1
0
mirror of https://git.FreeBSD.org/src.git synced 2025-01-30 16:51:41 +00:00

In kern_sendfile() adjust byte accounting of the file sending loop to

ignore the size of any headers that were passed with the sendfile(2)
system call.  Otherwise the file sent will be truncated by the header
size if the nbytes parameter was provided.  The bug doesn't show up
when either nbytes is zero, meaning send the whole file, or no header
iovec is provided.

Resolve a potential error aliasing of errors from the VM and sf_buf
parts and the protocol send parts where an error of the latter over-
writes one of the former.

Update comments.

The byte accounting bug wasn't seen in earlier because none of the popular
sendfile(2) consumers, Apache, lighttpd and our ftpd(8) use it in modes
that trigger it.  The varnish HTTP proxy makes full use of it and exposed
the problem.

Bug found by:	phk
Tested by:	phk
This commit is contained in:
Andre Oppermann 2007-05-19 20:50:59 +00:00
parent 98a0900ab4
commit 1c9dbd1567
Notes: svn2git 2020-12-20 02:59:44 +00:00
svn path=/head/; revision=169779

View File

@ -1819,8 +1819,8 @@ kern_sendfile(struct thread *td, struct sendfile_args *uap,
struct mbuf *m = NULL;
struct sf_buf *sf;
struct vm_page *pg;
off_t off, xfsize, sbytes = 0, rem = 0;
int error, mnw = 0;
off_t off, xfsize, fsbytes = 0, sbytes = 0, rem = 0;
int error, hdrlen = 0, mnw = 0;
int vfslocked;
NET_LOCK_GIANT();
@ -1916,6 +1916,7 @@ kern_sendfile(struct thread *td, struct sendfile_args *uap,
error = mnw ? EAGAIN : ENOBUFS;
goto out;
}
hdrlen = m_length(m, NULL);
}
}
@ -1933,7 +1934,7 @@ kern_sendfile(struct thread *td, struct sendfile_args *uap,
* The outer loop checks the state and available space of the socket
* and takes care of the overall progress.
*/
for (off = uap->offset; ; ) {
for (off = uap->offset, rem = uap->nbytes; ; ) {
int loopbytes = 0;
int space = 0;
int done = 0;
@ -1997,6 +1998,13 @@ kern_sendfile(struct thread *td, struct sendfile_args *uap,
}
SOCKBUF_UNLOCK(&so->so_snd);
/*
* Reduce space in the socket buffer by the size of
* the header mbuf chain.
* hdrlen is set to 0 after the first loop.
*/
space -= hdrlen;
/*
* Loop and construct maximum sized mbuf chain to be bulk
* dumped into socket buffer.
@ -2015,12 +2023,12 @@ kern_sendfile(struct thread *td, struct sendfile_args *uap,
pgoff = (vm_offset_t)(off & PAGE_MASK);
xfsize = omin(PAGE_SIZE - pgoff,
obj->un_pager.vnp.vnp_size - uap->offset -
sbytes - loopbytes);
fsbytes - loopbytes);
if (uap->nbytes)
rem = (uap->nbytes - sbytes - loopbytes);
rem = (uap->nbytes - fsbytes - loopbytes);
else
rem = obj->un_pager.vnp.vnp_size - uap->offset -
sbytes - loopbytes;
rem = obj->un_pager.vnp.vnp_size -
uap->offset - fsbytes - loopbytes;
xfsize = omin(rem, xfsize);
if (xfsize <= 0) {
VM_OBJECT_UNLOCK(obj);
@ -2038,9 +2046,8 @@ kern_sendfile(struct thread *td, struct sendfile_args *uap,
}
/*
* Attempt to look up the page.
* Allocate if not found or
* wait and loop if busy.
* Attempt to look up the page. Allocate
* if not found or wait and loop if busy.
*/
pindex = OFF_TO_IDX(off);
pg = vm_page_grab(obj, pindex, VM_ALLOC_NOBUSY |
@ -2161,7 +2168,7 @@ kern_sendfile(struct thread *td, struct sendfile_args *uap,
/* Add the buffer chain to the socket buffer. */
if (m != NULL) {
int mlen;
int mlen, err;
mlen = m_length(m, NULL);
SOCKBUF_LOCK(&so->so_snd);
@ -2171,10 +2178,27 @@ kern_sendfile(struct thread *td, struct sendfile_args *uap,
goto done;
}
SOCKBUF_UNLOCK(&so->so_snd);
error = (*so->so_proto->pr_usrreqs->pru_send)
/* Avoid error aliasing. */
err = (*so->so_proto->pr_usrreqs->pru_send)
(so, 0, m, NULL, NULL, td);
if (!error)
if (err == 0) {
/*
* We need two counters to get the
* file offset and nbytes to send
* right:
* - sbytes contains the total amount
* of bytes sent, including headers.
* - fsbytes contains the total amount
* of bytes sent from the file.
*/
sbytes += mlen;
fsbytes += mlen;
if (hdrlen) {
fsbytes -= hdrlen;
hdrlen = 0;
}
} else if (error == 0)
error = err;
m = NULL; /* pru_send always consumes */
}