From: Matthew Wilcox <willy@infradead.org>
To: Arnd Bergmann <arnd@kernel.org>
Cc: kernel test robot <lkp@intel.com>, Linux-MM <linux-mm@kvack.org>,
kbuild-all@lists.01.org,
clang-built-linux <clang-built-linux@googlegroups.com>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Linux FS-devel Mailing List <linux-fsdevel@vger.kernel.org>,
Michael Ellerman <mpe@ellerman.id.au>,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
Paul Mackerras <paulus@samba.org>,
linuxppc-dev <linuxppc-dev@lists.ozlabs.org>,
Linux ARM <linux-arm-kernel@lists.infradead.org>,
Jesper Dangaard Brouer <brouer@redhat.com>,
"David S. Miller" <davem@davemloft.net>
Subject: Re: Bogus struct page layout on 32-bit
Date: Sun, 11 Apr 2021 23:35:37 +0100 [thread overview]
Message-ID: <20210411223537.GF2531743@casper.infradead.org> (raw)
In-Reply-To: <CAK8P3a3uEGaEN-p06vFP+jwbFt3P=Bx4=aRN+kUyB4PcFPxLRg@mail.gmail.com>
On Sat, Apr 10, 2021 at 09:10:47PM +0200, Arnd Bergmann wrote:
> On Sat, Apr 10, 2021 at 4:44 AM Matthew Wilcox <willy@infradead.org> wrote:
> > + dma_addr_t dma_addr __packed;
> > };
> > struct { /* slab, slob and slub */
> > union {
> >
> > but I don't know if GCC is smart enough to realise that dma_addr is now
> > on an 8 byte boundary and it can use a normal instruction to access it,
> > or whether it'll do something daft like use byte loads to access it.
> >
> > We could also do:
> >
> > + dma_addr_t dma_addr __packed __aligned(sizeof(void *));
> >
> > and I see pahole, at least sees this correctly:
> >
> > struct {
> > long unsigned int _page_pool_pad; /* 4 4 */
> > dma_addr_t dma_addr __attribute__((__aligned__(4))); /* 8 8 */
> > } __attribute__((__packed__)) __attribute__((__aligned__(4)));
> >
> > This presumably affects any 32-bit architecture with a 64-bit phys_addr_t
> > / dma_addr_t. Advice, please?
>
> I've tried out what gcc would make of this: https://godbolt.org/z/aTEbxxbG3
>
> struct page {
> short a;
> struct {
> short b;
> long long c __attribute__((packed, aligned(2)));
> } __attribute__((packed));
> } __attribute__((aligned(8)));
>
> In this structure, 'c' is clearly aligned to eight bytes, and gcc does
> realize that
> it is safe to use the 'ldrd' instruction for 32-bit arm, which is forbidden on
> struct members with less than 4 byte alignment. However, it also complains
> that passing a pointer to 'c' into a function that expects a 'long long' is not
> allowed because alignof(c) is only '2' here.
>
> (I used 'short' here because I having a 64-bit member misaligned by four
> bytes wouldn't make a difference to the instructions on Arm, or any other
> 32-bit architecture I can think of, regardless of the ABI requirements).
So ... we could do this:
+++ b/include/linux/types.h
@@ -140,7 +140,7 @@ typedef u64 blkcnt_t;
* so they don't care about the size of the actual bus addresses.
*/
#ifdef CONFIG_ARCH_DMA_ADDR_T_64BIT
-typedef u64 dma_addr_t;
+typedef u64 __attribute__((aligned(sizeof(void *)))) dma_addr_t;
#else
typedef u32 dma_addr_t;
#endif
but I'm a little scared that this might have unintended consequences.
And Jesper points out that a big-endian 64-bit dma_addr_t can impersonate
a PageTail page, and we should solve that problem while we're at it.
So I don't think we should do this, but thought I should mention it as
a possibility.
next prev parent reply other threads:[~2021-04-11 22:36 UTC|newest]
Thread overview: 47+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-04-09 18:50 [PATCH v7 00/28] Memory Folios Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 01/28] mm: Optimise nth_page for contiguous memmap Matthew Wilcox (Oracle)
2021-04-12 6:08 ` Christoph Hellwig
2021-04-09 18:50 ` [PATCH v7 02/28] mm: Introduce struct folio Matthew Wilcox (Oracle)
2021-04-09 22:45 ` kernel test robot
2021-04-10 2:43 ` Bogus struct page layout on 32-bit Matthew Wilcox
2021-04-10 6:21 ` Jesper Dangaard Brouer
2021-04-10 8:52 ` Ilias Apalodimas
2021-04-10 14:06 ` Matthew Wilcox
2021-04-10 15:54 ` Russell King - ARM Linux admin
2021-04-16 9:26 ` Grygorii Strashko
2021-04-16 14:10 ` Arnd Bergmann
2021-04-17 13:08 ` David Laight
2021-04-10 14:17 ` David Laight
2021-04-10 19:10 ` Arnd Bergmann
2021-04-11 22:35 ` Matthew Wilcox [this message]
2021-04-10 2:51 ` [PATCH v7 02/28] mm: Introduce struct folio kernel test robot
2021-04-16 15:55 ` Matthew Wilcox
2021-04-19 9:06 ` Kirill A. Shutemov
2021-04-09 18:50 ` [PATCH v7 03/28] mm: Add folio_pgdat and folio_zone Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 04/28] mm/vmstat: Add functions to account folio statistics Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 05/28] mm/debug: Add VM_BUG_ON_FOLIO and VM_WARN_ON_ONCE_FOLIO Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 06/28] mm: Add folio reference count functions Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 07/28] mm: Add put_folio Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 08/28] mm: Add get_folio Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 09/28] mm: Create FolioFlags Matthew Wilcox (Oracle)
2021-04-19 13:25 ` Peter Zijlstra
2021-04-19 13:55 ` Matthew Wilcox
2021-04-09 18:50 ` [PATCH v7 10/28] mm: Handle per-folio private data Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 11/28] mm/filemap: Add folio_index, folio_file_page and folio_contains Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 12/28] mm/filemap: Add folio_next_index Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 13/28] mm/filemap: Add folio_offset and folio_file_offset Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 14/28] mm/util: Add folio_mapping and folio_file_mapping Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 15/28] mm: Add folio_mapcount Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 16/28] mm/memcg: Add folio wrappers for various functions Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 17/28] mm/filemap: Add unlock_folio Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 18/28] mm/filemap: Add lock_folio Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 19/28] mm/filemap: Add lock_folio_killable Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 20/28] mm/filemap: Add __lock_folio_async Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 21/28] mm/filemap: Add __lock_folio_or_retry Matthew Wilcox (Oracle)
2021-04-09 18:50 ` [PATCH v7 22/28] mm/filemap: Add wait_on_folio_locked Matthew Wilcox (Oracle)
2021-04-09 18:51 ` [PATCH v7 23/28] mm/filemap: Add end_folio_writeback Matthew Wilcox (Oracle)
2021-04-09 18:51 ` [PATCH v7 24/28] mm/writeback: Add wait_on_folio_writeback Matthew Wilcox (Oracle)
2021-04-09 18:51 ` [PATCH v7 25/28] mm/writeback: Add wait_for_stable_folio Matthew Wilcox (Oracle)
2021-04-09 18:51 ` [PATCH v7 26/28] mm/filemap: Convert wait_on_page_bit to wait_on_folio_bit Matthew Wilcox (Oracle)
2021-04-09 18:51 ` [PATCH v7 27/28] mm/filemap: Convert wake_up_page_bit to wake_up_folio_bit Matthew Wilcox (Oracle)
2021-04-09 18:51 ` [PATCH v7 28/28] mm/filemap: Convert page wait queues to be folios Matthew Wilcox (Oracle)
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=20210411223537.GF2531743@casper.infradead.org \
--to=willy@infradead.org \
--cc=arnd@kernel.org \
--cc=benh@kernel.crashing.org \
--cc=brouer@redhat.com \
--cc=clang-built-linux@googlegroups.com \
--cc=davem@davemloft.net \
--cc=kbuild-all@lists.01.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=lkp@intel.com \
--cc=mpe@ellerman.id.au \
--cc=paulus@samba.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).