DPDK patches and discussions
 help / color / mirror / Atom feed
From: Bruce Richardson <bruce.richardson@intel.com>
To: David Marchand <david.marchand@redhat.com>
Cc: <dev@dpdk.org>, Anatoly Burakov <anatoly.burakov@intel.com>,
	"Tyler Retzlaff" <roretzla@linux.microsoft.com>
Subject: Re: [PATCH v3 12/18] malloc: fix mp message alignment
Date: Tue, 8 Jul 2025 14:25:41 +0100	[thread overview]
Message-ID: <aG0cVcB55BF2mEDd@bricha3-mobl1.ger.corp.intel.com> (raw)
In-Reply-To: <CAJFAV8wj=EmaJnOVj4n8eVUR+BOAYhu38cJDu29Um4TkxFS9Tg@mail.gmail.com>

On Tue, Jul 08, 2025 at 02:46:04PM +0200, David Marchand wrote:
> On Tue, Jul 8, 2025 at 2:44 PM Bruce Richardson
> <bruce.richardson@intel.com> wrote:
> >
> > On Tue, Jul 08, 2025 at 02:28:16PM +0200, David Marchand wrote:
> > > Content (param[]) of received multiprocess messages are aligned with
> > > a 4 bytes constraint.
> > >
> > > Before patch:
> > > struct mp_msg_internal {
> > >  int type;                                                     /*   0     4 */
> > >  struct rte_mp_msg {
> > >   char name[64];                                               /*   4    64 */
> > >   /* --- cacheline 1 boundary (64 bytes) was 4 bytes ago --- */
> > >   int len_param;                                               /*  68     4 */
> > >   int num_fds;                                                 /*  72     4 */
> > >   /* typedef uint8_t -> __uint8_t */ unsigned char param[256]; /*  76   256 */
> > >   /* --- cacheline 5 boundary (320 bytes) was 12 bytes ago --- */
> > >   int fds[253];                                                /* 332  1012 */
> > >  } msg;                                                        /*   4  1340 */
> > >
> > >  /* size: 1344, cachelines: 21, members: 2 */
> > > };
> > >
> > > This results in many unaligned accesses for multiprocess malloc requests.
> > >
> > > Examples:
> > > ../lib/eal/common/malloc_mp.c:308:32: runtime error:
> > >       member access within misaligned address 0x7f7b35df4684 for type
> > >       'const struct malloc_mp_req', which requires 8 byte alignment
> > >
> > > ../lib/eal/common/malloc_mp.c:158:9: runtime error:
> > >       member access within misaligned address 0x7f36a535bb5c for type
> > >       'const struct malloc_mp_req', which requires 8 byte alignment
> > >
> > > ../lib/eal/common/malloc_mp.c:171:8: runtime error:
> > >       member access within misaligned address 0x7f4ba65f296c for type
> > >       'struct malloc_mp_req', which requires 8 byte alignment
> > >
> > > Signed-off-by: David Marchand <david.marchand@redhat.com>
> > > ---
> > >  lib/eal/common/eal_common_proc.c |  2 +-
> > >  lib/eal/common/malloc_mp.c       | 18 +++++++++---------
> > >  2 files changed, 10 insertions(+), 10 deletions(-)
> > >
> > > diff --git a/lib/eal/common/eal_common_proc.c b/lib/eal/common/eal_common_proc.c
> > > index 0dea787e38..3846c7178d 100644
> > > --- a/lib/eal/common/eal_common_proc.c
> > > +++ b/lib/eal/common/eal_common_proc.c
> > > @@ -62,7 +62,7 @@ enum mp_type {
> > >
> > >  struct mp_msg_internal {
> > >       int type;
> > > -     struct rte_mp_msg msg;
> > > +     alignas(8) struct rte_mp_msg msg;
> > >  };
> > >
> > >  struct async_request_param {
> > > diff --git a/lib/eal/common/malloc_mp.c b/lib/eal/common/malloc_mp.c
> > > index 9765277f5d..000c7f6b47 100644
> > > --- a/lib/eal/common/malloc_mp.c
> > > +++ b/lib/eal/common/malloc_mp.c
> > > @@ -148,7 +148,7 @@ get_unique_id(void)
> > >  static int
> > >  handle_sync(const struct rte_mp_msg *msg, const void *peer)
> > >  {
> > > -     struct rte_mp_msg reply;
> > > +     alignas(8) struct rte_mp_msg reply;
> > >       const struct malloc_mp_req *req =
> > >                       (const struct malloc_mp_req *)msg->param;
> >
> > This patch seems to have a lot of these definitions with alignas added to
> > them. Would it be simpler just to put the alignas inside the rte_mp_msg
> > definition?
> >
> > More specifically, if its the "uint8_t param" element that needs alignment,
> > how about changing that specific field to make it aligned?
> 
> We could probably enhance this, but I expect this breaks ABI.
> 
Is the multi-process message type part of the public ABI? I don't believe
we ever guaranteed multiprocess support working across versions of DPDK.

/Bruce

  reply	other threads:[~2025-07-08 13:26 UTC|newest]

Thread overview: 81+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-19  7:10 [PATCH 00/10] Run with UBSan in GHA David Marchand
2025-06-19  7:10 ` [PATCH 01/10] ci: save ccache on failure David Marchand
2025-06-25 12:16   ` Aaron Conole
2025-06-19  7:10 ` [PATCH 02/10] test/telemetry: fix test calling all commands David Marchand
2025-06-20  9:16   ` Bruce Richardson
2025-06-23  9:54   ` David Marchand
2025-06-19  7:10 ` [PATCH 03/10] test/mempool: fix test without stack driver David Marchand
2025-06-20  8:54   ` Andrew Rybchenko
2025-06-19  7:10 ` [PATCH 04/10] eal: fix plugin dir walk David Marchand
2025-06-20  9:19   ` Bruce Richardson
2025-06-23  9:41     ` David Marchand
2025-06-19  7:10 ` [PATCH 05/10] cmdline: fix port list parsing David Marchand
2025-06-20  9:58   ` Bruce Richardson
2025-06-23  9:40     ` David Marchand
2025-06-23 10:41       ` Bruce Richardson
2025-06-19  7:10 ` [PATCH 06/10] cmdline: fix highest bit " David Marchand
2025-06-20  9:21   ` Bruce Richardson
2025-06-23  9:32     ` David Marchand
2025-06-19  7:10 ` [PATCH 07/10] tailq: fix cast macro for null pointer David Marchand
2025-06-20  9:23   ` Bruce Richardson
2025-06-19  7:10 ` [PATCH 08/10] hash: fix unaligned access in predictable RSS David Marchand
2025-06-19  7:10 ` [PATCH 09/10] stack: fix unaligned accesses on 128-bit David Marchand
2025-06-19  7:10 ` [PATCH 10/10] build: support Undefined Behavior Sanitizer David Marchand
2025-06-25 12:17   ` Aaron Conole
2025-06-23 13:52 ` [PATCH v2 00/10] Run with UBSan in GHA David Marchand
2025-06-23 13:52   ` [PATCH v2 01/10] ci: save ccache on failure David Marchand
2025-06-23 13:52   ` [PATCH v2 02/10] test/telemetry: fix test calling all commands David Marchand
2025-06-24 15:59     ` Marat Khalili
2025-06-26  8:32       ` David Marchand
2025-06-26  9:51         ` Marat Khalili
2025-07-03 14:09           ` David Marchand
2025-07-03 15:08             ` Marat Khalili
2025-06-23 13:52   ` [PATCH v2 03/10] test/mempool: fix test without stack driver David Marchand
2025-06-24 16:21     ` Marat Khalili
2025-06-23 13:52   ` [PATCH v2 04/10] eal: fix plugin dir walk David Marchand
2025-06-25  8:43     ` Marat Khalili
2025-07-03 14:27       ` David Marchand
2025-06-23 13:52   ` [PATCH v2 05/10] cmdline: fix port list parsing David Marchand
2025-06-23 14:00     ` Bruce Richardson
2025-06-26  9:32     ` Marat Khalili
2025-06-23 13:52   ` [PATCH v2 06/10] cmdline: fix highest bit " David Marchand
2025-06-30 15:25     ` Marat Khalili
2025-06-23 13:52   ` [PATCH v2 07/10] tailq: fix cast macro for null pointer David Marchand
2025-06-30 16:06     ` Marat Khalili
2025-06-23 13:52   ` [PATCH v2 08/10] hash: fix unaligned access in predictable RSS David Marchand
2025-06-30 15:32     ` Bruce Richardson
2025-07-01  8:36     ` Konstantin Ananyev
2025-07-08  7:32       ` David Marchand
2025-06-23 13:52   ` [PATCH v2 09/10] stack: fix unaligned accesses on 128-bit David Marchand
2025-06-30 15:33     ` Bruce Richardson
2025-06-23 13:52   ` [PATCH v2 10/10] build: support Undefined Behavior Sanitizer David Marchand
2025-07-08 12:28 ` [PATCH v3 00/18] Run with UBSan in GHA David Marchand
2025-07-08 12:28   ` [PATCH v3 01/18] ci: save ccache on failure David Marchand
2025-07-08 12:28   ` [PATCH v3 02/18] test/telemetry: fix test calling all commands David Marchand
2025-07-08 12:28   ` [PATCH v3 03/18] test/mempool: fix test without stack driver David Marchand
2025-07-08 15:15     ` Morten Brørup
2025-07-08 12:28   ` [PATCH v3 04/18] eal: fix plugin dir walk David Marchand
2025-07-08 12:28   ` [PATCH v3 05/18] cmdline: fix port list parsing David Marchand
2025-07-08 12:28   ` [PATCH v3 06/18] cmdline: fix highest bit " David Marchand
2025-07-08 12:28   ` [PATCH v3 07/18] tailq: fix cast macro for null pointer David Marchand
2025-07-08 12:28   ` [PATCH v3 08/18] hash: fix unaligned access in predictable RSS David Marchand
2025-07-08 12:35     ` Medvedkin, Vladimir
2025-07-08 12:28   ` [PATCH v3 09/18] stack: fix unaligned accesses on 128-bit David Marchand
2025-07-08 15:41     ` Morten Brørup
2025-07-08 12:28   ` [PATCH v3 10/18] build: support Undefined Behavior Sanitizer David Marchand
2025-07-08 12:28   ` [PATCH v3 11/18] test/telemetry: catch errors in subshell David Marchand
2025-07-08 12:28   ` [PATCH v3 12/18] malloc: fix mp message alignment David Marchand
2025-07-08 12:44     ` Bruce Richardson
2025-07-08 12:46       ` David Marchand
2025-07-08 13:25         ` Bruce Richardson [this message]
2025-07-08 13:33           ` David Marchand
2025-07-08 12:28   ` [PATCH v3 13/18] graph: fix stats query with no node xstats David Marchand
2025-07-08 12:28   ` [PATCH v3 14/18] graph: fix unaligned access in stats David Marchand
2025-07-08 12:28   ` [PATCH v3 15/18] eventdev: fix listing timer adapters with telemetry David Marchand
2025-07-08 12:28   ` [PATCH v3 16/18] test/power: fix tests without power drivers David Marchand
2025-07-08 12:47     ` Bruce Richardson
2025-07-08 12:53       ` David Marchand
2025-07-08 13:26         ` Bruce Richardson
2025-07-08 12:28   ` [PATCH v3 17/18] test/raw: fix test without skeleton driver David Marchand
2025-07-08 12:48     ` Bruce Richardson
2025-07-08 12:28   ` [PATCH v3 18/18] ci: extend coverage with UBSan David Marchand

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=aG0cVcB55BF2mEDd@bricha3-mobl1.ger.corp.intel.com \
    --to=bruce.richardson@intel.com \
    --cc=anatoly.burakov@intel.com \
    --cc=david.marchand@redhat.com \
    --cc=dev@dpdk.org \
    --cc=roretzla@linux.microsoft.com \
    /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).