From: Suren Baghdasaryan <surenb@google.com>
To: Michal Hocko <mhocko@suse.com>
Cc: akpm@linux-foundation.org, kent.overstreet@linux.dev,
vbabka@suse.cz, hannes@cmpxchg.org, roman.gushchin@linux.dev,
mgorman@suse.de, dave@stgolabs.net, willy@infradead.org,
liam.howlett@oracle.com, corbet@lwn.net, void@manifault.com,
peterz@infradead.org, juri.lelli@redhat.com,
ldufour@linux.ibm.com, catalin.marinas@arm.com, will@kernel.org,
arnd@arndb.de, tglx@linutronix.de, mingo@redhat.com,
dave.hansen@linux.intel.com, x86@kernel.org, peterx@redhat.com,
david@redhat.com, axboe@kernel.dk, mcgrof@kernel.org,
masahiroy@kernel.org, nathan@kernel.org, dennis@kernel.org,
tj@kernel.org, muchun.song@linux.dev, rppt@kernel.org,
paulmck@kernel.org, pasha.tatashin@soleen.com,
yosryahmed@google.com, yuzhao@google.com, dhowells@redhat.com,
hughd@google.com, andreyknvl@gmail.com, keescook@chromium.org,
ndesaulniers@google.com, gregkh@linuxfoundation.org,
ebiggers@google.com, ytcoode@gmail.com,
vincent.guittot@linaro.org, dietmar.eggemann@arm.com,
rostedt@goodmis.org, bsegall@google.com, bristot@redhat.com,
vschneid@redhat.com, cl@linux.com, penberg@kernel.org,
iamjoonsoo.kim@lge.com, 42.hyeyoo@gmail.com, glider@google.com,
elver@google.com, dvyukov@google.com, shakeelb@google.com,
songmuchun@bytedance.com, jbaron@akamai.com, rientjes@google.com,
minchan@google.com, kaleshsingh@google.com,
kernel-team@android.com, linux-doc@vger.kernel.org,
linux-kernel@vger.kernel.org, iommu@lists.linux.dev,
linux-arch@vger.kernel.org, linux-fsdevel@vger.kernel.org,
linux-mm@kvack.org, linux-modules@vger.kernel.org,
kasan-dev@googlegroups.com, cgroups@vger.kernel.org
Subject: Re: [PATCH 00/40] Memory allocation profiling
Date: Thu, 4 May 2023 08:08:13 -0700 [thread overview]
Message-ID: <CAJuCfpEkV_+pAjxyEpMqY+x7buZhSpj5qDF6KubsS=ObrQKUZg@mail.gmail.com> (raw)
In-Reply-To: <ZFN1yswCd9wRgYPR@dhcp22.suse.cz>
On Thu, May 4, 2023 at 2:07 AM Michal Hocko <mhocko@suse.com> wrote:
>
> On Wed 03-05-23 08:09:28, Suren Baghdasaryan wrote:
> > On Wed, May 3, 2023 at 12:25 AM Michal Hocko <mhocko@suse.com> wrote:
> [...]
> > Thanks for summarizing!
> >
> > > At least those I find the most important:
> > > - This is a big change and it adds a significant maintenance burden
> > > because each allocation entry point needs to be handled specifically.
> > > The cost will grow with the intended coverage especially there when
> > > allocation is hidden in a library code.
> >
> > Do you mean with more allocations in the codebase more codetags will
> > be generated? Is that the concern?
>
> No. I am mostly concerned about the _maintenance_ overhead. For the
> bare tracking (without profiling and thus stack traces) only those
> allocations that are directly inlined into the consumer are really
> of any use. That increases the code impact of the tracing because any
> relevant allocation location has to go through the micro surgery.
>
> e.g. is it really interesting to know that there is a likely memory
> leak in seq_file proper doing and allocation? No as it is the specific
> implementation using seq_file that is leaking most likely. There are
> other examples like that See?
Yes, I see that. One level tracking does not provide all the
information needed to track such issues. Something more informative
would cost more. That's why our proposal is to have a light-weight
mechanism to get a high level picture and then be able to zoom into a
specific area using context capture. If you have ideas to improve
this, I'm open to suggestions.
>
> > Or maybe as you commented in
> > another patch that context capturing feature does not limit how many
> > stacks will be captured?
>
> That is a memory overhead which can be really huge and it would be nice
> to be more explicit about that in the cover letter. It is a downside for
> sure but not something that has a code maintenance impact and it is an
> opt-in so it can be enabled only when necessary.
You are right, I'll add that into the cover letter.
>
> Quite honestly, though, the more I look into context capturing part it
> seems to me that there is much more to be reconsidered there and if you
> really want to move forward with the code tagging part then you should
> drop that for now. It would make the whole series smaller and easier to
> digest.
Sure, I don't see an issue with removing that for now and refining the
mechanism before posting again.
>
> > > - It has been brought up that this is duplicating functionality already
> > > available via existing tracing infrastructure. You should make it very
> > > clear why that is not suitable for the job
> >
> > I experimented with using tracing with _RET_IP_ to implement this
> > accounting. The major issue is the _RET_IP_ to codetag lookup runtime
> > overhead which is orders of magnitude higher than proposed code
> > tagging approach. With code tagging proposal, that link is resolved at
> > compile time. Since we want this mechanism deployed in production, we
> > want to keep the overhead to the absolute minimum.
> > You asked me before how much overhead would be tolerable and the
> > answer will always be "as small as possible". This is especially true
> > for slab allocators which are ridiculously fast and regressing them
> > would be very noticable (due to the frequent use).
>
> It would have been more convincing if you had some numbers at hands.
> E.g. this is a typical workload we are dealing with. With the compile
> time tags we are able to learn this with that much of cost. With a dynamic
> tracing we are able to learn this much with that cost. See? As small as
> possible is a rather vague term that different people will have a very
> different idea about.
I'm rerunning my tests with the latest kernel to collect the
comparison data. I profiled these solutions before but the kernel
changed since then, so I need to update them.
>
> > There is another issue, which I think can be solved in a smart way but
> > will either affect performance or would require more memory. With the
> > tracing approach we don't know beforehand how many individual
> > allocation sites exist, so we have to allocate code tags (or similar
> > structures for counting) at runtime vs compile time. We can be smart
> > about it and allocate in batches or even preallocate more than we need
> > beforehand but, as I said, it will require some kind of compromise.
>
> I have tried our usual distribution config (only vmlinux without modules
> so the real impact will be larger as we build a lot of stuff into
> modules) just to get an idea:
> text data bss dec hex filename
> 28755345 17040322 19845124 65640791 3e99957 vmlinux.before
> 28867168 17571838 19386372 65825378 3ec6a62 vmlinux.after
>
> Less than 1% for text 3% for data. This is not all that terrible
> for an initial submission and a more dynamic approach could be added
> later. E.g. with a smaller pre-allocated hash table that could be
> expanded lazily. Anyway not something I would be losing sleep over. This
> can always be improved later on.
Ah, right. I should have mentioned this overhead too. Thanks for
keeping me honest.
> > I understand that code tagging creates additional maintenance burdens
> > but I hope it also produces enough benefits that people will want
> > this. The cost is also hopefully amortized when additional
> > applications like the ones we presented in RFC [1] are built using the
> > same framework.
>
> TBH I am much more concerned about the maintenance burden on the MM side
> than the actual code tagging itslef which is much more self contained. I
> haven't seen other potential applications of the same infrastructure and
> maybe the code impact would be much smaller than in the MM proper. Our
> allocator API is really hairy and convoluted.
Yes, other applications are much smaller and cleaner. MM allocation
code is quite complex indeed.
>
> > > - We already have page_owner infrastructure that provides allocation
> > > tracking data. Why it cannot be used/extended?
> >
> > 1. The overhead.
>
> Do you have any numbers?
Will post once my tests are completed.
>
> > 2. Covers only page allocators.
>
> Yes this sucks.
> >
> > I didn't think about extending the page_owner approach to slab
> > allocators but I suspect it would not be trivial. I don't see
> > attaching an owner to every slab object to be a scalable solution. The
> > overhead would again be of concern here.
>
> This would have been a nice argument to mention in the changelog so that
> we know that you have considered that option at least. Why should I (as
> a reviewer) wild guess that?
Sorry, It's hard to remember all the decisions, discussions and
conclusions when working on a feature over a long time period. I'll
include more information about that.
>
> > I should point out that there was one important technical concern
> > about lack of a kill switch for this feature, which was an issue for
> > distributions that can't disable the CONFIG flag. In this series we
> > addressed that concern.
>
> Thanks, that is certainly appreciated. I haven't looked deeper into that
> part but from the cover letter I have understood that CONFIG_MEM_ALLOC_PROFILING
> implies unconditional page_ext and therefore the memory overhead
> assosiated with that. There seems to be a killswitch nomem_profiling but
> from a quick look it doesn't seem to disable page_ext allocations. I
> might be missing something there of course. Having a highlevel
> describtion for that would be really nice as well.
Right, will add a description of that as well.
We eliminate the runtime overhead but not the memory one. However I
believe it's also doable using page_ext_operations.need callback. Will
look into it.
Thanks,
Suren.
>
> > [1] https://lore.kernel.org/all/20220830214919.53220-1-surenb@google.com/
>
> --
> Michal Hocko
> SUSE Labs
next prev parent reply other threads:[~2023-05-04 15:08 UTC|newest]
Thread overview: 160+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-01 16:54 [PATCH 00/40] Memory allocation profiling Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 01/40] lib/string_helpers: Drop space in string_get_size's output Suren Baghdasaryan
2023-05-01 18:13 ` Davidlohr Bueso
2023-05-01 19:35 ` Kent Overstreet
2023-05-01 19:57 ` Andy Shevchenko
2023-05-01 21:16 ` Kent Overstreet
2023-05-01 21:33 ` Liam R. Howlett
2023-05-02 0:11 ` Kent Overstreet
2023-05-02 0:53 ` Kent Overstreet
2023-05-02 2:22 ` James Bottomley
2023-05-02 3:17 ` Kent Overstreet
2023-05-02 5:33 ` Andy Shevchenko
2023-05-02 6:21 ` Kent Overstreet
2023-05-02 15:19 ` Andy Shevchenko
2023-05-03 2:07 ` Kent Overstreet
2023-05-03 6:30 ` Andy Shevchenko
2023-05-03 7:12 ` Kent Overstreet
2023-05-03 9:12 ` Andy Shevchenko
2023-05-03 9:16 ` Kent Overstreet
2023-05-02 11:42 ` James Bottomley
2023-05-02 22:50 ` Dave Chinner
2023-05-03 9:28 ` Vlastimil Babka
2023-05-03 9:44 ` Andy Shevchenko
2023-05-03 12:15 ` James Bottomley
2023-05-02 7:55 ` Jani Nikula
2023-05-01 16:54 ` [PATCH 02/40] scripts/kallysms: Always include __start and __stop symbols Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 03/40] fs: Convert alloc_inode_sb() to a macro Suren Baghdasaryan
2023-05-02 12:35 ` Petr Tesařík
2023-05-02 19:57 ` Kent Overstreet
2023-05-02 20:20 ` Petr Tesařík
2023-05-01 16:54 ` [PATCH 04/40] nodemask: Split out include/linux/nodemask_types.h Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 05/40] prandom: Remove unused include Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 06/40] lib/string.c: strsep_no_empty() Suren Baghdasaryan
2023-05-02 12:37 ` Petr Tesařík
2023-05-01 16:54 ` [PATCH 07/40] Lazy percpu counters Suren Baghdasaryan
2023-05-01 19:17 ` Randy Dunlap
2023-05-01 16:54 ` [PATCH 08/40] mm: introduce slabobj_ext to support slab object extensions Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 09/40] mm: introduce __GFP_NO_OBJ_EXT flag to selectively prevent slabobj_ext creation Suren Baghdasaryan
2023-05-02 12:50 ` Petr Tesařík
2023-05-02 18:33 ` Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 10/40] mm/slab: introduce SLAB_NO_OBJ_EXT to avoid obj_ext creation Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 11/40] mm: prevent slabobj_ext allocations for slabobj_ext and kmem_cache objects Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 12/40] slab: objext: introduce objext_flags as extension to page_memcg_data_flags Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 13/40] lib: code tagging framework Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 14/40] lib: code tagging module support Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 15/40] lib: prevent module unloading if memory is not freed Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 16/40] lib: code tagging query helper functions Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 17/40] lib: add allocation tagging support for memory allocation profiling Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 18/40] lib: introduce support for page allocation tagging Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 19/40] change alloc_pages name in dma_map_ops to avoid name conflicts Suren Baghdasaryan
2023-05-02 15:50 ` Petr Tesařík
2023-05-02 18:38 ` Suren Baghdasaryan
2023-05-02 20:09 ` Petr Tesařík
2023-05-02 20:18 ` Kent Overstreet
2023-05-02 20:24 ` Suren Baghdasaryan
2023-05-02 20:39 ` Petr Tesařík
2023-05-02 20:41 ` Suren Baghdasaryan
2023-05-03 16:25 ` Steven Rostedt
2023-05-03 18:03 ` Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 20/40] mm: enable page allocation tagging Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 21/40] mm/page_ext: enable early_page_ext when CONFIG_MEM_ALLOC_PROFILING_DEBUG=y Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 22/40] mm: create new codetag references during page splitting Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 23/40] lib: add codetag reference into slabobj_ext Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 24/40] mm/slab: add allocation accounting into slab allocation and free paths Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 25/40] mm/slab: enable slab allocation tagging for kmalloc and friends Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 26/40] mm/slub: Mark slab_free_freelist_hook() __always_inline Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 27/40] mempool: Hook up to memory allocation profiling Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 28/40] timekeeping: Fix a circular include dependency Suren Baghdasaryan
2023-05-02 15:50 ` Thomas Gleixner
2023-05-01 16:54 ` [PATCH 29/40] mm: percpu: Introduce pcpuobj_ext Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 30/40] mm: percpu: Add codetag reference into pcpuobj_ext Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 31/40] mm: percpu: enable per-cpu allocation tagging Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 32/40] arm64: Fix circular header dependency Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 33/40] move stack capture functionality into a separate function for reuse Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 34/40] lib: code tagging context capture support Suren Baghdasaryan
2023-05-03 7:35 ` Michal Hocko
2023-05-03 15:18 ` Suren Baghdasaryan
2023-05-03 15:26 ` Dave Hansen
2023-05-03 19:45 ` Suren Baghdasaryan
2023-05-04 8:04 ` Michal Hocko
2023-05-04 14:31 ` Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 35/40] lib: implement context capture support for tagged allocations Suren Baghdasaryan
2023-05-03 7:39 ` Michal Hocko
2023-05-03 15:24 ` Suren Baghdasaryan
2023-05-04 8:09 ` Michal Hocko
2023-05-04 16:22 ` Suren Baghdasaryan
2023-05-05 8:40 ` Michal Hocko
2023-05-05 18:10 ` Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 36/40] lib: add memory allocations report in show_mem() Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 37/40] codetag: debug: skip objext checking when it's for objext itself Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 38/40] codetag: debug: mark codetags for reserved pages as empty Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 39/40] codetag: debug: introduce OBJEXTS_ALLOC_FAIL to mark failed slab_ext allocations Suren Baghdasaryan
2023-05-01 16:54 ` [PATCH 40/40] MAINTAINERS: Add entries for code tagging and memory allocation profiling Suren Baghdasaryan
2023-05-01 17:47 ` [PATCH 00/40] Memory " Roman Gushchin
2023-05-01 18:08 ` Suren Baghdasaryan
2023-05-01 18:14 ` Roman Gushchin
2023-05-01 19:37 ` Kent Overstreet
2023-05-01 21:18 ` Roman Gushchin
2023-05-03 7:25 ` Michal Hocko
2023-05-03 7:34 ` Kent Overstreet
2023-05-03 7:51 ` Michal Hocko
2023-05-03 8:05 ` Kent Overstreet
2023-05-03 13:21 ` Steven Rostedt
2023-05-03 16:35 ` Tejun Heo
2023-05-03 17:42 ` Suren Baghdasaryan
2023-05-03 18:06 ` Tejun Heo
2023-05-03 17:44 ` Kent Overstreet
2023-05-03 17:51 ` Kent Overstreet
2023-05-03 18:24 ` Tejun Heo
2023-05-03 18:07 ` Johannes Weiner
2023-05-03 18:19 ` Tejun Heo
2023-05-03 18:40 ` Tejun Heo
2023-05-03 18:56 ` Kent Overstreet
2023-05-03 18:58 ` Tejun Heo
2023-05-03 19:09 ` Tejun Heo
2023-05-03 19:41 ` Suren Baghdasaryan
2023-05-03 19:48 ` Tejun Heo
2023-05-03 20:00 ` Tejun Heo
2023-05-03 20:14 ` Suren Baghdasaryan
2023-05-04 2:25 ` Tejun Heo
2023-05-04 3:33 ` Kent Overstreet
2023-05-04 3:33 ` Suren Baghdasaryan
2023-05-04 8:00 ` Petr Tesařík
2023-05-03 20:08 ` Suren Baghdasaryan
2023-05-03 20:11 ` Johannes Weiner
2023-05-04 2:16 ` Tejun Heo
2023-05-03 20:04 ` Andrey Ryabinin
2023-05-03 9:50 ` Petr Tesařík
2023-05-03 9:54 ` Kent Overstreet
2023-05-03 10:24 ` Petr Tesařík
2023-05-03 9:57 ` Kent Overstreet
2023-05-03 10:26 ` Petr Tesařík
2023-05-03 15:30 ` Kent Overstreet
2023-05-03 12:33 ` James Bottomley
2023-05-03 14:31 ` Suren Baghdasaryan
2023-05-03 15:28 ` Kent Overstreet
2023-05-03 15:37 ` Lorenzo Stoakes
2023-05-03 16:03 ` Kent Overstreet
2023-05-03 15:49 ` James Bottomley
2023-05-03 15:09 ` Suren Baghdasaryan
2023-05-03 16:28 ` Steven Rostedt
2023-05-03 17:40 ` Suren Baghdasaryan
2023-05-03 18:03 ` Steven Rostedt
2023-05-03 18:07 ` Suren Baghdasaryan
2023-05-03 18:12 ` Kent Overstreet
2023-05-04 9:07 ` Michal Hocko
2023-05-04 15:08 ` Suren Baghdasaryan [this message]
2023-05-07 10:27 ` Michal Hocko
2023-05-07 17:01 ` Kent Overstreet
2023-05-07 17:20 ` Kent Overstreet
2023-05-07 20:55 ` Steven Rostedt
2023-05-07 21:53 ` Kent Overstreet
2023-05-07 22:09 ` Steven Rostedt
2023-05-07 22:17 ` Kent Overstreet
2023-05-08 15:52 ` Petr Tesařík
2023-05-08 15:57 ` Kent Overstreet
2023-05-08 16:09 ` Petr Tesařík
2023-05-08 16:28 ` Kent Overstreet
2023-05-08 18:59 ` Petr Tesařík
2023-05-08 20:48 ` Kent Overstreet
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='CAJuCfpEkV_+pAjxyEpMqY+x7buZhSpj5qDF6KubsS=ObrQKUZg@mail.gmail.com' \
--to=surenb@google.com \
--cc=42.hyeyoo@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=andreyknvl@gmail.com \
--cc=arnd@arndb.de \
--cc=axboe@kernel.dk \
--cc=bristot@redhat.com \
--cc=bsegall@google.com \
--cc=catalin.marinas@arm.com \
--cc=cgroups@vger.kernel.org \
--cc=cl@linux.com \
--cc=corbet@lwn.net \
--cc=dave.hansen@linux.intel.com \
--cc=dave@stgolabs.net \
--cc=david@redhat.com \
--cc=dennis@kernel.org \
--cc=dhowells@redhat.com \
--cc=dietmar.eggemann@arm.com \
--cc=dvyukov@google.com \
--cc=ebiggers@google.com \
--cc=elver@google.com \
--cc=glider@google.com \
--cc=gregkh@linuxfoundation.org \
--cc=hannes@cmpxchg.org \
--cc=hughd@google.com \
--cc=iamjoonsoo.kim@lge.com \
--cc=iommu@lists.linux.dev \
--cc=jbaron@akamai.com \
--cc=juri.lelli@redhat.com \
--cc=kaleshsingh@google.com \
--cc=kasan-dev@googlegroups.com \
--cc=keescook@chromium.org \
--cc=kent.overstreet@linux.dev \
--cc=kernel-team@android.com \
--cc=ldufour@linux.ibm.com \
--cc=liam.howlett@oracle.com \
--cc=linux-arch@vger.kernel.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-modules@vger.kernel.org \
--cc=masahiroy@kernel.org \
--cc=mcgrof@kernel.org \
--cc=mgorman@suse.de \
--cc=mhocko@suse.com \
--cc=minchan@google.com \
--cc=mingo@redhat.com \
--cc=muchun.song@linux.dev \
--cc=nathan@kernel.org \
--cc=ndesaulniers@google.com \
--cc=pasha.tatashin@soleen.com \
--cc=paulmck@kernel.org \
--cc=penberg@kernel.org \
--cc=peterx@redhat.com \
--cc=peterz@infradead.org \
--cc=rientjes@google.com \
--cc=roman.gushchin@linux.dev \
--cc=rostedt@goodmis.org \
--cc=rppt@kernel.org \
--cc=shakeelb@google.com \
--cc=songmuchun@bytedance.com \
--cc=tglx@linutronix.de \
--cc=tj@kernel.org \
--cc=vbabka@suse.cz \
--cc=vincent.guittot@linaro.org \
--cc=void@manifault.com \
--cc=vschneid@redhat.com \
--cc=will@kernel.org \
--cc=willy@infradead.org \
--cc=x86@kernel.org \
--cc=yosryahmed@google.com \
--cc=ytcoode@gmail.com \
--cc=yuzhao@google.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).