linux.git/fs/proc/inode.c, branch v5.7

proc: faster open/read/close with "permanent" files

2020-04-07T17:43:42+00:00

Now that "struct proc_ops" exist we can start putting there stuff which
could not fly with VFS "struct file_operations"...

Most of fs/proc/inode.c file is dedicated to make open/read/.../close
reliable in the event of disappearing /proc entries which usually happens
if module is getting removed.  Files like /proc/cpuinfo which never
disappear simply do not need such protection.

Save 2 atomic ops, 1 allocation, 1 free per open/read/close sequence for such
"permanent" files.

Enable "permanent" flag for

	/proc/cpuinfo
	/proc/kmsg
	/proc/modules
	/proc/slabinfo
	/proc/stat
	/proc/sysvipc/*
	/proc/swaps

More will come once I figure out foolproof way to prevent out module
authors from marking their stuff "permanent" for performance reasons
when it is not.

This should help with scalability: benchmark is "read /proc/cpuinfo R times
by N threads scattered over the system".

	N	R	t, s (before)	t, s (after)
	-----------------------------------------------------
	64	4096	1.582458	1.530502	-3.2%
	256	4096	6.371926	6.125168	-3.9%
	1024	4096	25.64888	24.47528	-4.6%

Benchmark source:

#include 
#include 
#include 
#include 

#include 
#include 
#include 
#include 

const int NR_CPUS = sysconf(_SC_NPROCESSORS_ONLN);
int N;
const char *filename;
int R;

int xxx = 0;

int glue(int n)
{
	cpu_set_t m;
	CPU_ZERO(&m);
	CPU_SET(n, &m);
	return sched_setaffinity(0, sizeof(cpu_set_t), &m);
}

void f(int n)
{
	glue(n % NR_CPUS);

	while (*(volatile int *)&xxx == 0) {
	}

	for (int i = 0; i < R; i++) {
		int fd = open(filename, O_RDONLY);
		char buf[4096];
		ssize_t rv = read(fd, buf, sizeof(buf));
		asm volatile ("" :: "g" (rv));
		close(fd);
	}
}

int main(int argc, char *argv[])
{
	if (argc < 4) {
		std::cerr << "usage: " << argv[0] << ' ' << "N /proc/filename R
";
		return 1;
	}

	N = atoi(argv[1]);
	filename = argv[2];
	R = atoi(argv[3]);

	for (int i = 0; i < NR_CPUS; i++) {
		if (glue(i) == 0)
			break;
	}

	std::vector T;
	T.reserve(N);
	for (int i = 0; i < N; i++) {
		T.emplace_back(f, i);
	}

	auto t0 = std::chrono::system_clock::now();
	{
		*(volatile int *)&xxx = 1;
		for (auto& t: T) {
			t.join();
		}
	}
	auto t1 = std::chrono::system_clock::now();
	std::chrono::duration dt = t1 - t0;
	std::cout << dt.count() << '
';

	return 0;
}

P.S.:
Explicit randomization marker is added because adding non-function pointer
will silently disable structure layout randomization.

[akpm@linux-foundation.org: coding style fixes]
Reported-by: kbuild test robot 
Reported-by: Dan Carpenter 
Signed-off-by: Alexey Dobriyan 
Signed-off-by: Andrew Morton 
Cc: Al Viro 
Cc: Joe Perches 
Link: http://lkml.kernel.org/r/20200222201539.GA22576@avx2
Signed-off-by: Linus Torvalds

fs/proc/inode.c: annotate close_pdeo() for sparse

2020-04-07T17:43:42+00:00

Fix sparse locking imbalance warning:

	warning: context imbalance in close_pdeo() - unexpected unlock

Signed-off-by: Jules Irenge 
Signed-off-by: Alexey Dobriyan 
Signed-off-by: Andrew Morton 
Link: http://lkml.kernel.org/r/20200227201538.GA30462@avx2
Signed-off-by: Linus Torvalds

proc: Use a list of inodes to flush from proc

2020-02-24T16:14:44+00:00

Rework the flushing of proc to use a list of directory inodes that
need to be flushed.

The list is kept on struct pid not on struct task_struct, as there is
a fixed connection between proc inodes and pids but at least for the
case of de_thread the pid of a task_struct changes.

This removes the dependency on proc_mnt which allows for different
mounts of proc having different mount options even in the same pid
namespace and this allows for the removal of proc_mnt which will
trivially the first mount of proc to honor it's mount options.

This flushing remains an optimization.  The functions
pid_delete_dentry and pid_revalidate ensure that ordinary dcache
management will not attempt to use dentries past the point their
respective task has died.  When unused the shrinker will
eventually be able to remove these dentries.

There is a case in de_thread where proc_flush_pid can be
called early for a given pid.  Which winds up being
safe (if suboptimal) as this is just an optiimization.

Only pid directories are put on the list as the other
per pid files are children of those directories and
d_invalidate on the directory will get them as well.

So that the pid can be used during flushing it's reference count is
taken in release_task and dropped in proc_flush_pid.  Further the call
of proc_flush_pid is moved after the tasklist_lock is released in
release_task so that it is certain that the pid has already been
unhashed when flushing it taking place.  This removes a small race
where a dentry could recreated.

As struct pid is supposed to be small and I need a per pid lock
I reuse the only lock that currently exists in struct pid the
the wait_pidfd.lock.

The net result is that this adds all of this functionality
with just a little extra list management overhead and
a single extra pointer in struct pid.

v2: Initialize pid->inodes.  I somehow failed to get that
    initialization into the initial version of the patch.  A boot
    failure was reported by "kernel test robot ", and
    failure to initialize that pid->inodes matches all of the reported
    symptoms.

Signed-off-by: Eric W. Biederman

proc: Clear the pieces of proc_inode that proc_evict_inode cares about

2020-02-24T15:51:25+00:00

This just keeps everything tidier, and allows for using flags like
SLAB_TYPESAFE_BY_RCU where slabs are not always cleared before reuse.
I don't see reuse without reinitializing happening with the proc_inode
but I had a false alarm while reworking flushing of proc dentries and
indoes when a process dies that caused me to tidy this up.

The code is a little easier to follow and reason about this
way so I figured the changes might as well be kept.

Signed-off-by: "Eric W. Biederman"

proc: Use d_invalidate in proc_prune_siblings_dcache

2020-02-24T15:50:04+00:00

The function d_prune_aliases has the problem that it will only prune
aliases thare are completely unused.  It will not remove aliases for
the dcache or even think of removing mounts from the dcache.  For that
behavior d_invalidate is needed.

To use d_invalidate replace d_prune_aliases with d_find_alias followed
by d_invalidate and dput.

For completeness the directory and the non-directory cases are
separated because in theory (although not in currently in practice for
proc) directories can only ever have a single dentry while
non-directories can have hardlinks and thus multiple dentries.
As part of this separation use d_find_any_alias for directories
to spare d_find_alias the extra work of doing that.

Plus the differences between d_find_any_alias and d_find_alias makes
it clear why the directory and non-directory code and not share code.

To make it clear these routines now invalidate dentries rename
proc_prune_siblings_dache to proc_invalidate_siblings_dcache, and rename
proc_sys_prune_dcache proc_sys_invalidate_dcache.

V2: Split the directory and non-directory cases.  To make this
    code robust to future changes in proc.

Signed-off-by: "Eric W. Biederman"

proc: In proc_prune_siblings_dcache cache an aquired super block

2020-02-21T20:06:42+00:00

Because there are likely to be several sysctls in a row on the
same superblock cache the super_block after the count has
been raised and don't deactivate it until we are processing
another super_block.

Signed-off-by: "Eric W. Biederman"

proc: Generalize proc_sys_prune_dcache into proc_prune_siblings_dcache

2020-02-20T16:06:02+00:00

This prepares the way for allowing the pid part of proc to use this
dcache pruning code as well.

Signed-off-by: Eric W. Biederman

proc: Rename in proc_inode rename sysctl_inodes sibling_inodes

2020-02-20T14:14:32+00:00

I about to need and use the same functionality for pid based
inodes and there is no point in adding a second field when
this field is already here and serving the same purporse.

Just give the field a generic name so it is clear that
it is no longer sysctl specific.

Also for good measure initialize sibling_inodes when
proc_inode is initialized.

Signed-off-by: Eric W. Biederman

proc: decouple proc from VFS with "struct proc_ops"

2020-02-04T03:05:26+00:00

Currently core /proc code uses "struct file_operations" for custom hooks,
however, VFS doesn't directly call them.  Every time VFS expands
file_operations hook set, /proc code bloats for no reason.

Introduce "struct proc_ops" which contains only those hooks which /proc
allows to call into (open, release, read, write, ioctl, mmap, poll).  It
doesn't contain module pointer as well.

Save ~184 bytes per usage:

	add/remove: 26/26 grow/shrink: 1/4 up/down: 1922/-6674 (-4752)
	Function                                     old     new   delta
	sysvipc_proc_ops                               -      72     +72
				...
	config_gz_proc_ops                             -      72     +72
	proc_get_inode                               289     339     +50
	proc_reg_get_unmapped_area                   110     107      -3
	close_pdeo                                   227     224      -3
	proc_reg_open                                289     284      -5
	proc_create_data                              60      53      -7
	rt_cpu_seq_fops                              256       -    -256
				...
	default_affinity_proc_fops                   256       -    -256
	Total: Before=5430095, After=5425343, chg -0.09%

Link: http://lkml.kernel.org/r/20191225172228.GA13378@avx2
Signed-off-by: Alexey Dobriyan 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

fs/proc/inode.c: use typeof_member() macro

2019-07-17T02:23:21+00:00

Don't repeat function signatures twice.

This is a kind-of-precursor for "struct proc_ops".

Note:

	typeof(pde->proc_fops->...) ...;

can't be used because ->proc_fops is "const struct file_operations *".
"const" prevents assignment down the code and it can't be deleted in the
type system.

Link: http://lkml.kernel.org/r/20190529191110.GB5703@avx2
Signed-off-by: Alexey Dobriyan 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds