* [PATCH 1/1] alloc_tag: avoid execmem_vmap() when !MMU
@ 2024-10-28 20:29 Suren Baghdasaryan
2024-10-29 17:09 ` Mike Rapoport
0 siblings, 1 reply; 5+ messages in thread
From: Suren Baghdasaryan @ 2024-10-28 20:29 UTC (permalink / raw)
To: akpm
Cc: arnd, arnd, rppt, pasha.tatashin, mcgrof, song, mhiramat,
linux-mm, linux-kernel, surenb, kernel test robot
With CONFIG_MMU=n __get_vm_area_node() is not available and memory for the
allocation tags cannot be populated as needed. For this case, populate the
required memory at initialization time.
Fixes: 57bc3834fb6f ("alloc_tag: populate memory for module tags as needed")
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202410250808.dQGyYjlk-lkp@intel.com/
Closes: https://lore.kernel.org/oe-lkp/202410251525.9f85854d-oliver.sang@intel.com
Closes: https://lore.kernel.org/oe-kbuild-all/202410261016.IO7C6Cml-lkp@intel.com/
Closes: https://lore.kernel.org/oe-kbuild-all/202410270919.LebQlmxD-lkp@intel.com/
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Mike Rapoport (Microsoft) <rppt@kernel.org>
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
---
include/linux/execmem.h | 2 ++
lib/alloc_tag.c | 23 ++++++++++++++++++++++-
mm/execmem.c | 32 ++++++++++++++++----------------
3 files changed, 40 insertions(+), 17 deletions(-)
diff --git a/include/linux/execmem.h b/include/linux/execmem.h
index 5a5e2917f870..64130ae19690 100644
--- a/include/linux/execmem.h
+++ b/include/linux/execmem.h
@@ -139,6 +139,7 @@ void *execmem_alloc(enum execmem_type type, size_t size);
*/
void execmem_free(void *ptr);
+#ifdef CONFIG_MMU
/**
* execmem_vmap - create virtual mapping for EXECMEM_MODULE_DATA memory
* @size: size of the virtual mapping in bytes
@@ -148,6 +149,7 @@ void execmem_free(void *ptr);
* Return: the area descriptor on success or %NULL on failure.
*/
struct vm_struct *execmem_vmap(size_t size);
+#endif
/**
* execmem_update_copy - copy an update to executable memory
diff --git a/lib/alloc_tag.c b/lib/alloc_tag.c
index c1ddac2d29f0..1c74942e6dfd 100644
--- a/lib/alloc_tag.c
+++ b/lib/alloc_tag.c
@@ -268,7 +268,6 @@ void __init alloc_tag_sec_init(void)
#ifdef CONFIG_MODULES
static struct maple_tree mod_area_mt = MTREE_INIT(mod_area_mt, MT_FLAGS_ALLOC_RANGE);
-static struct vm_struct *vm_module_tags;
/* A dummy object used to indicate an unloaded module */
static struct module unloaded_mod;
/* A dummy object used to indicate a module prepended area */
@@ -391,6 +390,9 @@ static bool find_aligned_area(struct ma_state *mas, unsigned long section_size,
return false;
}
+#ifdef CONFIG_MMU
+static struct vm_struct *vm_module_tags;
+
static int vm_module_tags_populate(void)
{
unsigned long phys_size = vm_module_tags->nr_pages << PAGE_SHIFT;
@@ -417,6 +419,13 @@ static int vm_module_tags_populate(void)
return 0;
}
+#else
+static int vm_module_tags_populate(void)
+{
+ /* Memory was already allocated */
+ return 0;
+}
+#endif
static void *reserve_module_tags(struct module *mod, unsigned long size,
unsigned int prepend, unsigned long align)
@@ -561,6 +570,7 @@ static void replace_module(struct module *mod, struct module *new_mod)
static int __init alloc_mod_tags_mem(void)
{
+#ifdef CONFIG_MMU
/* Map space to copy allocation tags */
vm_module_tags = execmem_vmap(MODULE_ALLOC_TAG_VMAP_SIZE);
if (!vm_module_tags) {
@@ -578,6 +588,13 @@ static int __init alloc_mod_tags_mem(void)
}
module_tags.start_addr = (unsigned long)vm_module_tags->addr;
+#else
+ /* Allocate space to copy allocation tags */
+ module_tags.start_addr = (unsigned long)execmem_alloc(EXECMEM_MODULE_DATA,
+ MODULE_ALLOC_TAG_VMAP_SIZE);
+ if (!module_tags.start_addr)
+ return -ENOMEM;
+#endif
module_tags.end_addr = module_tags.start_addr + MODULE_ALLOC_TAG_VMAP_SIZE;
/* Ensure the base is alloc_tag aligned when required for indexing */
module_tags.start_addr = alloc_tag_align(module_tags.start_addr);
@@ -587,6 +604,7 @@ static int __init alloc_mod_tags_mem(void)
static void __init free_mod_tags_mem(void)
{
+#ifdef CONFIG_MMU
int i;
module_tags.start_addr = 0;
@@ -594,6 +612,9 @@ static void __init free_mod_tags_mem(void)
__free_page(vm_module_tags->pages[i]);
kfree(vm_module_tags->pages);
free_vm_area(vm_module_tags);
+#else
+ execmem_free((void *)module_tags.start_addr);
+#endif
}
#else /* CONFIG_MODULES */
diff --git a/mm/execmem.c b/mm/execmem.c
index 5c0f9f2d6f83..317b6a8d35be 100644
--- a/mm/execmem.c
+++ b/mm/execmem.c
@@ -64,6 +64,22 @@ static void *execmem_vmalloc(struct execmem_range *range, size_t size,
return p;
}
+
+struct vm_struct *execmem_vmap(size_t size)
+{
+ struct execmem_range *range = &execmem_info->ranges[EXECMEM_MODULE_DATA];
+ struct vm_struct *area;
+
+ area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC,
+ range->start, range->end, NUMA_NO_NODE,
+ GFP_KERNEL, __builtin_return_address(0));
+ if (!area && range->fallback_start)
+ area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC,
+ range->fallback_start, range->fallback_end,
+ NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0));
+
+ return area;
+}
#else
static void *execmem_vmalloc(struct execmem_range *range, size_t size,
pgprot_t pgprot, unsigned long vm_flags)
@@ -368,22 +384,6 @@ void execmem_free(void *ptr)
vfree(ptr);
}
-struct vm_struct *execmem_vmap(size_t size)
-{
- struct execmem_range *range = &execmem_info->ranges[EXECMEM_MODULE_DATA];
- struct vm_struct *area;
-
- area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC,
- range->start, range->end, NUMA_NO_NODE,
- GFP_KERNEL, __builtin_return_address(0));
- if (!area && range->fallback_start)
- area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC,
- range->fallback_start, range->fallback_end,
- NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0));
-
- return area;
-}
-
void *execmem_update_copy(void *dst, const void *src, size_t size)
{
return text_poke_copy(dst, src, size);
base-commit: 92409df3e00a17d6fd5bb27732fa186749725ed5
--
2.47.0.163.g1226f6d8fa-goog
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH 1/1] alloc_tag: avoid execmem_vmap() when !MMU 2024-10-28 20:29 [PATCH 1/1] alloc_tag: avoid execmem_vmap() when !MMU Suren Baghdasaryan @ 2024-10-29 17:09 ` Mike Rapoport 2024-10-29 20:03 ` Suren Baghdasaryan 0 siblings, 1 reply; 5+ messages in thread From: Mike Rapoport @ 2024-10-29 17:09 UTC (permalink / raw) To: Suren Baghdasaryan Cc: akpm, arnd, arnd, pasha.tatashin, mcgrof, song, mhiramat, linux-mm, linux-kernel, kernel test robot Hi Suren, On Mon, Oct 28, 2024 at 01:29:35PM -0700, Suren Baghdasaryan wrote: > With CONFIG_MMU=n __get_vm_area_node() is not available and memory for the > allocation tags cannot be populated as needed. For this case, populate the > required memory at initialization time. Do we care that much to have alloc_tags on !MMU configs? Your patch is great, but it would have been a lot easier to make MEM_ALLOC_PROFILING depend on MMU. > Fixes: 57bc3834fb6f ("alloc_tag: populate memory for module tags as needed") > Reported-by: kernel test robot <lkp@intel.com> > Closes: https://lore.kernel.org/oe-kbuild-all/202410250808.dQGyYjlk-lkp@intel.com/ > Closes: https://lore.kernel.org/oe-lkp/202410251525.9f85854d-oliver.sang@intel.com > Closes: https://lore.kernel.org/oe-kbuild-all/202410261016.IO7C6Cml-lkp@intel.com/ > Closes: https://lore.kernel.org/oe-kbuild-all/202410270919.LebQlmxD-lkp@intel.com/ > Cc: Arnd Bergmann <arnd@arndb.de> > Cc: Mike Rapoport (Microsoft) <rppt@kernel.org> > Signed-off-by: Suren Baghdasaryan <surenb@google.com> > --- > include/linux/execmem.h | 2 ++ > lib/alloc_tag.c | 23 ++++++++++++++++++++++- > mm/execmem.c | 32 ++++++++++++++++---------------- > 3 files changed, 40 insertions(+), 17 deletions(-) > > diff --git a/include/linux/execmem.h b/include/linux/execmem.h > index 5a5e2917f870..64130ae19690 100644 > --- a/include/linux/execmem.h > +++ b/include/linux/execmem.h > @@ -139,6 +139,7 @@ void *execmem_alloc(enum execmem_type type, size_t size); > */ > void execmem_free(void *ptr); > > +#ifdef CONFIG_MMU > /** > * execmem_vmap - create virtual mapping for EXECMEM_MODULE_DATA memory > * @size: size of the virtual mapping in bytes > @@ -148,6 +149,7 @@ void execmem_free(void *ptr); > * Return: the area descriptor on success or %NULL on failure. > */ > struct vm_struct *execmem_vmap(size_t size); > +#endif > > /** > * execmem_update_copy - copy an update to executable memory > diff --git a/lib/alloc_tag.c b/lib/alloc_tag.c > index c1ddac2d29f0..1c74942e6dfd 100644 > --- a/lib/alloc_tag.c > +++ b/lib/alloc_tag.c > @@ -268,7 +268,6 @@ void __init alloc_tag_sec_init(void) > #ifdef CONFIG_MODULES > > static struct maple_tree mod_area_mt = MTREE_INIT(mod_area_mt, MT_FLAGS_ALLOC_RANGE); > -static struct vm_struct *vm_module_tags; > /* A dummy object used to indicate an unloaded module */ > static struct module unloaded_mod; > /* A dummy object used to indicate a module prepended area */ > @@ -391,6 +390,9 @@ static bool find_aligned_area(struct ma_state *mas, unsigned long section_size, > return false; > } > > +#ifdef CONFIG_MMU > +static struct vm_struct *vm_module_tags; > + > static int vm_module_tags_populate(void) > { > unsigned long phys_size = vm_module_tags->nr_pages << PAGE_SHIFT; > @@ -417,6 +419,13 @@ static int vm_module_tags_populate(void) > > return 0; > } > +#else > +static int vm_module_tags_populate(void) > +{ > + /* Memory was already allocated */ > + return 0; > +} > +#endif > > static void *reserve_module_tags(struct module *mod, unsigned long size, > unsigned int prepend, unsigned long align) > @@ -561,6 +570,7 @@ static void replace_module(struct module *mod, struct module *new_mod) > > static int __init alloc_mod_tags_mem(void) > { > +#ifdef CONFIG_MMU > /* Map space to copy allocation tags */ > vm_module_tags = execmem_vmap(MODULE_ALLOC_TAG_VMAP_SIZE); > if (!vm_module_tags) { > @@ -578,6 +588,13 @@ static int __init alloc_mod_tags_mem(void) > } > > module_tags.start_addr = (unsigned long)vm_module_tags->addr; > +#else > + /* Allocate space to copy allocation tags */ > + module_tags.start_addr = (unsigned long)execmem_alloc(EXECMEM_MODULE_DATA, > + MODULE_ALLOC_TAG_VMAP_SIZE); > + if (!module_tags.start_addr) > + return -ENOMEM; > +#endif > module_tags.end_addr = module_tags.start_addr + MODULE_ALLOC_TAG_VMAP_SIZE; > /* Ensure the base is alloc_tag aligned when required for indexing */ > module_tags.start_addr = alloc_tag_align(module_tags.start_addr); > @@ -587,6 +604,7 @@ static int __init alloc_mod_tags_mem(void) > > static void __init free_mod_tags_mem(void) > { > +#ifdef CONFIG_MMU > int i; > > module_tags.start_addr = 0; > @@ -594,6 +612,9 @@ static void __init free_mod_tags_mem(void) > __free_page(vm_module_tags->pages[i]); > kfree(vm_module_tags->pages); > free_vm_area(vm_module_tags); > +#else > + execmem_free((void *)module_tags.start_addr); > +#endif > } > > #else /* CONFIG_MODULES */ > diff --git a/mm/execmem.c b/mm/execmem.c > index 5c0f9f2d6f83..317b6a8d35be 100644 > --- a/mm/execmem.c > +++ b/mm/execmem.c > @@ -64,6 +64,22 @@ static void *execmem_vmalloc(struct execmem_range *range, size_t size, > > return p; > } > + > +struct vm_struct *execmem_vmap(size_t size) > +{ > + struct execmem_range *range = &execmem_info->ranges[EXECMEM_MODULE_DATA]; > + struct vm_struct *area; > + > + area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > + range->start, range->end, NUMA_NO_NODE, > + GFP_KERNEL, __builtin_return_address(0)); > + if (!area && range->fallback_start) > + area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > + range->fallback_start, range->fallback_end, > + NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0)); > + > + return area; > +} > #else > static void *execmem_vmalloc(struct execmem_range *range, size_t size, > pgprot_t pgprot, unsigned long vm_flags) > @@ -368,22 +384,6 @@ void execmem_free(void *ptr) > vfree(ptr); > } > > -struct vm_struct *execmem_vmap(size_t size) > -{ > - struct execmem_range *range = &execmem_info->ranges[EXECMEM_MODULE_DATA]; > - struct vm_struct *area; > - > - area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > - range->start, range->end, NUMA_NO_NODE, > - GFP_KERNEL, __builtin_return_address(0)); > - if (!area && range->fallback_start) > - area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > - range->fallback_start, range->fallback_end, > - NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0)); > - > - return area; > -} > - > void *execmem_update_copy(void *dst, const void *src, size_t size) > { > return text_poke_copy(dst, src, size); > > base-commit: 92409df3e00a17d6fd5bb27732fa186749725ed5 > -- > 2.47.0.163.g1226f6d8fa-goog > -- Sincerely yours, Mike. ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 1/1] alloc_tag: avoid execmem_vmap() when !MMU 2024-10-29 17:09 ` Mike Rapoport @ 2024-10-29 20:03 ` Suren Baghdasaryan 2024-10-30 12:43 ` Mike Rapoport 0 siblings, 1 reply; 5+ messages in thread From: Suren Baghdasaryan @ 2024-10-29 20:03 UTC (permalink / raw) To: Mike Rapoport Cc: akpm, arnd, arnd, pasha.tatashin, mcgrof, song, mhiramat, linux-mm, linux-kernel, kernel test robot On Tue, Oct 29, 2024 at 10:13 AM Mike Rapoport <rppt@kernel.org> wrote: > > Hi Suren, > > On Mon, Oct 28, 2024 at 01:29:35PM -0700, Suren Baghdasaryan wrote: > > With CONFIG_MMU=n __get_vm_area_node() is not available and memory for the > > allocation tags cannot be populated as needed. For this case, populate the > > required memory at initialization time. > > Do we care that much to have alloc_tags on !MMU configs? > > Your patch is great, but it would have been a lot easier to make > MEM_ALLOC_PROFILING depend on MMU. I didn't want to prevent !MMU users from using allocation profiling but I'm not aware of any current !MMU users. Therefore I have no objections to making MEM_ALLOC_PROFILING dependent on MMU if we indeed don't care about !MMU && MEM_ALLOC_PROFILING configuration. I'll be posting v5 tomorrow with all these fixes, so please let me know which way you prefer. Thanks! > > > Fixes: 57bc3834fb6f ("alloc_tag: populate memory for module tags as needed") > > Reported-by: kernel test robot <lkp@intel.com> > > Closes: https://lore.kernel.org/oe-kbuild-all/202410250808.dQGyYjlk-lkp@intel.com/ > > Closes: https://lore.kernel.org/oe-lkp/202410251525.9f85854d-oliver.sang@intel.com > > Closes: https://lore.kernel.org/oe-kbuild-all/202410261016.IO7C6Cml-lkp@intel.com/ > > Closes: https://lore.kernel.org/oe-kbuild-all/202410270919.LebQlmxD-lkp@intel.com/ > > Cc: Arnd Bergmann <arnd@arndb.de> > > Cc: Mike Rapoport (Microsoft) <rppt@kernel.org> > > Signed-off-by: Suren Baghdasaryan <surenb@google.com> > > --- > > include/linux/execmem.h | 2 ++ > > lib/alloc_tag.c | 23 ++++++++++++++++++++++- > > mm/execmem.c | 32 ++++++++++++++++---------------- > > 3 files changed, 40 insertions(+), 17 deletions(-) > > > > diff --git a/include/linux/execmem.h b/include/linux/execmem.h > > index 5a5e2917f870..64130ae19690 100644 > > --- a/include/linux/execmem.h > > +++ b/include/linux/execmem.h > > @@ -139,6 +139,7 @@ void *execmem_alloc(enum execmem_type type, size_t size); > > */ > > void execmem_free(void *ptr); > > > > +#ifdef CONFIG_MMU > > /** > > * execmem_vmap - create virtual mapping for EXECMEM_MODULE_DATA memory > > * @size: size of the virtual mapping in bytes > > @@ -148,6 +149,7 @@ void execmem_free(void *ptr); > > * Return: the area descriptor on success or %NULL on failure. > > */ > > struct vm_struct *execmem_vmap(size_t size); > > +#endif > > > > /** > > * execmem_update_copy - copy an update to executable memory > > diff --git a/lib/alloc_tag.c b/lib/alloc_tag.c > > index c1ddac2d29f0..1c74942e6dfd 100644 > > --- a/lib/alloc_tag.c > > +++ b/lib/alloc_tag.c > > @@ -268,7 +268,6 @@ void __init alloc_tag_sec_init(void) > > #ifdef CONFIG_MODULES > > > > static struct maple_tree mod_area_mt = MTREE_INIT(mod_area_mt, MT_FLAGS_ALLOC_RANGE); > > -static struct vm_struct *vm_module_tags; > > /* A dummy object used to indicate an unloaded module */ > > static struct module unloaded_mod; > > /* A dummy object used to indicate a module prepended area */ > > @@ -391,6 +390,9 @@ static bool find_aligned_area(struct ma_state *mas, unsigned long section_size, > > return false; > > } > > > > +#ifdef CONFIG_MMU > > +static struct vm_struct *vm_module_tags; > > + > > static int vm_module_tags_populate(void) > > { > > unsigned long phys_size = vm_module_tags->nr_pages << PAGE_SHIFT; > > @@ -417,6 +419,13 @@ static int vm_module_tags_populate(void) > > > > return 0; > > } > > +#else > > +static int vm_module_tags_populate(void) > > +{ > > + /* Memory was already allocated */ > > + return 0; > > +} > > +#endif > > > > static void *reserve_module_tags(struct module *mod, unsigned long size, > > unsigned int prepend, unsigned long align) > > @@ -561,6 +570,7 @@ static void replace_module(struct module *mod, struct module *new_mod) > > > > static int __init alloc_mod_tags_mem(void) > > { > > +#ifdef CONFIG_MMU > > /* Map space to copy allocation tags */ > > vm_module_tags = execmem_vmap(MODULE_ALLOC_TAG_VMAP_SIZE); > > if (!vm_module_tags) { > > @@ -578,6 +588,13 @@ static int __init alloc_mod_tags_mem(void) > > } > > > > module_tags.start_addr = (unsigned long)vm_module_tags->addr; > > +#else > > + /* Allocate space to copy allocation tags */ > > + module_tags.start_addr = (unsigned long)execmem_alloc(EXECMEM_MODULE_DATA, > > + MODULE_ALLOC_TAG_VMAP_SIZE); > > + if (!module_tags.start_addr) > > + return -ENOMEM; > > +#endif > > module_tags.end_addr = module_tags.start_addr + MODULE_ALLOC_TAG_VMAP_SIZE; > > /* Ensure the base is alloc_tag aligned when required for indexing */ > > module_tags.start_addr = alloc_tag_align(module_tags.start_addr); > > @@ -587,6 +604,7 @@ static int __init alloc_mod_tags_mem(void) > > > > static void __init free_mod_tags_mem(void) > > { > > +#ifdef CONFIG_MMU > > int i; > > > > module_tags.start_addr = 0; > > @@ -594,6 +612,9 @@ static void __init free_mod_tags_mem(void) > > __free_page(vm_module_tags->pages[i]); > > kfree(vm_module_tags->pages); > > free_vm_area(vm_module_tags); > > +#else > > + execmem_free((void *)module_tags.start_addr); > > +#endif > > } > > > > #else /* CONFIG_MODULES */ > > diff --git a/mm/execmem.c b/mm/execmem.c > > index 5c0f9f2d6f83..317b6a8d35be 100644 > > --- a/mm/execmem.c > > +++ b/mm/execmem.c > > @@ -64,6 +64,22 @@ static void *execmem_vmalloc(struct execmem_range *range, size_t size, > > > > return p; > > } > > + > > +struct vm_struct *execmem_vmap(size_t size) > > +{ > > + struct execmem_range *range = &execmem_info->ranges[EXECMEM_MODULE_DATA]; > > + struct vm_struct *area; > > + > > + area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > > + range->start, range->end, NUMA_NO_NODE, > > + GFP_KERNEL, __builtin_return_address(0)); > > + if (!area && range->fallback_start) > > + area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > > + range->fallback_start, range->fallback_end, > > + NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0)); > > + > > + return area; > > +} > > #else > > static void *execmem_vmalloc(struct execmem_range *range, size_t size, > > pgprot_t pgprot, unsigned long vm_flags) > > @@ -368,22 +384,6 @@ void execmem_free(void *ptr) > > vfree(ptr); > > } > > > > -struct vm_struct *execmem_vmap(size_t size) > > -{ > > - struct execmem_range *range = &execmem_info->ranges[EXECMEM_MODULE_DATA]; > > - struct vm_struct *area; > > - > > - area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > > - range->start, range->end, NUMA_NO_NODE, > > - GFP_KERNEL, __builtin_return_address(0)); > > - if (!area && range->fallback_start) > > - area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > > - range->fallback_start, range->fallback_end, > > - NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0)); > > - > > - return area; > > -} > > - > > void *execmem_update_copy(void *dst, const void *src, size_t size) > > { > > return text_poke_copy(dst, src, size); > > > > base-commit: 92409df3e00a17d6fd5bb27732fa186749725ed5 > > -- > > 2.47.0.163.g1226f6d8fa-goog > > > > -- > Sincerely yours, > Mike. ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 1/1] alloc_tag: avoid execmem_vmap() when !MMU 2024-10-29 20:03 ` Suren Baghdasaryan @ 2024-10-30 12:43 ` Mike Rapoport 2024-10-30 23:42 ` Suren Baghdasaryan 0 siblings, 1 reply; 5+ messages in thread From: Mike Rapoport @ 2024-10-30 12:43 UTC (permalink / raw) To: Suren Baghdasaryan Cc: akpm, arnd, arnd, pasha.tatashin, mcgrof, song, mhiramat, linux-mm, linux-kernel, kernel test robot On Tue, Oct 29, 2024 at 01:03:51PM -0700, Suren Baghdasaryan wrote: > On Tue, Oct 29, 2024 at 10:13 AM Mike Rapoport <rppt@kernel.org> wrote: > > > > Hi Suren, > > > > On Mon, Oct 28, 2024 at 01:29:35PM -0700, Suren Baghdasaryan wrote: > > > With CONFIG_MMU=n __get_vm_area_node() is not available and memory for the > > > allocation tags cannot be populated as needed. For this case, populate the > > > required memory at initialization time. > > > > Do we care that much to have alloc_tags on !MMU configs? > > > > Your patch is great, but it would have been a lot easier to make > > MEM_ALLOC_PROFILING depend on MMU. > > I didn't want to prevent !MMU users from using allocation profiling > but I'm not aware of any current !MMU users. Therefore I have no > objections to making MEM_ALLOC_PROFILING dependent on MMU if we indeed > don't care about !MMU && MEM_ALLOC_PROFILING configuration. > I'll be posting v5 tomorrow with all these fixes, so please let me > know which way you prefer. Thanks! I think if nommu users would want that feature, we can add the ifdefery later > > > Fixes: 57bc3834fb6f ("alloc_tag: populate memory for module tags as needed") > > > Reported-by: kernel test robot <lkp@intel.com> > > > Closes: https://lore.kernel.org/oe-kbuild-all/202410250808.dQGyYjlk-lkp@intel.com/ > > > Closes: https://lore.kernel.org/oe-lkp/202410251525.9f85854d-oliver.sang@intel.com > > > Closes: https://lore.kernel.org/oe-kbuild-all/202410261016.IO7C6Cml-lkp@intel.com/ > > > Closes: https://lore.kernel.org/oe-kbuild-all/202410270919.LebQlmxD-lkp@intel.com/ > > > Cc: Arnd Bergmann <arnd@arndb.de> > > > Cc: Mike Rapoport (Microsoft) <rppt@kernel.org> > > > Signed-off-by: Suren Baghdasaryan <surenb@google.com> > > > --- > > > include/linux/execmem.h | 2 ++ > > > lib/alloc_tag.c | 23 ++++++++++++++++++++++- > > > mm/execmem.c | 32 ++++++++++++++++---------------- > > > 3 files changed, 40 insertions(+), 17 deletions(-) > > > > > > diff --git a/include/linux/execmem.h b/include/linux/execmem.h > > > index 5a5e2917f870..64130ae19690 100644 > > > --- a/include/linux/execmem.h > > > +++ b/include/linux/execmem.h > > > @@ -139,6 +139,7 @@ void *execmem_alloc(enum execmem_type type, size_t size); > > > */ > > > void execmem_free(void *ptr); > > > > > > +#ifdef CONFIG_MMU > > > /** > > > * execmem_vmap - create virtual mapping for EXECMEM_MODULE_DATA memory > > > * @size: size of the virtual mapping in bytes > > > @@ -148,6 +149,7 @@ void execmem_free(void *ptr); > > > * Return: the area descriptor on success or %NULL on failure. > > > */ > > > struct vm_struct *execmem_vmap(size_t size); > > > +#endif > > > > > > /** > > > * execmem_update_copy - copy an update to executable memory > > > diff --git a/lib/alloc_tag.c b/lib/alloc_tag.c > > > index c1ddac2d29f0..1c74942e6dfd 100644 > > > --- a/lib/alloc_tag.c > > > +++ b/lib/alloc_tag.c > > > @@ -268,7 +268,6 @@ void __init alloc_tag_sec_init(void) > > > #ifdef CONFIG_MODULES > > > > > > static struct maple_tree mod_area_mt = MTREE_INIT(mod_area_mt, MT_FLAGS_ALLOC_RANGE); > > > -static struct vm_struct *vm_module_tags; > > > /* A dummy object used to indicate an unloaded module */ > > > static struct module unloaded_mod; > > > /* A dummy object used to indicate a module prepended area */ > > > @@ -391,6 +390,9 @@ static bool find_aligned_area(struct ma_state *mas, unsigned long section_size, > > > return false; > > > } > > > > > > +#ifdef CONFIG_MMU > > > +static struct vm_struct *vm_module_tags; > > > + > > > static int vm_module_tags_populate(void) > > > { > > > unsigned long phys_size = vm_module_tags->nr_pages << PAGE_SHIFT; > > > @@ -417,6 +419,13 @@ static int vm_module_tags_populate(void) > > > > > > return 0; > > > } > > > +#else > > > +static int vm_module_tags_populate(void) > > > +{ > > > + /* Memory was already allocated */ > > > + return 0; > > > +} > > > +#endif > > > > > > static void *reserve_module_tags(struct module *mod, unsigned long size, > > > unsigned int prepend, unsigned long align) > > > @@ -561,6 +570,7 @@ static void replace_module(struct module *mod, struct module *new_mod) > > > > > > static int __init alloc_mod_tags_mem(void) > > > { > > > +#ifdef CONFIG_MMU > > > /* Map space to copy allocation tags */ > > > vm_module_tags = execmem_vmap(MODULE_ALLOC_TAG_VMAP_SIZE); > > > if (!vm_module_tags) { > > > @@ -578,6 +588,13 @@ static int __init alloc_mod_tags_mem(void) > > > } > > > > > > module_tags.start_addr = (unsigned long)vm_module_tags->addr; > > > +#else > > > + /* Allocate space to copy allocation tags */ > > > + module_tags.start_addr = (unsigned long)execmem_alloc(EXECMEM_MODULE_DATA, > > > + MODULE_ALLOC_TAG_VMAP_SIZE); > > > + if (!module_tags.start_addr) > > > + return -ENOMEM; > > > +#endif > > > module_tags.end_addr = module_tags.start_addr + MODULE_ALLOC_TAG_VMAP_SIZE; > > > /* Ensure the base is alloc_tag aligned when required for indexing */ > > > module_tags.start_addr = alloc_tag_align(module_tags.start_addr); > > > @@ -587,6 +604,7 @@ static int __init alloc_mod_tags_mem(void) > > > > > > static void __init free_mod_tags_mem(void) > > > { > > > +#ifdef CONFIG_MMU > > > int i; > > > > > > module_tags.start_addr = 0; > > > @@ -594,6 +612,9 @@ static void __init free_mod_tags_mem(void) > > > __free_page(vm_module_tags->pages[i]); > > > kfree(vm_module_tags->pages); > > > free_vm_area(vm_module_tags); > > > +#else > > > + execmem_free((void *)module_tags.start_addr); > > > +#endif > > > } > > > > > > #else /* CONFIG_MODULES */ > > > diff --git a/mm/execmem.c b/mm/execmem.c > > > index 5c0f9f2d6f83..317b6a8d35be 100644 > > > --- a/mm/execmem.c > > > +++ b/mm/execmem.c > > > @@ -64,6 +64,22 @@ static void *execmem_vmalloc(struct execmem_range *range, size_t size, > > > > > > return p; > > > } > > > + > > > +struct vm_struct *execmem_vmap(size_t size) > > > +{ > > > + struct execmem_range *range = &execmem_info->ranges[EXECMEM_MODULE_DATA]; > > > + struct vm_struct *area; > > > + > > > + area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > > > + range->start, range->end, NUMA_NO_NODE, > > > + GFP_KERNEL, __builtin_return_address(0)); > > > + if (!area && range->fallback_start) > > > + area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > > > + range->fallback_start, range->fallback_end, > > > + NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0)); > > > + > > > + return area; > > > +} > > > #else > > > static void *execmem_vmalloc(struct execmem_range *range, size_t size, > > > pgprot_t pgprot, unsigned long vm_flags) > > > @@ -368,22 +384,6 @@ void execmem_free(void *ptr) > > > vfree(ptr); > > > } > > > > > > -struct vm_struct *execmem_vmap(size_t size) > > > -{ > > > - struct execmem_range *range = &execmem_info->ranges[EXECMEM_MODULE_DATA]; > > > - struct vm_struct *area; > > > - > > > - area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > > > - range->start, range->end, NUMA_NO_NODE, > > > - GFP_KERNEL, __builtin_return_address(0)); > > > - if (!area && range->fallback_start) > > > - area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > > > - range->fallback_start, range->fallback_end, > > > - NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0)); > > > - > > > - return area; > > > -} > > > - > > > void *execmem_update_copy(void *dst, const void *src, size_t size) > > > { > > > return text_poke_copy(dst, src, size); > > > > > > base-commit: 92409df3e00a17d6fd5bb27732fa186749725ed5 > > > -- > > > 2.47.0.163.g1226f6d8fa-goog > > > > > > > -- > > Sincerely yours, > > Mike. -- Sincerely yours, Mike. ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 1/1] alloc_tag: avoid execmem_vmap() when !MMU 2024-10-30 12:43 ` Mike Rapoport @ 2024-10-30 23:42 ` Suren Baghdasaryan 0 siblings, 0 replies; 5+ messages in thread From: Suren Baghdasaryan @ 2024-10-30 23:42 UTC (permalink / raw) To: Mike Rapoport Cc: akpm, arnd, arnd, pasha.tatashin, mcgrof, song, mhiramat, linux-mm, linux-kernel, kernel test robot On Wed, Oct 30, 2024 at 5:46 AM Mike Rapoport <rppt@kernel.org> wrote: > > On Tue, Oct 29, 2024 at 01:03:51PM -0700, Suren Baghdasaryan wrote: > > On Tue, Oct 29, 2024 at 10:13 AM Mike Rapoport <rppt@kernel.org> wrote: > > > > > > Hi Suren, > > > > > > On Mon, Oct 28, 2024 at 01:29:35PM -0700, Suren Baghdasaryan wrote: > > > > With CONFIG_MMU=n __get_vm_area_node() is not available and memory for the > > > > allocation tags cannot be populated as needed. For this case, populate the > > > > required memory at initialization time. > > > > > > Do we care that much to have alloc_tags on !MMU configs? > > > > > > Your patch is great, but it would have been a lot easier to make > > > MEM_ALLOC_PROFILING depend on MMU. > > > > I didn't want to prevent !MMU users from using allocation profiling > > but I'm not aware of any current !MMU users. Therefore I have no > > objections to making MEM_ALLOC_PROFILING dependent on MMU if we indeed > > don't care about !MMU && MEM_ALLOC_PROFILING configuration. > > I'll be posting v5 tomorrow with all these fixes, so please let me > > know which way you prefer. Thanks! > > I think if nommu users would want that feature, we can add the ifdefery > later Sounds good. I'll add the dependency and simplify this fix in my next posting. Thanks! > > > > > Fixes: 57bc3834fb6f ("alloc_tag: populate memory for module tags as needed") > > > > Reported-by: kernel test robot <lkp@intel.com> > > > > Closes: https://lore.kernel.org/oe-kbuild-all/202410250808.dQGyYjlk-lkp@intel.com/ > > > > Closes: https://lore.kernel.org/oe-lkp/202410251525.9f85854d-oliver.sang@intel.com > > > > Closes: https://lore.kernel.org/oe-kbuild-all/202410261016.IO7C6Cml-lkp@intel.com/ > > > > Closes: https://lore.kernel.org/oe-kbuild-all/202410270919.LebQlmxD-lkp@intel.com/ > > > > Cc: Arnd Bergmann <arnd@arndb.de> > > > > Cc: Mike Rapoport (Microsoft) <rppt@kernel.org> > > > > Signed-off-by: Suren Baghdasaryan <surenb@google.com> > > > > --- > > > > include/linux/execmem.h | 2 ++ > > > > lib/alloc_tag.c | 23 ++++++++++++++++++++++- > > > > mm/execmem.c | 32 ++++++++++++++++---------------- > > > > 3 files changed, 40 insertions(+), 17 deletions(-) > > > > > > > > diff --git a/include/linux/execmem.h b/include/linux/execmem.h > > > > index 5a5e2917f870..64130ae19690 100644 > > > > --- a/include/linux/execmem.h > > > > +++ b/include/linux/execmem.h > > > > @@ -139,6 +139,7 @@ void *execmem_alloc(enum execmem_type type, size_t size); > > > > */ > > > > void execmem_free(void *ptr); > > > > > > > > +#ifdef CONFIG_MMU > > > > /** > > > > * execmem_vmap - create virtual mapping for EXECMEM_MODULE_DATA memory > > > > * @size: size of the virtual mapping in bytes > > > > @@ -148,6 +149,7 @@ void execmem_free(void *ptr); > > > > * Return: the area descriptor on success or %NULL on failure. > > > > */ > > > > struct vm_struct *execmem_vmap(size_t size); > > > > +#endif > > > > > > > > /** > > > > * execmem_update_copy - copy an update to executable memory > > > > diff --git a/lib/alloc_tag.c b/lib/alloc_tag.c > > > > index c1ddac2d29f0..1c74942e6dfd 100644 > > > > --- a/lib/alloc_tag.c > > > > +++ b/lib/alloc_tag.c > > > > @@ -268,7 +268,6 @@ void __init alloc_tag_sec_init(void) > > > > #ifdef CONFIG_MODULES > > > > > > > > static struct maple_tree mod_area_mt = MTREE_INIT(mod_area_mt, MT_FLAGS_ALLOC_RANGE); > > > > -static struct vm_struct *vm_module_tags; > > > > /* A dummy object used to indicate an unloaded module */ > > > > static struct module unloaded_mod; > > > > /* A dummy object used to indicate a module prepended area */ > > > > @@ -391,6 +390,9 @@ static bool find_aligned_area(struct ma_state *mas, unsigned long section_size, > > > > return false; > > > > } > > > > > > > > +#ifdef CONFIG_MMU > > > > +static struct vm_struct *vm_module_tags; > > > > + > > > > static int vm_module_tags_populate(void) > > > > { > > > > unsigned long phys_size = vm_module_tags->nr_pages << PAGE_SHIFT; > > > > @@ -417,6 +419,13 @@ static int vm_module_tags_populate(void) > > > > > > > > return 0; > > > > } > > > > +#else > > > > +static int vm_module_tags_populate(void) > > > > +{ > > > > + /* Memory was already allocated */ > > > > + return 0; > > > > +} > > > > +#endif > > > > > > > > static void *reserve_module_tags(struct module *mod, unsigned long size, > > > > unsigned int prepend, unsigned long align) > > > > @@ -561,6 +570,7 @@ static void replace_module(struct module *mod, struct module *new_mod) > > > > > > > > static int __init alloc_mod_tags_mem(void) > > > > { > > > > +#ifdef CONFIG_MMU > > > > /* Map space to copy allocation tags */ > > > > vm_module_tags = execmem_vmap(MODULE_ALLOC_TAG_VMAP_SIZE); > > > > if (!vm_module_tags) { > > > > @@ -578,6 +588,13 @@ static int __init alloc_mod_tags_mem(void) > > > > } > > > > > > > > module_tags.start_addr = (unsigned long)vm_module_tags->addr; > > > > +#else > > > > + /* Allocate space to copy allocation tags */ > > > > + module_tags.start_addr = (unsigned long)execmem_alloc(EXECMEM_MODULE_DATA, > > > > + MODULE_ALLOC_TAG_VMAP_SIZE); > > > > + if (!module_tags.start_addr) > > > > + return -ENOMEM; > > > > +#endif > > > > module_tags.end_addr = module_tags.start_addr + MODULE_ALLOC_TAG_VMAP_SIZE; > > > > /* Ensure the base is alloc_tag aligned when required for indexing */ > > > > module_tags.start_addr = alloc_tag_align(module_tags.start_addr); > > > > @@ -587,6 +604,7 @@ static int __init alloc_mod_tags_mem(void) > > > > > > > > static void __init free_mod_tags_mem(void) > > > > { > > > > +#ifdef CONFIG_MMU > > > > int i; > > > > > > > > module_tags.start_addr = 0; > > > > @@ -594,6 +612,9 @@ static void __init free_mod_tags_mem(void) > > > > __free_page(vm_module_tags->pages[i]); > > > > kfree(vm_module_tags->pages); > > > > free_vm_area(vm_module_tags); > > > > +#else > > > > + execmem_free((void *)module_tags.start_addr); > > > > +#endif > > > > } > > > > > > > > #else /* CONFIG_MODULES */ > > > > diff --git a/mm/execmem.c b/mm/execmem.c > > > > index 5c0f9f2d6f83..317b6a8d35be 100644 > > > > --- a/mm/execmem.c > > > > +++ b/mm/execmem.c > > > > @@ -64,6 +64,22 @@ static void *execmem_vmalloc(struct execmem_range *range, size_t size, > > > > > > > > return p; > > > > } > > > > + > > > > +struct vm_struct *execmem_vmap(size_t size) > > > > +{ > > > > + struct execmem_range *range = &execmem_info->ranges[EXECMEM_MODULE_DATA]; > > > > + struct vm_struct *area; > > > > + > > > > + area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > > > > + range->start, range->end, NUMA_NO_NODE, > > > > + GFP_KERNEL, __builtin_return_address(0)); > > > > + if (!area && range->fallback_start) > > > > + area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > > > > + range->fallback_start, range->fallback_end, > > > > + NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0)); > > > > + > > > > + return area; > > > > +} > > > > #else > > > > static void *execmem_vmalloc(struct execmem_range *range, size_t size, > > > > pgprot_t pgprot, unsigned long vm_flags) > > > > @@ -368,22 +384,6 @@ void execmem_free(void *ptr) > > > > vfree(ptr); > > > > } > > > > > > > > -struct vm_struct *execmem_vmap(size_t size) > > > > -{ > > > > - struct execmem_range *range = &execmem_info->ranges[EXECMEM_MODULE_DATA]; > > > > - struct vm_struct *area; > > > > - > > > > - area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > > > > - range->start, range->end, NUMA_NO_NODE, > > > > - GFP_KERNEL, __builtin_return_address(0)); > > > > - if (!area && range->fallback_start) > > > > - area = __get_vm_area_node(size, range->alignment, PAGE_SHIFT, VM_ALLOC, > > > > - range->fallback_start, range->fallback_end, > > > > - NUMA_NO_NODE, GFP_KERNEL, __builtin_return_address(0)); > > > > - > > > > - return area; > > > > -} > > > > - > > > > void *execmem_update_copy(void *dst, const void *src, size_t size) > > > > { > > > > return text_poke_copy(dst, src, size); > > > > > > > > base-commit: 92409df3e00a17d6fd5bb27732fa186749725ed5 > > > > -- > > > > 2.47.0.163.g1226f6d8fa-goog > > > > > > > > > > -- > > > Sincerely yours, > > > Mike. > > -- > Sincerely yours, > Mike. ^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2024-10-30 23:42 UTC | newest] Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2024-10-28 20:29 [PATCH 1/1] alloc_tag: avoid execmem_vmap() when !MMU Suren Baghdasaryan 2024-10-29 17:09 ` Mike Rapoport 2024-10-29 20:03 ` Suren Baghdasaryan 2024-10-30 12:43 ` Mike Rapoport 2024-10-30 23:42 ` Suren Baghdasaryan
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox