Date: Tue, 11 Jul 2023 00:49:49 GMT From: Konstantin Belousov <kib@FreeBSD.org> To: src-committers@FreeBSD.org, dev-commits-src-all@FreeBSD.org, dev-commits-src-main@FreeBSD.org Subject: git: 7e4c6b2163fb - main - msdosfs: zero partially valid extended cluster Message-ID: <202307110049.36B0nnxN082765@gitrepo.freebsd.org>
next in thread | raw e-mail | index | archive | help
The branch main has been updated by kib: URL: https://cgit.FreeBSD.org/src/commit/?id=7e4c6b2163fbed6be92b1e19f0eec4da973cfaec commit 7e4c6b2163fbed6be92b1e19f0eec4da973cfaec Author: Konstantin Belousov <kib@FreeBSD.org> AuthorDate: 2023-07-06 01:51:07 +0000 Commit: Konstantin Belousov <kib@FreeBSD.org> CommitDate: 2023-07-11 00:49:09 +0000 msdosfs: zero partially valid extended cluster It contains arbitrary garbage, which is not cleared by vfs_bio_clrbuf() which only zeroes invalid portions of the pages. Reported by: Maxim Suhanov <dfirblog@gmail.com> Discussed with: so Tested by: pho Sponsored by: The FreeBSD Foundation MFC after: 1 week --- sys/fs/msdosfs/msdosfs_denode.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/sys/fs/msdosfs/msdosfs_denode.c b/sys/fs/msdosfs/msdosfs_denode.c index 7acd90c8e713..ce7058b3ae60 100644 --- a/sys/fs/msdosfs/msdosfs_denode.c +++ b/sys/fs/msdosfs/msdosfs_denode.c @@ -498,6 +498,7 @@ deextend(struct denode *dep, u_long length, struct ucred *cred) struct msdosfsmount *pmp = dep->de_pmp; struct vnode *vp = DETOV(dep); struct buf *bp; + off_t eof_clusteroff; u_long count; int error; @@ -536,13 +537,19 @@ deextend(struct denode *dep, u_long length, struct ucred *cred) * B_CACHE | B_DELWRI but with invalid pages, and cannot be * neither written out nor validated. * - * Fix it by proactively clearing extended pages. + * Fix it by proactively clearing extended pages. Need to do + * both vfs_bio_clrbuf() to mark pages valid, and to zero + * actual buffer content which might exist in the tail of the + * already valid cluster. */ error = bread(vp, de_cluster(pmp, dep->de_FileSize), pmp->pm_bpcluster, NOCRED, &bp); if (error != 0) goto rewind; vfs_bio_clrbuf(bp); + eof_clusteroff = de_cn2off(pmp, de_cluster(pmp, dep->de_FileSize)); + vfs_bio_bzero_buf(bp, dep->de_FileSize - eof_clusteroff, + pmp->pm_bpcluster - dep->de_FileSize + eof_clusteroff); if (!DOINGASYNC(vp)) (void)bwrite(bp); else if (vm_page_count_severe() || buf_dirty_count_severe())
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?202307110049.36B0nnxN082765>