samba - Feb 2020 - vfs fruit disk_free fails on tmsize overflow with macOS Time Machine

Thanks torch ? I tried those suggestions but nothing changed. 

I did however manage to find an existing bug which matches my symptoms:
https://bugzilla.samba.org/show_bug.cgi?id=13622

Fortunately it mentions a fix, but that was never implemented, so I created a
new branch from v4-11-test on the gitlab shadow copy and tried the fix there.
After compiling, my tests on Raspbian show that this does indeed fix the issue
for me.

The diff for the patch is

--- source3/modules/vfs_fruit.c 2019-10-24 15:15:00.000000000 +0000
+++ source3/modules/vfs_fruit.c 2020-02-18 19:54:00.000000000 +0000
@@ -6971,18 +6971,7 @@
return true;
}

- if (bandsize > SIZE_MAX/nbands) {
- DBG_ERR("tmsize overflow: bandsize [%zu] nbands [%zu]\n",
- bandsize, nbands);
- return false;
- }
- tm_size = bandsize * nbands;
-
- if (state->total_size + tm_size < state->total_size) {
- DBG_ERR("tmsize overflow: bandsize [%zu] nbands [%zu]\n",
- bandsize, nbands);
- return false;
- }
+ tm_size = (off_t)bandsize * (off_t)nbands;

state->total_size += tm_size;


and I have the branch ready at
https://gitlab.com/artmg/samba/-/blob/artmg-tmsize-overflow-fix/
Should I just raise a PR from this, or do I need to go through some additional
process?

I welcome advice on contributing
Thanks

On Mon, Feb 24, 2020 at 10:26:38AM +0000, Art MG via samba
wrote:
> Thanks torch ? I tried those suggestions but nothing changed. 
> 
> I did however manage to find an existing bug which matches my symptoms:
https://bugzilla.samba.org/show_bug.cgi?id=13622
> 
> Fortunately it mentions a fix, but that was never implemented, so I created
a new branch from v4-11-test on the gitlab shadow copy and tried the fix there.
After compiling, my tests on Raspbian show that this does indeed fix the issue
for me.
> 
> The diff for the patch is
> 
> --- source3/modules/vfs_fruit.c 2019-10-24 15:15:00.000000000 +0000
> +++ source3/modules/vfs_fruit.c 2020-02-18 19:54:00.000000000 +0000
> @@ -6971,18 +6971,7 @@
> return true;
> }
> 
> - if (bandsize > SIZE_MAX/nbands) {
> - DBG_ERR("tmsize overflow: bandsize [%zu] nbands [%zu]\n",
> - bandsize, nbands);
> - return false;
> - }
> - tm_size = bandsize * nbands;
> -
> - if (state->total_size + tm_size < state->total_size) {
> - DBG_ERR("tmsize overflow: bandsize [%zu] nbands [%zu]\n",
> - bandsize, nbands);
> - return false;
> - }
> + tm_size = (off_t)bandsize * (off_t)nbands;
> 
> state->total_size += tm_size;
> 
> 
> and I have the branch ready at
https://gitlab.com/artmg/samba/-/blob/artmg-tmsize-overflow-fix/
> Should I just raise a PR from this, or do I need to go through some
additional process?
Raise a PR from this at gitlab please.

Having said that, your fix looks like it's
removing a bunch of overflow safety checks,
so I'm not sure it's the right one.