Hi Zi,
FYI, the error/warning was bisected to this commit, please ignore it if it's irrelevant.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
head: fff5a5e7f528b2ed2c335991399a766c2cf01103
commit: 0192445cb2f7ed1cd7a95a0fc8c7645480baba25 arch: mm: rename FORCE_MAX_ZONEORDER to ARCH_FORCE_MAX_ORDER
date: 6 months ago
config: sparc64-buildonly-randconfig-r002-20230322 (https://download.01.org/0day-ci/archive/20230323/[email protected]/config)
compiler: sparc64-linux-gcc (GCC) 12.1.0
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
# https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=0192445cb2f7ed1cd7a95a0fc8c7645480baba25
git remote add linus https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
git fetch --no-tags linus master
git checkout 0192445cb2f7ed1cd7a95a0fc8c7645480baba25
# save the config file
mkdir build_dir && cp config build_dir/.config
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=sparc64 olddefconfig
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=sparc64 SHELL=/bin/bash arch/sparc/mm/ drivers/base/regmap/ drivers/media/usb/em28xx/
If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <[email protected]>
| Link: https://lore.kernel.org/oe-kbuild-all/[email protected]/
All error/warnings (new ones prefixed by >>):
In file included from include/linux/gfp.h:7,
from include/linux/slab.h:15,
from arch/sparc/mm/tsb.c:9:
include/linux/mmzone.h:636:33: error: size of array 'free_area' is negative
636 | struct free_area free_area[MAX_ORDER];
| ^~~~~~~~~
arch/sparc/mm/tsb.c: In function 'tsb_grow':
>> arch/sparc/mm/tsb.c:405:39: error: left shift count is negative [-Werror=shift-count-negative]
405 | if (max_tsb_size > (PAGE_SIZE << MAX_ORDER))
| ^~
arch/sparc/mm/tsb.c:406:43: error: left shift count is negative [-Werror=shift-count-negative]
406 | max_tsb_size = (PAGE_SIZE << MAX_ORDER);
| ^~
cc1: all warnings being treated as errors
--
In file included from include/linux/gfp.h:7,
from include/linux/slab.h:15,
from drivers/base/regmap/regmap-debugfs.c:9:
include/linux/mmzone.h:636:33: error: size of array 'free_area' is negative
636 | struct free_area free_area[MAX_ORDER];
| ^~~~~~~~~
drivers/base/regmap/regmap-debugfs.c: In function 'regmap_read_debugfs':
>> drivers/base/regmap/regmap-debugfs.c:229:32: warning: left shift count is negative [-Wshift-count-negative]
229 | if (count > (PAGE_SIZE << (MAX_ORDER - 1)))
| ^~
drivers/base/regmap/regmap-debugfs.c:230:35: warning: left shift count is negative [-Wshift-count-negative]
230 | count = PAGE_SIZE << (MAX_ORDER - 1);
| ^~
drivers/base/regmap/regmap-debugfs.c: In function 'regmap_reg_ranges_read_file':
drivers/base/regmap/regmap-debugfs.c:376:32: warning: left shift count is negative [-Wshift-count-negative]
376 | if (count > (PAGE_SIZE << (MAX_ORDER - 1)))
| ^~
drivers/base/regmap/regmap-debugfs.c:377:35: warning: left shift count is negative [-Wshift-count-negative]
377 | count = PAGE_SIZE << (MAX_ORDER - 1);
| ^~
--
In file included from include/linux/gfp.h:7,
from include/linux/xarray.h:15,
from include/linux/radix-tree.h:21,
from include/linux/idr.h:15,
from include/linux/kernfs.h:12,
from include/linux/sysfs.h:16,
from include/linux/kobject.h:20,
from include/linux/of.h:17,
from include/linux/irqdomain.h:35,
from include/linux/acpi.h:13,
from include/linux/i2c.h:13,
from drivers/media/usb/em28xx/em28xx.h:22,
from drivers/media/usb/em28xx/em28xx-audio.c:14:
include/linux/mmzone.h:636:33: error: size of array 'free_area' is negative
636 | struct free_area free_area[MAX_ORDER];
| ^~~~~~~~~
drivers/media/usb/em28xx/em28xx-audio.c: In function 'em28xx_audio_init':
>> drivers/media/usb/em28xx/em28xx-audio.c:808:29: warning: variable 'devnr' set but not used [-Wunused-but-set-variable]
808 | static int devnr;
| ^~~~~
drivers/media/usb/em28xx/em28xx-audio.c: At top level:
drivers/media/usb/em28xx/em28xx-audio.c:47:12: warning: 'index' defined but not used [-Wunused-variable]
47 | static int index[SNDRV_CARDS] = SNDRV_DEFAULT_IDX;
| ^~~~~
vim +405 arch/sparc/mm/tsb.c
0871420fad5844 arch/sparc/mm/tsb.c David S. Miller 2008-11-16 379
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 380 /* When the RSS of an address space exceeds tsb_rss_limit for a TSB,
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 381 * do_sparc64_fault() invokes this routine to try and grow it.
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 382 *
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 383 * When we reach the maximum TSB size supported, we stick ~0UL into
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 384 * tsb_rss_limit for that TSB so the grow checks in do_sparc64_fault()
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 385 * will not trigger any longer.
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 386 *
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 387 * The TSB can be anywhere from 8K to 1MB in size, in increasing powers
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 388 * of two. The TSB must be aligned to it's size, so f.e. a 512K TSB
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 389 * must be 512K aligned. It also must be physically contiguous, so we
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 390 * cannot use vmalloc().
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 391 *
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 392 * The idea here is to grow the TSB when the RSS of the process approaches
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 393 * the number of entries that the current TSB can hold at once. Currently,
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 394 * we trigger when the RSS hits 3/4 of the TSB capacity.
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 395 */
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 396 void tsb_grow(struct mm_struct *mm, unsigned long tsb_index, unsigned long rss)
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 397 {
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 398 unsigned long max_tsb_size = 1 * 1024 * 1024;
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 399 unsigned long new_size, old_size, flags;
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 400 struct tsb *old_tsb, *new_tsb;
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 401 unsigned long new_cache_index, old_cache_index;
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 402 unsigned long new_rss_limit;
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 403 gfp_t gfp_flags;
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 404
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 @405 if (max_tsb_size > (PAGE_SIZE << MAX_ORDER))
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 406 max_tsb_size = (PAGE_SIZE << MAX_ORDER);
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 407
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 408 new_cache_index = 0;
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 409 for (new_size = 8192; new_size < max_tsb_size; new_size <<= 1UL) {
0871420fad5844 arch/sparc/mm/tsb.c David S. Miller 2008-11-16 410 new_rss_limit = tsb_size_to_rss_limit(new_size);
0871420fad5844 arch/sparc/mm/tsb.c David S. Miller 2008-11-16 411 if (new_rss_limit > rss)
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 412 break;
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 413 new_cache_index++;
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 414 }
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 415
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 416 if (new_size == max_tsb_size)
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 417 new_rss_limit = ~0UL;
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 418
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 419 retry_tsb_alloc:
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 420 gfp_flags = GFP_KERNEL;
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 421 if (new_size > (PAGE_SIZE * 2))
a55ee1ff751f88 arch/sparc/mm/tsb.c David S. Miller 2013-02-19 422 gfp_flags |= __GFP_NOWARN | __GFP_NORETRY;
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 423
1f261ef53ba066 arch/sparc64/mm/tsb.c David S. Miller 2008-03-19 424 new_tsb = kmem_cache_alloc_node(tsb_caches[new_cache_index],
1f261ef53ba066 arch/sparc64/mm/tsb.c David S. Miller 2008-03-19 425 gfp_flags, numa_node_id());
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 426 if (unlikely(!new_tsb)) {
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 427 /* Not being able to fork due to a high-order TSB
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 428 * allocation failure is very bad behavior. Just back
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 429 * down to a 0-order allocation and force no TSB
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 430 * growing for this address space.
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 431 */
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 432 if (mm->context.tsb_block[tsb_index].tsb == NULL &&
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 433 new_cache_index > 0) {
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 434 new_cache_index = 0;
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 435 new_size = 8192;
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 436 new_rss_limit = ~0UL;
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 437 goto retry_tsb_alloc;
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 438 }
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 439
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 440 /* If we failed on a TSB grow, we are under serious
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 441 * memory pressure so don't try to grow any more.
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 442 */
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 443 if (mm->context.tsb_block[tsb_index].tsb != NULL)
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 444 mm->context.tsb_block[tsb_index].tsb_rss_limit = ~0UL;
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 445 return;
b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 446 }
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 447
8b234274418d6d arch/sparc64/mm/tsb.c David S. Miller 2006-02-17 448 /* Mark all tags as invalid. */
bb8646d8340fa7 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 449 tsb_init(new_tsb, new_size);
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 450
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 451 /* Ok, we are about to commit the changes. If we are
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 452 * growing an existing TSB the locking is very tricky,
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 453 * so WATCH OUT!
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 454 *
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 455 * We have to hold mm->context.lock while committing to the
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 456 * new TSB, this synchronizes us with processors in
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 457 * flush_tsb_user() and switch_mm() for this address space.
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 458 *
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 459 * But even with that lock held, processors run asynchronously
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 460 * accessing the old TSB via TLB miss handling. This is OK
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 461 * because those actions are just propagating state from the
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 462 * Linux page tables into the TSB, page table mappings are not
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 463 * being changed. If a real fault occurs, the processor will
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 464 * synchronize with us when it hits flush_tsb_user(), this is
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 465 * also true for the case where vmscan is modifying the page
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 466 * tables. The only thing we need to be careful with is to
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 467 * skip any locked TSB entries during copy_tsb().
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 468 *
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 469 * When we finish committing to the new TSB, we have to drop
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 470 * the lock and ask all other cpus running this address space
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 471 * to run tsb_context_switch() to see the new TSB table.
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 472 */
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 473 spin_lock_irqsave(&mm->context.lock, flags);
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 474
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 475 old_tsb = mm->context.tsb_block[tsb_index].tsb;
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 476 old_cache_index =
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 477 (mm->context.tsb_block[tsb_index].tsb_reg_val & 0x7UL);
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 478 old_size = (mm->context.tsb_block[tsb_index].tsb_nentries *
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 479 sizeof(struct tsb));
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 480
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 481
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 482 /* Handle multiple threads trying to grow the TSB at the same time.
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 483 * One will get in here first, and bump the size and the RSS limit.
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 484 * The others will get in here next and hit this check.
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 485 */
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 486 if (unlikely(old_tsb &&
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 487 (rss < mm->context.tsb_block[tsb_index].tsb_rss_limit))) {
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 488 spin_unlock_irqrestore(&mm->context.lock, flags);
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 489
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 490 kmem_cache_free(tsb_caches[new_cache_index], new_tsb);
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 491 return;
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 492 }
8b234274418d6d arch/sparc64/mm/tsb.c David S. Miller 2006-02-17 493
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 494 mm->context.tsb_block[tsb_index].tsb_rss_limit = new_rss_limit;
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 495
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 496 if (old_tsb) {
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 497 extern void copy_tsb(unsigned long old_tsb_base,
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 498 unsigned long old_tsb_size,
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 499 unsigned long new_tsb_base,
654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 500 unsigned long new_tsb_size,
654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 501 unsigned long page_size_shift);
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 502 unsigned long old_tsb_base = (unsigned long) old_tsb;
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 503 unsigned long new_tsb_base = (unsigned long) new_tsb;
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 504
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 505 if (tlb_type == cheetah_plus || tlb_type == hypervisor) {
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 506 old_tsb_base = __pa(old_tsb_base);
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 507 new_tsb_base = __pa(new_tsb_base);
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 508 }
654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 509 copy_tsb(old_tsb_base, old_size, new_tsb_base, new_size,
654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 510 tsb_index == MM_TSB_BASE ?
654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 511 PAGE_SHIFT : REAL_HPAGE_SHIFT);
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 512 }
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 513
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 514 mm->context.tsb_block[tsb_index].tsb = new_tsb;
dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 515 setup_tsb_params(mm, tsb_index, new_size);
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 516
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 517 spin_unlock_irqrestore(&mm->context.lock, flags);
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 518
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 519 /* If old_tsb is NULL, we're being invoked for the first time
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 520 * from init_new_context().
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 521 */
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 522 if (old_tsb) {
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 523 /* Reload it on the local cpu. */
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 524 tsb_context_switch(mm);
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 525
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 526 /* Now force other processors to do the same. */
a3cf5e6b6f2548 arch/sparc64/mm/tsb.c David S. Miller 2008-08-03 527 preempt_disable();
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 528 smp_tsb_sync(mm);
a3cf5e6b6f2548 arch/sparc64/mm/tsb.c David S. Miller 2008-08-03 529 preempt_enable();
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 530
7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 531 /* Now it is safe to free the old tsb. */
9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 532 kmem_cache_free(tsb_caches[old_cache_index], old_tsb);
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 533 }
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 534 }
bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 535
:::::: The code at line 405 was first introduced by commit
:::::: bd40791e1d289d807b8580abe1f117e9c62894e4 [SPARC64]: Dynamically grow TSB in response to RSS growth.
:::::: TO: David S. Miller <[email protected]>
:::::: CC: David S. Miller <[email protected]>
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests
On 23 Mar 2023, at 6:40, kernel test robot wrote:
> Hi Zi,
>
> FYI, the error/warning was bisected to this commit, please ignore it if it's irrelevant.
>
> tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> head: fff5a5e7f528b2ed2c335991399a766c2cf01103
> commit: 0192445cb2f7ed1cd7a95a0fc8c7645480baba25 arch: mm: rename FORCE_MAX_ZONEORDER to ARCH_FORCE_MAX_ORDER
> date: 6 months ago
> config: sparc64-buildonly-randconfig-r002-20230322 (https://download.01.org/0day-ci/archive/20230323/[email protected]/config)
Cc Mike, since he is working on ARCH_FORCE_MAX_ORDER.
This random config sets CONFIG_ARCH_FORCE_MAX_ORDER to -1, causing this error. I am not sure if we need to always set a range for ARCH_FORCE_MAX_ORDER or random config generator should be changed to not use non-positive values.
BTW, this issue pre-exists my changes.
> compiler: sparc64-linux-gcc (GCC) 12.1.0
> 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
> # https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=0192445cb2f7ed1cd7a95a0fc8c7645480baba25
> git remote add linus https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
> git fetch --no-tags linus master
> git checkout 0192445cb2f7ed1cd7a95a0fc8c7645480baba25
> # save the config file
> mkdir build_dir && cp config build_dir/.config
> COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=sparc64 olddefconfig
> COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=sparc64 SHELL=/bin/bash arch/sparc/mm/ drivers/base/regmap/ drivers/media/usb/em28xx/
>
> If you fix the issue, kindly add following tag where applicable
> | Reported-by: kernel test robot <[email protected]>
> | Link: https://lore.kernel.org/oe-kbuild-all/[email protected]/
>
> All error/warnings (new ones prefixed by >>):
>
> In file included from include/linux/gfp.h:7,
> from include/linux/slab.h:15,
> from arch/sparc/mm/tsb.c:9:
> include/linux/mmzone.h:636:33: error: size of array 'free_area' is negative
> 636 | struct free_area free_area[MAX_ORDER];
> | ^~~~~~~~~
> arch/sparc/mm/tsb.c: In function 'tsb_grow':
>>> arch/sparc/mm/tsb.c:405:39: error: left shift count is negative [-Werror=shift-count-negative]
> 405 | if (max_tsb_size > (PAGE_SIZE << MAX_ORDER))
> | ^~
> arch/sparc/mm/tsb.c:406:43: error: left shift count is negative [-Werror=shift-count-negative]
> 406 | max_tsb_size = (PAGE_SIZE << MAX_ORDER);
> | ^~
> cc1: all warnings being treated as errors
> --
> In file included from include/linux/gfp.h:7,
> from include/linux/slab.h:15,
> from drivers/base/regmap/regmap-debugfs.c:9:
> include/linux/mmzone.h:636:33: error: size of array 'free_area' is negative
> 636 | struct free_area free_area[MAX_ORDER];
> | ^~~~~~~~~
> drivers/base/regmap/regmap-debugfs.c: In function 'regmap_read_debugfs':
>>> drivers/base/regmap/regmap-debugfs.c:229:32: warning: left shift count is negative [-Wshift-count-negative]
> 229 | if (count > (PAGE_SIZE << (MAX_ORDER - 1)))
> | ^~
> drivers/base/regmap/regmap-debugfs.c:230:35: warning: left shift count is negative [-Wshift-count-negative]
> 230 | count = PAGE_SIZE << (MAX_ORDER - 1);
> | ^~
> drivers/base/regmap/regmap-debugfs.c: In function 'regmap_reg_ranges_read_file':
> drivers/base/regmap/regmap-debugfs.c:376:32: warning: left shift count is negative [-Wshift-count-negative]
> 376 | if (count > (PAGE_SIZE << (MAX_ORDER - 1)))
> | ^~
> drivers/base/regmap/regmap-debugfs.c:377:35: warning: left shift count is negative [-Wshift-count-negative]
> 377 | count = PAGE_SIZE << (MAX_ORDER - 1);
> | ^~
> --
> In file included from include/linux/gfp.h:7,
> from include/linux/xarray.h:15,
> from include/linux/radix-tree.h:21,
> from include/linux/idr.h:15,
> from include/linux/kernfs.h:12,
> from include/linux/sysfs.h:16,
> from include/linux/kobject.h:20,
> from include/linux/of.h:17,
> from include/linux/irqdomain.h:35,
> from include/linux/acpi.h:13,
> from include/linux/i2c.h:13,
> from drivers/media/usb/em28xx/em28xx.h:22,
> from drivers/media/usb/em28xx/em28xx-audio.c:14:
> include/linux/mmzone.h:636:33: error: size of array 'free_area' is negative
> 636 | struct free_area free_area[MAX_ORDER];
> | ^~~~~~~~~
> drivers/media/usb/em28xx/em28xx-audio.c: In function 'em28xx_audio_init':
>>> drivers/media/usb/em28xx/em28xx-audio.c:808:29: warning: variable 'devnr' set but not used [-Wunused-but-set-variable]
> 808 | static int devnr;
> | ^~~~~
> drivers/media/usb/em28xx/em28xx-audio.c: At top level:
> drivers/media/usb/em28xx/em28xx-audio.c:47:12: warning: 'index' defined but not used [-Wunused-variable]
> 47 | static int index[SNDRV_CARDS] = SNDRV_DEFAULT_IDX;
> | ^~~~~
>
>
> vim +405 arch/sparc/mm/tsb.c
>
> 0871420fad5844 arch/sparc/mm/tsb.c David S. Miller 2008-11-16 379
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 380 /* When the RSS of an address space exceeds tsb_rss_limit for a TSB,
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 381 * do_sparc64_fault() invokes this routine to try and grow it.
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 382 *
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 383 * When we reach the maximum TSB size supported, we stick ~0UL into
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 384 * tsb_rss_limit for that TSB so the grow checks in do_sparc64_fault()
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 385 * will not trigger any longer.
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 386 *
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 387 * The TSB can be anywhere from 8K to 1MB in size, in increasing powers
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 388 * of two. The TSB must be aligned to it's size, so f.e. a 512K TSB
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 389 * must be 512K aligned. It also must be physically contiguous, so we
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 390 * cannot use vmalloc().
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 391 *
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 392 * The idea here is to grow the TSB when the RSS of the process approaches
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 393 * the number of entries that the current TSB can hold at once. Currently,
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 394 * we trigger when the RSS hits 3/4 of the TSB capacity.
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 395 */
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 396 void tsb_grow(struct mm_struct *mm, unsigned long tsb_index, unsigned long rss)
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 397 {
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 398 unsigned long max_tsb_size = 1 * 1024 * 1024;
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 399 unsigned long new_size, old_size, flags;
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 400 struct tsb *old_tsb, *new_tsb;
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 401 unsigned long new_cache_index, old_cache_index;
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 402 unsigned long new_rss_limit;
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 403 gfp_t gfp_flags;
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 404
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 @405 if (max_tsb_size > (PAGE_SIZE << MAX_ORDER))
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 406 max_tsb_size = (PAGE_SIZE << MAX_ORDER);
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 407
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 408 new_cache_index = 0;
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 409 for (new_size = 8192; new_size < max_tsb_size; new_size <<= 1UL) {
> 0871420fad5844 arch/sparc/mm/tsb.c David S. Miller 2008-11-16 410 new_rss_limit = tsb_size_to_rss_limit(new_size);
> 0871420fad5844 arch/sparc/mm/tsb.c David S. Miller 2008-11-16 411 if (new_rss_limit > rss)
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 412 break;
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 413 new_cache_index++;
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 414 }
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 415
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 416 if (new_size == max_tsb_size)
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 417 new_rss_limit = ~0UL;
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 418
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 419 retry_tsb_alloc:
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 420 gfp_flags = GFP_KERNEL;
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 421 if (new_size > (PAGE_SIZE * 2))
> a55ee1ff751f88 arch/sparc/mm/tsb.c David S. Miller 2013-02-19 422 gfp_flags |= __GFP_NOWARN | __GFP_NORETRY;
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 423
> 1f261ef53ba066 arch/sparc64/mm/tsb.c David S. Miller 2008-03-19 424 new_tsb = kmem_cache_alloc_node(tsb_caches[new_cache_index],
> 1f261ef53ba066 arch/sparc64/mm/tsb.c David S. Miller 2008-03-19 425 gfp_flags, numa_node_id());
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 426 if (unlikely(!new_tsb)) {
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 427 /* Not being able to fork due to a high-order TSB
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 428 * allocation failure is very bad behavior. Just back
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 429 * down to a 0-order allocation and force no TSB
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 430 * growing for this address space.
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 431 */
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 432 if (mm->context.tsb_block[tsb_index].tsb == NULL &&
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 433 new_cache_index > 0) {
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 434 new_cache_index = 0;
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 435 new_size = 8192;
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 436 new_rss_limit = ~0UL;
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 437 goto retry_tsb_alloc;
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 438 }
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 439
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 440 /* If we failed on a TSB grow, we are under serious
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 441 * memory pressure so don't try to grow any more.
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 442 */
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 443 if (mm->context.tsb_block[tsb_index].tsb != NULL)
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 444 mm->context.tsb_block[tsb_index].tsb_rss_limit = ~0UL;
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 445 return;
> b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 446 }
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 447
> 8b234274418d6d arch/sparc64/mm/tsb.c David S. Miller 2006-02-17 448 /* Mark all tags as invalid. */
> bb8646d8340fa7 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 449 tsb_init(new_tsb, new_size);
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 450
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 451 /* Ok, we are about to commit the changes. If we are
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 452 * growing an existing TSB the locking is very tricky,
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 453 * so WATCH OUT!
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 454 *
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 455 * We have to hold mm->context.lock while committing to the
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 456 * new TSB, this synchronizes us with processors in
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 457 * flush_tsb_user() and switch_mm() for this address space.
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 458 *
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 459 * But even with that lock held, processors run asynchronously
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 460 * accessing the old TSB via TLB miss handling. This is OK
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 461 * because those actions are just propagating state from the
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 462 * Linux page tables into the TSB, page table mappings are not
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 463 * being changed. If a real fault occurs, the processor will
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 464 * synchronize with us when it hits flush_tsb_user(), this is
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 465 * also true for the case where vmscan is modifying the page
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 466 * tables. The only thing we need to be careful with is to
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 467 * skip any locked TSB entries during copy_tsb().
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 468 *
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 469 * When we finish committing to the new TSB, we have to drop
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 470 * the lock and ask all other cpus running this address space
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 471 * to run tsb_context_switch() to see the new TSB table.
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 472 */
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 473 spin_lock_irqsave(&mm->context.lock, flags);
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 474
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 475 old_tsb = mm->context.tsb_block[tsb_index].tsb;
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 476 old_cache_index =
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 477 (mm->context.tsb_block[tsb_index].tsb_reg_val & 0x7UL);
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 478 old_size = (mm->context.tsb_block[tsb_index].tsb_nentries *
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 479 sizeof(struct tsb));
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 480
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 481
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 482 /* Handle multiple threads trying to grow the TSB at the same time.
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 483 * One will get in here first, and bump the size and the RSS limit.
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 484 * The others will get in here next and hit this check.
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 485 */
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 486 if (unlikely(old_tsb &&
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 487 (rss < mm->context.tsb_block[tsb_index].tsb_rss_limit))) {
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 488 spin_unlock_irqrestore(&mm->context.lock, flags);
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 489
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 490 kmem_cache_free(tsb_caches[new_cache_index], new_tsb);
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 491 return;
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 492 }
> 8b234274418d6d arch/sparc64/mm/tsb.c David S. Miller 2006-02-17 493
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 494 mm->context.tsb_block[tsb_index].tsb_rss_limit = new_rss_limit;
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 495
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 496 if (old_tsb) {
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 497 extern void copy_tsb(unsigned long old_tsb_base,
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 498 unsigned long old_tsb_size,
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 499 unsigned long new_tsb_base,
> 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 500 unsigned long new_tsb_size,
> 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 501 unsigned long page_size_shift);
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 502 unsigned long old_tsb_base = (unsigned long) old_tsb;
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 503 unsigned long new_tsb_base = (unsigned long) new_tsb;
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 504
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 505 if (tlb_type == cheetah_plus || tlb_type == hypervisor) {
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 506 old_tsb_base = __pa(old_tsb_base);
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 507 new_tsb_base = __pa(new_tsb_base);
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 508 }
> 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 509 copy_tsb(old_tsb_base, old_size, new_tsb_base, new_size,
> 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 510 tsb_index == MM_TSB_BASE ?
> 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 511 PAGE_SHIFT : REAL_HPAGE_SHIFT);
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 512 }
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 513
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 514 mm->context.tsb_block[tsb_index].tsb = new_tsb;
> dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 515 setup_tsb_params(mm, tsb_index, new_size);
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 516
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 517 spin_unlock_irqrestore(&mm->context.lock, flags);
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 518
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 519 /* If old_tsb is NULL, we're being invoked for the first time
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 520 * from init_new_context().
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 521 */
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 522 if (old_tsb) {
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 523 /* Reload it on the local cpu. */
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 524 tsb_context_switch(mm);
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 525
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 526 /* Now force other processors to do the same. */
> a3cf5e6b6f2548 arch/sparc64/mm/tsb.c David S. Miller 2008-08-03 527 preempt_disable();
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 528 smp_tsb_sync(mm);
> a3cf5e6b6f2548 arch/sparc64/mm/tsb.c David S. Miller 2008-08-03 529 preempt_enable();
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 530
> 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 531 /* Now it is safe to free the old tsb. */
> 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 532 kmem_cache_free(tsb_caches[old_cache_index], old_tsb);
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 533 }
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 534 }
> bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 535
>
> :::::: The code at line 405 was first introduced by commit
> :::::: bd40791e1d289d807b8580abe1f117e9c62894e4 [SPARC64]: Dynamically grow TSB in response to RSS growth.
>
> :::::: TO: David S. Miller <[email protected]>
> :::::: CC: David S. Miller <[email protected]>
>
> --
> 0-DAY CI Kernel Test Service
> https://github.com/intel/lkp-tests
--
Best Regards,
Yan, Zi
On Thu, Mar 23, 2023 at 07:50:01AM -0400, Zi Yan wrote:
> On 23 Mar 2023, at 6:40, kernel test robot wrote:
>
> > Hi Zi,
> >
> > FYI, the error/warning was bisected to this commit, please ignore it if it's irrelevant.
> >
> > tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> > head: fff5a5e7f528b2ed2c335991399a766c2cf01103
> > commit: 0192445cb2f7ed1cd7a95a0fc8c7645480baba25 arch: mm: rename FORCE_MAX_ZONEORDER to ARCH_FORCE_MAX_ORDER
> > date: 6 months ago
> > config: sparc64-buildonly-randconfig-r002-20230322 (https://download.01.org/0day-ci/archive/20230323/[email protected]/config)
>
> Cc Mike, since he is working on ARCH_FORCE_MAX_ORDER.
>
> This random config sets CONFIG_ARCH_FORCE_MAX_ORDER to -1, causing this error. I am not sure if we need to always set a range for ARCH_FORCE_MAX_ORDER or random config generator should be changed to not use non-positive values.
I believe it's the same issue as with Kirill's MAX_ORDER change [1].
Apparently kbuild used an old config as the template and renaming of
Kconfig option caused it to be set to -1.
_
[1] https://lore.kernel.org/all/[email protected]/
> BTW, this issue pre-exists my changes.
>
> > compiler: sparc64-linux-gcc (GCC) 12.1.0
> > 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
> > # https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=0192445cb2f7ed1cd7a95a0fc8c7645480baba25
> > git remote add linus https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
> > git fetch --no-tags linus master
> > git checkout 0192445cb2f7ed1cd7a95a0fc8c7645480baba25
> > # save the config file
> > mkdir build_dir && cp config build_dir/.config
> > COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=sparc64 olddefconfig
> > COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=sparc64 SHELL=/bin/bash arch/sparc/mm/ drivers/base/regmap/ drivers/media/usb/em28xx/
> >
> > If you fix the issue, kindly add following tag where applicable
> > | Reported-by: kernel test robot <[email protected]>
> > | Link: https://lore.kernel.org/oe-kbuild-all/[email protected]/
> >
> > All error/warnings (new ones prefixed by >>):
> >
> > In file included from include/linux/gfp.h:7,
> > from include/linux/slab.h:15,
> > from arch/sparc/mm/tsb.c:9:
> > include/linux/mmzone.h:636:33: error: size of array 'free_area' is negative
> > 636 | struct free_area free_area[MAX_ORDER];
> > | ^~~~~~~~~
> > arch/sparc/mm/tsb.c: In function 'tsb_grow':
> >>> arch/sparc/mm/tsb.c:405:39: error: left shift count is negative [-Werror=shift-count-negative]
> > 405 | if (max_tsb_size > (PAGE_SIZE << MAX_ORDER))
> > | ^~
> > arch/sparc/mm/tsb.c:406:43: error: left shift count is negative [-Werror=shift-count-negative]
> > 406 | max_tsb_size = (PAGE_SIZE << MAX_ORDER);
> > | ^~
> > cc1: all warnings being treated as errors
> > --
> > In file included from include/linux/gfp.h:7,
> > from include/linux/slab.h:15,
> > from drivers/base/regmap/regmap-debugfs.c:9:
> > include/linux/mmzone.h:636:33: error: size of array 'free_area' is negative
> > 636 | struct free_area free_area[MAX_ORDER];
> > | ^~~~~~~~~
> > drivers/base/regmap/regmap-debugfs.c: In function 'regmap_read_debugfs':
> >>> drivers/base/regmap/regmap-debugfs.c:229:32: warning: left shift count is negative [-Wshift-count-negative]
> > 229 | if (count > (PAGE_SIZE << (MAX_ORDER - 1)))
> > | ^~
> > drivers/base/regmap/regmap-debugfs.c:230:35: warning: left shift count is negative [-Wshift-count-negative]
> > 230 | count = PAGE_SIZE << (MAX_ORDER - 1);
> > | ^~
> > drivers/base/regmap/regmap-debugfs.c: In function 'regmap_reg_ranges_read_file':
> > drivers/base/regmap/regmap-debugfs.c:376:32: warning: left shift count is negative [-Wshift-count-negative]
> > 376 | if (count > (PAGE_SIZE << (MAX_ORDER - 1)))
> > | ^~
> > drivers/base/regmap/regmap-debugfs.c:377:35: warning: left shift count is negative [-Wshift-count-negative]
> > 377 | count = PAGE_SIZE << (MAX_ORDER - 1);
> > | ^~
> > --
> > In file included from include/linux/gfp.h:7,
> > from include/linux/xarray.h:15,
> > from include/linux/radix-tree.h:21,
> > from include/linux/idr.h:15,
> > from include/linux/kernfs.h:12,
> > from include/linux/sysfs.h:16,
> > from include/linux/kobject.h:20,
> > from include/linux/of.h:17,
> > from include/linux/irqdomain.h:35,
> > from include/linux/acpi.h:13,
> > from include/linux/i2c.h:13,
> > from drivers/media/usb/em28xx/em28xx.h:22,
> > from drivers/media/usb/em28xx/em28xx-audio.c:14:
> > include/linux/mmzone.h:636:33: error: size of array 'free_area' is negative
> > 636 | struct free_area free_area[MAX_ORDER];
> > | ^~~~~~~~~
> > drivers/media/usb/em28xx/em28xx-audio.c: In function 'em28xx_audio_init':
> >>> drivers/media/usb/em28xx/em28xx-audio.c:808:29: warning: variable 'devnr' set but not used [-Wunused-but-set-variable]
> > 808 | static int devnr;
> > | ^~~~~
> > drivers/media/usb/em28xx/em28xx-audio.c: At top level:
> > drivers/media/usb/em28xx/em28xx-audio.c:47:12: warning: 'index' defined but not used [-Wunused-variable]
> > 47 | static int index[SNDRV_CARDS] = SNDRV_DEFAULT_IDX;
> > | ^~~~~
> >
> >
> > vim +405 arch/sparc/mm/tsb.c
> >
> > 0871420fad5844 arch/sparc/mm/tsb.c David S. Miller 2008-11-16 379
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 380 /* When the RSS of an address space exceeds tsb_rss_limit for a TSB,
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 381 * do_sparc64_fault() invokes this routine to try and grow it.
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 382 *
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 383 * When we reach the maximum TSB size supported, we stick ~0UL into
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 384 * tsb_rss_limit for that TSB so the grow checks in do_sparc64_fault()
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 385 * will not trigger any longer.
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 386 *
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 387 * The TSB can be anywhere from 8K to 1MB in size, in increasing powers
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 388 * of two. The TSB must be aligned to it's size, so f.e. a 512K TSB
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 389 * must be 512K aligned. It also must be physically contiguous, so we
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 390 * cannot use vmalloc().
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 391 *
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 392 * The idea here is to grow the TSB when the RSS of the process approaches
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 393 * the number of entries that the current TSB can hold at once. Currently,
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 394 * we trigger when the RSS hits 3/4 of the TSB capacity.
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 395 */
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 396 void tsb_grow(struct mm_struct *mm, unsigned long tsb_index, unsigned long rss)
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 397 {
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 398 unsigned long max_tsb_size = 1 * 1024 * 1024;
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 399 unsigned long new_size, old_size, flags;
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 400 struct tsb *old_tsb, *new_tsb;
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 401 unsigned long new_cache_index, old_cache_index;
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 402 unsigned long new_rss_limit;
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 403 gfp_t gfp_flags;
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 404
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 @405 if (max_tsb_size > (PAGE_SIZE << MAX_ORDER))
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 406 max_tsb_size = (PAGE_SIZE << MAX_ORDER);
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 407
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 408 new_cache_index = 0;
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 409 for (new_size = 8192; new_size < max_tsb_size; new_size <<= 1UL) {
> > 0871420fad5844 arch/sparc/mm/tsb.c David S. Miller 2008-11-16 410 new_rss_limit = tsb_size_to_rss_limit(new_size);
> > 0871420fad5844 arch/sparc/mm/tsb.c David S. Miller 2008-11-16 411 if (new_rss_limit > rss)
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 412 break;
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 413 new_cache_index++;
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 414 }
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 415
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 416 if (new_size == max_tsb_size)
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 417 new_rss_limit = ~0UL;
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 418
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 419 retry_tsb_alloc:
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 420 gfp_flags = GFP_KERNEL;
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 421 if (new_size > (PAGE_SIZE * 2))
> > a55ee1ff751f88 arch/sparc/mm/tsb.c David S. Miller 2013-02-19 422 gfp_flags |= __GFP_NOWARN | __GFP_NORETRY;
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 423
> > 1f261ef53ba066 arch/sparc64/mm/tsb.c David S. Miller 2008-03-19 424 new_tsb = kmem_cache_alloc_node(tsb_caches[new_cache_index],
> > 1f261ef53ba066 arch/sparc64/mm/tsb.c David S. Miller 2008-03-19 425 gfp_flags, numa_node_id());
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 426 if (unlikely(!new_tsb)) {
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 427 /* Not being able to fork due to a high-order TSB
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 428 * allocation failure is very bad behavior. Just back
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 429 * down to a 0-order allocation and force no TSB
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 430 * growing for this address space.
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 431 */
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 432 if (mm->context.tsb_block[tsb_index].tsb == NULL &&
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 433 new_cache_index > 0) {
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 434 new_cache_index = 0;
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 435 new_size = 8192;
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 436 new_rss_limit = ~0UL;
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 437 goto retry_tsb_alloc;
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 438 }
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 439
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 440 /* If we failed on a TSB grow, we are under serious
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 441 * memory pressure so don't try to grow any more.
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 442 */
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 443 if (mm->context.tsb_block[tsb_index].tsb != NULL)
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 444 mm->context.tsb_block[tsb_index].tsb_rss_limit = ~0UL;
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 445 return;
> > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 446 }
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 447
> > 8b234274418d6d arch/sparc64/mm/tsb.c David S. Miller 2006-02-17 448 /* Mark all tags as invalid. */
> > bb8646d8340fa7 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 449 tsb_init(new_tsb, new_size);
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 450
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 451 /* Ok, we are about to commit the changes. If we are
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 452 * growing an existing TSB the locking is very tricky,
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 453 * so WATCH OUT!
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 454 *
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 455 * We have to hold mm->context.lock while committing to the
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 456 * new TSB, this synchronizes us with processors in
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 457 * flush_tsb_user() and switch_mm() for this address space.
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 458 *
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 459 * But even with that lock held, processors run asynchronously
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 460 * accessing the old TSB via TLB miss handling. This is OK
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 461 * because those actions are just propagating state from the
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 462 * Linux page tables into the TSB, page table mappings are not
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 463 * being changed. If a real fault occurs, the processor will
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 464 * synchronize with us when it hits flush_tsb_user(), this is
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 465 * also true for the case where vmscan is modifying the page
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 466 * tables. The only thing we need to be careful with is to
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 467 * skip any locked TSB entries during copy_tsb().
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 468 *
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 469 * When we finish committing to the new TSB, we have to drop
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 470 * the lock and ask all other cpus running this address space
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 471 * to run tsb_context_switch() to see the new TSB table.
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 472 */
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 473 spin_lock_irqsave(&mm->context.lock, flags);
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 474
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 475 old_tsb = mm->context.tsb_block[tsb_index].tsb;
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 476 old_cache_index =
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 477 (mm->context.tsb_block[tsb_index].tsb_reg_val & 0x7UL);
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 478 old_size = (mm->context.tsb_block[tsb_index].tsb_nentries *
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 479 sizeof(struct tsb));
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 480
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 481
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 482 /* Handle multiple threads trying to grow the TSB at the same time.
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 483 * One will get in here first, and bump the size and the RSS limit.
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 484 * The others will get in here next and hit this check.
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 485 */
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 486 if (unlikely(old_tsb &&
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 487 (rss < mm->context.tsb_block[tsb_index].tsb_rss_limit))) {
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 488 spin_unlock_irqrestore(&mm->context.lock, flags);
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 489
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 490 kmem_cache_free(tsb_caches[new_cache_index], new_tsb);
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 491 return;
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 492 }
> > 8b234274418d6d arch/sparc64/mm/tsb.c David S. Miller 2006-02-17 493
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 494 mm->context.tsb_block[tsb_index].tsb_rss_limit = new_rss_limit;
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 495
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 496 if (old_tsb) {
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 497 extern void copy_tsb(unsigned long old_tsb_base,
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 498 unsigned long old_tsb_size,
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 499 unsigned long new_tsb_base,
> > 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 500 unsigned long new_tsb_size,
> > 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 501 unsigned long page_size_shift);
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 502 unsigned long old_tsb_base = (unsigned long) old_tsb;
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 503 unsigned long new_tsb_base = (unsigned long) new_tsb;
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 504
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 505 if (tlb_type == cheetah_plus || tlb_type == hypervisor) {
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 506 old_tsb_base = __pa(old_tsb_base);
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 507 new_tsb_base = __pa(new_tsb_base);
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 508 }
> > 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 509 copy_tsb(old_tsb_base, old_size, new_tsb_base, new_size,
> > 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 510 tsb_index == MM_TSB_BASE ?
> > 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 511 PAGE_SHIFT : REAL_HPAGE_SHIFT);
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 512 }
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 513
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 514 mm->context.tsb_block[tsb_index].tsb = new_tsb;
> > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 515 setup_tsb_params(mm, tsb_index, new_size);
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 516
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 517 spin_unlock_irqrestore(&mm->context.lock, flags);
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 518
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 519 /* If old_tsb is NULL, we're being invoked for the first time
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 520 * from init_new_context().
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 521 */
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 522 if (old_tsb) {
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 523 /* Reload it on the local cpu. */
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 524 tsb_context_switch(mm);
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 525
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 526 /* Now force other processors to do the same. */
> > a3cf5e6b6f2548 arch/sparc64/mm/tsb.c David S. Miller 2008-08-03 527 preempt_disable();
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 528 smp_tsb_sync(mm);
> > a3cf5e6b6f2548 arch/sparc64/mm/tsb.c David S. Miller 2008-08-03 529 preempt_enable();
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 530
> > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 531 /* Now it is safe to free the old tsb. */
> > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 532 kmem_cache_free(tsb_caches[old_cache_index], old_tsb);
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 533 }
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 534 }
> > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 535
> >
> > :::::: The code at line 405 was first introduced by commit
> > :::::: bd40791e1d289d807b8580abe1f117e9c62894e4 [SPARC64]: Dynamically grow TSB in response to RSS growth.
> >
> > :::::: TO: David S. Miller <[email protected]>
> > :::::: CC: David S. Miller <[email protected]>
> >
> > --
> > 0-DAY CI Kernel Test Service
> > https://github.com/intel/lkp-tests
>
> --
> Best Regards,
> Yan, Zi
--
Sincerely yours,
Mike.
On Thu, Mar 23, 2023 at 02:13:22PM +0200, Mike Rapoport wrote:
> On Thu, Mar 23, 2023 at 07:50:01AM -0400, Zi Yan wrote:
> > On 23 Mar 2023, at 6:40, kernel test robot wrote:
> >
> > > Hi Zi,
> > >
> > > FYI, the error/warning was bisected to this commit, please ignore it if it's irrelevant.
> > >
> > > tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> > > head: fff5a5e7f528b2ed2c335991399a766c2cf01103
> > > commit: 0192445cb2f7ed1cd7a95a0fc8c7645480baba25 arch: mm: rename FORCE_MAX_ZONEORDER to ARCH_FORCE_MAX_ORDER
> > > date: 6 months ago
> > > config: sparc64-buildonly-randconfig-r002-20230322 (https://download.01.org/0day-ci/archive/20230323/[email protected]/config)
> >
> > Cc Mike, since he is working on ARCH_FORCE_MAX_ORDER.
> >
> > This random config sets CONFIG_ARCH_FORCE_MAX_ORDER to -1, causing this error. I am not sure if we need to always set a range for ARCH_FORCE_MAX_ORDER or random config generator should be changed to not use non-positive values.
>
> I believe it's the same issue as with Kirill's MAX_ORDER change [1].
> Apparently kbuild used an old config as the template and renaming of
> Kconfig option caused it to be set to -1.
Sorry for the false positive, -1 is wrongly set here, we will look into
it and resolve the bug.
> _
> [1] https://lore.kernel.org/all/[email protected]/
>
> > BTW, this issue pre-exists my changes.
> >
> > > compiler: sparc64-linux-gcc (GCC) 12.1.0
> > > 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
> > > # https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=0192445cb2f7ed1cd7a95a0fc8c7645480baba25
> > > git remote add linus https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
> > > git fetch --no-tags linus master
> > > git checkout 0192445cb2f7ed1cd7a95a0fc8c7645480baba25
> > > # save the config file
> > > mkdir build_dir && cp config build_dir/.config
> > > COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=sparc64 olddefconfig
> > > COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-12.1.0 make.cross W=1 O=build_dir ARCH=sparc64 SHELL=/bin/bash arch/sparc/mm/ drivers/base/regmap/ drivers/media/usb/em28xx/
> > >
> > > If you fix the issue, kindly add following tag where applicable
> > > | Reported-by: kernel test robot <[email protected]>
> > > | Link: https://lore.kernel.org/oe-kbuild-all/[email protected]/
> > >
> > > All error/warnings (new ones prefixed by >>):
> > >
> > > In file included from include/linux/gfp.h:7,
> > > from include/linux/slab.h:15,
> > > from arch/sparc/mm/tsb.c:9:
> > > include/linux/mmzone.h:636:33: error: size of array 'free_area' is negative
> > > 636 | struct free_area free_area[MAX_ORDER];
> > > | ^~~~~~~~~
> > > arch/sparc/mm/tsb.c: In function 'tsb_grow':
> > >>> arch/sparc/mm/tsb.c:405:39: error: left shift count is negative [-Werror=shift-count-negative]
> > > 405 | if (max_tsb_size > (PAGE_SIZE << MAX_ORDER))
> > > | ^~
> > > arch/sparc/mm/tsb.c:406:43: error: left shift count is negative [-Werror=shift-count-negative]
> > > 406 | max_tsb_size = (PAGE_SIZE << MAX_ORDER);
> > > | ^~
> > > cc1: all warnings being treated as errors
> > > --
> > > In file included from include/linux/gfp.h:7,
> > > from include/linux/slab.h:15,
> > > from drivers/base/regmap/regmap-debugfs.c:9:
> > > include/linux/mmzone.h:636:33: error: size of array 'free_area' is negative
> > > 636 | struct free_area free_area[MAX_ORDER];
> > > | ^~~~~~~~~
> > > drivers/base/regmap/regmap-debugfs.c: In function 'regmap_read_debugfs':
> > >>> drivers/base/regmap/regmap-debugfs.c:229:32: warning: left shift count is negative [-Wshift-count-negative]
> > > 229 | if (count > (PAGE_SIZE << (MAX_ORDER - 1)))
> > > | ^~
> > > drivers/base/regmap/regmap-debugfs.c:230:35: warning: left shift count is negative [-Wshift-count-negative]
> > > 230 | count = PAGE_SIZE << (MAX_ORDER - 1);
> > > | ^~
> > > drivers/base/regmap/regmap-debugfs.c: In function 'regmap_reg_ranges_read_file':
> > > drivers/base/regmap/regmap-debugfs.c:376:32: warning: left shift count is negative [-Wshift-count-negative]
> > > 376 | if (count > (PAGE_SIZE << (MAX_ORDER - 1)))
> > > | ^~
> > > drivers/base/regmap/regmap-debugfs.c:377:35: warning: left shift count is negative [-Wshift-count-negative]
> > > 377 | count = PAGE_SIZE << (MAX_ORDER - 1);
> > > | ^~
> > > --
> > > In file included from include/linux/gfp.h:7,
> > > from include/linux/xarray.h:15,
> > > from include/linux/radix-tree.h:21,
> > > from include/linux/idr.h:15,
> > > from include/linux/kernfs.h:12,
> > > from include/linux/sysfs.h:16,
> > > from include/linux/kobject.h:20,
> > > from include/linux/of.h:17,
> > > from include/linux/irqdomain.h:35,
> > > from include/linux/acpi.h:13,
> > > from include/linux/i2c.h:13,
> > > from drivers/media/usb/em28xx/em28xx.h:22,
> > > from drivers/media/usb/em28xx/em28xx-audio.c:14:
> > > include/linux/mmzone.h:636:33: error: size of array 'free_area' is negative
> > > 636 | struct free_area free_area[MAX_ORDER];
> > > | ^~~~~~~~~
> > > drivers/media/usb/em28xx/em28xx-audio.c: In function 'em28xx_audio_init':
> > >>> drivers/media/usb/em28xx/em28xx-audio.c:808:29: warning: variable 'devnr' set but not used [-Wunused-but-set-variable]
> > > 808 | static int devnr;
> > > | ^~~~~
> > > drivers/media/usb/em28xx/em28xx-audio.c: At top level:
> > > drivers/media/usb/em28xx/em28xx-audio.c:47:12: warning: 'index' defined but not used [-Wunused-variable]
> > > 47 | static int index[SNDRV_CARDS] = SNDRV_DEFAULT_IDX;
> > > | ^~~~~
> > >
> > >
> > > vim +405 arch/sparc/mm/tsb.c
> > >
> > > 0871420fad5844 arch/sparc/mm/tsb.c David S. Miller 2008-11-16 379
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 380 /* When the RSS of an address space exceeds tsb_rss_limit for a TSB,
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 381 * do_sparc64_fault() invokes this routine to try and grow it.
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 382 *
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 383 * When we reach the maximum TSB size supported, we stick ~0UL into
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 384 * tsb_rss_limit for that TSB so the grow checks in do_sparc64_fault()
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 385 * will not trigger any longer.
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 386 *
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 387 * The TSB can be anywhere from 8K to 1MB in size, in increasing powers
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 388 * of two. The TSB must be aligned to it's size, so f.e. a 512K TSB
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 389 * must be 512K aligned. It also must be physically contiguous, so we
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 390 * cannot use vmalloc().
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 391 *
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 392 * The idea here is to grow the TSB when the RSS of the process approaches
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 393 * the number of entries that the current TSB can hold at once. Currently,
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 394 * we trigger when the RSS hits 3/4 of the TSB capacity.
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 395 */
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 396 void tsb_grow(struct mm_struct *mm, unsigned long tsb_index, unsigned long rss)
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 397 {
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 398 unsigned long max_tsb_size = 1 * 1024 * 1024;
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 399 unsigned long new_size, old_size, flags;
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 400 struct tsb *old_tsb, *new_tsb;
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 401 unsigned long new_cache_index, old_cache_index;
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 402 unsigned long new_rss_limit;
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 403 gfp_t gfp_flags;
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 404
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 @405 if (max_tsb_size > (PAGE_SIZE << MAX_ORDER))
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 406 max_tsb_size = (PAGE_SIZE << MAX_ORDER);
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 407
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 408 new_cache_index = 0;
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 409 for (new_size = 8192; new_size < max_tsb_size; new_size <<= 1UL) {
> > > 0871420fad5844 arch/sparc/mm/tsb.c David S. Miller 2008-11-16 410 new_rss_limit = tsb_size_to_rss_limit(new_size);
> > > 0871420fad5844 arch/sparc/mm/tsb.c David S. Miller 2008-11-16 411 if (new_rss_limit > rss)
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 412 break;
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 413 new_cache_index++;
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 414 }
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 415
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 416 if (new_size == max_tsb_size)
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 417 new_rss_limit = ~0UL;
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 418
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 419 retry_tsb_alloc:
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 420 gfp_flags = GFP_KERNEL;
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 421 if (new_size > (PAGE_SIZE * 2))
> > > a55ee1ff751f88 arch/sparc/mm/tsb.c David S. Miller 2013-02-19 422 gfp_flags |= __GFP_NOWARN | __GFP_NORETRY;
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 423
> > > 1f261ef53ba066 arch/sparc64/mm/tsb.c David S. Miller 2008-03-19 424 new_tsb = kmem_cache_alloc_node(tsb_caches[new_cache_index],
> > > 1f261ef53ba066 arch/sparc64/mm/tsb.c David S. Miller 2008-03-19 425 gfp_flags, numa_node_id());
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 426 if (unlikely(!new_tsb)) {
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 427 /* Not being able to fork due to a high-order TSB
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 428 * allocation failure is very bad behavior. Just back
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 429 * down to a 0-order allocation and force no TSB
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 430 * growing for this address space.
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 431 */
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 432 if (mm->context.tsb_block[tsb_index].tsb == NULL &&
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 433 new_cache_index > 0) {
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 434 new_cache_index = 0;
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 435 new_size = 8192;
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 436 new_rss_limit = ~0UL;
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 437 goto retry_tsb_alloc;
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 438 }
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 439
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 440 /* If we failed on a TSB grow, we are under serious
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 441 * memory pressure so don't try to grow any more.
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 442 */
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 443 if (mm->context.tsb_block[tsb_index].tsb != NULL)
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 444 mm->context.tsb_block[tsb_index].tsb_rss_limit = ~0UL;
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 445 return;
> > > b52439c22c63db arch/sparc64/mm/tsb.c David S. Miller 2006-03-17 446 }
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 447
> > > 8b234274418d6d arch/sparc64/mm/tsb.c David S. Miller 2006-02-17 448 /* Mark all tags as invalid. */
> > > bb8646d8340fa7 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 449 tsb_init(new_tsb, new_size);
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 450
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 451 /* Ok, we are about to commit the changes. If we are
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 452 * growing an existing TSB the locking is very tricky,
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 453 * so WATCH OUT!
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 454 *
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 455 * We have to hold mm->context.lock while committing to the
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 456 * new TSB, this synchronizes us with processors in
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 457 * flush_tsb_user() and switch_mm() for this address space.
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 458 *
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 459 * But even with that lock held, processors run asynchronously
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 460 * accessing the old TSB via TLB miss handling. This is OK
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 461 * because those actions are just propagating state from the
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 462 * Linux page tables into the TSB, page table mappings are not
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 463 * being changed. If a real fault occurs, the processor will
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 464 * synchronize with us when it hits flush_tsb_user(), this is
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 465 * also true for the case where vmscan is modifying the page
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 466 * tables. The only thing we need to be careful with is to
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 467 * skip any locked TSB entries during copy_tsb().
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 468 *
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 469 * When we finish committing to the new TSB, we have to drop
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 470 * the lock and ask all other cpus running this address space
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 471 * to run tsb_context_switch() to see the new TSB table.
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 472 */
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 473 spin_lock_irqsave(&mm->context.lock, flags);
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 474
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 475 old_tsb = mm->context.tsb_block[tsb_index].tsb;
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 476 old_cache_index =
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 477 (mm->context.tsb_block[tsb_index].tsb_reg_val & 0x7UL);
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 478 old_size = (mm->context.tsb_block[tsb_index].tsb_nentries *
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 479 sizeof(struct tsb));
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 480
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 481
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 482 /* Handle multiple threads trying to grow the TSB at the same time.
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 483 * One will get in here first, and bump the size and the RSS limit.
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 484 * The others will get in here next and hit this check.
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 485 */
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 486 if (unlikely(old_tsb &&
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 487 (rss < mm->context.tsb_block[tsb_index].tsb_rss_limit))) {
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 488 spin_unlock_irqrestore(&mm->context.lock, flags);
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 489
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 490 kmem_cache_free(tsb_caches[new_cache_index], new_tsb);
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 491 return;
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 492 }
> > > 8b234274418d6d arch/sparc64/mm/tsb.c David S. Miller 2006-02-17 493
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 494 mm->context.tsb_block[tsb_index].tsb_rss_limit = new_rss_limit;
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 495
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 496 if (old_tsb) {
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 497 extern void copy_tsb(unsigned long old_tsb_base,
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 498 unsigned long old_tsb_size,
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 499 unsigned long new_tsb_base,
> > > 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 500 unsigned long new_tsb_size,
> > > 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 501 unsigned long page_size_shift);
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 502 unsigned long old_tsb_base = (unsigned long) old_tsb;
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 503 unsigned long new_tsb_base = (unsigned long) new_tsb;
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 504
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 505 if (tlb_type == cheetah_plus || tlb_type == hypervisor) {
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 506 old_tsb_base = __pa(old_tsb_base);
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 507 new_tsb_base = __pa(new_tsb_base);
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 508 }
> > > 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 509 copy_tsb(old_tsb_base, old_size, new_tsb_base, new_size,
> > > 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 510 tsb_index == MM_TSB_BASE ?
> > > 654f4807624a65 arch/sparc/mm/tsb.c Mike Kravetz 2017-06-02 511 PAGE_SHIFT : REAL_HPAGE_SHIFT);
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 512 }
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 513
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 514 mm->context.tsb_block[tsb_index].tsb = new_tsb;
> > > dcc1e8dd88d4bc arch/sparc64/mm/tsb.c David S. Miller 2006-03-22 515 setup_tsb_params(mm, tsb_index, new_size);
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 516
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 517 spin_unlock_irqrestore(&mm->context.lock, flags);
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 518
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 519 /* If old_tsb is NULL, we're being invoked for the first time
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 520 * from init_new_context().
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 521 */
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 522 if (old_tsb) {
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 523 /* Reload it on the local cpu. */
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 524 tsb_context_switch(mm);
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 525
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 526 /* Now force other processors to do the same. */
> > > a3cf5e6b6f2548 arch/sparc64/mm/tsb.c David S. Miller 2008-08-03 527 preempt_disable();
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 528 smp_tsb_sync(mm);
> > > a3cf5e6b6f2548 arch/sparc64/mm/tsb.c David S. Miller 2008-08-03 529 preempt_enable();
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 530
> > > 7a1ac5264108fc arch/sparc64/mm/tsb.c David S. Miller 2006-03-16 531 /* Now it is safe to free the old tsb. */
> > > 9b4006dcf6a8c4 arch/sparc64/mm/tsb.c David S. Miller 2006-03-18 532 kmem_cache_free(tsb_caches[old_cache_index], old_tsb);
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 533 }
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 534 }
> > > bd40791e1d289d arch/sparc64/mm/tsb.c David S. Miller 2006-01-31 535
> > >
> > > :::::: The code at line 405 was first introduced by commit
> > > :::::: bd40791e1d289d807b8580abe1f117e9c62894e4 [SPARC64]: Dynamically grow TSB in response to RSS growth.
> > >
> > > :::::: TO: David S. Miller <[email protected]>
> > > :::::: CC: David S. Miller <[email protected]>
> > >
> > > --
> > > 0-DAY CI Kernel Test Service
> > > https://github.com/intel/lkp-tests
> >
> > --
> > Best Regards,
> > Yan, Zi
>
>
>
> --
> Sincerely yours,
> Mike.
>