linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Linus Torvalds <torvalds@linux-foundation.org>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: agordeev@linux.ibm.com, Linux-MM <linux-mm@kvack.org>,
	mm-commits@vger.kernel.org
Subject: Re: [patch 24/54] mm/mmap.c: do not allow mappings outside of allowed limits
Date: Mon, 8 Jun 2020 10:50:59 -0700	[thread overview]
Message-ID: <CAHk-=wg2W8mFkGRKaT__HVaUOwED+EjdiFxV10mauYVag9J_hQ@mail.gmail.com> (raw)
In-Reply-To: <20200608044121.stuEeVRhM%akpm@linux-foundation.org>

On Sun, Jun 7, 2020 at 9:41 PM Andrew Morton <akpm@linux-foundation.org> wrote:
>
> One can set a lowest possible address in /proc/sys/vm/mmap_min_addr
> and mmap below that bound nevertheless.

Well, /proc/sys/vm/mmap_min_addr actually sets "dac_mmap_min_addr"
directly, and then indirectly sets mmap_min_addr with slightly
different rules.

> It is possible to request a fixed mapping address below mmap_min_addr and
> succeed.  This update adds early checks of mmap_min_addr and mmap_end
> boundaries and fixes the above issue.
>
> Apart from it is wrong I am not aware of any existing issue.

Hmm. I think your patch is generally fine, although a few things worries me:

 - the "mmap_end" check seems wrong. It should be something like

        if (len > mmap_end || addr > mmap_end-len)

   shouldn't it?

 - the limit we _do_ test is that "dac_mmap_min_addr", and we allow
lower limits for CAP_SYS_RAWIO

 - I think this will break vm86 mode on 32-bit x86. Have you tested that?

In other words, I think the reason we don't do that hard check of
mmap_min_addr is exactly that vm86 issue. If we were to force a hard
check, we're now making it impossible to use vm86 mode.

So I'm dropping this patch, because it is not clear that this was
fully thought through.

And if it *was* intentional, and people knew about the vm86 issues,
and the thing about dac_mmap_min_addr and the check in
cap_mmap_addr(), then it should be mentioned in the commit message.

That said, I'd be more than willing to move the "cap_mmap_addr()"
check into the mm layer and make it a whole lot more obvious.

And I'd also be more than willing to really make the difference
between "dac_mmap_min_addr" and "mmap_min_addr" actually meaningful,
and really enforce "mmap_min_addr", because right now it's a
half-baked feature that isn't actually used for anything but hinting
and the grow-down issue.

And that, in turn, is because of those vm86 issues, but also because
we've historically had programs that wanted some legacy behavior and
did a mmap() of a zero-page at the 0 address (because that's how some
old Unix environments worked - I htink legacy parisc or similar).

            Linus


  reply	other threads:[~2020-06-08 17:51 UTC|newest]

Thread overview: 57+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20200607212615.b050e41fac139a1e16fe00bd@linux-foundation.org>
2020-06-08  4:40 ` [patch 01/54] mm/page_idle.c: skip offline pages Andrew Morton
2020-06-08  4:40 ` [patch 02/54] ipc/msg: add missing annotation for freeque() Andrew Morton
2020-06-08  4:40 ` [patch 03/54] ipc/namespace.c: use a work queue to free_ipc Andrew Morton
2020-06-08  4:40 ` [patch 04/54] dynamic_debug: add an option to enable dynamic debug for modules only Andrew Morton
2020-06-08  4:58   ` 答复: " 翟京 (Orson Zhai)
2020-06-08  4:40 ` [patch 05/54] kernel: add panic_on_taint Andrew Morton
2020-06-08  4:40 ` [patch 06/54] xarray.h: correct return code documentation for xa_store_{bh,irq}() Andrew Morton
2020-06-08  4:40 ` [patch 07/54] kernel/sysctl: support setting sysctl parameters from kernel command line Andrew Morton
2020-06-08  4:40 ` [patch 08/54] kernel/sysctl: support handling command line aliases Andrew Morton
2020-06-08  4:40 ` [patch 09/54] kernel/hung_task convert hung_task_panic boot parameter to sysctl Andrew Morton
2020-06-08  4:40 ` [patch 10/54] tools/testing/selftests/sysctl/sysctl.sh: support CONFIG_TEST_SYSCTL=y Andrew Morton
2020-06-08  4:40 ` [patch 11/54] lib/test_sysctl: support testing of sysctl. boot parameter Andrew Morton
2020-06-08  4:40 ` [patch 12/54] kernel/watchdog.c: convert {soft/hard}lockup boot parameters to sysctl aliases Andrew Morton
2020-06-08  4:40 ` [patch 13/54] kernel/hung_task.c: introduce sysctl to print all traces when a hung task is detected Andrew Morton
2020-06-08  4:40 ` [patch 14/54] panic: add sysctl to dump all CPUs backtraces on oops event Andrew Morton
2020-06-08  4:40 ` [patch 15/54] kernel/sysctl.c: ignore out-of-range taint bits introduced via kernel.tainted Andrew Morton
2020-06-08  4:40 ` [patch 16/54] mm/gup.c: convert to use get_user_{page|pages}_fast_only() Andrew Morton
2020-06-08  4:40 ` [patch 17/54] mm/gup: update pin_user_pages.rst for "case 3" (mmu notifiers) Andrew Morton
2020-06-08  4:41 ` [patch 18/54] mm/gup: introduce pin_user_pages_locked() Andrew Morton
2020-06-08  4:41 ` [patch 19/54] mm/gup: frame_vector: convert get_user_pages() --> pin_user_pages() Andrew Morton
2020-06-08  4:41 ` [patch 20/54] mm/gup: documentation fix for pin_user_pages*() APIs Andrew Morton
2020-06-08  4:41 ` [patch 21/54] docs: mm/gup: pin_user_pages.rst: add a "case 5" Andrew Morton
2020-06-08  4:41 ` [patch 22/54] vhost: convert get_user_pages() --> pin_user_pages() Andrew Morton
2020-06-08  4:41 ` [patch 23/54] mm/mmap.c: add more sanity checks to get_unmapped_area() Andrew Morton
2020-06-08  4:41 ` [patch 24/54] mm/mmap.c: do not allow mappings outside of allowed limits Andrew Morton
2020-06-08 17:50   ` Linus Torvalds [this message]
2020-06-08 17:55     ` Linus Torvalds
2020-06-08  4:41 ` [patch 25/54] arm: fix the flush_icache_range arguments in set_fiq_handler Andrew Morton
2020-06-08  4:41 ` [patch 26/54] nds32: unexport flush_icache_page Andrew Morton
2020-06-08  4:41 ` [patch 27/54] powerpc: unexport flush_icache_user_range Andrew Morton
2020-06-08  4:41 ` [patch 28/54] unicore32: remove flush_cache_user_range Andrew Morton
2020-06-08  4:41 ` [patch 29/54] asm-generic: fix the inclusion guards for cacheflush.h Andrew Morton
2020-06-08  4:41 ` [patch 30/54] asm-generic: don't include <linux/mm.h> in cacheflush.h Andrew Morton
2020-06-08  4:41 ` [patch 31/54] asm-generic: improve the flush_dcache_page stub Andrew Morton
2020-06-08  4:41 ` [patch 32/54] alpha: use asm-generic/cacheflush.h Andrew Morton
2020-06-08  4:41 ` [patch 33/54] arm64: " Andrew Morton
2020-06-08  4:41 ` [patch 34/54] c6x: " Andrew Morton
2020-06-08  4:41 ` [patch 35/54] hexagon: " Andrew Morton
2020-06-08  4:42 ` [patch 36/54] ia64: " Andrew Morton
2020-06-08  4:42 ` [patch 37/54] microblaze: " Andrew Morton
2020-06-08  4:42 ` [patch 38/54] m68knommu: " Andrew Morton
2020-06-08  4:42 ` [patch 39/54] openrisc: " Andrew Morton
2020-06-08  4:42 ` [patch 40/54] powerpc: " Andrew Morton
2020-06-08  4:42 ` [patch 41/54] riscv: " Andrew Morton
2020-06-08  4:42 ` [patch 42/54] arm,sparc,unicore32: remove flush_icache_user_range Andrew Morton
2020-06-08  4:42 ` [patch 43/54] mm: rename flush_icache_user_range to flush_icache_user_page Andrew Morton
2020-06-08  4:42 ` [patch 44/54] asm-generic: add a flush_icache_user_range stub Andrew Morton
2020-06-08  4:42 ` [patch 45/54] sh: implement flush_icache_user_range Andrew Morton
2020-06-08  4:42 ` [patch 46/54] xtensa: " Andrew Morton
2020-06-08  4:42 ` [patch 47/54] arm: rename flush_cache_user_range to flush_icache_user_range Andrew Morton
2020-06-08  4:42 ` [patch 48/54] m68k: implement flush_icache_user_range Andrew Morton
2020-06-08  4:42 ` [patch 49/54] exec: only build read_code when needed Andrew Morton
2020-06-08  4:42 ` [patch 50/54] exec: use flush_icache_user_range in read_code Andrew Morton
2020-06-08  4:42 ` [patch 51/54] binfmt_flat: use flush_icache_user_range Andrew Morton
2020-06-08  4:42 ` [patch 52/54] nommu: use flush_icache_user_range in brk and mmap Andrew Morton
2020-06-08  4:42 ` [patch 53/54] module: move the set_fs hack for flush_icache_range to m68k Andrew Morton
2020-06-08  4:42 ` [patch 54/54] doc: cgroup: update note about conditions when oom killer is invoked Andrew Morton

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CAHk-=wg2W8mFkGRKaT__HVaUOwED+EjdiFxV10mauYVag9J_hQ@mail.gmail.com' \
    --to=torvalds@linux-foundation.org \
    --cc=agordeev@linux.ibm.com \
    --cc=akpm@linux-foundation.org \
    --cc=linux-mm@kvack.org \
    --cc=mm-commits@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).