| Message ID | 1401839519-2153-1-git-send-email-peter.maydell@linaro.org |
|---|---|
| State | Superseded |
| Headers | show |
On Wed, 2014-06-04 at 00:51 +0100, Peter Maydell wrote: > Trying to override the implementations of g_malloc and g_free is > a really bad idea -- it means statically linked builds fail to > link (because of the multiple definitions provided by this file > and by glib), and non-statically linked builds segfault as soon > as they try to do anything more complicated than printing the > usage message. Remove these overridden versions and just use > the glib ones. > > This is sufficient that bsd-user can run basic x86-64 > binaries on OpenBSD again; FreeBSD and NetBSD seem to have > further issues. > > Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed by: Sean Bruno sbruno@freebsd.org > --- > The FreeBSD QEMU bsd-user tree took this same approach. > > Blue, I'm planning to put this together with a few other bsd-user > patches (and further submissions from the FreeBSD folk as they > get them into shape) into a 'bsd-user.next' queue. Let me > know if you'd rather do that yourself (as the listed bsd-user > maintainer). My current queue is at: > https://git.linaro.org/people/peter.maydell/qemu-arm.git/shortlog/refs/heads/bsd-user.next > (though not all of those have been through review yet). > > bsd-user/mmap.c | 60 --------------------------------------------------------- > 1 file changed, 60 deletions(-) > > diff --git a/bsd-user/mmap.c b/bsd-user/mmap.c > index aae8ea1..092bf7f 100644 > --- a/bsd-user/mmap.c > +++ b/bsd-user/mmap.c > @@ -74,66 +74,6 @@ void mmap_unlock(void) > } > #endif > > -static void *bsd_vmalloc(size_t size) > -{ > - void *p; > - mmap_lock(); > - /* Use map and mark the pages as used. */ > - p = mmap(NULL, size, PROT_READ | PROT_WRITE, > - MAP_PRIVATE | MAP_ANON, -1, 0); > - > - if (h2g_valid(p)) { > - /* Allocated region overlaps guest address space. > - This may recurse. */ > - abi_ulong addr = h2g(p); > - page_set_flags(addr & TARGET_PAGE_MASK, TARGET_PAGE_ALIGN(addr + size), > - PAGE_RESERVED); > - } > - > - mmap_unlock(); > - return p; > -} > - > -void *g_malloc(size_t size) > -{ > - char * p; > - size += 16; > - p = bsd_vmalloc(size); > - *(size_t *)p = size; > - return p + 16; > -} > - > -/* We use map, which is always zero initialized. */ > -void * g_malloc0(size_t size) > -{ > - return g_malloc(size); > -} > - > -void g_free(void *ptr) > -{ > - /* FIXME: We should unmark the reserved pages here. However this gets > - complicated when one target page spans multiple host pages, so we > - don't bother. */ > - size_t *p; > - p = (size_t *)((char *)ptr - 16); > - munmap(p, *p); > -} > - > -void *g_realloc(void *ptr, size_t size) > -{ > - size_t old_size, copy; > - void *new_ptr; > - > - if (!ptr) > - return g_malloc(size); > - old_size = *(size_t *)((char *)ptr - 16); > - copy = old_size < size ? old_size : size; > - new_ptr = g_malloc(size); > - memcpy(new_ptr, ptr, copy); > - g_free(ptr); > - return new_ptr; > -} > - > /* NOTE: all the constants are the HOST ones, but addresses are target. */ > int target_mprotect(abi_ulong start, abi_ulong len, int prot) > {
On 3 June 2014 19:51, Peter Maydell <peter.maydell@linaro.org> wrote: > Trying to override the implementations of g_malloc and g_free is > a really bad idea -- it means statically linked builds fail to > link (because of the multiple definitions provided by this file > and by glib), and non-statically linked builds segfault as soon > as they try to do anything more complicated than printing the > usage message. Remove these overridden versions and just use > the glib ones. > > This is sufficient that bsd-user can run basic x86-64 > binaries on OpenBSD again; FreeBSD and NetBSD seem to have > further issues. > > Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Ed Maste <emaste@freebsd.org>
diff --git a/bsd-user/mmap.c b/bsd-user/mmap.c index aae8ea1..092bf7f 100644 --- a/bsd-user/mmap.c +++ b/bsd-user/mmap.c @@ -74,66 +74,6 @@ void mmap_unlock(void) } #endif -static void *bsd_vmalloc(size_t size) -{ - void *p; - mmap_lock(); - /* Use map and mark the pages as used. */ - p = mmap(NULL, size, PROT_READ | PROT_WRITE, - MAP_PRIVATE | MAP_ANON, -1, 0); - - if (h2g_valid(p)) { - /* Allocated region overlaps guest address space. - This may recurse. */ - abi_ulong addr = h2g(p); - page_set_flags(addr & TARGET_PAGE_MASK, TARGET_PAGE_ALIGN(addr + size), - PAGE_RESERVED); - } - - mmap_unlock(); - return p; -} - -void *g_malloc(size_t size) -{ - char * p; - size += 16; - p = bsd_vmalloc(size); - *(size_t *)p = size; - return p + 16; -} - -/* We use map, which is always zero initialized. */ -void * g_malloc0(size_t size) -{ - return g_malloc(size); -} - -void g_free(void *ptr) -{ - /* FIXME: We should unmark the reserved pages here. However this gets - complicated when one target page spans multiple host pages, so we - don't bother. */ - size_t *p; - p = (size_t *)((char *)ptr - 16); - munmap(p, *p); -} - -void *g_realloc(void *ptr, size_t size) -{ - size_t old_size, copy; - void *new_ptr; - - if (!ptr) - return g_malloc(size); - old_size = *(size_t *)((char *)ptr - 16); - copy = old_size < size ? old_size : size; - new_ptr = g_malloc(size); - memcpy(new_ptr, ptr, copy); - g_free(ptr); - return new_ptr; -} - /* NOTE: all the constants are the HOST ones, but addresses are target. */ int target_mprotect(abi_ulong start, abi_ulong len, int prot) {
Trying to override the implementations of g_malloc and g_free is a really bad idea -- it means statically linked builds fail to link (because of the multiple definitions provided by this file and by glib), and non-statically linked builds segfault as soon as they try to do anything more complicated than printing the usage message. Remove these overridden versions and just use the glib ones. This is sufficient that bsd-user can run basic x86-64 binaries on OpenBSD again; FreeBSD and NetBSD seem to have further issues. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- The FreeBSD QEMU bsd-user tree took this same approach. Blue, I'm planning to put this together with a few other bsd-user patches (and further submissions from the FreeBSD folk as they get them into shape) into a 'bsd-user.next' queue. Let me know if you'd rather do that yourself (as the listed bsd-user maintainer). My current queue is at: https://git.linaro.org/people/peter.maydell/qemu-arm.git/shortlog/refs/heads/bsd-user.next (though not all of those have been through review yet). bsd-user/mmap.c | 60 --------------------------------------------------------- 1 file changed, 60 deletions(-)