From: Jacob Keller <jacob.e.keller@intel.com>
To: Jeff King <peff@peff.net>
Cc: <git@vger.kernel.org>
Subject: Re: [PATCH 4/4] Makefile: turn on NO_MMAP when building with LSan
Date: Fri, 6 Mar 2026 01:17:24 -0800 [thread overview]
Message-ID: <796110ee-d795-4445-9d82-7026370a88cf@intel.com> (raw)
In-Reply-To: <20260305231305.GD2901305@coredump.intra.peff.net>
On 3/5/2026 3:13 PM, Jeff King wrote:
> The past few commits fixed some cases where we leak memory allocated by
> mmap(). Building with SANITIZE=leak doesn't detect these because it
> covers only heap buffers allocated by malloc().
>
> But if we build with NO_MMAP, our compat mmap() implementation will
> allocate a heap buffer and pread() into it. And thus Lsan will detect
> these leaks for free.
>
> Using NO_MMAP is less performant, of course, since we have to use extra
> memory and read in the whole file, rather than faulting in pages from
> disk. But LSan builds are already slow, and this doesn't make them
> measurably worse. Getting extra coverage for our leak-checking is worth
> it.
>
> Signed-off-by: Jeff King <peff@peff.net>
> ---
> Makefile | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/Makefile b/Makefile
> index f3264d0a37..4cf1afd395 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -1600,6 +1600,7 @@ BASIC_CFLAGS += -DSHA1DC_FORCE_ALIGNED_ACCESS
> endif
> ifneq ($(filter leak,$(SANITIZERS)),)
> BASIC_CFLAGS += -O0
> +NO_MMAP = CatchMapLeaks
> SANITIZE_LEAK = YesCompiledWithIt
> endif
> ifneq ($(filter address,$(SANITIZERS)),)
Should this patch also affect the meson.build?
There is the following in meson.build:
if host_machine.system() == 'windows'
libgit_c_args += '-DUSE_WIN32_MMAP'
else
checkfuncs += {
# provided by compat/mingw.c.
'unsetenv' : ['unsetenv.c'],
# provided by compat/mingw.c.
'getpagesize' : [],
}
if get_option('b_sanitize').contains('address')
libgit_c_args += '-DNO_MMAP'
libgit_sources += 'compat/mmap.c'
else
checkfuncs += { 'mmap': ['mmap.c'] }
endif
endif
This probably needs to also check if it contains leak, no?
Also I think this might be somewhat less flexible than Make since you
can't forcibly enable mmap even with sanitizers enabled. I suppose thats
not a big deal since enabling sanitizers already has a high cost.
Thanks,
Jake
next prev parent reply other threads:[~2026-03-06 9:17 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-05 20:51 memory leak when cloning a repository Jacob Keller
2026-03-05 22:02 ` Jeff King
2026-03-05 23:03 ` [PATCH 0/4] plugging some mmap() leaks Jeff King
2026-03-05 23:08 ` [PATCH 1/4] check_connected(): delay opening new_pack Jeff King
2026-03-05 23:18 ` Jacob Keller
2026-03-05 23:09 ` [PATCH 2/4] check_connected(): fix leak of pack-index mmap Jeff King
2026-03-05 23:20 ` Jacob Keller
2026-03-05 23:12 ` [PATCH 3/4] pack-revindex: avoid double-loading .rev files Jeff King
2026-03-05 23:13 ` [PATCH 4/4] Makefile: turn on NO_MMAP when building with LSan Jeff King
2026-03-06 9:17 ` Jacob Keller [this message]
2026-03-06 16:25 ` [PATCH 5/4] meson: " Jeff King
2026-03-06 18:00 ` Ramsay Jones
2026-03-07 1:14 ` [PATCH 4/4] Makefile: " Junio C Hamano
2026-03-07 2:24 ` [PATCH 3.5/4] object-file: fix mmap() leak in odb_source_loose_read_object_stream() Jeff King
2026-03-07 5:35 ` Junio C Hamano
2026-03-10 12:23 ` Patrick Steinhardt
2026-03-06 4:37 ` [PATCH 0/4] plugging some mmap() leaks Ramsay Jones
2026-03-06 16:21 ` Jeff King
2026-03-06 17:49 ` Ramsay Jones
2026-03-06 18:37 ` Junio C Hamano
2026-03-06 18:55 ` Ramsay Jones
2026-03-06 22:05 ` Junio C Hamano
2026-03-06 23:25 ` Ramsay Jones
2026-03-07 1:15 ` Junio C Hamano
2026-03-05 23:16 ` memory leak when cloning a repository Jacob Keller
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=796110ee-d795-4445-9d82-7026370a88cf@intel.com \
--to=jacob.e.keller@intel.com \
--cc=git@vger.kernel.org \
--cc=peff@peff.net \
/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