[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH v3 03/16] util: add memfd helpers
From: |
Marc-André Lureau |
Subject: |
Re: [Qemu-devel] [PATCH v3 03/16] util: add memfd helpers |
Date: |
Tue, 29 Sep 2015 11:25:04 -0400 (EDT) |
----- Original Message -----
> On Thu, Aug 06, 2015 at 02:40:39PM +0200, address@hidden wrote:
> > From: Marc-André Lureau <address@hidden>
> >
> > Add qemu_memfd_alloc/free() helpers.
> >
> > The function helps to allocate and seal a memfd, and implements an
> > open/unlink/mmap fallback for system that do not support memfd.
> >
> > Signed-off-by: Marc-André Lureau <address@hidden>
> > ---
> > include/qemu/memfd.h | 4 +++
> > util/memfd.c | 74
> > ++++++++++++++++++++++++++++++++++++++++++++++++++--
> > 2 files changed, 76 insertions(+), 2 deletions(-)
> >
> > diff --git a/include/qemu/memfd.h b/include/qemu/memfd.h
> > index 8b1fe6a..950fb88 100644
> > --- a/include/qemu/memfd.h
> > +++ b/include/qemu/memfd.h
> > @@ -17,4 +17,8 @@
> > #define F_SEAL_WRITE 0x0008 /* prevent writes */
> > #endif
> >
> > +void *qemu_memfd_alloc(const char *name, size_t size, unsigned int seals,
> > + int *fd);
> > +void qemu_memfd_free(void *ptr, size_t size, int fd);
> > +
> > #endif /* QEMU_MEMFD_H */
> > diff --git a/util/memfd.c b/util/memfd.c
> > index a98d57e..8b2b785 100644
> > --- a/util/memfd.c
> > +++ b/util/memfd.c
> > @@ -27,6 +27,14 @@
> >
> > #include "config-host.h"
> >
> > +#include <glib.h>
> > +#include <glib/gprintf.h>
> > +
> > +#include <stdio.h>
> > +#include <stdlib.h>
> > +#include <fcntl.h>
> > +#include <sys/mman.h>
> > +
> > #include "qemu/memfd.h"
> >
> > #ifdef CONFIG_MEMFD
> > @@ -44,13 +52,75 @@
> > #define MFD_ALLOW_SEALING 0x0002U
> > #endif
> >
> > -static inline int memfd_create(const char *name, unsigned int flags)
> > +static int memfd_create(const char *name, unsigned int flags)
> > {
> > return syscall(__NR_memfd_create, name, flags);
> > }
> > #else /* !LINUX */
> > -static inline int memfd_create(const char *name, unsigned int flags)
> > +static int memfd_create(const char *name, unsigned int flags)
> > {
> > return -1;
> > }
> > #endif
> > +
> > +void *qemu_memfd_alloc(const char *name, size_t size, unsigned int seals,
> > + int *fd)
> > +{
> > + void *ptr;
> > + int mfd;
> > +
> > + mfd = memfd_create(name, MFD_ALLOW_SEALING|MFD_CLOEXEC);
>
>
> Hmm. Does this interact correctly with the -mem-prealloc flag?
It's unrelated imho. It's helper here.
In the rest of the series, it's used at runtime when migrating with variable
size (today code doesn't prealloc that either)
>
> > + if (mfd != -1) {
> > + if (ftruncate(mfd, size) == -1) {
>
> Any limitations on size?
not that I know (reading memfd_create)
>
> > + perror("ftruncate");
> > + close(mfd);
> > + return NULL;
> > + }
> > +
> > + if (fcntl(mfd, F_ADD_SEALS, seals) == -1) {
> > + perror("fcntl");
> > + close(mfd);
> > + return NULL;
> > + }
> > + } else {
> > + const char *tmpdir = getenv("TMPDIR");
> > + gchar *fname;
> > +
> > + tmpdir = tmpdir ? tmpdir : "/tmp";
> > +
> > + fname = g_strdup_printf("%s/memfd-XXXXXX", tmpdir);
>
> This means there's now work to be done to set up selinux
> to allow QEMU creating memfd under /tmp.
doesn't sound unreasonable to me
>
> Maybe it's better to just fail gracefully for now.
it's a fallback, but sure we can remove it and add it back later if needed
> > + mfd = mkstemp(fname);
> > + unlink(fname);
> > + g_free(fname);
> > +
> > + if (mfd == -1) {
> > + perror("mkstemp");
> > + return NULL;
> > + }
> > +
> > + if (ftruncate(mfd, size) == -1) {
> > + perror("ftruncate");
> > + close(mfd);
> > + return NULL;
> > + }
> > + }
> > +
> > + ptr = mmap(0, size, PROT_READ|PROT_WRITE, MAP_SHARED, mfd, 0);
>
> Pls add space around | here and elsewhere.
>
ok
>
> > + if (ptr == MAP_FAILED) {
> > + perror("mmap");
> > + close(mfd);
> > + return NULL;
> > + }
> > +
> > + *fd = mfd;
> > + return ptr;
> > +}
> > +
> > +void qemu_memfd_free(void *ptr, size_t size, int fd)
> > +{
> > + if (ptr) {
> > + munmap(ptr, size);
> > + }
> > +
> > + close(fd);
>
> I notice you close fd unconditionally, but it's only returned
> on success above. So this will close an uninitialized one?
Ok, I'll add a -1 check