From: Stephen Hemminger <stephen@networkplumber.org>
To: Daniel Gregory <daniel.gregory@bytedance.com>
Cc: Ruifeng Wang <ruifeng.wang@arm.com>,
dev@dpdk.org, Punit Agrawal <punit.agrawal@bytedance.com>,
Liang Ma <liangma@bytedance.com>
Subject: Re: [PATCH] eal/arm: replace RTE_BUILD_BUG on non-constant
Date: Fri, 3 May 2024 17:56:24 -0700 [thread overview]
Message-ID: <20240503175624.54dd0a72@hermes.local> (raw)
In-Reply-To: <20240503094605.GB3883350@ste-uk-lab-gw>
On Fri, 3 May 2024 10:46:05 +0100
Daniel Gregory <daniel.gregory@bytedance.com> wrote:
> On Thu, May 02, 2024 at 02:48:26PM -0700, Stephen Hemminger wrote:
> > There are already constant checks like this elsewhere in the file.
>
> Yes, but they're in macros, rather than inlined functions, so my
> understanding was that at compile time, macro expansion has put the
> memorder constant in the _Static_assert call as opposed to still being
> a function parameter in the inline definition.
Gcc and clang are smart enough that it is possible to use the internal
__builtin_constant_p() in the function. Some examples in DPDK:
static __rte_always_inline int
rte_mempool_do_generic_get(struct rte_mempool *mp, void **obj_table,
unsigned int n, struct rte_mempool_cache *cache)
{
int ret;
unsigned int remaining;
uint32_t index, len;
void **cache_objs;
/* No cache provided */
if (unlikely(cache == NULL)) {
remaining = n;
goto driver_dequeue;
}
/* The cache is a stack, so copy will be in reverse order. */
cache_objs = &cache->objs[cache->len];
if (__extension__(__builtin_constant_p(n)) && n <= cache->len) {
It should be possible to use RTE_BUILD_BUG_ON() or static_assert here.
Changing a compile time check into a runtime check means that buggy programs
blow up much later and in a confusing manner. And it impacts all code that
is doing a spin lock, certainly one of the hottest paths in DPDK.
next prev parent reply other threads:[~2024-05-04 0:56 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-05-02 14:21 Daniel Gregory
2024-05-02 16:20 ` Stephen Hemminger
2024-05-02 17:44 ` Daniel Gregory
2024-05-02 18:27 ` Stephen Hemminger
2024-05-02 21:48 ` Stephen Hemminger
2024-05-03 9:46 ` Daniel Gregory
2024-05-04 0:56 ` Stephen Hemminger [this message]
2024-05-09 11:02 ` Daniel Gregory
2024-05-03 13:32 ` David Marchand
2024-05-03 14:21 ` Daniel Gregory
2024-05-03 18:27 ` [PATCH v2] " Daniel Gregory
2024-05-03 18:30 ` Daniel Gregory
2024-05-04 0:59 ` Stephen Hemminger
2024-05-06 9:30 ` Ruifeng Wang
2024-05-11 17:00 ` Wathsala Wathawana Vithanage
2024-05-04 1:02 ` [PATCH] " Stephen Hemminger
2024-05-09 11:11 ` Daniel Gregory
2024-05-09 16:47 ` Tyler Retzlaff
2024-05-11 16:48 ` Wathsala Wathawana Vithanage
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=20240503175624.54dd0a72@hermes.local \
--to=stephen@networkplumber.org \
--cc=daniel.gregory@bytedance.com \
--cc=dev@dpdk.org \
--cc=liangma@bytedance.com \
--cc=punit.agrawal@bytedance.com \
--cc=ruifeng.wang@arm.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).