Re: [PATCH v3 7/9] bsd-user/mmap.c: Don't mmap fd == -1 independently from MAP_ANON flag

[Kyle Evans]

On Fri, Oct 8, 2021 at 4:29 PM Warner Losh <imp@bsdimp.com> wrote:
>
> From: Guy Yur <guyyur@gmail.com>
>
> Switch checks for !(flags & MAP_ANONYMOUS) with checks for fd != -1.
> MAP_STACK and MAP_GUARD both require fd == -1 and don't require mapping
> the fd either. Add analysis from Guy Yur detailing the different cases
> for MAP_GUARD and MAP_STACK.
>
> Signed-off-by: Guy Yur <guyyur@gmail.com>
> [ partially merged before, finishing the job and documenting origin]
> Signed-off-by: Warner Losh <imp@bsdimp.com>
> ---
>  bsd-user/mmap.c | 30 +++++++++++++++++++++++++-----
>  1 file changed, 25 insertions(+), 5 deletions(-)
>
> diff --git a/bsd-user/mmap.c b/bsd-user/mmap.c
> index face98573f..4ecd949a10 100644
> --- a/bsd-user/mmap.c
> +++ b/bsd-user/mmap.c
> @@ -127,7 +127,27 @@ error:
>      return ret;
>  }
>
> -/* map an incomplete host page */
> +/*
> + * map an incomplete host page
> + *
> + * mmap_frag can be called with a valid fd, if flags doesn't contain one of
> + * MAP_ANON, MAP_STACK, MAP_GUARD. If we need to map a page in those cases, 
> we
> + * pass fd == -1. However, if flags contains MAP_GUARD then MAP_ANON cannot 
> be
> + * added.
> + *
> + * * If fd is valid (not -1) we want to map the pages with MAP_ANON.
> + * * If flags contains MAP_GUARD we don't want to add MAP_ANON because it
> + *   will be rejected.  See kern_mmap's enforcing of constraints for 
> MAP_GUARD
> + *   in sys/vm/vm_mmap.c.
> + * * If flags contains MAP_ANON it doesn't matter if we add it or not.
> + * * If flags contains MAP_STACK, mmap adds MAP_ANON when called so doesn't
> + *   matter if we add it or not either. See enforcing of constraints for
> + *   MAP_STACK in kern_mmap.
> + *
> + * Don't add MAP_ANON for the flags that use fd == -1 without specifying the
> + * flags directly, with the assumption that future flags that require fd == 
> -1
> + * will also not require MAP_ANON.
> + */
>  static int mmap_frag(abi_ulong real_start,
>                       abi_ulong start, abi_ulong end,
>                       int prot, int flags, int fd, abi_ulong offset)
> @@ -147,9 +167,9 @@ static int mmap_frag(abi_ulong real_start,
>      }
>
>      if (prot1 == 0) {
> -        /* no page was there, so we allocate one */
> +        /* no page was there, so we allocate one. See also above. */
>          void *p = mmap(host_start, qemu_host_page_size, prot,
> -                       flags | MAP_ANON, -1, 0);
> +                       flags | ((fd != -1) ? MAP_ANON : 0), -1, 0);
>          if (p == MAP_FAILED)
>              return -1;
>          prot1 = prot;
> @@ -157,7 +177,7 @@ static int mmap_frag(abi_ulong real_start,
>      prot1 &= PAGE_BITS;
>
>      prot_new = prot | prot1;
> -    if (!(flags & MAP_ANON)) {
> +    if (fd != -1) {
>          /* msync() won't work here, so we return an error if write is
>             possible while it is a shared mapping */
>          if ((flags & TARGET_BSD_MAP_FLAGMASK) == MAP_SHARED &&
> @@ -565,7 +585,7 @@ abi_long target_mmap(abi_ulong start, abi_ulong len, int 
> prot,
>           * worst case: we cannot map the file because the offset is not
>           * aligned, so we read it
>           */
> -        if (!(flags & MAP_ANON) &&
> +        if (fd != -1 &&
>              (offset & ~qemu_host_page_mask) != (start & 
> ~qemu_host_page_mask)) {
>              /*
>               * msync() won't work here, so we return an error if write is
> --
> 2.32.0
>

Reviewed-by: Kyle Evans <kevans@FreeBSD.org>