The patch titled
Subject: exofs: avoid VLA in structures
has been added to the -mm tree. Its filename is
exofs-avoid-vla-in-structures.patch
This patch should soon appear at
http://ozlabs.org/~akpm/mmots/broken-out/exofs-avoid-vla-in-structures.patch
and later at
http://ozlabs.org/~akpm/mmotm/broken-out/exofs-avoid-vla-in-structures.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: Kees Cook <keescook@xxxxxxxxxxxx>
Subject: exofs: avoid VLA in structures
On the quest to remove all VLAs from the kernel[1] this adjusts several
cases where allocation is made after an array of structures that points
back into the allocation. The allocations are changed to perform explicit
calculations instead of using a Variable Length Array in a structure.
Additionally, this lets Clang compile this code now, since Clang does not
support VLAIS[2].
[1] https://lkml.org/lkml/2018/3/7/621
[2] https://lkml.org/lkml/2013/9/23/500
Link: http://lkml.kernel.org/r/20180327203904.GA1151@beast
Signed-off-by: Kees Cook <keescook@xxxxxxxxxxxx>
Cc: Boaz Harrosh <ooo@xxxxxxxxxxxxxxx>
Cc: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
fs/exofs/ore.c | 82 ++++++++++++++++++++++++------------------
fs/exofs/ore_raid.c | 73 +++++++++++++++++++++++++++----------
fs/exofs/super.c | 23 +++++------
3 files changed, 113 insertions(+), 65 deletions(-)
diff -puN fs/exofs/ore.c~exofs-avoid-vla-in-structures fs/exofs/ore.c
--- a/fs/exofs/ore.c~exofs-avoid-vla-in-structures
+++ a/fs/exofs/ore.c
@@ -146,68 +146,82 @@ int _ore_get_io_state(struct ore_layout
struct ore_io_state **pios)
{
struct ore_io_state *ios;
- struct page **pages;
- struct osd_sg_entry *sgilist;
+ size_t size_ios, size_extra, size_total;
+ void *ios_extra;
+
+ /*
+ * The desired layout looks like this, with the extra_allocation
+ * items pointed at from fields within ios or per_dev:
+
struct __alloc_all_io_state {
struct ore_io_state ios;
struct ore_per_dev_state per_dev[numdevs];
union {
struct osd_sg_entry sglist[sgs_per_dev * numdevs];
struct page *pages[num_par_pages];
- };
- } *_aios;
+ } extra_allocation;
+ } whole_allocation;
+
+ */
+
+ /* This should never happen, so abort early if it ever does. */
+ if (sgs_per_dev && num_par_pages) {
+ ORE_DBGMSG("Tried to use both pages and sglist\n");
+ *pios = NULL;
+ return -EINVAL;
+ }
+
+ if (numdevs > (INT_MAX - sizeof(*ios)) /
+ sizeof(struct ore_per_dev_state))
+ return -ENOMEM;
+ size_ios = sizeof(*ios) + sizeof(struct ore_per_dev_state) * numdevs;
- if (likely(sizeof(*_aios) <= PAGE_SIZE)) {
- _aios = kzalloc(sizeof(*_aios), GFP_KERNEL);
- if (unlikely(!_aios)) {
- ORE_DBGMSG("Failed kzalloc bytes=%zd\n",
- sizeof(*_aios));
+ if (sgs_per_dev * numdevs > INT_MAX / sizeof(struct osd_sg_entry))
+ return -ENOMEM;
+ if (num_par_pages > INT_MAX / sizeof(struct page *))
+ return -ENOMEM;
+ size_extra = max(sizeof(struct osd_sg_entry) * (sgs_per_dev * numdevs),
+ sizeof(struct page *) * num_par_pages);
+
+ size_total = size_ios + size_extra;
+
+ if (likely(size_total <= PAGE_SIZE)) {
+ ios = kzalloc(size_total, GFP_KERNEL);
+ if (unlikely(!ios)) {
+ ORE_DBGMSG("Failed kzalloc bytes=%zd\n", size_total);
*pios = NULL;
return -ENOMEM;
}
- pages = num_par_pages ? _aios->pages : NULL;
- sgilist = sgs_per_dev ? _aios->sglist : NULL;
- ios = &_aios->ios;
+ ios_extra = (char *)ios + size_ios;
} else {
- struct __alloc_small_io_state {
- struct ore_io_state ios;
- struct ore_per_dev_state per_dev[numdevs];
- } *_aio_small;
- union __extra_part {
- struct osd_sg_entry sglist[sgs_per_dev * numdevs];
- struct page *pages[num_par_pages];
- } *extra_part;
-
- _aio_small = kzalloc(sizeof(*_aio_small), GFP_KERNEL);
- if (unlikely(!_aio_small)) {
+ ios = kzalloc(size_ios, GFP_KERNEL);
+ if (unlikely(!ios)) {
ORE_DBGMSG("Failed alloc first part bytes=%zd\n",
- sizeof(*_aio_small));
+ size_ios);
*pios = NULL;
return -ENOMEM;
}
- extra_part = kzalloc(sizeof(*extra_part), GFP_KERNEL);
- if (unlikely(!extra_part)) {
+ ios_extra = kzalloc(size_extra, GFP_KERNEL);
+ if (unlikely(!ios_extra)) {
ORE_DBGMSG("Failed alloc second part bytes=%zd\n",
- sizeof(*extra_part));
- kfree(_aio_small);
+ size_extra);
+ kfree(ios);
*pios = NULL;
return -ENOMEM;
}
- pages = num_par_pages ? extra_part->pages : NULL;
- sgilist = sgs_per_dev ? extra_part->sglist : NULL;
/* In this case the per_dev[0].sgilist holds the pointer to
* be freed
*/
- ios = &_aio_small->ios;
ios->extra_part_alloc = true;
}
- if (pages) {
- ios->parity_pages = pages;
+ if (num_par_pages) {
+ ios->parity_pages = ios_extra;
ios->max_par_pages = num_par_pages;
}
- if (sgilist) {
+ if (sgs_per_dev) {
+ struct osd_sg_entry *sgilist = ios_extra;
unsigned d;
for (d = 0; d < numdevs; ++d) {
diff -puN fs/exofs/ore_raid.c~exofs-avoid-vla-in-structures fs/exofs/ore_raid.c
--- a/fs/exofs/ore_raid.c~exofs-avoid-vla-in-structures
+++ a/fs/exofs/ore_raid.c
@@ -71,6 +71,11 @@ static int _sp2d_alloc(unsigned pages_in
{
struct __stripe_pages_2d *sp2d;
unsigned data_devs = group_width - parity;
+
+ /*
+ * Desired allocation layout is, though when larger than PAGE_SIZE,
+ * each struct __alloc_1p_arrays is separately allocated:
+
struct _alloc_all_bytes {
struct __alloc_stripe_pages_2d {
struct __stripe_pages_2d sp2d;
@@ -82,55 +87,85 @@ static int _sp2d_alloc(unsigned pages_in
char page_is_read[data_devs];
} __a1pa[pages_in_unit];
} *_aab;
+
struct __alloc_1p_arrays *__a1pa;
struct __alloc_1p_arrays *__a1pa_end;
- const unsigned sizeof__a1pa = sizeof(_aab->__a1pa[0]);
+
+ */
+
+ char *__a1pa;
+ char *__a1pa_end;
+
+ const size_t sizeof_stripe_pages_2d =
+ sizeof(struct __stripe_pages_2d) +
+ sizeof(struct __1_page_stripe) * pages_in_unit;
+ const size_t sizeof__a1pa =
+ ALIGN(sizeof(struct page *) * (2 * group_width) + data_devs,
+ sizeof(void *));
+ const size_t sizeof__a1pa_arrays = sizeof__a1pa * pages_in_unit;
+ const size_t alloc_total = sizeof_stripe_pages_2d +
+ sizeof__a1pa_arrays;
+
unsigned num_a1pa, alloc_size, i;
/* FIXME: check these numbers in ore_verify_layout */
- BUG_ON(sizeof(_aab->__asp2d) > PAGE_SIZE);
+ BUG_ON(sizeof_stripe_pages_2d > PAGE_SIZE);
BUG_ON(sizeof__a1pa > PAGE_SIZE);
- if (sizeof(*_aab) > PAGE_SIZE) {
- num_a1pa = (PAGE_SIZE - sizeof(_aab->__asp2d)) / sizeof__a1pa;
- alloc_size = sizeof(_aab->__asp2d) + sizeof__a1pa * num_a1pa;
+ /*
+ * If alloc_total would be larger than PAGE_SIZE, only allocate
+ * as many a1pa items as would fill the rest of the page, instead
+ * of the full pages_in_unit count.
+ */
+ if (alloc_total > PAGE_SIZE) {
+ num_a1pa = (PAGE_SIZE - sizeof_stripe_pages_2d) / sizeof__a1pa;
+ alloc_size = sizeof_stripe_pages_2d + sizeof__a1pa * num_a1pa;
} else {
num_a1pa = pages_in_unit;
- alloc_size = sizeof(*_aab);
+ alloc_size = alloc_total;
}
- _aab = kzalloc(alloc_size, GFP_KERNEL);
- if (unlikely(!_aab)) {
+ *psp2d = sp2d = kzalloc(alloc_size, GFP_KERNEL);
+ if (unlikely(!sp2d)) {
ORE_DBGMSG("!! Failed to alloc sp2d size=%d\n", alloc_size);
return -ENOMEM;
}
+ /* From here Just call _sp2d_free */
- sp2d = &_aab->__asp2d.sp2d;
- *psp2d = sp2d; /* From here Just call _sp2d_free */
-
- __a1pa = _aab->__a1pa;
- __a1pa_end = __a1pa + num_a1pa;
+ /* Find start of a1pa area. */
+ __a1pa = (char *)sp2d + sizeof_stripe_pages_2d;
+ /* Find end of the _allocated_ a1pa area. */
+ __a1pa_end = __a1pa + alloc_size;
+ /* Allocate additionally needed a1pa items in PAGE_SIZE chunks. */
for (i = 0; i < pages_in_unit; ++i) {
if (unlikely(__a1pa >= __a1pa_end)) {
num_a1pa = min_t(unsigned, PAGE_SIZE / sizeof__a1pa,
pages_in_unit - i);
+ alloc_size = sizeof__a1pa * num_a1pa;
- __a1pa = kcalloc(num_a1pa, sizeof__a1pa, GFP_KERNEL);
+ __a1pa = kzalloc(alloc_size, GFP_KERNEL);
if (unlikely(!__a1pa)) {
ORE_DBGMSG("!! Failed to _alloc_1p_arrays=%d\n",
num_a1pa);
return -ENOMEM;
}
- __a1pa_end = __a1pa + num_a1pa;
+ __a1pa_end = __a1pa + alloc_size;
/* First *pages is marked for kfree of the buffer */
sp2d->_1p_stripes[i].alloc = true;
}
- sp2d->_1p_stripes[i].pages = __a1pa->pages;
- sp2d->_1p_stripes[i].scribble = __a1pa->scribble ;
- sp2d->_1p_stripes[i].page_is_read = __a1pa->page_is_read;
- ++__a1pa;
+ /*
+ * Attach all _lp_stripes pointers to the allocation for
+ * it which was either part of the original PAGE_SIZE
+ * allocation or the subsequent allocation in this loop.
+ */
+ sp2d->_1p_stripes[i].pages = (void *)__a1pa;
+ sp2d->_1p_stripes[i].scribble =
+ sp2d->_1p_stripes[i].pages + group_width;
+ sp2d->_1p_stripes[i].page_is_read =
+ (char *)(sp2d->_1p_stripes[i].scribble + group_width);
+ __a1pa += sizeof__a1pa;
}
sp2d->parity = parity;
diff -puN fs/exofs/super.c~exofs-avoid-vla-in-structures fs/exofs/super.c
--- a/fs/exofs/super.c~exofs-avoid-vla-in-structures
+++ a/fs/exofs/super.c
@@ -549,27 +549,26 @@ static int exofs_devs_2_odi(struct exofs
static int __alloc_dev_table(struct exofs_sb_info *sbi, unsigned numdevs,
struct exofs_dev **peds)
{
- struct __alloc_ore_devs_and_exofs_devs {
- /* Twice bigger table: See exofs_init_comps() and comment at
- * exofs_read_lookup_dev_table()
- */
- struct ore_dev *oreds[numdevs * 2 - 1];
- struct exofs_dev eds[numdevs];
- } *aoded;
+ /* Twice bigger table: See exofs_init_comps() and comment at
+ * exofs_read_lookup_dev_table()
+ */
+ size_t numores = numdevs * 2 - 1;
struct exofs_dev *eds;
unsigned i;
- aoded = kzalloc(sizeof(*aoded), GFP_KERNEL);
- if (unlikely(!aoded)) {
+ sbi->oc.ods = kzalloc(numores * sizeof(struct ore_dev *) +
+ numdevs * sizeof(struct exofs_dev), GFP_KERNEL);
+ if (unlikely(!sbi->oc.ods)) {
EXOFS_ERR("ERROR: failed allocating Device array[%d]\n",
numdevs);
return -ENOMEM;
}
- sbi->oc.ods = aoded->oreds;
- *peds = eds = aoded->eds;
+ /* Start of allocated struct exofs_dev entries */
+ *peds = eds = (void *)sbi->oc.ods[numores];
+ /* Initialize pointers into struct exofs_dev */
for (i = 0; i < numdevs; ++i)
- aoded->oreds[i] = &eds[i].ored;
+ sbi->oc.ods[i] = &eds[i].ored;
return 0;
}
_
Patches currently in -mm which might be from keescook@xxxxxxxxxxxx are
taint-convert-to-indexed-initialization.patch
taint-consolidate-documentation.patch
taint-add-taint-for-randstruct.patch
kernelh-introduce-const_max-for-vla-removal.patch
remove-false-positive-vlas-when-using-max.patch
test_bitmap-do-not-accidentally-use-stack-vla.patch
fork-unconditionally-clear-stack-on-fork.patch
exec-pass-stack-rlimit-into-mm-layout-functions.patch
exec-introduce-finalize_exec-before-start_thread.patch
exec-pin-stack-limit-during-exec.patch
exofs-avoid-vla-in-structures.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
