| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Mon, 23 Aug 2021 16:35:18 +0200
From: Christophe Leroy <christophe.leroy@...roup.eu>
To: kernel test robot <lkp@...el.com>
Cc: clang-built-linux@...glegroups.com, kbuild-all@...ts.01.org,
linux-kernel@...r.kernel.org, Michael Ellerman <mpe@...erman.id.au>
Subject: Re: arch/powerpc/kernel/signal_32.c:297:2: warning: Value stored to
'msr' is never read [clang-analyzer-deadcode.DeadStores]
Le 23/08/2021 à 12:59, kernel test robot a écrit :
> tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> head: e22ce8eb631bdc47a4a4ea7ecf4e4ba499db4f93
> commit: ef75e73182949a94bde169a774de1b62ae21fbbc powerpc/signal32: Transform save_user_regs() and save_tm_user_regs() in 'unsafe' version
> date: 9 months ago
> config: powerpc-randconfig-c003-20210821 (attached as .config)
> compiler: clang version 14.0.0 (https://github.com/llvm/llvm-project d9c5613e856cf2addfbf892fc4c1ce9ef9feceaa)
> reproduce (this is a W=1 build):
> wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
> chmod +x ~/bin/make.cross
> # install powerpc cross compiling tool for clang build
> # apt-get install binutils-powerpc-linux-gnu
> # https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=ef75e73182949a94bde169a774de1b62ae21fbbc
> git remote add linus https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
> git fetch --no-tags linus master
> git checkout ef75e73182949a94bde169a774de1b62ae21fbbc
> # save the attached .config to linux build tree
> COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross ARCH=powerpc clang-analyzer
>
> If you fix the issue, kindly add following tag as appropriate
> Reported-by: kernel test robot <lkp@...el.com>
>
>
> clang-analyzer warnings: (new ones prefixed by >>)
> ^
>>> arch/powerpc/kernel/signal_32.c:297:2: warning: Value stored to 'msr' is never read [clang-analyzer-deadcode.DeadStores]
> msr &= ~MSR_VSX;
> ^ ~~~~~~~~
> arch/powerpc/kernel/signal_32.c:297:2: note: Value stored to 'msr' is never read
> msr &= ~MSR_VSX;
> ^ ~~~~~~~~
This is wrong.
msr is used at line 326:
ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 326
unsafe_put_user(msr, &frame->mc_gregs[PT_MSR], failed);
> arch/powerpc/kernel/signal_32.c:765:16: warning: Value stored to 'msr' during its initialization is never read [clang-analyzer-deadcode.DeadStores]
> unsigned long msr = regs->msr;
> ^~~ ~~~~~~~~~
> arch/powerpc/kernel/signal_32.c:765:16: note: Value stored to 'msr' during its initialization is never read
> unsigned long msr = regs->msr;
> ^~~ ~~~~~~~~~
> arch/powerpc/kernel/signal_32.c:869:16: warning: Value stored to 'msr' during its initialization is never read [clang-analyzer-deadcode.DeadStores]
> unsigned long msr = regs->msr;
> ^~~ ~~~~~~~~~
> arch/powerpc/kernel/signal_32.c:869:16: note: Value stored to 'msr' during its initialization is never read
> unsigned long msr = regs->msr;
> ^~~ ~~~~~~~~~
>
> vim +/msr +297 arch/powerpc/kernel/signal_32.c
>
> 968c4fccd1bb8b arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 262
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 263 static int save_user_regs_unsafe(struct pt_regs *regs, struct mcontext __user *frame,
> 8d33001dd650b8 arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 264 struct mcontext __user *tm_frame, int ctx_has_vsx_region)
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 265 {
> 9e7511861c4f8d arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 266 unsigned long msr = regs->msr;
> 9e7511861c4f8d arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 267
> c6e6771b87d4e3 arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 268 /* save general registers */
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 269 unsafe_save_general_regs(regs, frame, failed);
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 270
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 271 #ifdef CONFIG_ALTIVEC
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 272 /* save altivec registers */
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 273 if (current->thread.used_vr) {
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 274 unsafe_copy_to_user(&frame->mc_vregs, ¤t->thread.vr_state,
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 275 ELF_NVRREG * sizeof(vector128), failed);
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 276 /* set MSR_VEC in the saved MSR value to indicate that
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 277 frame->mc_vregs contains valid data */
> 9e7511861c4f8d arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 278 msr |= MSR_VEC;
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 279 }
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 280 /* else assert((regs->msr & MSR_VEC) == 0) */
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 281
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 282 /* We always copy to/from vrsave, it's 0 if we don't have or don't
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 283 * use altivec. Since VSCR only contains 32 bits saved in the least
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 284 * significant bits of a vector, we "cheat" and stuff VRSAVE in the
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 285 * most significant bits of that same vector. --BenH
> 408a7e08b2112f arch/powerpc/kernel/signal_32.c Paul Mackerras 2013-08-05 286 * Note that the current VRSAVE value is in the SPR at this point.
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 287 */
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 288 unsafe_put_user(current->thread.vrsave, (u32 __user *)&frame->mc_vregs[32],
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 289 failed);
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 290 #endif /* CONFIG_ALTIVEC */
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 291 unsafe_copy_fpr_to_user(&frame->mc_fregs, current, failed);
> ec67ad82814bee arch/powerpc/kernel/signal_32.c Michael Neuling 2013-11-25 292
> ec67ad82814bee arch/powerpc/kernel/signal_32.c Michael Neuling 2013-11-25 293 /*
> ec67ad82814bee arch/powerpc/kernel/signal_32.c Michael Neuling 2013-11-25 294 * Clear the MSR VSX bit to indicate there is no valid state attached
> ec67ad82814bee arch/powerpc/kernel/signal_32.c Michael Neuling 2013-11-25 295 * to this context, except in the specific case below where we set it.
> ec67ad82814bee arch/powerpc/kernel/signal_32.c Michael Neuling 2013-11-25 296 */
> ec67ad82814bee arch/powerpc/kernel/signal_32.c Michael Neuling 2013-11-25 @297 msr &= ~MSR_VSX;
> 6a274c08f2f4df arch/powerpc/kernel/signal_32.c Michael Neuling 2008-07-02 298 #ifdef CONFIG_VSX
> ce48b2100785e5 arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 299 /*
> ce48b2100785e5 arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 300 * Copy VSR 0-31 upper half from thread_struct to local
> ce48b2100785e5 arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 301 * buffer, then write that to userspace. Also set MSR_VSX in
> ce48b2100785e5 arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 302 * the saved MSR value to indicate that frame->mc_vregs
> ce48b2100785e5 arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 303 * contains valid data
> ce48b2100785e5 arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 304 */
> 16c29d180becc5 arch/powerpc/kernel/signal_32.c Michael Neuling 2008-10-23 305 if (current->thread.used_vsr && ctx_has_vsx_region) {
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 306 unsafe_copy_vsx_to_user(&frame->mc_vsregs, current, failed);
> ce48b2100785e5 arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 307 msr |= MSR_VSX;
> ec67ad82814bee arch/powerpc/kernel/signal_32.c Michael Neuling 2013-11-25 308 }
> c6e6771b87d4e3 arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 309 #endif /* CONFIG_VSX */
> 81e7009ea46c95 arch/powerpc/kernel/signal_32.c Stephen Rothwell 2005-10-18 310 #ifdef CONFIG_SPE
> 81e7009ea46c95 arch/powerpc/kernel/signal_32.c Stephen Rothwell 2005-10-18 311 /* save spe registers */
> 81e7009ea46c95 arch/powerpc/kernel/signal_32.c Stephen Rothwell 2005-10-18 312 if (current->thread.used_spe) {
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 313 unsafe_copy_to_user(&frame->mc_vregs, current->thread.evr,
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 314 ELF_NEVRREG * sizeof(u32), failed);
> 81e7009ea46c95 arch/powerpc/kernel/signal_32.c Stephen Rothwell 2005-10-18 315 /* set MSR_SPE in the saved MSR value to indicate that
> 81e7009ea46c95 arch/powerpc/kernel/signal_32.c Stephen Rothwell 2005-10-18 316 frame->mc_vregs contains valid data */
> 9e7511861c4f8d arch/powerpc/kernel/signal_32.c Michael Neuling 2008-06-25 317 msr |= MSR_SPE;
> 81e7009ea46c95 arch/powerpc/kernel/signal_32.c Stephen Rothwell 2005-10-18 318 }
> 81e7009ea46c95 arch/powerpc/kernel/signal_32.c Stephen Rothwell 2005-10-18 319 /* else assert((regs->msr & MSR_SPE) == 0) */
> 81e7009ea46c95 arch/powerpc/kernel/signal_32.c Stephen Rothwell 2005-10-18 320
> 81e7009ea46c95 arch/powerpc/kernel/signal_32.c Stephen Rothwell 2005-10-18 321 /* We always copy to/from spefscr */
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 322 unsafe_put_user(current->thread.spefscr,
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 323 (u32 __user *)&frame->mc_vregs + ELF_NEVRREG, failed);
> 81e7009ea46c95 arch/powerpc/kernel/signal_32.c Stephen Rothwell 2005-10-18 324 #endif /* CONFIG_SPE */
> 81e7009ea46c95 arch/powerpc/kernel/signal_32.c Stephen Rothwell 2005-10-18 325
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 326 unsafe_put_user(msr, &frame->mc_gregs[PT_MSR], failed);
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 327
> 1d25f11fdbcc53 arch/powerpc/kernel/signal_32.c Michael Neuling 2013-06-09 328 /* We need to write 0 the MSR top 32 bits in the tm frame so that we
> 1d25f11fdbcc53 arch/powerpc/kernel/signal_32.c Michael Neuling 2013-06-09 329 * can check it on the restore to see if TM is active
> 1d25f11fdbcc53 arch/powerpc/kernel/signal_32.c Michael Neuling 2013-06-09 330 */
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 331 if (tm_frame)
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 332 unsafe_put_user(0, &tm_frame->mc_gregs[PT_MSR], failed);
> 1d25f11fdbcc53 arch/powerpc/kernel/signal_32.c Michael Neuling 2013-06-09 333
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 334 return 0;
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 335
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 336 failed:
> ef75e73182949a arch/powerpc/kernel/signal_32.c Christophe Leroy 2020-08-18 337 return 1;
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 338 }
> ^1da177e4c3f41 arch/ppc64/kernel/signal32.c Linus Torvalds 2005-04-16 339
>
> :::::: The code at line 297 was first introduced by commit
> :::::: ec67ad82814bee92251fd963bf01c7a173856555 powerpc/signals: Improved mark VSX not saved with small contexts fix
>
> :::::: TO: Michael Neuling <mikey@...ling.org>
> :::::: CC: Benjamin Herrenschmidt <benh@...nel.crashing.org>
>
> ---
> 0-DAY CI Kernel Test Service, Intel Corporation
> https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
>
Powered by blists - more mailing lists