GIFT
  blkcg: use radix tree to index blkgs from blkcg
  blkcg: fix blkcg->css ref leak in __blkg_lookup_create()
  block: fix elvpriv allocation failure handling
  block: collapse blk_alloc_request() into get_request()
  blkcg: collapse blkcg_policy_ops into blkcg_policy
  blkcg: embed struct blkg_policy_data in policy specific data
  blkcg: mass rename of blkcg API
  blkcg: style cleanups for blk-cgroup.h
  blkcg: remove blkio_group->path[]
  blkcg: blkg_rwstat_read() was missing inline
  blkcg: shoot down blkgs if all policies are deactivated
  blkcg: drop stuff unused after per-queue policy activation update
  blkcg: implement per-queue policy activation
  blkcg: add request_queue->root_blkg
  blkcg: make request_queue bypassing on allocation
  blkcg: make sure blkg_lookup() returns %NULL if @q is bypassing
  blkcg: make blkg_conf_prep() take @pol and return with queue lock held
  blkcg: remove static policy ID enums
  ...
		
	
			
		
			
				
	
	
		
			241 lines
		
	
	
	
		
			5 KiB
			
		
	
	
	
		
			C
		
	
	
	
	
	
			
		
		
	
	
			241 lines
		
	
	
	
		
			5 KiB
			
		
	
	
	
		
			C
		
	
	
	
	
	
/*
 | 
						|
 * fs/ioprio.c
 | 
						|
 *
 | 
						|
 * Copyright (C) 2004 Jens Axboe <axboe@kernel.dk>
 | 
						|
 *
 | 
						|
 * Helper functions for setting/querying io priorities of processes. The
 | 
						|
 * system calls closely mimmick getpriority/setpriority, see the man page for
 | 
						|
 * those. The prio argument is a composite of prio class and prio data, where
 | 
						|
 * the data argument has meaning within that class. The standard scheduling
 | 
						|
 * classes have 8 distinct prio levels, with 0 being the highest prio and 7
 | 
						|
 * being the lowest.
 | 
						|
 *
 | 
						|
 * IOW, setting BE scheduling class with prio 2 is done ala:
 | 
						|
 *
 | 
						|
 * unsigned int prio = (IOPRIO_CLASS_BE << IOPRIO_CLASS_SHIFT) | 2;
 | 
						|
 *
 | 
						|
 * ioprio_set(PRIO_PROCESS, pid, prio);
 | 
						|
 *
 | 
						|
 * See also Documentation/block/ioprio.txt
 | 
						|
 *
 | 
						|
 */
 | 
						|
#include <linux/gfp.h>
 | 
						|
#include <linux/kernel.h>
 | 
						|
#include <linux/export.h>
 | 
						|
#include <linux/ioprio.h>
 | 
						|
#include <linux/blkdev.h>
 | 
						|
#include <linux/capability.h>
 | 
						|
#include <linux/syscalls.h>
 | 
						|
#include <linux/security.h>
 | 
						|
#include <linux/pid_namespace.h>
 | 
						|
 | 
						|
int set_task_ioprio(struct task_struct *task, int ioprio)
 | 
						|
{
 | 
						|
	int err;
 | 
						|
	struct io_context *ioc;
 | 
						|
	const struct cred *cred = current_cred(), *tcred;
 | 
						|
 | 
						|
	rcu_read_lock();
 | 
						|
	tcred = __task_cred(task);
 | 
						|
	if (!uid_eq(tcred->uid, cred->euid) &&
 | 
						|
	    !uid_eq(tcred->uid, cred->uid) && !capable(CAP_SYS_NICE)) {
 | 
						|
		rcu_read_unlock();
 | 
						|
		return -EPERM;
 | 
						|
	}
 | 
						|
	rcu_read_unlock();
 | 
						|
 | 
						|
	err = security_task_setioprio(task, ioprio);
 | 
						|
	if (err)
 | 
						|
		return err;
 | 
						|
 | 
						|
	ioc = get_task_io_context(task, GFP_ATOMIC, NUMA_NO_NODE);
 | 
						|
	if (ioc) {
 | 
						|
		ioc->ioprio = ioprio;
 | 
						|
		put_io_context(ioc);
 | 
						|
	}
 | 
						|
 | 
						|
	return err;
 | 
						|
}
 | 
						|
EXPORT_SYMBOL_GPL(set_task_ioprio);
 | 
						|
 | 
						|
SYSCALL_DEFINE3(ioprio_set, int, which, int, who, int, ioprio)
 | 
						|
{
 | 
						|
	int class = IOPRIO_PRIO_CLASS(ioprio);
 | 
						|
	int data = IOPRIO_PRIO_DATA(ioprio);
 | 
						|
	struct task_struct *p, *g;
 | 
						|
	struct user_struct *user;
 | 
						|
	struct pid *pgrp;
 | 
						|
	kuid_t uid;
 | 
						|
	int ret;
 | 
						|
 | 
						|
	switch (class) {
 | 
						|
		case IOPRIO_CLASS_RT:
 | 
						|
			if (!capable(CAP_SYS_ADMIN))
 | 
						|
				return -EPERM;
 | 
						|
			/* fall through, rt has prio field too */
 | 
						|
		case IOPRIO_CLASS_BE:
 | 
						|
			if (data >= IOPRIO_BE_NR || data < 0)
 | 
						|
				return -EINVAL;
 | 
						|
 | 
						|
			break;
 | 
						|
		case IOPRIO_CLASS_IDLE:
 | 
						|
			break;
 | 
						|
		case IOPRIO_CLASS_NONE:
 | 
						|
			if (data)
 | 
						|
				return -EINVAL;
 | 
						|
			break;
 | 
						|
		default:
 | 
						|
			return -EINVAL;
 | 
						|
	}
 | 
						|
 | 
						|
	ret = -ESRCH;
 | 
						|
	rcu_read_lock();
 | 
						|
	switch (which) {
 | 
						|
		case IOPRIO_WHO_PROCESS:
 | 
						|
			if (!who)
 | 
						|
				p = current;
 | 
						|
			else
 | 
						|
				p = find_task_by_vpid(who);
 | 
						|
			if (p)
 | 
						|
				ret = set_task_ioprio(p, ioprio);
 | 
						|
			break;
 | 
						|
		case IOPRIO_WHO_PGRP:
 | 
						|
			if (!who)
 | 
						|
				pgrp = task_pgrp(current);
 | 
						|
			else
 | 
						|
				pgrp = find_vpid(who);
 | 
						|
			do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
 | 
						|
				ret = set_task_ioprio(p, ioprio);
 | 
						|
				if (ret)
 | 
						|
					break;
 | 
						|
			} while_each_pid_thread(pgrp, PIDTYPE_PGID, p);
 | 
						|
			break;
 | 
						|
		case IOPRIO_WHO_USER:
 | 
						|
			uid = make_kuid(current_user_ns(), who);
 | 
						|
			if (!uid_valid(uid))
 | 
						|
				break;
 | 
						|
			if (!who)
 | 
						|
				user = current_user();
 | 
						|
			else
 | 
						|
				user = find_user(uid);
 | 
						|
 | 
						|
			if (!user)
 | 
						|
				break;
 | 
						|
 | 
						|
			do_each_thread(g, p) {
 | 
						|
				if (!uid_eq(task_uid(p), uid))
 | 
						|
					continue;
 | 
						|
				ret = set_task_ioprio(p, ioprio);
 | 
						|
				if (ret)
 | 
						|
					goto free_uid;
 | 
						|
			} while_each_thread(g, p);
 | 
						|
free_uid:
 | 
						|
			if (who)
 | 
						|
				free_uid(user);
 | 
						|
			break;
 | 
						|
		default:
 | 
						|
			ret = -EINVAL;
 | 
						|
	}
 | 
						|
 | 
						|
	rcu_read_unlock();
 | 
						|
	return ret;
 | 
						|
}
 | 
						|
 | 
						|
static int get_task_ioprio(struct task_struct *p)
 | 
						|
{
 | 
						|
	int ret;
 | 
						|
 | 
						|
	ret = security_task_getioprio(p);
 | 
						|
	if (ret)
 | 
						|
		goto out;
 | 
						|
	ret = IOPRIO_PRIO_VALUE(IOPRIO_CLASS_NONE, IOPRIO_NORM);
 | 
						|
	if (p->io_context)
 | 
						|
		ret = p->io_context->ioprio;
 | 
						|
out:
 | 
						|
	return ret;
 | 
						|
}
 | 
						|
 | 
						|
int ioprio_best(unsigned short aprio, unsigned short bprio)
 | 
						|
{
 | 
						|
	unsigned short aclass = IOPRIO_PRIO_CLASS(aprio);
 | 
						|
	unsigned short bclass = IOPRIO_PRIO_CLASS(bprio);
 | 
						|
 | 
						|
	if (aclass == IOPRIO_CLASS_NONE)
 | 
						|
		aclass = IOPRIO_CLASS_BE;
 | 
						|
	if (bclass == IOPRIO_CLASS_NONE)
 | 
						|
		bclass = IOPRIO_CLASS_BE;
 | 
						|
 | 
						|
	if (aclass == bclass)
 | 
						|
		return min(aprio, bprio);
 | 
						|
	if (aclass > bclass)
 | 
						|
		return bprio;
 | 
						|
	else
 | 
						|
		return aprio;
 | 
						|
}
 | 
						|
 | 
						|
SYSCALL_DEFINE2(ioprio_get, int, which, int, who)
 | 
						|
{
 | 
						|
	struct task_struct *g, *p;
 | 
						|
	struct user_struct *user;
 | 
						|
	struct pid *pgrp;
 | 
						|
	kuid_t uid;
 | 
						|
	int ret = -ESRCH;
 | 
						|
	int tmpio;
 | 
						|
 | 
						|
	rcu_read_lock();
 | 
						|
	switch (which) {
 | 
						|
		case IOPRIO_WHO_PROCESS:
 | 
						|
			if (!who)
 | 
						|
				p = current;
 | 
						|
			else
 | 
						|
				p = find_task_by_vpid(who);
 | 
						|
			if (p)
 | 
						|
				ret = get_task_ioprio(p);
 | 
						|
			break;
 | 
						|
		case IOPRIO_WHO_PGRP:
 | 
						|
			if (!who)
 | 
						|
				pgrp = task_pgrp(current);
 | 
						|
			else
 | 
						|
				pgrp = find_vpid(who);
 | 
						|
			do_each_pid_thread(pgrp, PIDTYPE_PGID, p) {
 | 
						|
				tmpio = get_task_ioprio(p);
 | 
						|
				if (tmpio < 0)
 | 
						|
					continue;
 | 
						|
				if (ret == -ESRCH)
 | 
						|
					ret = tmpio;
 | 
						|
				else
 | 
						|
					ret = ioprio_best(ret, tmpio);
 | 
						|
			} while_each_pid_thread(pgrp, PIDTYPE_PGID, p);
 | 
						|
			break;
 | 
						|
		case IOPRIO_WHO_USER:
 | 
						|
			uid = make_kuid(current_user_ns(), who);
 | 
						|
			if (!who)
 | 
						|
				user = current_user();
 | 
						|
			else
 | 
						|
				user = find_user(uid);
 | 
						|
 | 
						|
			if (!user)
 | 
						|
				break;
 | 
						|
 | 
						|
			do_each_thread(g, p) {
 | 
						|
				if (!uid_eq(task_uid(p), user->uid))
 | 
						|
					continue;
 | 
						|
				tmpio = get_task_ioprio(p);
 | 
						|
				if (tmpio < 0)
 | 
						|
					continue;
 | 
						|
				if (ret == -ESRCH)
 | 
						|
					ret = tmpio;
 | 
						|
				else
 | 
						|
					ret = ioprio_best(ret, tmpio);
 | 
						|
			} while_each_thread(g, p);
 | 
						|
 | 
						|
			if (who)
 | 
						|
				free_uid(user);
 | 
						|
			break;
 | 
						|
		default:
 | 
						|
			ret = -EINVAL;
 | 
						|
	}
 | 
						|
 | 
						|
	rcu_read_unlock();
 | 
						|
	return ret;
 | 
						|
}
 |