Building the Linux kernel with Clang and LLVM
 help / color / mirror / Atom feed
From: kernel test robot <lkp@intel.com>
Cc: oe-kbuild-all@lists.linux.dev, llvm@lists.linux.dev
Subject: arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
Date: Fri, 25 Oct 2024 05:26:19 +0800	[thread overview]
Message-ID: <202410250537.v3GaLBau-lkp@intel.com> (raw)

CC: linux-kernel@vger.kernel.org
TO: "Jason A. Donenfeld" <zx2c4@kernel.org>
CC: Herbert Xu <herbert@gondor.apana.org.au>

Hi Jason,

FYI, the error/warning still remains.

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
head:   c2ee9f594da826bea183ed14f2cc029c719bf4da
commit: 07b586fe06625b0b610dc3d3a969c51913d143d4 crypto: x86/curve25519 - replace with formally verified implementation
date:   4 years, 8 months ago
config: x86_64-buildonly-randconfig-004-20241025 (https://download.01.org/0day-ci/archive/20241025/202410250537.v3GaLBau-lkp@intel.com/config)
compiler: clang version 19.1.2 (https://github.com/llvm/llvm-project 7ba7d8e2f7b6445b60679da826210cdde29eaf8b)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20241025/202410250537.v3GaLBau-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/202410250537.v3GaLBau-lkp@intel.com/

All errors (new ones prefixed by >>):

   In file included from arch/x86/crypto/curve25519-x86_64.c:7:
   In file included from include/crypto/curve25519.h:9:
   In file included from include/crypto/algapi.h:13:
   In file included from include/linux/skbuff.h:17:
   In file included from include/linux/bvec.h:13:
   In file included from include/linux/mm.h:1289:
   include/linux/vmstat.h:412:36: warning: arithmetic between different enumeration types ('enum node_stat_item' and 'enum lru_list') [-Wenum-enum-conversion]
     412 |         return node_stat_name(NR_LRU_BASE + lru) + 3; // skip "nr_"
         |                               ~~~~~~~~~~~ ^ ~~~
   arch/x86/crypto/curve25519-x86_64.c:951:6: warning: variable 'carry0' set but not used [-Wunused-but-set-variable]
     951 |         u64 carry0;
         |             ^
   arch/x86/crypto/curve25519-x86_64.c:954:6: warning: variable 'carry' set but not used [-Wunused-but-set-variable]
     954 |         u64 carry;
         |             ^
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
     518 |                 "  movq 0(%1), %%rdx;"                                       /* f[0] */
         |                 ^
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
>> arch/x86/crypto/curve25519-x86_64.c:518:3: error: inline assembly requires more registers than available
   fatal error: too many errors emitted, stopping now [-ferror-limit=]
   3 warnings and 20 errors generated.


vim +518 arch/x86/crypto/curve25519-x86_64.c

   509	
   510	/* Computes the square of a field element: out <- f * f
   511	 * Uses the 8-element buffer tmp for intermediate results */
   512	static inline void fsqr(u64 *out, const u64 *f, u64 *tmp)
   513	{
   514		asm volatile(
   515			/* Compute the raw multiplication: tmp <- f * f */
   516	
   517			/* Step 1: Compute all partial products */
 > 518			"  movq 0(%1), %%rdx;"                                       /* f[0] */
   519			"  mulxq 8(%1), %%r8, %%r14;"      "  xor %%r15, %%r15;"     /* f[1]*f[0] */
   520			"  mulxq 16(%1), %%r9, %%r10;"     "  adcx %%r14, %%r9;"     /* f[2]*f[0] */
   521			"  mulxq 24(%1), %%rax, %%rcx;"    "  adcx %%rax, %%r10;"    /* f[3]*f[0] */
   522			"  movq 24(%1), %%rdx;"                                      /* f[3] */
   523			"  mulxq 8(%1), %%r11, %%r12;"     "  adcx %%rcx, %%r11;"    /* f[1]*f[3] */
   524			"  mulxq 16(%1), %%rax, %%r13;"    "  adcx %%rax, %%r12;"    /* f[2]*f[3] */
   525			"  movq 8(%1), %%rdx;"             "  adcx %%r15, %%r13;"    /* f1 */
   526			"  mulxq 16(%1), %%rax, %%rcx;"    "  mov $0, %%r14;"        /* f[2]*f[1] */
   527	
   528			/* Step 2: Compute two parallel carry chains */
   529			"  xor %%r15, %%r15;"
   530			"  adox %%rax, %%r10;"
   531			"  adcx %%r8, %%r8;"
   532			"  adox %%rcx, %%r11;"
   533			"  adcx %%r9, %%r9;"
   534			"  adox %%r15, %%r12;"
   535			"  adcx %%r10, %%r10;"
   536			"  adox %%r15, %%r13;"
   537			"  adcx %%r11, %%r11;"
   538			"  adox %%r15, %%r14;"
   539			"  adcx %%r12, %%r12;"
   540			"  adcx %%r13, %%r13;"
   541			"  adcx %%r14, %%r14;"
   542	
   543			/* Step 3: Compute intermediate squares */
   544			"  movq 0(%1), %%rdx;"     "  mulx %%rdx, %%rax, %%rcx;"    /* f[0]^2 */
   545			                           "  movq %%rax, 0(%0);"
   546			"  add %%rcx, %%r8;"       "  movq %%r8, 8(%0);"
   547			"  movq 8(%1), %%rdx;"     "  mulx %%rdx, %%rax, %%rcx;"    /* f[1]^2 */
   548			"  adcx %%rax, %%r9;"      "  movq %%r9, 16(%0);"
   549			"  adcx %%rcx, %%r10;"     "  movq %%r10, 24(%0);"
   550			"  movq 16(%1), %%rdx;"    "  mulx %%rdx, %%rax, %%rcx;"    /* f[2]^2 */
   551			"  adcx %%rax, %%r11;"     "  movq %%r11, 32(%0);"
   552			"  adcx %%rcx, %%r12;"     "  movq %%r12, 40(%0);"
   553			"  movq 24(%1), %%rdx;"    "  mulx %%rdx, %%rax, %%rcx;"    /* f[3]^2 */
   554			"  adcx %%rax, %%r13;"     "  movq %%r13, 48(%0);"
   555			"  adcx %%rcx, %%r14;"     "  movq %%r14, 56(%0);"
   556	
   557			/* Line up pointers */
   558			"  mov %0, %1;"
   559			"  mov %2, %0;"
   560	
   561			/* Wrap the result back into the field */
   562	
   563			/* Step 1: Compute dst + carry == tmp_hi * 38 + tmp_lo */
   564			"  mov $38, %%rdx;"
   565			"  mulxq 32(%1), %%r8, %%r13;"
   566			"  xor %%rcx, %%rcx;"
   567			"  adoxq 0(%1), %%r8;"
   568			"  mulxq 40(%1), %%r9, %%r12;"
   569			"  adcx %%r13, %%r9;"
   570			"  adoxq 8(%1), %%r9;"
   571			"  mulxq 48(%1), %%r10, %%r13;"
   572			"  adcx %%r12, %%r10;"
   573			"  adoxq 16(%1), %%r10;"
   574			"  mulxq 56(%1), %%r11, %%rax;"
   575			"  adcx %%r13, %%r11;"
   576			"  adoxq 24(%1), %%r11;"
   577			"  adcx %%rcx, %%rax;"
   578			"  adox %%rcx, %%rax;"
   579			"  imul %%rdx, %%rax;"
   580	
   581			/* Step 2: Fold the carry back into dst */
   582			"  add %%rax, %%r8;"
   583			"  adcx %%rcx, %%r9;"
   584			"  movq %%r9, 8(%0);"
   585			"  adcx %%rcx, %%r10;"
   586			"  movq %%r10, 16(%0);"
   587			"  adcx %%rcx, %%r11;"
   588			"  movq %%r11, 24(%0);"
   589	
   590			/* Step 3: Fold the carry bit back in; guaranteed not to carry at this point */
   591			"  mov $0, %%rax;"
   592			"  cmovc %%rdx, %%rax;"
   593			"  add %%rax, %%r8;"
   594			"  movq %%r8, 0(%0);"
   595		: "+&r" (tmp), "+&r" (f), "+&r" (out)
   596		:
   597		: "%rax", "%rcx", "%rdx", "%r8", "%r9", "%r10", "%r11", "%r12", "%r13", "%r14", "%r15", "memory", "cc"
   598		);
   599	}
   600	

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki

                 reply	other threads:[~2024-10-24 21:26 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=202410250537.v3GaLBau-lkp@intel.com \
    --to=lkp@intel.com \
    --cc=llvm@lists.linux.dev \
    --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