Messages in this thread | | | Date | Wed, 3 Apr 2019 15:36:48 +0300 | From | Andy Shevchenko <> | Subject | Re: [PATCH 1/5] lib: make bitmap_parselist_user() a wrapper on bitmap_parselist() |
| |
On Wed, Apr 03, 2019 at 02:17:39PM +0200, Rasmus Villemoes wrote: > On 03/04/2019 13.17, Andy Shevchenko wrote: > > On Wed, Apr 03, 2019 at 02:15:22PM +0300, Andy Shevchenko wrote: > >> On Wed, Apr 03, 2019 at 07:45:36AM +0300, Yury Norov wrote: > >>> Currently we parse user data byte after byte which leads to > >>> overcomplification of parsing algorithm. The only user of > >>> bitmap_parselist_user() is not performance-critical, and so we > >>> can copy user data to kernel buffer and simply call > >>> bitmap_parselist(). This rework lets us unify and simplify > >>> bitmap_parselist() and bitmap_parselist_user(), which is done > >>> in the following patch. > >> > >>> + buf = kmalloc(ulen + 1, GFP_KERNEL); > >>> + if (!buf) > >>> + return -ENOMEM; > >>> + > >>> + buf[ulen] = 0; > >>> + > >>> + ret = copy_from_user(buf, ubuf, ulen); > >>> + if (ret) > >>> + goto out; > >> > >> Why not memdup_user() ? > > > > Even more precisely (for strings) strndup_user(). > > > > But the user buffer is not nul-terminated, i.e. it's not a string.
Missed that. I stand corrected.
> What > you want is memdup_user_nul() - take a length-delimited user buffer and > turn it into a nul-terminated string in kernel memory.
Good, we have a helper.
-- With Best Regards, Andy Shevchenko
| |