[linux-next:master 7646/8170] arch/powerpc/kexec/crash.c:488:2-8: WARNING: NULL check before some freeing functions is not needed.

[linux-next:master 7646/8170] arch/powerpc/kexec/crash.c:488:2-8: WARNING: NULL check before some freeing functions is not needed.

[kernel test robot]

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
head:   5e4f84f18c4ee9b0ccdc19e39b7de41df21699dd
commit: 849599b702ef8977fcd5b2f27c61ef773c42bb88 [7646/8170] powerpc/crash: add crash memory hotplug support
config: powerpc-randconfig-r061-20240426 (https://download.01.org/0day-ci/archive/20240426/202404261048.skfV5DDB-lkp@intel.com/config)
compiler: powerpc-linux-gcc (GCC) 13.2.0

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/202404261048.skfV5DDB-lkp@intel.com/

cocci warnings: (new ones prefixed by >>)
>> arch/powerpc/kexec/crash.c:488:2-8: WARNING: NULL check before some freeing functions is not needed.

vim +488 arch/powerpc/kexec/crash.c

   418	
   419	/**
   420	 * update_crash_elfcorehdr() - Recreate the elfcorehdr and replace it with old
   421	 *			       elfcorehdr in the kexec segment array.
   422	 * @image: the active struct kimage
   423	 * @mn: struct memory_notify data handler
   424	 */
   425	static void update_crash_elfcorehdr(struct kimage *image, struct memory_notify *mn)
   426	{
   427		int ret;
   428		struct crash_mem *cmem = NULL;
   429		struct kexec_segment *ksegment;
   430		void *ptr, *mem, *elfbuf = NULL;
   431		unsigned long elfsz, memsz, base_addr, size;
   432	
   433		ksegment = &image->segment[image->elfcorehdr_index];
   434		mem = (void *) ksegment->mem;
   435		memsz = ksegment->memsz;
   436	
   437		ret = get_crash_memory_ranges(&cmem);
   438		if (ret) {
   439			pr_err("Failed to get crash mem range\n");
   440			return;
   441		}
   442	
   443		/*
   444		 * The hot unplugged memory is part of crash memory ranges,
   445		 * remove it here.
   446		 */
   447		if (image->hp_action == KEXEC_CRASH_HP_REMOVE_MEMORY) {
   448			base_addr = PFN_PHYS(mn->start_pfn);
   449			size = mn->nr_pages * PAGE_SIZE;
   450			ret = remove_mem_range(&cmem, base_addr, size);
   451			if (ret) {
   452				pr_err("Failed to remove hot-unplugged memory from crash memory ranges\n");
   453				goto out;
   454			}
   455		}
   456	
   457		ret = crash_prepare_elf64_headers(cmem, false, &elfbuf, &elfsz);
   458		if (ret) {
   459			pr_err("Failed to prepare elf header\n");
   460			goto out;
   461		}
   462	
   463		/*
   464		 * It is unlikely that kernel hit this because elfcorehdr kexec
   465		 * segment (memsz) is built with addition space to accommodate growing
   466		 * number of crash memory ranges while loading the kdump kernel. It is
   467		 * Just to avoid any unforeseen case.
   468		 */
   469		if (elfsz > memsz) {
   470			pr_err("Updated crash elfcorehdr elfsz %lu > memsz %lu", elfsz, memsz);
   471			goto out;
   472		}
   473	
   474		ptr = __va(mem);
   475		if (ptr) {
   476			/* Temporarily invalidate the crash image while it is replaced */
   477			xchg(&kexec_crash_image, NULL);
   478	
   479			/* Replace the old elfcorehdr with newly prepared elfcorehdr */
   480			memcpy((void *)ptr, elfbuf, elfsz);
   481	
   482			/* The crash image is now valid once again */
   483			xchg(&kexec_crash_image, image);
   484		}
   485	out:
   486		kvfree(cmem);
   487		if (elfbuf)
 > 488			kvfree(elfbuf);
   489	}
   490	

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

Re: [linux-next:master 7646/8170] arch/powerpc/kexec/crash.c:488:2-8: WARNING: NULL check before some freeing functions is not needed.

[Sourabh Jain]

On 26/04/24 07:57, kernel test robot wrote:
> tree:   https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
> head:   5e4f84f18c4ee9b0ccdc19e39b7de41df21699dd
> commit: 849599b702ef8977fcd5b2f27c61ef773c42bb88 [7646/8170] powerpc/crash: add crash memory hotplug support
> config: powerpc-randconfig-r061-20240426 (https://download.01.org/0day-ci/archive/20240426/202404261048.skfV5DDB-lkp@intel.com/config)
> compiler: powerpc-linux-gcc (GCC) 13.2.0
>
> 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/202404261048.skfV5DDB-lkp@intel.com/
>
> cocci warnings: (new ones prefixed by >>)
>>> arch/powerpc/kexec/crash.c:488:2-8: WARNING: NULL check before some freeing functions is not needed.
> vim +488 arch/powerpc/kexec/crash.c
>
>     418	
>     419	/**
>     420	 * update_crash_elfcorehdr() - Recreate the elfcorehdr and replace it with old
>     421	 *			       elfcorehdr in the kexec segment array.
>     422	 * @image: the active struct kimage
>     423	 * @mn: struct memory_notify data handler
>     424	 */
>     425	static void update_crash_elfcorehdr(struct kimage *image, struct memory_notify *mn)
>     426	{
>     427		int ret;
>     428		struct crash_mem *cmem = NULL;
>     429		struct kexec_segment *ksegment;
>     430		void *ptr, *mem, *elfbuf = NULL;
>     431		unsigned long elfsz, memsz, base_addr, size;
>     432	
>     433		ksegment = &image->segment[image->elfcorehdr_index];
>     434		mem = (void *) ksegment->mem;
>     435		memsz = ksegment->memsz;
>     436	
>     437		ret = get_crash_memory_ranges(&cmem);
>     438		if (ret) {
>     439			pr_err("Failed to get crash mem range\n");
>     440			return;
>     441		}
>     442	
>     443		/*
>     444		 * The hot unplugged memory is part of crash memory ranges,
>     445		 * remove it here.
>     446		 */
>     447		if (image->hp_action == KEXEC_CRASH_HP_REMOVE_MEMORY) {
>     448			base_addr = PFN_PHYS(mn->start_pfn);
>     449			size = mn->nr_pages * PAGE_SIZE;
>     450			ret = remove_mem_range(&cmem, base_addr, size);
>     451			if (ret) {
>     452				pr_err("Failed to remove hot-unplugged memory from crash memory ranges\n");
>     453				goto out;
>     454			}
>     455		}
>     456	
>     457		ret = crash_prepare_elf64_headers(cmem, false, &elfbuf, &elfsz);
>     458		if (ret) {
>     459			pr_err("Failed to prepare elf header\n");
>     460			goto out;
>     461		}
>     462	
>     463		/*
>     464		 * It is unlikely that kernel hit this because elfcorehdr kexec
>     465		 * segment (memsz) is built with addition space to accommodate growing
>     466		 * number of crash memory ranges while loading the kdump kernel. It is
>     467		 * Just to avoid any unforeseen case.
>     468		 */
>     469		if (elfsz > memsz) {
>     470			pr_err("Updated crash elfcorehdr elfsz %lu > memsz %lu", elfsz, memsz);
>     471			goto out;
>     472		}
>     473	
>     474		ptr = __va(mem);
>     475		if (ptr) {
>     476			/* Temporarily invalidate the crash image while it is replaced */
>     477			xchg(&kexec_crash_image, NULL);
>     478	
>     479			/* Replace the old elfcorehdr with newly prepared elfcorehdr */
>     480			memcpy((void *)ptr, elfbuf, elfsz);
>     481	
>     482			/* The crash image is now valid once again */
>     483			xchg(&kexec_crash_image, image);
>     484		}
>     485	out:
>     486		kvfree(cmem);
>     487		if (elfbuf)
>   > 488			kvfree(elfbuf);
>     489	}
>     490	
Thanks for the report.

I will fix this and send next version.

- Sourabh Jain

Re: [linux-next:master 7646/8170] arch/powerpc/kexec/crash.c:488:2-8: WARNING: NULL check before some freeing functions is not needed.

[Sourabh Jain]

Hello Michael and Stephen,

I fixed the below build warning in v19:

v19 patch series link: 
https://lore.kernel.org/all/20240426060728.559753-1-sourabhjain@linux.ibm.com/

Only 6/6 patch has been adjusted; all other patches remain unchanged.

Please consider the updated patch series. Apologies for inconvenience.

Thanks,
Sourabh Jain


On 26/04/24 07:57, kernel test robot wrote:
> tree:   https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
> head:   5e4f84f18c4ee9b0ccdc19e39b7de41df21699dd
> commit: 849599b702ef8977fcd5b2f27c61ef773c42bb88 [7646/8170] powerpc/crash: add crash memory hotplug support
> config: powerpc-randconfig-r061-20240426 (https://download.01.org/0day-ci/archive/20240426/202404261048.skfV5DDB-lkp@intel.com/config)
> compiler: powerpc-linux-gcc (GCC) 13.2.0
>
> 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/202404261048.skfV5DDB-lkp@intel.com/
>
> cocci warnings: (new ones prefixed by >>)
>>> arch/powerpc/kexec/crash.c:488:2-8: WARNING: NULL check before some freeing functions is not needed.
> vim +488 arch/powerpc/kexec/crash.c
>
>     418	
>     419	/**
>     420	 * update_crash_elfcorehdr() - Recreate the elfcorehdr and replace it with old
>     421	 *			       elfcorehdr in the kexec segment array.
>     422	 * @image: the active struct kimage
>     423	 * @mn: struct memory_notify data handler
>     424	 */
>     425	static void update_crash_elfcorehdr(struct kimage *image, struct memory_notify *mn)
>     426	{
>     427		int ret;
>     428		struct crash_mem *cmem = NULL;
>     429		struct kexec_segment *ksegment;
>     430		void *ptr, *mem, *elfbuf = NULL;
>     431		unsigned long elfsz, memsz, base_addr, size;
>     432	
>     433		ksegment = &image->segment[image->elfcorehdr_index];
>     434		mem = (void *) ksegment->mem;
>     435		memsz = ksegment->memsz;
>     436	
>     437		ret = get_crash_memory_ranges(&cmem);
>     438		if (ret) {
>     439			pr_err("Failed to get crash mem range\n");
>     440			return;
>     441		}
>     442	
>     443		/*
>     444		 * The hot unplugged memory is part of crash memory ranges,
>     445		 * remove it here.
>     446		 */
>     447		if (image->hp_action == KEXEC_CRASH_HP_REMOVE_MEMORY) {
>     448			base_addr = PFN_PHYS(mn->start_pfn);
>     449			size = mn->nr_pages * PAGE_SIZE;
>     450			ret = remove_mem_range(&cmem, base_addr, size);
>     451			if (ret) {
>     452				pr_err("Failed to remove hot-unplugged memory from crash memory ranges\n");
>     453				goto out;
>     454			}
>     455		}
>     456	
>     457		ret = crash_prepare_elf64_headers(cmem, false, &elfbuf, &elfsz);
>     458		if (ret) {
>     459			pr_err("Failed to prepare elf header\n");
>     460			goto out;
>     461		}
>     462	
>     463		/*
>     464		 * It is unlikely that kernel hit this because elfcorehdr kexec
>     465		 * segment (memsz) is built with addition space to accommodate growing
>     466		 * number of crash memory ranges while loading the kdump kernel. It is
>     467		 * Just to avoid any unforeseen case.
>     468		 */
>     469		if (elfsz > memsz) {
>     470			pr_err("Updated crash elfcorehdr elfsz %lu > memsz %lu", elfsz, memsz);
>     471			goto out;
>     472		}
>     473	
>     474		ptr = __va(mem);
>     475		if (ptr) {
>     476			/* Temporarily invalidate the crash image while it is replaced */
>     477			xchg(&kexec_crash_image, NULL);
>     478	
>     479			/* Replace the old elfcorehdr with newly prepared elfcorehdr */
>     480			memcpy((void *)ptr, elfbuf, elfsz);
>     481	
>     482			/* The crash image is now valid once again */
>     483			xchg(&kexec_crash_image, image);
>     484		}
>     485	out:
>     486		kvfree(cmem);
>     487		if (elfbuf)
>   > 488			kvfree(elfbuf);
>     489	}
>     490	

Re: [linux-next:master 7646/8170] arch/powerpc/kexec/crash.c:488:2-8: WARNING: NULL check before some freeing functions is not needed.

[Michael Ellerman]

Sourabh Jain <sourabhjain@linux.ibm.com> writes:
> Hello Michael and Stephen,
>
> I fixed the below build warning in v19:
>
> v19 patch series link: 
> https://lore.kernel.org/all/20240426060728.559753-1-sourabhjain@linux.ibm.com/
>
> Only 6/6 patch has been adjusted; all other patches remain unchanged.

I don't want to rebase the branch just for that tiny change.

If you want to fix it, send an incremental patch on top of v18.

cheers

Re: [linux-next:master 7646/8170] arch/powerpc/kexec/crash.c:488:2-8: WARNING: NULL check before some freeing functions is not needed.

[Sourabh Jain]

Hello Michael and Stephen,

On 02/05/24 18:53, Michael Ellerman wrote:
> Sourabh Jain <sourabhjain@linux.ibm.com> writes:
>> Hello Michael and Stephen,
>>
>> I fixed the below build warning in v19:
>>
>> v19 patch series link:
>> https://lore.kernel.org/all/20240426060728.559753-1-sourabhjain@linux.ibm.com/
>>
>> Only 6/6 patch has been adjusted; all other patches remain unchanged.
> I don't want to rebase the branch just for that tiny change.
>
> If you want to fix it, send an incremental patch on top of v18.

The below patch:
https://lore.kernel.org/all/20240502182040.774759-1-sourabhjain@linux.ibm.com/

fixes the following build warning:
https://lore.kernel.org/oe-kbuild-all/202404261048.skfV5DDB-lkp@intel.com/

The fix patch is rebased on the linux-next/master branch.

Thanks,
Sourabh Jain