Commit 25e6bae3 authored by Yan, Zheng's avatar Yan, Zheng Committed by Sage Weil
Browse files

ceph: use pagelist to present MDS request data 



Current code uses page array to present MDS request data. Pages in the
array are allocated/freed by caller of ceph_mdsc_do_request(). If request
is interrupted, the pages can be freed while they are still being used by
the request message.

The fix is use pagelist to present MDS request data. Pagelist is
reference counted.

Signed-off-by: default avatarYan, Zheng <zyan@redhat.com>
Reviewed-by: default avatarSage Weil <sage@redhat.com>
parent e4339d28

fs/ceph/mds_client.c

+9 −5
Original line number Original line Diff line number Diff line
@@ -543,6 +543,8 @@ void ceph_mdsc_release_request(struct kref *kref) 
	}

	}

	kfree(req->r_path1);

	kfree(req->r_path1);

	kfree(req->r_path2);

	kfree(req->r_path2);

	if (req->r_pagelist)

		ceph_pagelist_release(req->r_pagelist);

	put_request_session(req);

	put_request_session(req);

	ceph_unreserve_caps(req->r_mdsc, &req->r_caps_reservation);

	ceph_unreserve_caps(req->r_mdsc, &req->r_caps_reservation);

	kfree(req);

	kfree(req);
@@ -1916,13 +1918,15 @@ static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc, 
	msg->front.iov_len = p - msg->front.iov_base;

	msg->front.iov_len = p - msg->front.iov_base;

	msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);

	msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);





	if (req->r_data_len) {

	if (req->r_pagelist) {

		/* outbound data set only by ceph_sync_setxattr() */

		struct ceph_pagelist *pagelist = req->r_pagelist;

		BUG_ON(!req->r_pages);

		atomic_inc(&pagelist->refcnt);

		ceph_msg_data_add_pages(msg, req->r_pages, req->r_data_len, 0);

		ceph_msg_data_add_pagelist(msg, pagelist);

		msg->hdr.data_len = cpu_to_le32(pagelist->length);

	} else {

		msg->hdr.data_len = 0;

	}

	}





	msg->hdr.data_len = cpu_to_le32(req->r_data_len);

	msg->hdr.data_off = cpu_to_le16(0);

	msg->hdr.data_off = cpu_to_le16(0);





out_free2:

out_free2:

fs/ceph/mds_client.h

+1 −3
Original line number Original line Diff line number Diff line
@@ -202,9 +202,7 @@ struct ceph_mds_request { 
	bool r_direct_is_hash;  /* true if r_direct_hash is valid */

	bool r_direct_is_hash;  /* true if r_direct_hash is valid */





	/* data payload is used for xattr ops */

	/* data payload is used for xattr ops */

	struct page **r_pages;

	struct ceph_pagelist *r_pagelist;

	int r_num_pages;

	int r_data_len;





	/* what caps shall we drop? */

	/* what caps shall we drop? */

	int r_inode_drop, r_inode_unless;

	int r_inode_drop, r_inode_unless;

fs/ceph/xattr.c

+18 −31
Original line number Original line Diff line number Diff line 
#include <linux/ceph/ceph_debug.h>

#include <linux/ceph/ceph_debug.h>

#include <linux/ceph/pagelist.h>





#include "super.h"

#include "super.h"

#include "mds_client.h"

#include "mds_client.h"
@@ -850,35 +851,25 @@ static int ceph_sync_setxattr(struct dentry *dentry, const char *name, 
	struct ceph_inode_info *ci = ceph_inode(inode);

	struct ceph_inode_info *ci = ceph_inode(inode);

	struct ceph_mds_request *req;

	struct ceph_mds_request *req;

	struct ceph_mds_client *mdsc = fsc->mdsc;

	struct ceph_mds_client *mdsc = fsc->mdsc;

	struct ceph_pagelist *pagelist = NULL;

	int err;

	int err;

	int i, nr_pages;



	struct page **pages = NULL;

	if (value) {

	void *kaddr;

		/* copy value into pagelist */



		pagelist = kmalloc(sizeof(*pagelist), GFP_NOFS);

	/* copy value into some pages */

		if (!pagelist)

	nr_pages = calc_pages_for(0, size);

	if (nr_pages) {

		pages = kmalloc(sizeof(pages[0])*nr_pages, GFP_NOFS);

		if (!pages)

			return -ENOMEM;

			return -ENOMEM;

		err = -ENOMEM;



		for (i = 0; i < nr_pages; i++) {

		ceph_pagelist_init(pagelist);

			pages[i] = __page_cache_alloc(GFP_NOFS);

		err = ceph_pagelist_append(pagelist, value, size);

			if (!pages[i]) {

		if (err)

				nr_pages = i;

			goto out;

			goto out;

			}

	} else {

			kaddr = kmap(pages[i]);

		flags |= CEPH_XATTR_REMOVE;

			memcpy(kaddr, value + i*PAGE_CACHE_SIZE,

			       min(PAGE_CACHE_SIZE, size-i*PAGE_CACHE_SIZE));

		}

	}

	}





	dout("setxattr value=%.*s\n", (int)size, value);

	dout("setxattr value=%.*s\n", (int)size, value);





	if (!value)

		flags |= CEPH_XATTR_REMOVE;



	/* do request */

	/* do request */

	req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_SETXATTR,

	req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_SETXATTR,

				       USE_AUTH_MDS);

				       USE_AUTH_MDS);
@@ -893,9 +884,8 @@ static int ceph_sync_setxattr(struct dentry *dentry, const char *name, 
	req->r_args.setxattr.flags = cpu_to_le32(flags);

	req->r_args.setxattr.flags = cpu_to_le32(flags);

	req->r_path2 = kstrdup(name, GFP_NOFS);

	req->r_path2 = kstrdup(name, GFP_NOFS);





	req->r_pages = pages;

	req->r_pagelist = pagelist;

	req->r_num_pages = nr_pages;

	pagelist = NULL;

	req->r_data_len = size;





	dout("xattr.ver (before): %lld\n", ci->i_xattrs.version);

	dout("xattr.ver (before): %lld\n", ci->i_xattrs.version);

	err = ceph_mdsc_do_request(mdsc, NULL, req);

	err = ceph_mdsc_do_request(mdsc, NULL, req);
@@ -903,11 +893,8 @@ static int ceph_sync_setxattr(struct dentry *dentry, const char *name, 
	dout("xattr.ver (after): %lld\n", ci->i_xattrs.version);

	dout("xattr.ver (after): %lld\n", ci->i_xattrs.version);





out:

out:

	if (pages) {

	if (pagelist)

		for (i = 0; i < nr_pages; i++)

		ceph_pagelist_release(pagelist);

			__free_page(pages[i]);

		kfree(pages);

	}

	return err;

	return err;

}

}

CRA Git | Maintained and supported by SUSTech CRA and CCSE