sched_ext: Maintain an online cid mask in the scheduler arenagrafted

Schedulers on the default cid mapping treat [0, nr_online_cids) as the
online set and restart on hotplug. Schedulers that install their own mapping
with scx_bpf_cid_override() have no way to learn which cids are online: the
count no longer identifies members and the CPU-form cpumask is unusable from
cid programs. This is an obvious hole in the cid API.

Add scx_bpf_online_cmask(), a kernel-maintained cmask in the scheduler's
arena, allocated alongside the per-CPU scratch masks and populated after the
cid mapping is finalized and before ops.init(), for child schedulers too.
The pointer stays valid through ops.exit() with no reference to take. It is
the arena offset as a void pointer, the same form struct_ops arena arguments
arrive in. The verifier types the void return as a scalar for the program's
arena cast.

The mask follows the SCX hotplug notifications: seeded from cpu_active_mask
and updated before ops.cid_online/offline() runs, so it lags cpu_online_mask
only inside a hotplug transition. Updates walk the scheduler list under the
lock that also serializes unlinking. Reads are live, not atomic snapshots.
Root initialization excludes hotplug.

v2: Reworded the getter kerneldoc (Andrea Righi).

Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Andrea Righi <arighi@nvidia.com>

1 files changed, 225 insertions, 0 deletions

diff --git a/fs/squashfs/dir.c b/fs/squashfs/dir.c
new file mode 100644
index 000000000..cd3598bd0
--- /dev/null
+++ b/fs/squashfs/dir.c
@@ -0,0 +1,225 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@squashfs.org.uk>
+ *
+ * dir.c
+ */
+
+/*
+ * This file implements code to read directories from disk.
+ *
+ * See namei.c for a description of directory organisation on disk.
+ */
+
+#include <linux/fs.h>
+#include <linux/filelock.h>
+#include <linux/vfs.h>
+#include <linux/slab.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+static const unsigned char squashfs_filetype_table[] = {
+	DT_UNKNOWN, DT_DIR, DT_REG, DT_LNK, DT_BLK, DT_CHR, DT_FIFO, DT_SOCK
+};
+
+/*
+ * Lookup offset (f_pos) in the directory index, returning the
+ * metadata block containing it.
+ *
+ * If we get an error reading the index then return the part of the index
+ * (if any) we have managed to read - the index isn't essential, just
+ * quicker.
+ */
+static int get_dir_index_using_offset(struct super_block *sb,
+	u64 *next_block, int *next_offset, u64 index_start, int index_offset,
+	int i_count, u64 f_pos)
+{
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	int err, i, index, length = 0;
+	unsigned int size;
+	struct squashfs_dir_index dir_index;
+
+	TRACE("Entered get_dir_index_using_offset, i_count %d, f_pos %lld\n",
+					i_count, f_pos);
+
+	/*
+	 * Translate from external f_pos to the internal f_pos.  This
+	 * is offset by 3 because we invent "." and ".." entries which are
+	 * not actually stored in the directory.
+	 */
+	if (f_pos <= 3)
+		return f_pos;
+	f_pos -= 3;
+
+	for (i = 0; i < i_count; i++) {
+		err = squashfs_read_metadata(sb, &dir_index, &index_start,
+				&index_offset, sizeof(dir_index));
+		if (err < 0)
+			break;
+
+		index = le32_to_cpu(dir_index.index);
+		if (index > f_pos)
+			/*
+			 * Found the index we're looking for.
+			 */
+			break;
+
+		size = le32_to_cpu(dir_index.size) + 1;
+
+		/* size should never be larger than SQUASHFS_NAME_LEN */
+		if (size > SQUASHFS_NAME_LEN)
+			break;
+
+		err = squashfs_read_metadata(sb, NULL, &index_start,
+				&index_offset, size);
+		if (err < 0)
+			break;
+
+		length = index;
+		*next_block = le32_to_cpu(dir_index.start_block) +
+					msblk->directory_table;
+	}
+
+	*next_offset = (length + *next_offset) % SQUASHFS_METADATA_SIZE;
+
+	/*
+	 * Translate back from internal f_pos to external f_pos.
+	 */
+	return length + 3;
+}
+
+
+static int squashfs_readdir(struct file *file, struct dir_context *ctx)
+{
+	struct inode *inode = file_inode(file);
+	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
+	u64 block = squashfs_i(inode)->start + msblk->directory_table;
+	int offset = squashfs_i(inode)->offset, length, err;
+	unsigned int inode_number, dir_count, size, type;
+	struct squashfs_dir_header dirh;
+	struct squashfs_dir_entry *dire;
+
+	TRACE("Entered squashfs_readdir [%llx:%x]\n", block, offset);
+
+	dire = kmalloc(sizeof(*dire) + SQUASHFS_NAME_LEN + 1, GFP_KERNEL);
+	if (dire == NULL) {
+		ERROR("Failed to allocate squashfs_dir_entry\n");
+		goto finish;
+	}
+
+	/*
+	 * Return "." and  ".." entries as the first two filenames in the
+	 * directory.  To maximise compression these two entries are not
+	 * stored in the directory, and so we invent them here.
+	 *
+	 * It also means that the external f_pos is offset by 3 from the
+	 * on-disk directory f_pos.
+	 */
+	while (ctx->pos < 3) {
+		char *name;
+		int i_ino;
+
+		if (ctx->pos == 0) {
+			name = ".";
+			size = 1;
+			i_ino = inode->i_ino;
+		} else {
+			name = "..";
+			size = 2;
+			i_ino = squashfs_i(inode)->parent;
+		}
+
+		if (!dir_emit(ctx, name, size, i_ino,
+				squashfs_filetype_table[1]))
+			goto finish;
+
+		ctx->pos += size;
+	}
+
+	length = get_dir_index_using_offset(inode->i_sb, &block, &offset,
+				squashfs_i(inode)->dir_idx_start,
+				squashfs_i(inode)->dir_idx_offset,
+				squashfs_i(inode)->dir_idx_cnt,
+				ctx->pos);
+
+	while (length < i_size_read(inode)) {
+		/*
+		 * Read directory header
+		 */
+		err = squashfs_read_metadata(inode->i_sb, &dirh, &block,
+					&offset, sizeof(dirh));
+		if (err < 0)
+			goto failed_read;
+
+		length += sizeof(dirh);
+
+		dir_count = le32_to_cpu(dirh.count) + 1;
+
+		if (dir_count > SQUASHFS_DIR_COUNT)
+			goto failed_read;
+
+		while (dir_count--) {
+			/*
+			 * Read directory entry.
+			 */
+			err = squashfs_read_metadata(inode->i_sb, dire, &block,
+					&offset, sizeof(*dire));
+			if (err < 0)
+				goto failed_read;
+
+			size = le16_to_cpu(dire->size) + 1;
+
+			/* size should never be larger than SQUASHFS_NAME_LEN */
+			if (size > SQUASHFS_NAME_LEN)
+				goto failed_read;
+
+			err = squashfs_read_metadata(inode->i_sb, dire->name,
+					&block, &offset, size);
+			if (err < 0)
+				goto failed_read;
+
+			length += sizeof(*dire) + size;
+
+			if (ctx->pos >= length)
+				continue;
+
+			dire->name[size] = '\0';
+			inode_number = le32_to_cpu(dirh.inode_number) +
+				((short) le16_to_cpu(dire->inode_number));
+			type = le16_to_cpu(dire->type);
+
+			if (type > SQUASHFS_MAX_DIR_TYPE)
+				goto failed_read;
+
+			if (!dir_emit(ctx, dire->name, size,
+					inode_number,
+					squashfs_filetype_table[type]))
+				goto finish;
+
+			ctx->pos = length;
+		}
+	}
+
+finish:
+	kfree(dire);
+	return 0;
+
+failed_read:
+	ERROR("Unable to read directory block [%llx:%x]\n", block, offset);
+	kfree(dire);
+	return 0;
+}
+
+
+const struct file_operations squashfs_dir_ops = {
+	.read = generic_read_dir,
+	.iterate_shared = squashfs_readdir,
+	.llseek = generic_file_llseek,
+	.setlease = generic_setlease,
+};