afs: Zero uninitialized uio structs

In several places in the code, we allocate a 'struct uio' on the
stack, or allocate one from non-zeroed memory. In most of these
places, we initialize the structure by assigning individual fields to
certain values. However, this leaves any remaining fields assigned to
random garbage, if there are any additional fields in the struct uio
that we don't know about.

One such platform is Solaris, which has a field called uio_extflg,
which exists in Solaris 11, Solaris 10, and possibly further back.
One of the flags defined for this field in Solaris 11 is UIO_XUIO,
which indicates that the structure is actually an xuio_t, which is
larger than a normal uio_t and contains additional fields. So when we
allocate a uio on the stack without initializing it, it can randomly
appear to be an xuio_t, depending on what garbage was on the stack at
the time. An xuio_t is a kind of extensible structure, which is used
for things like async I/O or DMA, that kind of thing.

One of the places we make use of such a uio_t is in afs_ustrategy,
which we go through for cache reads and writes on most Unix platforms
(but not Linux). When handling a read (reading from the disk cache
into a mapped page), a copy of our stack-allocated uio eventually gets
passed to VOP_READ. So VOP_READ for the cache filesystem will randomly
interpret our uio_t as an xuio_t.

In many scenarios, this (amazingly) does not cause any problems, since
generally, Solaris code will not notice if something is flagged as an
xuio_t, unless it is specifically written to handle specific xuio_t
types. ZFS is one of the apparent few filesystem implementations that
can handle xuio_t's, and will detect and specially handle a
UIOTYPE_ZEROCOPY xuio_t differently than a regular uio_t.

If ZFS gets a UIOTYPE_ZEROCOPY xuio_t, it appears to ignore the uio
buffers passed in, and supplies its own buffers from its cache. This
means that our VOP_READ request will return success, and act like it
serviced the read just fine. However, the actual buffer that we passed
in will remain untouched, and so we will return the page to the VFS
filled with garbage data.

The way this typically manifests is that seemingly random pages will
contain random data. This seems to happen very rarely, though it may
not always be obvious what is going on when this occurs.

It is also worth noting that the above description on Solaris only
happens with Solaris 11 and newer, and only with a ZFS disk cache.
Anything older than Solaris 11 does not have the xuio_t framework
(though other uio_extflg values can cause performance degradations),
and all known non-ZFS local disk filesystems do not interpret special
xuio_t structures (networked filesystems might have xuio_t handling,
but they shouldn't be used for a cache).

Bugs similar to this may also exist on other Unix clients, but at
least this specific scenario should not occur on Linux (since we don't
use afs_ustrategy), and newer Darwin (since we get a uio allocated for
us).

To fix this, zero out the entire uio structure before we use it, for
all instances where we allocate a uio from the stack or from
non-zeroed memory. Also zero out the accompanying iovec in many
places, just to be safe. Some of these may not actually need to be
zeroed (since we do actually initialize the whole thing, or a platform
doesn't have any additional unknown uio fields), but it seems
worthwhile to err on the side of caution.

Thanks to Oracle for their assistance on this issue, and thanks to the
organization experiencing this issue for their patience and
persistence.

Change-Id: I0eae0b49a70aee19f3a9ec118b03cfb3a6bd03a3
Reviewed-on: http://gerrit.openafs.org/11705
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Perry Ruiter <pruiter@sinenomine.net>
Reviewed-by: Daria Brashear <shadow@your-file-system.com>

src/afs/AIX/osi_misc.c

@@ -97,6 +97,9 @@ gop_rdwr(rw, vp, base, len, offset, segflg, unit, aresid)
     struct iovec uiovector;
     int code;
 
+    memset(&uio_struct, 0, sizeof(uio_struct));
+    memset(&uiovector, 0, sizeof(uiovector));
+
     AFS_STATCNT(gop_rdwr);
     /* Set up the uio structure */
     uiovector.iov_base = (caddr_t) base;

src/afs/AIX/osi_vnodeops.c

@@ -643,6 +643,9 @@ afs_gn_fclear(struct vnode *vp,
     static int fclear_init = 0;
     struct vcache *avc = VTOAFS(vp);
 
+    memset(&uio, 0, sizeof(uio));
+    memset(&iov, 0, sizeof(iov));
+
     AFS_STATCNT(afs_gn_fclear);
     if (!fclear_init) {
 	memset(zero_buffer, 0, PAGESIZE);
@@ -919,6 +922,9 @@ afs_vm_rdwr(struct vnode *vp,
 	    struct iovec tvec[16];	/* Should have access to #define */
 	    afs_int32 tsize;
 
+	    memset(&tuio, 0, sizeof(tuio));
+	    memset(&tvec, 0, sizeof(tvec));
+
 	    mixed = 1;
 	    finalOffset = xfrOffset + xfrSize;
 	    tsize = (afs_size_t) (xfrOffset + xfrSize - afs_vmMappingEnd);
@@ -1086,6 +1092,9 @@ afs_vm_rdwr(struct vnode *vp,
 	    struct uio tuio;
 	    struct iovec tvec[16];	/* Should have access to #define */
 
+	    memset(&tuio, 0, sizeof(tuio));
+	    memset(&tvec, 0, sizeof(tvec));
+
 	    /* Purge dirty chunks of file if there are too many dirty chunks.
 	     * Inside the write loop, we only do this at a chunk boundary.
 	     * Clean up partial chunk if necessary at end of loop.

src/afs/DARWIN/osi_vnodeops.c

@@ -837,17 +837,20 @@ afs_vop_pagein(ap)
     int flags = ap->a_flags;
     struct ucred *cred;
     vm_offset_t ioaddr;
+    int code;
+    struct vcache *tvc = VTOAFS(vp);
+    int nocommit = flags & UPL_NOCOMMIT;
 #ifdef AFS_DARWIN80_ENV
     struct uio *uio;
 #else
     struct uio auio;
     struct iovec aiov;
     struct uio *uio = &auio;
-#endif
-    int nocommit = flags & UPL_NOCOMMIT;
 
-    int code;
-    struct vcache *tvc = VTOAFS(vp);
+    memset(&auio, 0, sizeof(auio));
+    memset(&aiov, 0, sizeof(aiov));
+#endif
+
 #ifndef AFS_DARWIN80_ENV
     if (UBCINVALID(vp)) {
 #if DIAGNOSTIC
@@ -982,18 +985,21 @@ afs_vop_pageout(ap)
     int flags = ap->a_flags;
     struct ucred *cred;
     vm_offset_t ioaddr;
+    int nocommit = flags & UPL_NOCOMMIT;
+    int iosize;
+    int code;
+    struct vcache *tvc = VTOAFS(vp);
 #ifdef AFS_DARWIN80_ENV
     struct uio *uio;
 #else
     struct uio auio;
     struct iovec aiov;
     struct uio *uio = &auio;
-#endif
-    int nocommit = flags & UPL_NOCOMMIT;
-    int iosize;
 
-    int code;
-    struct vcache *tvc = VTOAFS(vp);
+    memset(&auio, 0, sizeof(auio));
+    memset(&aiov, 0, sizeof(aiov));
+#endif
+
 #ifndef AFS_DARWIN80_ENV
     if (UBCINVALID(vp)) {
 #if DIAGNOSTIC

src/afs/FBSD/osi_vnodeops.c

@@ -801,6 +801,9 @@ afs_vop_getpages(struct vop_getpages_args *ap)
     struct vnode *vp;
     struct vcache *avc;
 
+    memset(&uio, 0, sizeof(uio));
+    memset(&iov, 0, sizeof(iov));
+
     vp = ap->a_vp;
     avc = VTOAFS(vp);
     if ((object = vp->v_object) == NULL) {
@@ -993,6 +996,9 @@ afs_vop_putpages(struct vop_putpages_args *ap)
     struct vnode *vp;
     struct vcache *avc;
 
+    memset(&uio, 0, sizeof(uio));
+    memset(&iov, 0, sizeof(iov));
+
     vp = ap->a_vp;
     avc = VTOAFS(vp);
     /* Perhaps these two checks should just be KASSERTs instead... */

src/afs/HPUX/osi_vnodeops.c

@@ -222,6 +222,9 @@ afs_bread(vp, lbn, bpp)
     struct iovec iov;
     struct uio uio;
 
+    memset(&uio, 0, sizeof(uio));
+    memset(&iov, 0, sizeof(iov));
+
     AFS_STATCNT(afs_bread);
     fsbsize = vp->v_vfsp->vfs_bsize;
     offset = lbn * fsbsize;
@@ -2085,6 +2088,9 @@ afs_readdir(vp, uiop, cred)
     dir_off_t offset;
     uint64_t tmp_offset;
 
+    memset(&auio, 0, sizeof(auio));
+    memset(&aiov, 0, sizeof(aiov));
+
     count = uiop->uio_resid;
     /* Allocate temporary space for format conversion */
     ibuf = kmem_alloc(2 * count);	/* overkill - fix later */
@@ -2151,6 +2157,9 @@ afs_readdir3(vp, uiop, cred)
     int count, outcount;
     dir_off_t offset;
 
+    memset(&auio, 0, sizeof(auio));
+    memset(&aiov, 0, sizeof(aiov));
+
     count = uiop->uio_resid;
     /* Allocate temporary space for format conversion */
     ibuf = kmem_alloc(2 * count);	/* overkill - fix later */
@@ -2521,6 +2530,9 @@ afs_hp_strategy(bp)
     extern caddr_t hdl_kmap_bp();
     struct kthread *t = u.u_kthreadp;
 
+    memset(&tuio, 0, sizeof(tuio));
+    memset(&tiovec, 0, sizeof(tiovec));
+
     AFS_STATCNT(afs_hp_strategy);
     /*
      * hdl_kmap_bp() saves "b_bcount" and restores it in hdl_remap_bp() after

src/afs/LINUX/osi_file.c

@@ -208,6 +208,9 @@ afs_osi_Read(struct osi_file *afile, int offset, void *aptr,
     struct iovec iov;
     afs_int32 code;
 
+    memset(&auio, 0, sizeof(auio));
+    memset(&iov, 0, sizeof(iov));
+
     AFS_STATCNT(osi_Read);
 
     /*
@@ -249,6 +252,9 @@ afs_osi_Write(struct osi_file *afile, afs_int32 offset, void *aptr,
     struct iovec iov;
     afs_int32 code;
 
+    memset(&auio, 0, sizeof(auio));
+    memset(&iov, 0, sizeof(iov));
+
     AFS_STATCNT(osi_Write);
 
     if (!afile) {

src/afs/LINUX/osi_vnodeops.c

@@ -1855,6 +1855,9 @@ afs_linux_ireadlink(struct inode *ip, char *target, int maxlen, uio_seg_t seg)
     struct uio tuio;
     struct iovec iov;
 
+    memset(&tuio, 0, sizeof(tuio));
+    memset(&iov, 0, sizeof(iov));
+
     setup_uio(&tuio, &iov, target, (afs_offs_t) 0, maxlen, UIO_READ, seg);
     code = afs_readlink(VTOAFS(ip), &tuio, credp);
     crfree(credp);
@@ -2608,6 +2611,9 @@ afs_linux_page_writeback(struct inode *ip, struct page *pp,
     struct iovec iovec;
     int f_flags = 0;
 
+    memset(&tuio, 0, sizeof(tuio));
+    memset(&iovec, 0, sizeof(iovec));
+
     buffer = kmap(pp) + offset;
     base = page_offset(pp) + offset;
 

src/afs/LINUX24/osi_file.c

@@ -190,6 +190,9 @@ afs_osi_Read(struct osi_file *afile, int offset, void *aptr,
     struct iovec iov;
     afs_int32 code;
 
+    memset(&auio, 0, sizeof(auio));
+    memset(&iov, 0, sizeof(iov));
+
     AFS_STATCNT(osi_Read);
 
     /*
@@ -231,6 +234,9 @@ afs_osi_Write(struct osi_file *afile, afs_int32 offset, void *aptr,
     struct iovec iov;
     afs_int32 code;
 
+    memset(&auio, 0, sizeof(auio));
+    memset(&iov, 0, sizeof(iov));
+
     AFS_STATCNT(osi_Write);
 
     if (!afile) {

src/afs/LINUX24/osi_vnodeops.c

@@ -1546,6 +1546,9 @@ afs_linux_ireadlink(struct inode *ip, char *target, int maxlen, uio_seg_t seg)
     struct uio tuio;
     struct iovec iov;
 
+    memset(&tuio, 0, sizeof(tuio));
+    memset(&iov, 0, sizeof(iov));
+
     setup_uio(&tuio, &iov, target, (afs_offs_t) 0, maxlen, UIO_READ, seg);
     code = afs_readlink(VTOAFS(ip), &tuio, credp);
     crfree(credp);
@@ -1814,6 +1817,9 @@ afs_linux_writepage_sync(struct inode *ip, struct page *pp,
     struct iovec iovec;
     int f_flags = 0;
 
+    memset(&tuio, 0, sizeof(tuio));
+    memset(&iovec, 0, sizeof(iovec));
+
     buffer = kmap(pp) + offset;
     base = (((loff_t) pp->index) << PAGE_CACHE_SHIFT)  + offset;
 
@@ -1923,6 +1929,9 @@ afs_linux_updatepage(struct file *fp, struct page *pp, unsigned long offset,
     struct uio tuio;
     struct iovec iovec;
 
+    memset(&tuio, 0, sizeof(tuio));
+    memset(&iovec, 0, sizeof(iovec));
+
     set_bit(PG_locked, &pp->flags);
 
     credp = crref();

src/afs/OBSD/osi_vnodeops.c

@@ -1033,6 +1033,9 @@ afs_obsd_strategy(void *v)
     long len = abp->b_bcount;
     int code;
 
+    memset(&tuio, 0, sizeof(tuio));
+    memset(&tiovec, 0, sizeof(tiovec));
+
     AFS_STATCNT(afs_strategy);
 
     tuio.afsio_iov = tiovec;

src/afs/VNOPS/afs_vnop_strategy.c

@@ -46,6 +46,9 @@ int afs_ustrategy(struct buf *abp)
     afs_ucred_t *credp = u.u_cred;
 #endif
 
+    memset(&tuio, 0, sizeof(tuio));
+    memset(&tiovec, 0, sizeof(tiovec));
+
     AFS_STATCNT(afs_ustrategy);
 #ifdef	AFS_AIX41_ENV
     /*

src/afs/afs_osi_uio.c

@@ -88,10 +88,12 @@ afsio_partialcopy(struct uio *auio, size_t len) {
     char *space;
     struct uio *newuio;
     struct iovec *newvec;
+    size_t space_len = sizeof(struct uio) +
+                       sizeof(struct iovec) * AFS_MAXIOVCNT;
 
     /* Allocate a block that can contain both the UIO and the iovec */
-    space = osi_AllocSmallSpace(sizeof(struct uio) +
-				sizeof(struct iovec) * AFS_MAXIOVCNT);
+    space = osi_AllocSmallSpace(space_len);
+    memset(space, 0, space_len);
 
     newuio = (struct uio *) space;
     newvec = (struct iovec *) (space + sizeof(struct uio));