From: Nathan Chancellor <nathan@kernel.org>
To: Vincent Mailhol <mailhol.vincent@wanadoo.fr>,
Andrew Morton <akpm@linux-foundation.org>
Cc: kernel test robot <lkp@intel.com>,
llvm@lists.linux.dev, oe-kbuild-all@lists.linux.dev,
Linux Memory Management List <linux-mm@kvack.org>
Subject: Re: [linux-next:master 2260/2462] drivers/gpu/drm/xe/xe_gt_idle.c:127:33: error: static assertion expression is not an integral constant expression
Date: Tue, 17 Dec 2024 10:34:48 -0700 [thread overview]
Message-ID: <20241217173448.GA2647780@ax162> (raw)
In-Reply-To: <CAMZ6RqJXWfnNebqxZZL4ANOOTjnLPfVOOdG0Rr6TM+jarJdQZA@mail.gmail.com>
Andrew,
Top posting for visibility. Vincent requested dropping
build_bugh-more-user-friendly-error-messages-in-build_bug_on_zero.patch
because it breaks the build for older clang versions in certain drivers
and I am still seeing those failures in the latest -next. Can you please
drop this patch?
Cheers,
Nathan
On Mon, Dec 09, 2024 at 02:01:49PM +0900, Vincent Mailhol wrote:
> On Mon. 9 Dec. 2024 at 13:34, kernel test robot <lkp@intel.com> wrote:
> > tree: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
> > head: ebe1b11614e079c5e366ce9bd3c8f44ca0fbcc1b
> > commit: 86c4e589d5c8fbb5092de0052b89402e86706515 [2260/2462] build_bug.h: more user friendly error messages in BUILD_BUG_ON_ZERO()
> > config: s390-randconfig-001-20241208 (https://download.01.org/0day-ci/archive/20241208/202412080849.sPp82jSi-lkp@intel.com/config)
> > compiler: clang version 15.0.7 (https://github.com/llvm/llvm-project 8dfdcc7b7bf66834a761bd8de445840ef68e4d1a)
> > reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20241208/202412080849.sPp82jSi-lkp@intel.com/reproduce)
> >
> > If you fix the issue in a separate patch/commit (i.e. not just a new version of
> > the same patch/commit), kindly add following tags
> > | Reported-by: kernel test robot <lkp@intel.com>
> > | Closes: https://lore.kernel.org/oe-kbuild-all/202412080849.sPp82jSi-lkp@intel.com/
> >
> > All errors (new ones prefixed by >>):
> >
> > >> drivers/gpu/drm/xe/xe_gt_idle.c:127:33: error: static assertion expression is not an integral constant expression
> > gtidle->powergate_enable |= (VDN_HCP_POWERGATE_ENABLE(j) |
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~
> > drivers/gpu/drm/xe/regs/xe_gt_regs.h:342:40: note: expanded from macro 'VDN_HCP_POWERGATE_ENABLE'
> > #define VDN_HCP_POWERGATE_ENABLE(n) REG_BIT(3 + 2 * (n))
> > ^~~~~~~~~~~~~~~~~~~~
> > drivers/gpu/drm/xe/compat-i915-headers/../../i915/i915_reg_defs.h:22:9: note: expanded from macro 'REG_BIT'
> > BUILD_BUG_ON_ZERO(__is_constexpr(__n) && \
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:16:35: note: expanded from macro 'BUILD_BUG_ON_ZERO'
> > #define BUILD_BUG_ON_ZERO(e, ...) __BUILD_BUG_ON_ZERO(e, ##__VA_ARGS__, #e)
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:17:37: note: expanded from macro '__BUILD_BUG_ON_ZERO'
> > #define __BUILD_BUG_ON_ZERO(e, msg) __BUILD_BUG_ON_ZERO_MSG(e, msg " is true")
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/compiler.h:245:77: note: expanded from macro '__BUILD_BUG_ON_ZERO_MSG'
> > #define __BUILD_BUG_ON_ZERO_MSG(e, msg) ((int)sizeof(struct {_Static_assert(!(e), msg);}))
> > ^~~~
> > drivers/gpu/drm/xe/xe_gt_idle.c:128:12: error: static assertion expression is not an integral constant expression
> > VDN_MFXVDENC_POWERGATE_ENABLE(j));
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
> So, here, j is not an integer constant expression. When passed to
> REG_BIT(), it will eventually expands to:
>
> BUILD_BUG_ON_ZERO(__is_constexpr(j) && (j < 0 || j > 31))
>
> which is a compile time constant but not an integer constant
> expression. It was constant enough for the previous definition of
> BUILD_BUG_ON_ZERO() which relied on bitfields, but not enough constant
> for _Static_assert().
>
> The reason I did not notice that is because I have another series in
> my tree which solves this issue, precisely, this patch:
>
> https://lore.kernel.org/all/20241203-is_constexpr-refactor-v1-8-4e4cbaecc216@wanadoo.fr/
>
> @Andrew, maybe you can remove this patch from your tree. I thought it
> was standalone, but it is not. I will resend once the above series get
> merged.
>
> > drivers/gpu/drm/xe/regs/xe_gt_regs.h:343:44: note: expanded from macro 'VDN_MFXVDENC_POWERGATE_ENABLE'
> > #define VDN_MFXVDENC_POWERGATE_ENABLE(n) REG_BIT(4 + 2 * (n))
> > ^~~~~~~~~~~~~~~~~~~~
> > drivers/gpu/drm/xe/compat-i915-headers/../../i915/i915_reg_defs.h:22:9: note: expanded from macro 'REG_BIT'
> > BUILD_BUG_ON_ZERO(__is_constexpr(__n) && \
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:16:35: note: expanded from macro 'BUILD_BUG_ON_ZERO'
> > #define BUILD_BUG_ON_ZERO(e, ...) __BUILD_BUG_ON_ZERO(e, ##__VA_ARGS__, #e)
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:17:37: note: expanded from macro '__BUILD_BUG_ON_ZERO'
> > #define __BUILD_BUG_ON_ZERO(e, msg) __BUILD_BUG_ON_ZERO_MSG(e, msg " is true")
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/compiler.h:245:77: note: expanded from macro '__BUILD_BUG_ON_ZERO_MSG'
> > #define __BUILD_BUG_ON_ZERO_MSG(e, msg) ((int)sizeof(struct {_Static_assert(!(e), msg);}))
> > ^~~~
> > 2 errors generated.
> > --
> > >> drivers/gpu/drm/xe/xe_reg_whitelist.c:139:22: error: static assertion expression is not an integral constant expression
> > range_start = reg & REG_GENMASK(25, range_bit);
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~
> > drivers/gpu/drm/xe/compat-i915-headers/../../i915/i915_reg_defs.h:49:9: note: expanded from macro 'REG_GENMASK'
> > BUILD_BUG_ON_ZERO(__is_constexpr(__high) && \
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:16:35: note: expanded from macro 'BUILD_BUG_ON_ZERO'
> > #define BUILD_BUG_ON_ZERO(e, ...) __BUILD_BUG_ON_ZERO(e, ##__VA_ARGS__, #e)
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:17:37: note: expanded from macro '__BUILD_BUG_ON_ZERO'
> > #define __BUILD_BUG_ON_ZERO(e, msg) __BUILD_BUG_ON_ZERO_MSG(e, msg " is true")
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/compiler.h:245:77: note: expanded from macro '__BUILD_BUG_ON_ZERO_MSG'
> > #define __BUILD_BUG_ON_ZERO_MSG(e, msg) ((int)sizeof(struct {_Static_assert(!(e), msg);}))
> > ^~~~
> > drivers/gpu/drm/xe/xe_reg_whitelist.c:140:28: error: static assertion expression is not an integral constant expression
> > range_end = range_start | REG_GENMASK(range_bit, 0);
> > ^~~~~~~~~~~~~~~~~~~~~~~~~
> > drivers/gpu/drm/xe/compat-i915-headers/../../i915/i915_reg_defs.h:49:9: note: expanded from macro 'REG_GENMASK'
> > BUILD_BUG_ON_ZERO(__is_constexpr(__high) && \
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:16:35: note: expanded from macro 'BUILD_BUG_ON_ZERO'
> > #define BUILD_BUG_ON_ZERO(e, ...) __BUILD_BUG_ON_ZERO(e, ##__VA_ARGS__, #e)
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:17:37: note: expanded from macro '__BUILD_BUG_ON_ZERO'
> > #define __BUILD_BUG_ON_ZERO(e, msg) __BUILD_BUG_ON_ZERO_MSG(e, msg " is true")
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/compiler.h:245:77: note: expanded from macro '__BUILD_BUG_ON_ZERO_MSG'
> > #define __BUILD_BUG_ON_ZERO_MSG(e, msg) ((int)sizeof(struct {_Static_assert(!(e), msg);}))
> > ^~~~
> > 2 errors generated.
> > --
> > >> drivers/gpu/drm/xe/xe_irq.c:303:22: error: static assertion expression is not an integral constant expression
> > if (!(master_ctl & GT_DW_IRQ(bank)))
> > ^~~~~~~~~~~~~~~
> > drivers/gpu/drm/xe/regs/xe_irq_regs.h:22:27: note: expanded from macro 'GT_DW_IRQ'
> > #define GT_DW_IRQ(x) REG_BIT(x)
> > ^~~~~~~~~~
> > drivers/gpu/drm/xe/compat-i915-headers/../../i915/i915_reg_defs.h:22:9: note: expanded from macro 'REG_BIT'
> > BUILD_BUG_ON_ZERO(__is_constexpr(__n) && \
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:16:35: note: expanded from macro 'BUILD_BUG_ON_ZERO'
> > #define BUILD_BUG_ON_ZERO(e, ...) __BUILD_BUG_ON_ZERO(e, ##__VA_ARGS__, #e)
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:17:37: note: expanded from macro '__BUILD_BUG_ON_ZERO'
> > #define __BUILD_BUG_ON_ZERO(e, msg) __BUILD_BUG_ON_ZERO_MSG(e, msg " is true")
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/compiler.h:245:77: note: expanded from macro '__BUILD_BUG_ON_ZERO_MSG'
> > #define __BUILD_BUG_ON_ZERO_MSG(e, msg) ((int)sizeof(struct {_Static_assert(!(e), msg);}))
> > ^~~~
> > drivers/gpu/drm/xe/xe_irq.c:436:26: error: static assertion expression is not an integral constant expression
> > if ((master_tile_ctl & DG1_MSTR_TILE(tile->id)) == 0)
> > ^~~~~~~~~~~~~~~~~~~~~~~
> > drivers/gpu/drm/xe/regs/xe_irq_regs.h:16:30: note: expanded from macro 'DG1_MSTR_TILE'
> > #define DG1_MSTR_TILE(t) REG_BIT(t)
> > ^~~~~~~~~~
> > drivers/gpu/drm/xe/compat-i915-headers/../../i915/i915_reg_defs.h:22:9: note: expanded from macro 'REG_BIT'
> > BUILD_BUG_ON_ZERO(__is_constexpr(__n) && \
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:16:35: note: expanded from macro 'BUILD_BUG_ON_ZERO'
> > #define BUILD_BUG_ON_ZERO(e, ...) __BUILD_BUG_ON_ZERO(e, ##__VA_ARGS__, #e)
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:17:37: note: expanded from macro '__BUILD_BUG_ON_ZERO'
> > #define __BUILD_BUG_ON_ZERO(e, msg) __BUILD_BUG_ON_ZERO_MSG(e, msg " is true")
> > ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > include/linux/compiler.h:245:77: note: expanded from macro '__BUILD_BUG_ON_ZERO_MSG'
> > #define __BUILD_BUG_ON_ZERO_MSG(e, msg) ((int)sizeof(struct {_Static_assert(!(e), msg);}))
> > ^~~~
> > 2 errors generated.
> >
> >
> > vim +127 drivers/gpu/drm/xe/xe_gt_idle.c
> >
> > 1c2097bbde107e drivers/gpu/drm/xe/xe_gt_idle_sysfs.c Riana Tauro 2023-06-23 97
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 98 void xe_gt_idle_enable_pg(struct xe_gt *gt)
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 99 {
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 100 struct xe_device *xe = gt_to_xe(gt);
> > c2bf07dd0bbce1 drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-09-06 101 struct xe_gt_idle *gtidle = >->gtidle;
> > 50089a9534e258 drivers/gpu/drm/xe/xe_gt_idle.c Matt Roper 2024-09-10 102 struct xe_mmio *mmio = >->mmio;
> > c2bf07dd0bbce1 drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-09-06 103 u32 vcs_mask, vecs_mask;
> > a66c19895396e6 drivers/gpu/drm/xe/xe_gt_idle.c Himal Prasad Ghimiray 2024-10-14 104 unsigned int fw_ref;
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 105 int i, j;
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 106
> > ecab82af278733 drivers/gpu/drm/xe/xe_gt_idle.c Michal Wajdeczko 2024-06-19 107 if (IS_SRIOV_VF(xe))
> > ecab82af278733 drivers/gpu/drm/xe/xe_gt_idle.c Michal Wajdeczko 2024-06-19 108 return;
> > ecab82af278733 drivers/gpu/drm/xe/xe_gt_idle.c Michal Wajdeczko 2024-06-19 109
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 110 /* Disable CPG for PVC */
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 111 if (xe->info.platform == XE_PVC)
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 112 return;
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 113
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 114 xe_device_assert_mem_access(gt_to_xe(gt));
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 115
> > c2bf07dd0bbce1 drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-09-06 116 vcs_mask = xe_hw_engine_mask_per_class(gt, XE_ENGINE_CLASS_VIDEO_DECODE);
> > c2bf07dd0bbce1 drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-09-06 117 vecs_mask = xe_hw_engine_mask_per_class(gt, XE_ENGINE_CLASS_VIDEO_ENHANCE);
> > c2bf07dd0bbce1 drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-09-06 118
> > c2bf07dd0bbce1 drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-09-06 119 if (vcs_mask || vecs_mask)
> > c2bf07dd0bbce1 drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-09-06 120 gtidle->powergate_enable = MEDIA_POWERGATE_ENABLE;
> > c2bf07dd0bbce1 drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-09-06 121
> > c2bf07dd0bbce1 drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-09-06 122 if (!xe_gt_is_media_type(gt))
> > c2bf07dd0bbce1 drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-09-06 123 gtidle->powergate_enable |= RENDER_POWERGATE_ENABLE;
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 124
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 125 for (i = XE_HW_ENGINE_VCS0, j = 0; i <= XE_HW_ENGINE_VCS7; ++i, ++j) {
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 126 if ((gt->info.engine_mask & BIT(i)))
> > c2bf07dd0bbce1 drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-09-06 @127 gtidle->powergate_enable |= (VDN_HCP_POWERGATE_ENABLE(j) |
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 128 VDN_MFXVDENC_POWERGATE_ENABLE(j));
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 129 }
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 130
> > a66c19895396e6 drivers/gpu/drm/xe/xe_gt_idle.c Himal Prasad Ghimiray 2024-10-14 131 fw_ref = xe_force_wake_get(gt_to_fw(gt), XE_FW_GT);
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 132 if (xe->info.skip_guc_pc) {
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 133 /*
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 134 * GuC sets the hysteresis value when GuC PC is enabled
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 135 * else set it to 25 (25 * 1.28us)
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 136 */
> > 50089a9534e258 drivers/gpu/drm/xe/xe_gt_idle.c Matt Roper 2024-09-10 137 xe_mmio_write32(mmio, MEDIA_POWERGATE_IDLE_HYSTERESIS, 25);
> > 50089a9534e258 drivers/gpu/drm/xe/xe_gt_idle.c Matt Roper 2024-09-10 138 xe_mmio_write32(mmio, RENDER_POWERGATE_IDLE_HYSTERESIS, 25);
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 139 }
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 140
> > 50089a9534e258 drivers/gpu/drm/xe/xe_gt_idle.c Matt Roper 2024-09-10 141 xe_mmio_write32(mmio, POWERGATE_ENABLE, gtidle->powergate_enable);
> > a66c19895396e6 drivers/gpu/drm/xe/xe_gt_idle.c Himal Prasad Ghimiray 2024-10-14 142 xe_force_wake_put(gt_to_fw(gt), fw_ref);
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 143 }
> > 38e8c4184ea0ee drivers/gpu/drm/xe/xe_gt_idle.c Riana Tauro 2024-05-24 144
> >
> > :::::: The code at line 127 was first introduced by commit
> > :::::: c2bf07dd0bbce1f318b73b525e21fbc6d67a3a94 drm/xe/xe_gt_idle: modify powergate enable condition
> >
> > :::::: TO: Riana Tauro <riana.tauro@intel.com>
> > :::::: CC: Rodrigo Vivi <rodrigo.vivi@intel.com>
> >
> > --
> > 0-DAY CI Kernel Test Service
> > https://github.com/intel/lkp-tests/wiki
prev parent reply other threads:[~2024-12-17 17:34 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-12-09 4:34 kernel test robot
2024-12-09 5:01 ` Vincent Mailhol
2024-12-17 17:34 ` Nathan Chancellor [this message]
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=20241217173448.GA2647780@ax162 \
--to=nathan@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=linux-mm@kvack.org \
--cc=lkp@intel.com \
--cc=llvm@lists.linux.dev \
--cc=mailhol.vincent@wanadoo.fr \
--cc=oe-kbuild-all@lists.linux.dev \
/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