platform/kernel/linux-starfive.git
2 years agofs/jffs2: fix comments mentioning i_mutex
hongnanli [Fri, 11 Feb 2022 03:19:28 +0000 (11:19 +0800)]
fs/jffs2: fix comments mentioning i_mutex

inode->i_mutex has been replaced with inode->i_rwsem long ago. Fix
comments still mentioning i_mutex.

Signed-off-by: hongnanli <hongnan.li@linux.alibaba.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubi: fastmap: Return error code if memory allocation fails in add_aeb()
Zhihao Cheng [Mon, 27 Dec 2021 03:22:42 +0000 (11:22 +0800)]
ubi: fastmap: Return error code if memory allocation fails in add_aeb()

Abort fastmap scanning and return error code if memory allocation fails
in add_aeb(). Otherwise ubi will get wrong peb statistics information
after scanning.

Fixes: dbb7d2a88d2a7b ("UBI: Add fastmap core")
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Fix to add refcount once page is set private
Zhihao Cheng [Mon, 27 Dec 2021 03:22:41 +0000 (11:22 +0800)]
ubifs: Fix to add refcount once page is set private

MM defined the rule [1] very clearly that once page was set with PG_private
flag, we should increment the refcount in that page, also main flows like
pageout(), migrate_page() will assume there is one additional page
reference count if page_has_private() returns true. Otherwise, we may
get a BUG in page migration:

  page:0000000080d05b9d refcount:-1 mapcount:0 mapping:000000005f4d82a8
  index:0xe2 pfn:0x14c12
  aops:ubifs_file_address_operations [ubifs] ino:8f1 dentry name:"f30e"
  flags: 0x1fffff80002405(locked|uptodate|owner_priv_1|private|node=0|
  zone=1|lastcpupid=0x1fffff)
  page dumped because: VM_BUG_ON_PAGE(page_count(page) != 0)
  ------------[ cut here ]------------
  kernel BUG at include/linux/page_ref.h:184!
  invalid opcode: 0000 [#1] SMP
  CPU: 3 PID: 38 Comm: kcompactd0 Not tainted 5.15.0-rc5
  RIP: 0010:migrate_page_move_mapping+0xac3/0xe70
  Call Trace:
    ubifs_migrate_page+0x22/0xc0 [ubifs]
    move_to_new_page+0xb4/0x600
    migrate_pages+0x1523/0x1cc0
    compact_zone+0x8c5/0x14b0
    kcompactd+0x2bc/0x560
    kthread+0x18c/0x1e0
    ret_from_fork+0x1f/0x30

Before the time, we should make clean a concept, what does refcount means
in page gotten from grab_cache_page_write_begin(). There are 2 situations:
Situation 1: refcount is 3, page is created by __page_cache_alloc.
  TYPE_A - the write process is using this page
  TYPE_B - page is assigned to one certain mapping by calling
   __add_to_page_cache_locked()
  TYPE_C - page is added into pagevec list corresponding current cpu by
   calling lru_cache_add()
Situation 2: refcount is 2, page is gotten from the mapping's tree
  TYPE_B - page has been assigned to one certain mapping
  TYPE_A - the write process is using this page (by calling
   page_cache_get_speculative())
Filesystem releases one refcount by calling put_page() in xxx_write_end(),
the released refcount corresponds to TYPE_A (write task is using it). If
there are any processes using a page, page migration process will skip the
page by judging whether expected_page_refs() equals to page refcount.

The BUG is caused by following process:
    PA(cpu 0)                           kcompactd(cpu 1)
compact_zone
ubifs_write_begin
  page_a = grab_cache_page_write_begin
    add_to_page_cache_lru
      lru_cache_add
        pagevec_add // put page into cpu 0's pagevec
  (refcnf = 3, for page creation process)
ubifs_write_end
  SetPagePrivate(page_a) // doesn't increase page count !
  unlock_page(page_a)
  put_page(page_a)  // refcnt = 2
[...]

    PB(cpu 0)
filemap_read
  filemap_get_pages
    add_to_page_cache_lru
      lru_cache_add
        __pagevec_lru_add // traverse all pages in cpu 0's pagevec
  __pagevec_lru_add_fn
    SetPageLRU(page_a)
isolate_migratepages
                                  isolate_migratepages_block
    get_page_unless_zero(page_a)
    // refcnt = 3
                                      list_add(page_a, from_list)
migrate_pages(from_list)
  __unmap_and_move
    move_to_new_page
      ubifs_migrate_page(page_a)
        migrate_page_move_mapping
  expected_page_refs get 3
                                  (migration[1] + mapping[1] + private[1])
 release_pages
   put_page_testzero(page_a) // refcnt = 3
                                          page_ref_freeze  // refcnt = 0
     page_ref_dec_and_test(0 - 1 = -1)
                                          page_ref_unfreeze
                                            VM_BUG_ON_PAGE(-1 != 0, page)

UBIFS doesn't increase the page refcount after setting private flag, which
leads to page migration task believes the page is not used by any other
processes, so the page is migrated. This causes concurrent accessing on
page refcount between put_page() called by other process(eg. read process
calls lru_cache_add) and page_ref_unfreeze() called by migration task.

Actually zhangjun has tried to fix this problem [2] by recalculating page
refcnt in ubifs_migrate_page(). It's better to follow MM rules [1], because
just like Kirill suggested in [2], we need to check all users of
page_has_private() helper. Like f2fs does in [3], fix it by adding/deleting
refcount when setting/clearing private for a page. BTW, according to [4],
we set 'page->private' as 1 because ubifs just simply SetPagePrivate().
And, [5] provided a common helper to set/clear page private, ubifs can
use this helper following the example of iomap, afs, btrfs, etc.

Jump [6] to find a reproducer.

[1] https://lore.kernel.org/lkml/2b19b3c4-2bc4-15fa-15cc-27a13e5c7af1@aol.com
[2] https://www.spinics.net/lists/linux-mtd/msg04018.html
[3] http://lkml.iu.edu/hypermail/linux/kernel/1903.0/03313.html
[4] https://lore.kernel.org/linux-f2fs-devel/20210422154705.GO3596236@casper.infradead.org
[5] https://lore.kernel.org/all/20200517214718.468-1-guoqing.jiang@cloud.ionos.com
[6] https://bugzilla.kernel.org/show_bug.cgi?id=214961

Fixes: 1e51764a3c2ac0 ("UBIFS: add new flash file system")
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Fix read out-of-bounds in ubifs_wbuf_write_nolock()
Zhihao Cheng [Mon, 27 Dec 2021 03:22:40 +0000 (11:22 +0800)]
ubifs: Fix read out-of-bounds in ubifs_wbuf_write_nolock()

Function ubifs_wbuf_write_nolock() may access buf out of bounds in
following process:

ubifs_wbuf_write_nolock():
  aligned_len = ALIGN(len, 8);   // Assume len = 4089, aligned_len = 4096
  if (aligned_len <= wbuf->avail) ... // Not satisfy
  if (wbuf->used) {
    ubifs_leb_write()  // Fill some data in avail wbuf
    len -= wbuf->avail;   // len is still not 8-bytes aligned
    aligned_len -= wbuf->avail;
  }
  n = aligned_len >> c->max_write_shift;
  if (n) {
    n <<= c->max_write_shift;
    err = ubifs_leb_write(c, wbuf->lnum, buf + written,
                          wbuf->offs, n);
    // n > len, read out of bounds less than 8(n-len) bytes
  }

, which can be catched by KASAN:
  =========================================================
  BUG: KASAN: slab-out-of-bounds in ecc_sw_hamming_calculate+0x1dc/0x7d0
  Read of size 4 at addr ffff888105594ff8 by task kworker/u8:4/128
  Workqueue: writeback wb_workfn (flush-ubifs_0_0)
  Call Trace:
    kasan_report.cold+0x81/0x165
    nand_write_page_swecc+0xa9/0x160
    ubifs_leb_write+0xf2/0x1b0 [ubifs]
    ubifs_wbuf_write_nolock+0x421/0x12c0 [ubifs]
    write_head+0xdc/0x1c0 [ubifs]
    ubifs_jnl_write_inode+0x627/0x960 [ubifs]
    wb_workfn+0x8af/0xb80

Function ubifs_wbuf_write_nolock() accepts that parameter 'len' is not 8
bytes aligned, the 'len' represents the true length of buf (which is
allocated in 'ubifs_jnl_xxx', eg. ubifs_jnl_write_inode), so
ubifs_wbuf_write_nolock() must handle the length read from 'buf' carefully
to write leb safely.

Fetch a reproducer in [Link].

Fixes: 1e51764a3c2ac0 ("UBIFS: add new flash file system")
Link: https://bugzilla.kernel.org/show_bug.cgi?id=214785
Reported-by: Chengsong Ke <kechengsong@huawei.com>
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: setflags: Make dirtied_ino_d 8 bytes aligned
Zhihao Cheng [Mon, 27 Dec 2021 03:22:39 +0000 (11:22 +0800)]
ubifs: setflags: Make dirtied_ino_d 8 bytes aligned

Make 'ui->data_len' aligned with 8 bytes before it is assigned to
dirtied_ino_d. Since 8871d84c8f8b0c6b("ubifs: convert to fileattr")
applied, 'setflags()' only affects regular files and directories, only
xattr inode, symlink inode and special inode(pipe/char_dev/block_dev)
have none- zero 'ui->data_len' field, so assertion
'!(req->dirtied_ino_d & 7)' cannot fail in ubifs_budget_space().
To avoid assertion fails in future evolution(eg. setflags can operate
special inodes), it's better to make dirtied_ino_d 8 bytes aligned,
after all aligned size is still zero for regular files.

Fixes: 1e51764a3c2ac05a ("UBIFS: add new flash file system")
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Rectify space amount budget for mkdir/tmpfile operations
Zhihao Cheng [Mon, 27 Dec 2021 03:22:38 +0000 (11:22 +0800)]
ubifs: Rectify space amount budget for mkdir/tmpfile operations

UBIFS should make sure the flash has enough space to store dirty (Data
that is newer than disk) data (in memory), space budget is exactly
designed to do that. If space budget calculates less data than we need,
'make_reservation()' will do more work(return -ENOSPC if no free space
lelf, sometimes we can see "cannot reserve xxx bytes in jhead xxx, error
-28" in ubifs error messages) with ubifs inodes locked, which may effect
other syscalls.

A simple way to decide how much space do we need when make a budget:
See how much space is needed by 'make_reservation()' in ubifs_jnl_xxx()
function according to corresponding operation.

It's better to report ENOSPC in ubifs_budget_space(), as early as we can.

Fixes: 474b93704f32163 ("ubifs: Implement O_TMPFILE")
Fixes: 1e51764a3c2ac05 ("UBIFS: add new flash file system")
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Fix 'ui->dirty' race between do_tmpfile() and writeback work
Zhihao Cheng [Mon, 27 Dec 2021 03:22:37 +0000 (11:22 +0800)]
ubifs: Fix 'ui->dirty' race between do_tmpfile() and writeback work

'ui->dirty' is not protected by 'ui_mutex' in function do_tmpfile() which
may race with ubifs_write_inode[wb_workfn] to access/update 'ui->dirty',
finally dirty space is released twice.

open(O_TMPFILE)                wb_workfn
do_tmpfile
  ubifs_budget_space(ino_req = { .dirtied_ino = 1})
  d_tmpfile // mark inode(tmpfile) dirty
  ubifs_jnl_update // without holding tmpfile's ui_mutex
    mark_inode_clean(ui)
      if (ui->dirty)
        ubifs_release_dirty_inode_budget(ui)  // release first time
                                   ubifs_write_inode
     mutex_lock(&ui->ui_mutex)
                                     ubifs_release_dirty_inode_budget(ui)
     // release second time
     mutex_unlock(&ui->ui_mutex)
      ui->dirty = 0

Run generic/476 can reproduce following message easily
(See reproducer in [Link]):

  UBIFS error (ubi0:0 pid 2578): ubifs_assert_failed [ubifs]: UBIFS assert
  failed: c->bi.dd_growth >= 0, in fs/ubifs/budget.c:554
  UBIFS warning (ubi0:0 pid 2578): ubifs_ro_mode [ubifs]: switched to
  read-only mode, error -22
  Workqueue: writeback wb_workfn (flush-ubifs_0_0)
  Call Trace:
    ubifs_ro_mode+0x54/0x60 [ubifs]
    ubifs_assert_failed+0x4b/0x80 [ubifs]
    ubifs_release_budget+0x468/0x5a0 [ubifs]
    ubifs_release_dirty_inode_budget+0x53/0x80 [ubifs]
    ubifs_write_inode+0x121/0x1f0 [ubifs]
    ...
    wb_workfn+0x283/0x7b0

Fix it by holding tmpfile ubifs inode lock during ubifs_jnl_update().
Similar problem exists in whiteout renaming, but previous fix("ubifs:
Rename whiteout atomically") has solved the problem.

Fixes: 474b93704f32163 ("ubifs: Implement O_TMPFILE")
Link: https://bugzilla.kernel.org/show_bug.cgi?id=214765
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Rename whiteout atomically
Zhihao Cheng [Mon, 27 Dec 2021 03:22:36 +0000 (11:22 +0800)]
ubifs: Rename whiteout atomically

Currently, rename whiteout has 3 steps:
  1. create tmpfile(which associates old dentry to tmpfile inode) for
     whiteout, and store tmpfile to disk
  2. link whiteout, associate whiteout inode to old dentry agagin and
     store old dentry, old inode, new dentry on disk
  3. writeback dirty whiteout inode to disk

Suddenly power-cut or error occurring(eg. ENOSPC returned by budget,
memory allocation failure) during above steps may cause kinds of problems:
  Problem 1: ENOSPC returned by whiteout space budget (before step 2),
     old dentry will disappear after rename syscall, whiteout file
     cannot be found either.

     ls dir  // we get file, whiteout
     rename(dir/file, dir/whiteout, REANME_WHITEOUT)
     ENOSPC = ubifs_budget_space(&wht_req) // return
     ls dir  // empty (no file, no whiteout)
  Problem 2: Power-cut happens before step 3, whiteout inode with 'nlink=1'
     is not stored on disk, whiteout dentry(old dentry) is written
     on disk, whiteout file is lost on next mount (We get "dead
     directory entry" after executing 'ls -l' on whiteout file).

Now, we use following 3 steps to finish rename whiteout:
  1. create an in-mem inode with 'nlink = 1' as whiteout
  2. ubifs_jnl_rename (Write on disk to finish associating old dentry to
     whiteout inode, associating new dentry with old inode)
  3. iput(whiteout)

Rely writing in-mem inode on disk by ubifs_jnl_rename() to finish rename
whiteout, which avoids middle disk state caused by suddenly power-cut
and error occurring.

Fixes: 9e0a1fff8db56ea ("ubifs: Implement RENAME_WHITEOUT")
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Add missing iput if do_tmpfile() failed in rename whiteout
Zhihao Cheng [Mon, 27 Dec 2021 03:22:35 +0000 (11:22 +0800)]
ubifs: Add missing iput if do_tmpfile() failed in rename whiteout

whiteout inode should be put when do_tmpfile() failed if inode has been
initialized. Otherwise we will get following warning during umount:
  UBIFS error (ubi0:0 pid 1494): ubifs_assert_failed [ubifs]: UBIFS
  assert failed: c->bi.dd_growth == 0, in fs/ubifs/super.c:1930
  VFS: Busy inodes after unmount of ubifs. Self-destruct in 5 seconds.

Fixes: 9e0a1fff8db56ea ("ubifs: Implement RENAME_WHITEOUT")
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Suggested-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Fix wrong number of inodes locked by ui_mutex in ubifs_inode comment
Zhihao Cheng [Mon, 27 Dec 2021 03:22:34 +0000 (11:22 +0800)]
ubifs: Fix wrong number of inodes locked by ui_mutex in ubifs_inode comment

Since 9ec64962afb1702f75b("ubifs: Implement RENAME_EXCHANGE") and
9e0a1fff8db56eaaebb("ubifs: Implement RENAME_WHITEOUT") are applied,
ubifs_rename locks and changes 4 ubifs inodes, correct the comment
for ui_mutex in ubifs_inode.

Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Fix deadlock in concurrent rename whiteout and inode writeback
Zhihao Cheng [Mon, 27 Dec 2021 03:22:33 +0000 (11:22 +0800)]
ubifs: Fix deadlock in concurrent rename whiteout and inode writeback

Following hung tasks:
[   77.028764] task:kworker/u8:4    state:D stack:    0 pid:  132
[   77.028820] Call Trace:
[   77.029027]  schedule+0x8c/0x1b0
[   77.029067]  mutex_lock+0x50/0x60
[   77.029074]  ubifs_write_inode+0x68/0x1f0 [ubifs]
[   77.029117]  __writeback_single_inode+0x43c/0x570
[   77.029128]  writeback_sb_inodes+0x259/0x740
[   77.029148]  wb_writeback+0x107/0x4d0
[   77.029163]  wb_workfn+0x162/0x7b0

[   92.390442] task:aa              state:D stack:    0 pid: 1506
[   92.390448] Call Trace:
[   92.390458]  schedule+0x8c/0x1b0
[   92.390461]  wb_wait_for_completion+0x82/0xd0
[   92.390469]  __writeback_inodes_sb_nr+0xb2/0x110
[   92.390472]  writeback_inodes_sb_nr+0x14/0x20
[   92.390476]  ubifs_budget_space+0x705/0xdd0 [ubifs]
[   92.390503]  do_rename.cold+0x7f/0x187 [ubifs]
[   92.390549]  ubifs_rename+0x8b/0x180 [ubifs]
[   92.390571]  vfs_rename+0xdb2/0x1170
[   92.390580]  do_renameat2+0x554/0x770

, are caused by concurrent rename whiteout and inode writeback processes:
rename_whiteout(Thread 1)         wb_workfn(Thread2)
ubifs_rename
  do_rename
    lock_4_inodes (Hold ui_mutex)
    ubifs_budget_space
      make_free_space
        shrink_liability
  __writeback_inodes_sb_nr
    bdi_split_work_to_wbs (Queue new wb work)
      wb_do_writeback(wb work)
__writeback_single_inode
          ubifs_write_inode
            LOCK(ui_mutex)
   ↑
      wb_wait_for_completion (Wait wb work) <-- deadlock!

Reproducer (Detail program in [Link]):
  1. SYS_renameat2("/mp/dir/file", "/mp/dir/whiteout", RENAME_WHITEOUT)
  2. Consume out of space before kernel(mdelay) doing budget for whiteout

Fix it by doing whiteout space budget before locking ubifs inodes.
BTW, it also fixes wrong goto tag 'out_release' in whiteout budget
error handling path(It should at least recover dir i_size and unlock
4 ubifs inodes).

Fixes: 9e0a1fff8db56ea ("ubifs: Implement RENAME_WHITEOUT")
Link: https://bugzilla.kernel.org/show_bug.cgi?id=214733
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: rename_whiteout: Fix double free for whiteout_ui->data
Zhihao Cheng [Mon, 27 Dec 2021 03:22:32 +0000 (11:22 +0800)]
ubifs: rename_whiteout: Fix double free for whiteout_ui->data

'whiteout_ui->data' will be freed twice if space budget fail for
rename whiteout operation as following process:

rename_whiteout
  dev = kmalloc
  whiteout_ui->data = dev
  kfree(whiteout_ui->data)  // Free first time
  iput(whiteout)
    ubifs_free_inode
      kfree(ui->data)     // Double free!

KASAN reports:
==================================================================
BUG: KASAN: double-free or invalid-free in ubifs_free_inode+0x4f/0x70
Call Trace:
  kfree+0x117/0x490
  ubifs_free_inode+0x4f/0x70 [ubifs]
  i_callback+0x30/0x60
  rcu_do_batch+0x366/0xac0
  __do_softirq+0x133/0x57f

Allocated by task 1506:
  kmem_cache_alloc_trace+0x3c2/0x7a0
  do_rename+0x9b7/0x1150 [ubifs]
  ubifs_rename+0x106/0x1f0 [ubifs]
  do_syscall_64+0x35/0x80

Freed by task 1506:
  kfree+0x117/0x490
  do_rename.cold+0x53/0x8a [ubifs]
  ubifs_rename+0x106/0x1f0 [ubifs]
  do_syscall_64+0x35/0x80

The buggy address belongs to the object at ffff88810238bed8 which
belongs to the cache kmalloc-8 of size 8
==================================================================

Let ubifs_free_inode() free 'whiteout_ui->data'. BTW, delete unused
assignment 'whiteout_ui->data_len = 0', process 'ubifs_evict_inode()
-> ubifs_jnl_delete_inode() -> ubifs_jnl_write_inode()' doesn't need it
(because 'inc_nlink(whiteout)' won't be excuted by 'goto out_release',
 and the nlink of whiteout inode is 0).

Fixes: 9e0a1fff8db56ea ("ubifs: Implement RENAME_WHITEOUT")
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubi: Fix race condition between ctrl_cdev_ioctl and ubi_cdev_ioctl
Baokun Li [Fri, 5 Nov 2021 09:30:22 +0000 (17:30 +0800)]
ubi: Fix race condition between ctrl_cdev_ioctl and ubi_cdev_ioctl

Hulk Robot reported a KASAN report about use-after-free:
 ==================================================================
 BUG: KASAN: use-after-free in __list_del_entry_valid+0x13d/0x160
 Read of size 8 at addr ffff888035e37d98 by task ubiattach/1385
 [...]
 Call Trace:
  klist_dec_and_del+0xa7/0x4a0
  klist_put+0xc7/0x1a0
  device_del+0x4d4/0xed0
  cdev_device_del+0x1a/0x80
  ubi_attach_mtd_dev+0x2951/0x34b0 [ubi]
  ctrl_cdev_ioctl+0x286/0x2f0 [ubi]

 Allocated by task 1414:
  device_add+0x60a/0x18b0
  cdev_device_add+0x103/0x170
  ubi_create_volume+0x1118/0x1a10 [ubi]
  ubi_cdev_ioctl+0xb7f/0x1ba0 [ubi]

 Freed by task 1385:
  cdev_device_del+0x1a/0x80
  ubi_remove_volume+0x438/0x6c0 [ubi]
  ubi_cdev_ioctl+0xbf4/0x1ba0 [ubi]
 [...]
 ==================================================================

The lock held by ctrl_cdev_ioctl is ubi_devices_mutex, but the lock held
by ubi_cdev_ioctl is ubi->device_mutex. Therefore, the two locks can be
concurrent.

ctrl_cdev_ioctl contains two operations: ubi_attach and ubi_detach.
ubi_detach is bug-free because it uses reference counting to prevent
concurrency. However, uif_init and uif_close in ubi_attach may race with
ubi_cdev_ioctl.

uif_init will race with ubi_cdev_ioctl as in the following stack.
           cpu1                   cpu2                  cpu3
_______________________|________________________|______________________
ctrl_cdev_ioctl
 ubi_attach_mtd_dev
  uif_init
                           ubi_cdev_ioctl
                            ubi_create_volume
                             cdev_device_add
   ubi_add_volume
   // sysfs exist
   kill_volumes
                                                    ubi_cdev_ioctl
                                                     ubi_remove_volume
                                                      cdev_device_del
                                                       // first free
    ubi_free_volume
     cdev_del
     // double free
   cdev_device_del

And uif_close will race with ubi_cdev_ioctl as in the following stack.
           cpu1                   cpu2                  cpu3
_______________________|________________________|______________________
ctrl_cdev_ioctl
 ubi_attach_mtd_dev
  uif_init
                           ubi_cdev_ioctl
                            ubi_create_volume
                             cdev_device_add
  ubi_debugfs_init_dev
  //error goto out_uif;
  uif_close
   kill_volumes
                                                    ubi_cdev_ioctl
                                                     ubi_remove_volume
                                                      cdev_device_del
                                                       // first free
    ubi_free_volume
    // double free

The cause of this problem is that commit 714fb87e8bc0 make device
"available" before it becomes accessible via sysfs. Therefore, we
roll back the modification. We will fix the race condition between
ubi device creation and udev by removing ubi_get_device in
vol_attribute_show and dev_attribute_show.This avoids accessing
uninitialized ubi_devices[ubi_num].

ubi_get_device is used to prevent devices from being deleted during
sysfs execution. However, now kernfs ensures that devices will not
be deleted before all reference counting are released.
The key process is shown in the following stack.

device_del
  device_remove_attrs
    device_remove_groups
      sysfs_remove_groups
        sysfs_remove_group
          remove_files
            kernfs_remove_by_name
              kernfs_remove_by_name_ns
                __kernfs_remove
                  kernfs_drain

Fixes: 714fb87e8bc0 ("ubi: Fix race condition between ubi device creation and udev")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agojffs2: GC deadlock reading a page that is used in jffs2_write_begin()
Kyeong Yoo [Tue, 4 Jul 2017 04:22:38 +0000 (16:22 +1200)]
jffs2: GC deadlock reading a page that is used in jffs2_write_begin()

GC task can deadlock in read_cache_page() because it may attempt
to release a page that is actually allocated by another task in
jffs2_write_begin().
The reason is that in jffs2_write_begin() there is a small window
a cache page is allocated for use but not set Uptodate yet.

This ends up with a deadlock between two tasks:
1) A task (e.g. file copy)
   - jffs2_write_begin() locks a cache page
   - jffs2_write_end() tries to lock "alloc_sem" from
 jffs2_reserve_space() <-- STUCK
2) GC task (jffs2_gcd_mtd3)
   - jffs2_garbage_collect_pass() locks "alloc_sem"
   - try to lock the same cache page in read_cache_page() <-- STUCK

So to avoid this deadlock, hold "alloc_sem" in jffs2_write_begin()
while reading data in a cache page.

Signed-off-by: Kyeong Yoo <kyeong.yoo@alliedtelesis.co.nz>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: read-only if LEB may always be taken in ubifs_garbage_collect
Baokun Li [Mon, 15 Nov 2021 01:31:44 +0000 (09:31 +0800)]
ubifs: read-only if LEB may always be taken in ubifs_garbage_collect

If ubifs_garbage_collect_leb() returns -EAGAIN and ubifs_return_leb
returns error, a LEB will always has a "taken" flag. In this case,
set the ubifs to read-only to prevent a worse situation.

Signed-off-by: Baokun Li <libaokun1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: fix double return leb in ubifs_garbage_collect
Baokun Li [Mon, 15 Nov 2021 01:31:43 +0000 (09:31 +0800)]
ubifs: fix double return leb in ubifs_garbage_collect

If ubifs_garbage_collect_leb() returns -EAGAIN and enters the "out"
branch, ubifs_return_leb will execute twice on the same lnum. This
can cause data loss in concurrency situations.

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: fix slab-out-of-bounds in ubifs_change_lp
Baokun Li [Mon, 15 Nov 2021 01:31:42 +0000 (09:31 +0800)]
ubifs: fix slab-out-of-bounds in ubifs_change_lp

Hulk Robot reported a KASAN report about slab-out-of-bounds:
 ==================================================================
 BUG: KASAN: slab-out-of-bounds in ubifs_change_lp+0x3a9/0x1390 [ubifs]
 Read of size 8 at addr ffff888101c961f8 by task fsstress/1068
 [...]
 Call Trace:
  check_memory_region+0x1c1/0x1e0
  ubifs_change_lp+0x3a9/0x1390 [ubifs]
  ubifs_change_one_lp+0x170/0x220 [ubifs]
  ubifs_garbage_collect+0x7f9/0xda0 [ubifs]
  ubifs_budget_space+0xfe4/0x1bd0 [ubifs]
  ubifs_write_begin+0x528/0x10c0 [ubifs]

 Allocated by task 1068:
  kmemdup+0x25/0x50
  ubifs_lpt_lookup_dirty+0x372/0xb00 [ubifs]
  ubifs_update_one_lp+0x46/0x260 [ubifs]
  ubifs_tnc_end_commit+0x98b/0x1720 [ubifs]
  do_commit+0x6cb/0x1950 [ubifs]
  ubifs_run_commit+0x15a/0x2b0 [ubifs]
  ubifs_budget_space+0x1061/0x1bd0 [ubifs]
  ubifs_write_begin+0x528/0x10c0 [ubifs]
 [...]
 ==================================================================

In ubifs_garbage_collect(), if ubifs_find_dirty_leb returns an error,
lp is an uninitialized variable. But lp.num might be used in the out
branch, which is a random value. If the value is -1 or another value
that can pass the check, soob may occur in the ubifs_change_lp() in
the following procedure.

To solve this problem, we initialize lp.lnum to -1, and then initialize
it correctly in ubifs_find_dirty_leb, which is not equal to -1, and
ubifs_return_leb is executed only when lp.lnum != -1.

if find a retained or indexing LEB and continue to next loop, but break
before find another LEB, the "taken" flag of this LEB will be cleaned
in ubi_return_lebi(). This bug has also been fixed in this patch.

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: fix snprintf() length check
Dan Carpenter [Tue, 9 Nov 2021 11:50:51 +0000 (14:50 +0300)]
ubifs: fix snprintf() length check

The snprintf() function returns the number of bytes (not including the
NUL terminator) which would have been printed if there were enough
space.  So it can be greater than UBIFS_DFS_DIR_LEN.  And actually if
it equals UBIFS_DFS_DIR_LEN then that's okay so this check is too
strict.

Fixes: 9a620291fc01 ("ubifs: Export filesystem error counters")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Document sysfs nodes
Stefan Schaeckeler [Thu, 28 Oct 2021 07:44:04 +0000 (00:44 -0700)]
ubifs: Document sysfs nodes

Add documentation for the new sysfs nodes

 /sys/fs/ubifs/ubiX_Y/errors_magic
 /sys/fs/ubifs/ubiX_Y/errors_node
 /sys/fs/ubifs/ubiX_Y/errors_crc

Signed-off-by: Stefan Schaeckeler <sschaeck@cisco.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Export filesystem error counters
Stefan Schaeckeler [Sun, 10 Oct 2021 04:22:39 +0000 (21:22 -0700)]
ubifs: Export filesystem error counters

Not all ubifs filesystem errors are propagated to userspace.

Export bad magic, bad node and crc errors via sysfs. This allows userspace
to notice filesystem errors:

 /sys/fs/ubifs/ubiX_Y/errors_magic
 /sys/fs/ubifs/ubiX_Y/errors_node
 /sys/fs/ubifs/ubiX_Y/errors_crc

The counters are reset to 0 with a remount.

Signed-off-by: Stefan Schaeckeler <sschaeck@cisco.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Error path in ubifs_remount_rw() seems to wrongly free write buffers
Petr Cvachoucek [Mon, 30 Aug 2021 19:20:37 +0000 (21:20 +0200)]
ubifs: Error path in ubifs_remount_rw() seems to wrongly free write buffers

it seems freeing the write buffers in the error path of the
ubifs_remount_rw() is wrong. It leads later to a kernel oops like this:

[10016.431274] UBIFS (ubi0:0): start fixing up free space
[10090.810042] UBIFS (ubi0:0): free space fixup complete
[10090.814623] UBIFS error (ubi0:0 pid 512): ubifs_remount_fs: cannot
spawn "ubifs_bgt0_0", error -4
[10101.915108] UBIFS (ubi0:0): background thread "ubifs_bgt0_0" started,
PID 517
[10105.275498] Unable to handle kernel NULL pointer dereference at
virtual address 0000000000000030
[10105.284352] Mem abort info:
[10105.287160]   ESR = 0x96000006
[10105.290252]   EC = 0x25: DABT (current EL), IL = 32 bits
[10105.295592]   SET = 0, FnV = 0
[10105.298652]   EA = 0, S1PTW = 0
[10105.301848] Data abort info:
[10105.304723]   ISV = 0, ISS = 0x00000006
[10105.308573]   CM = 0, WnR = 0
[10105.311564] user pgtable: 4k pages, 48-bit VAs, pgdp=00000000f03d1000
[10105.318034] [0000000000000030] pgd=00000000f6cee003,
pud=00000000f4884003, pmd=0000000000000000
[10105.326783] Internal error: Oops: 96000006 [#1] PREEMPT SMP
[10105.332355] Modules linked in: ath10k_pci ath10k_core ath mac80211
libarc4 cfg80211 nvme nvme_core cryptodev(O)
[10105.342468] CPU: 3 PID: 518 Comm: touch Tainted: G           O
5.4.3 #1
[10105.349517] Hardware name: HYPEX CPU (DT)
[10105.353525] pstate: 40000005 (nZcv daif -PAN -UAO)
[10105.358324] pc : atomic64_try_cmpxchg_acquire.constprop.22+0x8/0x34
[10105.364596] lr : mutex_lock+0x1c/0x34
[10105.368253] sp : ffff000075633aa0
[10105.371563] x29: ffff000075633aa0 x28: 0000000000000001
[10105.376874] x27: ffff000076fa80c8 x26: 0000000000000004
[10105.382185] x25: 0000000000000030 x24: 0000000000000000
[10105.387495] x23: 0000000000000000 x22: 0000000000000038
[10105.392807] x21: 000000000000000c x20: ffff000076fa80c8
[10105.398119] x19: ffff000076fa8000 x18: 0000000000000000
[10105.403429] x17: 0000000000000000 x16: 0000000000000000
[10105.408741] x15: 0000000000000000 x14: fefefefefefefeff
[10105.414052] x13: 0000000000000000 x12: 0000000000000fe0
[10105.419364] x11: 0000000000000fe0 x10: ffff000076709020
[10105.424675] x9 : 0000000000000000 x8 : 00000000000000a0
[10105.429986] x7 : ffff000076fa80f4 x6 : 0000000000000030
[10105.435297] x5 : 0000000000000000 x4 : 0000000000000000
[10105.440609] x3 : 0000000000000000 x2 : ffff00006f276040
[10105.445920] x1 : ffff000075633ab8 x0 : 0000000000000030
[10105.451232] Call trace:
[10105.453676]  atomic64_try_cmpxchg_acquire.constprop.22+0x8/0x34
[10105.459600]  ubifs_garbage_collect+0xb4/0x334
[10105.463956]  ubifs_budget_space+0x398/0x458
[10105.468139]  ubifs_create+0x50/0x180
[10105.471712]  path_openat+0x6a0/0x9b0
[10105.475284]  do_filp_open+0x34/0x7c
[10105.478771]  do_sys_open+0x78/0xe4
[10105.482170]  __arm64_sys_openat+0x1c/0x24
[10105.486180]  el0_svc_handler+0x84/0xc8
[10105.489928]  el0_svc+0x8/0xc
[10105.492808] Code: 52800013 17fffffb d2800003 f9800011 (c85ffc05)
[10105.498903] ---[ end trace 46b721d93267a586 ]---

To reproduce the problem:

1. Filesystem initially mounted read-only, free space fixup flag set.

2. mount -o remount,rw <mountpoint>

3. it takes some time (free space fixup running)
    ... try to terminate running mount by CTRL-C
    ... does not respond, only after free space fixup is complete
    ... then "ubifs_remount_fs: cannot spawn "ubifs_bgt0_0", error -4"

4. mount -o remount,rw <mountpoint>
    ... now finished instantly (fixup already done).

5. Create file or just unmount the filesystem and we get the oops.

Cc: <stable@vger.kernel.org>
Fixes: b50b9f408502 ("UBIFS: do not free write-buffers when in R/O mode")
Signed-off-by: Petr Cvachoucek <cvachoucek@gmail.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Make use of the helper macro kthread_run()
Cai Huoqing [Thu, 21 Oct 2021 08:43:43 +0000 (16:43 +0800)]
ubifs: Make use of the helper macro kthread_run()

Repalce kthread_create/wake_up_process() with kthread_run()
to simplify the code.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubi: Fix a mistake in comment
Kai Song [Tue, 5 Oct 2021 06:56:55 +0000 (14:56 +0800)]
ubi: Fix a mistake in comment

Fixes: 2a734bb8d502 ("UBI: use debugfs for the extra checks knobs")
There is a mistake in docstrings, it should be ubi_debugfs_exit_dev
instead of dbg_debug_exit_dev.

Signed-off-by: Kai Song <songkai01@inspur.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoubifs: Fix spelling mistakes
Alexander Dahl [Mon, 20 Sep 2021 07:05:55 +0000 (09:05 +0200)]
ubifs: Fix spelling mistakes

Found with `codespell -i 3 -w fs/ubifs/**` and proof reading that parts.

Signed-off-by: Alexander Dahl <ada@thorsis.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2 years agoLinux 5.16-rc6
Linus Torvalds [Sun, 19 Dec 2021 22:14:33 +0000 (14:14 -0800)]
Linux 5.16-rc6

2 years agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Sun, 19 Dec 2021 20:44:03 +0000 (12:44 -0800)]
Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull kvm fixes from Paolo Bonzini:
 "Two small fixes, one of which was being worked around in selftests"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: x86: Retry page fault if MMU reload is pending and root has no sp
  KVM: selftests: vmx_pmu_msrs_test: Drop tests mangling guest visible CPUIDs
  KVM: x86: Drop guest CPUID check for host initiated writes to MSR_IA32_PERF_CAPABILITIES

2 years agoMerge tag 'block-5.16-2021-12-19' of git://git.kernel.dk/linux-block
Linus Torvalds [Sun, 19 Dec 2021 20:38:53 +0000 (12:38 -0800)]
Merge tag 'block-5.16-2021-12-19' of git://git.kernel.dk/linux-block

Pull block revert from Jens Axboe:
 "It turns out that the fix for not hammering on the delayed work timer
  too much caused a performance regression for BFQ, so let's revert the
  change for now.

  I've got some ideas on how to fix it appropriately, but they should
  wait for 5.17"

* tag 'block-5.16-2021-12-19' of git://git.kernel.dk/linux-block:
  Revert "block: reduce kblockd_mod_delayed_work_on() CPU consumption"

2 years agoMerge tag 'irq_urgent_for_v5.16_rc6' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 19 Dec 2021 20:28:46 +0000 (12:28 -0800)]
Merge tag 'irq_urgent_for_v5.16_rc6' of git://git./linux/kernel/git/tip/tip

Pull irq fixes from Borislav Petkov:

 - Clear the PCI_MSIX_FLAGS_MASKALL bit too on the error path so that it
   is restored to its reset state

 - Mask MSI-X vectors late on the init path in order to handle
   out-of-spec Marvell NVME devices which apparently look at the MSI-X
   mask even when MSI-X is disabled

* tag 'irq_urgent_for_v5.16_rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  PCI/MSI: Clear PCI_MSIX_FLAGS_MASKALL on error
  PCI/MSI: Mask MSI-X vectors only on success

2 years agoMerge tag 'timers_urgent_for_v5.16_rc6' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 19 Dec 2021 20:23:18 +0000 (12:23 -0800)]
Merge tag 'timers_urgent_for_v5.16_rc6' of git://git./linux/kernel/git/tip/tip

Pull timer fix from Borislav Petkov:

 - Make sure the CLOCK_REALTIME to CLOCK_MONOTONIC offset is never
   positive

* tag 'timers_urgent_for_v5.16_rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  timekeeping: Really make sure wall_to_monotonic isn't positive

2 years agoMerge tag 'locking_urgent_for_v5.16_rc6' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 19 Dec 2021 20:17:26 +0000 (12:17 -0800)]
Merge tag 'locking_urgent_for_v5.16_rc6' of git://git./linux/kernel/git/tip/tip

Pull locking fix from Borislav Petkov:

 - Fix the rtmutex condition checking when the optimistic spinning of a
   waiter needs to be terminated

* tag 'locking_urgent_for_v5.16_rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  locking/rtmutex: Fix incorrect condition in rtmutex_spin_on_owner()

2 years agoMerge tag 'core_urgent_for_v5.16_rc6' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 19 Dec 2021 19:46:54 +0000 (11:46 -0800)]
Merge tag 'core_urgent_for_v5.16_rc6' of git://git./linux/kernel/git/tip/tip

Pull signal handlign fix from Borislav Petkov:

 - Prevent lock contention on the new sigaltstack lock on the
   common-case path, when no changes have been made to the alternative
   signal stack.

* tag 'core_urgent_for_v5.16_rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  signal: Skip the altstack update when not needed

2 years agoMerge tag 'mips-fixes_5.16_3' of git://git.kernel.org/pub/scm/linux/kernel/git/mips...
Linus Torvalds [Sun, 19 Dec 2021 19:40:11 +0000 (11:40 -0800)]
Merge tag 'mips-fixes_5.16_3' of git://git./linux/kernel/git/mips/linux

Pull MIPS fix from Thomas Bogendoerfer:

 - only enable pci_remap_iospace() for Ralink devices

* tag 'mips-fixes_5.16_3' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:
  MIPS: Only define pci_remap_iospace() for Ralink

2 years agoMerge tag 'powerpc-5.16-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc...
Linus Torvalds [Sun, 19 Dec 2021 19:31:14 +0000 (11:31 -0800)]
Merge tag 'powerpc-5.16-4' of git://git./linux/kernel/git/powerpc/linux

Pull powerpc fixes from Michael Ellerman:
 "Fix a recently introduced oops at boot on 85xx in some configurations.

  Fix crashes when loading some livepatch modules with
  STRICT_MODULE_RWX.

  Thanks to Joe Lawrence, Russell Currey, and Xiaoming Ni"

* tag 'powerpc-5.16-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  powerpc/module_64: Fix livepatching for RO modules
  powerpc/85xx: Fix oops when CONFIG_FSL_PMC=n

2 years agoMerge tag '5.16-rc5-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sun, 19 Dec 2021 19:23:02 +0000 (11:23 -0800)]
Merge tag '5.16-rc5-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6

Pull cifs fixes from Steve French:
 "Two cifs/smb3 fixes, one fscache related, and one mount parsing
  related for stable"

* tag '5.16-rc5-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: sanitize multiple delimiters in prepath
  cifs: ignore resource_id while getting fscache super cookie

2 years agoKVM: x86: Retry page fault if MMU reload is pending and root has no sp
Sean Christopherson [Thu, 9 Dec 2021 06:05:46 +0000 (06:05 +0000)]
KVM: x86: Retry page fault if MMU reload is pending and root has no sp

Play nice with a NULL shadow page when checking for an obsolete root in
the page fault handler by flagging the page fault as stale if there's no
shadow page associated with the root and KVM_REQ_MMU_RELOAD is pending.
Invalidating memslots, which is the only case where _all_ roots need to
be reloaded, requests all vCPUs to reload their MMUs while holding
mmu_lock for lock.

The "special" roots, e.g. pae_root when KVM uses PAE paging, are not
backed by a shadow page.  Running with TDP disabled or with nested NPT
explodes spectaculary due to dereferencing a NULL shadow page pointer.

Skip the KVM_REQ_MMU_RELOAD check if there is a valid shadow page for the
root.  Zapping shadow pages in response to guest activity, e.g. when the
guest frees a PGD, can trigger KVM_REQ_MMU_RELOAD even if the current
vCPU isn't using the affected root.  I.e. KVM_REQ_MMU_RELOAD can be seen
with a completely valid root shadow page.  This is a bit of a moot point
as KVM currently unloads all roots on KVM_REQ_MMU_RELOAD, but that will
be cleaned up in the future.

Fixes: a955cad84cda ("KVM: x86/mmu: Retry page fault if root is invalidated by memslot update")
Cc: stable@vger.kernel.org
Cc: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20211209060552.2956723-2-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agoKVM: selftests: vmx_pmu_msrs_test: Drop tests mangling guest visible CPUIDs
Vitaly Kuznetsov [Thu, 16 Dec 2021 16:52:12 +0000 (17:52 +0100)]
KVM: selftests: vmx_pmu_msrs_test: Drop tests mangling guest visible CPUIDs

Host initiated writes to MSR_IA32_PERF_CAPABILITIES should not depend
on guest visible CPUIDs and (incorrect) KVM logic implementing it is
about to change. Also, KVM_SET_CPUID{,2} after KVM_RUN is now forbidden
and causes test to fail.

Reported-by: kernel test robot <oliver.sang@intel.com>
Fixes: feb627e8d6f6 ("KVM: x86: Forbid KVM_SET_CPUID{,2} after KVM_RUN")
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Message-Id: <20211216165213.338923-2-vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agoKVM: x86: Drop guest CPUID check for host initiated writes to MSR_IA32_PERF_CAPABILITIES
Vitaly Kuznetsov [Thu, 16 Dec 2021 16:52:13 +0000 (17:52 +0100)]
KVM: x86: Drop guest CPUID check for host initiated writes to MSR_IA32_PERF_CAPABILITIES

The ability to write to MSR_IA32_PERF_CAPABILITIES from the host should
not depend on guest visible CPUID entries, even if just to allow
creating/restoring guest MSRs and CPUIDs in any sequence.

Fixes: 27461da31089 ("KVM: x86/pmu: Support full width counting")
Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Message-Id: <20211216165213.338923-3-vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2 years agoRevert "block: reduce kblockd_mod_delayed_work_on() CPU consumption"
Jens Axboe [Sun, 19 Dec 2021 14:58:44 +0000 (07:58 -0700)]
Revert "block: reduce kblockd_mod_delayed_work_on() CPU consumption"

This reverts commit cb2ac2912a9ca7d3d26291c511939a41361d2d83.

Alex and the kernel test robot report that this causes a significant
performance regression with BFQ. I can reproduce that result, so let's
revert this one as we're close to -rc6 and we there's no point in trying
to rush a fix.

Link: https://lore.kernel.org/linux-block/1639853092.524jxfaem2.none@localhost/
Link: https://lore.kernel.org/lkml/20211219141852.GH14057@xsang-OptiPlex-9020/
Reported-by: Alex Xu (Hello71) <alex_y_xu@yahoo.ca>
Reported-by: kernel test robot <oliver.sang@intel.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2 years agoMerge tag 'tty-5.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Sat, 18 Dec 2021 21:23:55 +0000 (13:23 -0800)]
Merge tag 'tty-5.16-rc6' of git://git./linux/kernel/git/gregkh/tty

Pull tty/serial fixes from Greg KH:
 "Here are two small tty/serial fixes for 5.16-rc6.  They include:

   - n_hdlc fix for syzbot reported problem that you were previously
     copied on.

   - 8250_fintek driver fix that resolved a console problem by removing
     a previous change.

  Both have been in linux-next with no reported issues"

* tag 'tty-5.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  serial: 8250_fintek: Fix garbled text for console
  tty: n_hdlc: make n_hdlc_tty_wakeup() asynchronous

2 years agoMerge tag 'usb-5.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Sat, 18 Dec 2021 21:16:43 +0000 (13:16 -0800)]
Merge tag 'usb-5.16-rc6' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are a number of small USB driver fixes for reported problems.
  They include:

   - dwc2 driver fixes

   - xhci driver fixes

   - cdnsp driver fixes

   - typec driver fix

   - gadget u_ether driver fix

   - new quirk additions

   - usb gadget endpoint calculation fix

   - usb serial new device ids

   - revert of a xhci-dbg change that broke early debug booting

  All changes, except for the revert, have been in linux-next with no
  reported problems. The revert was from yesterday, and it was reported
  by the developers affected that it resolved their problem"

* tag 'usb-5.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  Revert "usb: early: convert to readl_poll_timeout_atomic()"
  usb: typec: tcpm: fix tcpm unregister port but leave a pending timer
  usb: cdnsp: Fix lack of spin_lock_irqsave/spin_lock_restore
  USB: NO_LPM quirk Lenovo USB-C to Ethernet Adapher(RTL8153-04)
  usb: xhci: Extend support for runtime power management for AMD's Yellow carp.
  usb: dwc2: fix STM ID/VBUS detection startup delay in dwc2_driver_probe
  USB: gadget: bRequestType is a bitfield, not a enum
  USB: serial: option: add Telit FN990 compositions
  USB: serial: cp210x: fix CP2105 GPIO registration
  usb: cdnsp: Fix incorrect status for control request
  usb: cdnsp: Fix issue in cdnsp_log_ep trace event
  usb: cdnsp: Fix incorrect calling of cdnsp_died function
  usb: xhci-mtk: fix list_del warning when enable list debug
  usb: gadget: u_ether: fix race in setting MAC address in setup phase

2 years agoMerge tag 'perf-tools-fixes-for-v5.16-2021-12-18' of git://git.kernel.org/pub/scm...
Linus Torvalds [Sat, 18 Dec 2021 19:53:14 +0000 (11:53 -0800)]
Merge tag 'perf-tools-fixes-for-v5.16-2021-12-18' of git://git./linux/kernel/git/acme/linux

Pull perf tools fixes from Arnaldo Carvalho de Melo:

 - Fix segfaults in 'perf inject' related to usage of unopened files

 - The return value of hashmap__new() should be checked using IS_ERR()

* tag 'perf-tools-fixes-for-v5.16-2021-12-18' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux:
  perf inject: Fix segfault due to perf_data__fd() without open
  perf inject: Fix segfault due to close without open
  perf expr: Fix missing check for return value of hashmap__new()

2 years agoperf inject: Fix segfault due to perf_data__fd() without open
Adrian Hunter [Mon, 13 Dec 2021 08:48:29 +0000 (10:48 +0200)]
perf inject: Fix segfault due to perf_data__fd() without open

The fixed commit attempts to get the output file descriptor even if the
file was never opened e.g.

  $ perf record uname
  Linux
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 0.002 MB perf.data (7 samples) ]
  $ perf inject -i perf.data --vm-time-correlation=dry-run
  Segmentation fault (core dumped)
  $ gdb --quiet perf
  Reading symbols from perf...
  (gdb) r inject -i perf.data --vm-time-correlation=dry-run
  Starting program: /home/ahunter/bin/perf inject -i perf.data --vm-time-correlation=dry-run
  [Thread debugging using libthread_db enabled]
  Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".

  Program received signal SIGSEGV, Segmentation fault.
  __GI___fileno (fp=0x0) at fileno.c:35
  35      fileno.c: No such file or directory.
  (gdb) bt
  #0  __GI___fileno (fp=0x0) at fileno.c:35
  #1  0x00005621e48dd987 in perf_data__fd (data=0x7fff4c68bd08) at util/data.h:72
  #2  perf_data__fd (data=0x7fff4c68bd08) at util/data.h:69
  #3  cmd_inject (argc=<optimized out>, argv=0x7fff4c69c1f0) at builtin-inject.c:1017
  #4  0x00005621e4936783 in run_builtin (p=0x5621e4ee6878 <commands+600>, argc=4, argv=0x7fff4c69c1f0) at perf.c:313
  #5  0x00005621e4897d5c in handle_internal_command (argv=<optimized out>, argc=<optimized out>) at perf.c:365
  #6  run_argv (argcp=<optimized out>, argv=<optimized out>) at perf.c:409
  #7  main (argc=4, argv=0x7fff4c69c1f0) at perf.c:539
  (gdb)

Fixes: 0ae03893623dd1dd ("perf tools: Pass a fd to perf_file_header__read_pipe()")
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Riccardo Mancini <rickyman7@gmail.com>
Cc: stable@vger.kernel.org
Link: http://lore.kernel.org/lkml/20211213084829.114772-3-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2 years agoperf inject: Fix segfault due to close without open
Adrian Hunter [Mon, 13 Dec 2021 08:48:28 +0000 (10:48 +0200)]
perf inject: Fix segfault due to close without open

The fixed commit attempts to close inject.output even if it was never
opened e.g.

  $ perf record uname
  Linux
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 0.002 MB perf.data (7 samples) ]
  $ perf inject -i perf.data --vm-time-correlation=dry-run
  Segmentation fault (core dumped)
  $ gdb --quiet perf
  Reading symbols from perf...
  (gdb) r inject -i perf.data --vm-time-correlation=dry-run
  Starting program: /home/ahunter/bin/perf inject -i perf.data --vm-time-correlation=dry-run
  [Thread debugging using libthread_db enabled]
  Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".

  Program received signal SIGSEGV, Segmentation fault.
  0x00007eff8afeef5b in _IO_new_fclose (fp=0x0) at iofclose.c:48
  48      iofclose.c: No such file or directory.
  (gdb) bt
  #0  0x00007eff8afeef5b in _IO_new_fclose (fp=0x0) at iofclose.c:48
  #1  0x0000557fc7b74f92 in perf_data__close (data=data@entry=0x7ffcdafa6578) at util/data.c:376
  #2  0x0000557fc7a6b807 in cmd_inject (argc=<optimized out>, argv=<optimized out>) at builtin-inject.c:1085
  #3  0x0000557fc7ac4783 in run_builtin (p=0x557fc8074878 <commands+600>, argc=4, argv=0x7ffcdafb6a60) at perf.c:313
  #4  0x0000557fc7a25d5c in handle_internal_command (argv=<optimized out>, argc=<optimized out>) at perf.c:365
  #5  run_argv (argcp=<optimized out>, argv=<optimized out>) at perf.c:409
  #6  main (argc=4, argv=0x7ffcdafb6a60) at perf.c:539
  (gdb)

Fixes: 02e6246f5364d526 ("perf inject: Close inject.output on exit")
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Riccardo Mancini <rickyman7@gmail.com>
Cc: stable@vger.kernel.org
Link: http://lore.kernel.org/lkml/20211213084829.114772-2-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2 years agoperf expr: Fix missing check for return value of hashmap__new()
Miaoqian Lin [Sun, 12 Dec 2021 06:25:02 +0000 (06:25 +0000)]
perf expr: Fix missing check for return value of hashmap__new()

The hashmap__new() function may return ERR_PTR(-ENOMEM) when malloc()
fails, add IS_ERR() checking for ctx->ids.

Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lore.kernel.org/lkml/20211212062504.25841-1-linmq006@gmail.com
[ s/kfree()/free()/ and add missing linux/err.h include ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2 years agolocking/rtmutex: Fix incorrect condition in rtmutex_spin_on_owner()
Zqiang [Fri, 17 Dec 2021 07:42:07 +0000 (15:42 +0800)]
locking/rtmutex: Fix incorrect condition in rtmutex_spin_on_owner()

Optimistic spinning needs to be terminated when the spinning waiter is not
longer the top waiter on the lock, but the condition is negated. It
terminates if the waiter is the top waiter, which is defeating the whole
purpose.

Fixes: c3123c431447 ("locking/rtmutex: Dont dereference waiter lockless")
Signed-off-by: Zqiang <qiang1.zhang@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20211217074207.77425-1-qiang1.zhang@intel.com
2 years agoMerge tag 'libata-5.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal...
Linus Torvalds [Sat, 18 Dec 2021 01:24:53 +0000 (17:24 -0800)]
Merge tag 'libata-5.16-rc6' of git://git./linux/kernel/git/dlemoal/libata

Pull libata fix from Damien Le Moal:
 "A single fix for this cycle:

   - Check that ATA16 passthrough commands that do not transfer any data
     have a DMA direction set to DMA_NONE (From George)"

* tag 'libata-5.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/libata:
  libata: if T_LENGTH is zero, dma direction should be DMA_NONE

2 years agoMerge tag 'zonefs-5.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal...
Linus Torvalds [Sat, 18 Dec 2021 01:19:51 +0000 (17:19 -0800)]
Merge tag 'zonefs-5.16-rc6' of git://git./linux/kernel/git/dlemoal/zonefs

Pull zonefs fixes from Damien Le Moal:
 "One fix and one trivial update for rc6:

   - Add MODULE_ALIAS_FS to get automatic module loading on mount
     (Naohiro)

   - Update Damien's email address in the MAINTAINERS file (me)"

* tag 'zonefs-5.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/zonefs:
  MAITAINERS: Change zonefs maintainer email address
  zonefs: add MODULE_ALIAS_FS

2 years agocifs: sanitize multiple delimiters in prepath
Thiago Rafael Becker [Fri, 17 Dec 2021 18:20:22 +0000 (15:20 -0300)]
cifs: sanitize multiple delimiters in prepath

mount.cifs can pass a device with multiple delimiters in it. This will
cause rename(2) to fail with ENOENT.

V2:
  - Make sanitize_path more readable.
  - Fix multiple delimiters between UNC and prepath.
  - Avoid a memory leak if a bad user starts putting a lot of delimiters
    in the path on purpose.

BugLink: https://bugzilla.redhat.com/show_bug.cgi?id=2031200
Fixes: 24e0a1eff9e2 ("cifs: switch to new mount api")
Cc: stable@vger.kernel.org # 5.11+
Acked-by: Ronnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: Thiago Rafael Becker <trbecker@gmail.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2 years agocifs: ignore resource_id while getting fscache super cookie
Shyam Prasad N [Wed, 8 Dec 2021 16:33:19 +0000 (16:33 +0000)]
cifs: ignore resource_id while getting fscache super cookie

We have a cyclic dependency between fscache super cookie
and root inode cookie. The super cookie relies on
tcon->resource_id, which gets populated from the root inode
number. However, fetching the root inode initializes inode
cookie as a child of super cookie, which is yet to be populated.

resource_id is only used as auxdata to check the validity of
super cookie. We can completely avoid setting resource_id to
remove the circular dependency. Since vol creation time and
vol serial numbers are used for auxdata, we should be fine.
Additionally, there will be auxiliary data check for each
inode cookie as well.

Fixes: 5bf91ef03d98 ("cifs: wait for tcon resource_id before getting fscache super")
CC: David Howells <dhowells@redhat.com>
Signed-off-by: Shyam Prasad N <sprasad@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
2 years agotimekeeping: Really make sure wall_to_monotonic isn't positive
Yu Liao [Mon, 13 Dec 2021 13:57:27 +0000 (21:57 +0800)]
timekeeping: Really make sure wall_to_monotonic isn't positive

Even after commit e1d7ba873555 ("time: Always make sure wall_to_monotonic
isn't positive") it is still possible to make wall_to_monotonic positive
by running the following code:

    int main(void)
    {
        struct timespec time;

        clock_gettime(CLOCK_MONOTONIC, &time);
        time.tv_nsec = 0;
        clock_settime(CLOCK_REALTIME, &time);
        return 0;
    }

The reason is that the second parameter of timespec64_compare(), ts_delta,
may be unnormalized because the delta is calculated with an open coded
substraction which causes the comparison of tv_sec to yield the wrong
result:

  wall_to_monotonic = { .tv_sec = -10, .tv_nsec =  900000000 }
  ts_delta      = { .tv_sec =  -9, .tv_nsec = -900000000 }

That makes timespec64_compare() claim that wall_to_monotonic < ts_delta,
but actually the result should be wall_to_monotonic > ts_delta.

After normalization, the result of timespec64_compare() is correct because
the tv_sec comparison is not longer misleading:

  wall_to_monotonic = { .tv_sec = -10, .tv_nsec =  900000000 }
  ts_delta      = { .tv_sec = -10, .tv_nsec =  100000000 }

Use timespec64_sub() to ensure that ts_delta is normalized, which fixes the
issue.

Fixes: e1d7ba873555 ("time: Always make sure wall_to_monotonic isn't positive")
Signed-off-by: Yu Liao <liaoyu15@huawei.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20211213135727.1656662-1-liaoyu15@huawei.com
2 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Fri, 17 Dec 2021 21:55:03 +0000 (13:55 -0800)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fix from James Bottomley:
 "One driver fix: the pm8001 has never actually worked on a system with
  an IOMMU and this fixes that use case"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: pm8001: Fix phys_to_virt() usage on dma_addr_t

2 years agoMerge tag 'for-5.16-rc5-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave...
Linus Torvalds [Fri, 17 Dec 2021 21:50:58 +0000 (13:50 -0800)]
Merge tag 'for-5.16-rc5-tag' of git://git./linux/kernel/git/kdave/linux

Pull btrfs fixes from David Sterba:
 "A few more fixes, almost all error handling one-liners and for stable.

   - regression fix in directory logging items

   - regression fix of extent buffer status bits handling after an error

   - fix memory leak in error handling path in tree-log

   - fix freeing invalid anon device number when handling errors during
     subvolume creation

   - fix warning when freeing leaf after subvolume creation failure

   - fix missing blkdev put in device scan error handling

   - fix invalid delayed ref after subvolume creation failure"

* tag 'for-5.16-rc5-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: fix missing blkdev_put() call in btrfs_scan_one_device()
  btrfs: fix warning when freeing leaf after subvolume creation failure
  btrfs: fix invalid delayed ref after subvolume creation failure
  btrfs: check WRITE_ERR when trying to read an extent buffer
  btrfs: fix missing last dir item offset update when logging directory
  btrfs: fix double free of anon_dev after failure to create subvolume
  btrfs: fix memory leak in __add_inode_ref()

2 years agoMerge tag 'selinux-pr-20211217' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 17 Dec 2021 20:03:14 +0000 (12:03 -0800)]
Merge tag 'selinux-pr-20211217' of git://git./linux/kernel/git/pcmoore/selinux

Pull selinux fix from Paul Moore:
 "Another small SELinux fix for v5.16 to ensure that we don't block on
  memory allocations while holding a spinlock.

  This passes all our tests without problem"

* tag 'selinux-pr-20211217' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux:
  selinux: fix sleeping function called from invalid context

2 years agoMerge tag 'riscv-for-linus-5.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 17 Dec 2021 19:55:35 +0000 (11:55 -0800)]
Merge tag 'riscv-for-linus-5.16-rc6' of git://git./linux/kernel/git/riscv/linux

Pull RISC-V fixes from Palmer Dabbelt:

 - A handful of DT updates for the SiFive HiFive Unmatched, that fix the
   regulator handling. These should stop some warning spew.

 - A pair of fixes for both the SiFive Hifive Unleashed and Unmatched,
   that correctly hook up the MMC card detect signal.

* tag 'riscv-for-linus-5.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux:
  riscv: dts: sifive unmatched: Link the tmp451 with its power supply
  riscv: dts: sifive unmatched: Fix regulator for board rev3
  riscv: dts: sifive unmatched: Expose the PMIC sub-functions
  riscv: dts: sifive unmatched: Expose the board ID eeprom
  riscv: dts: sifive unmatched: Name gpio lines
  riscv: dts: unmatched: Add gpio card detect to mmc-spi-slot
  riscv: dts: unleashed: Add gpio card detect to mmc-spi-slot

2 years agoMerge tag 'block-5.16-2021-12-17' of git://git.kernel.dk/linux-block
Linus Torvalds [Fri, 17 Dec 2021 19:46:07 +0000 (11:46 -0800)]
Merge tag 'block-5.16-2021-12-17' of git://git.kernel.dk/linux-block

Pull block fixes from Jens Axboe:

 - Fix for hammering on the delayed run queue timer (me)

 - bcache regression fix for this merge window (Lin)

 - Fix a divide-by-zero in the blk-iocost code (Tejun)

* tag 'block-5.16-2021-12-17' of git://git.kernel.dk/linux-block:
  bcache: fix NULL pointer reference in cached_dev_detach_finish
  block: reduce kblockd_mod_delayed_work_on() CPU consumption
  iocost: Fix divide-by-zero on donation from low hweight cgroup

2 years agoMerge tag 'io_uring-5.16-2021-12-17' of git://git.kernel.dk/linux-block
Linus Torvalds [Fri, 17 Dec 2021 19:31:46 +0000 (11:31 -0800)]
Merge tag 'io_uring-5.16-2021-12-17' of git://git.kernel.dk/linux-block

Pull io_uring fix from Jens Axboe:
 "Just a single fix, fixing an issue with the worker creation change
  that was merged last week"

* tag 'io_uring-5.16-2021-12-17' of git://git.kernel.dk/linux-block:
  io-wq: drop wqe lock before creating new worker

2 years agoMerge tag 'dmaengine-fix-5.16' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul...
Linus Torvalds [Fri, 17 Dec 2021 19:25:50 +0000 (11:25 -0800)]
Merge tag 'dmaengine-fix-5.16' of git://git./linux/kernel/git/vkoul/dmaengine

Pull dmaengine fixes from Vinod Koul:
 "A bunch of driver fixes, notably:

   - uninit variable fix for dw-axi-dmac driver

   - return value check dw-edma driver

   - calling wq quiesce inside spinlock and missed completion for idxd
     driver

   - mod alias fix for st_fdma driver"

* tag 'dmaengine-fix-5.16' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine:
  dmaengine: st_fdma: fix MODULE_ALIAS
  dmaengine: idxd: fix missed completion on abort path
  dmaengine: ti: k3-udma: Fix smatch warnings
  dmaengine: idxd: fix calling wq quiesce inside spinlock
  dmaengine: dw-edma: Fix return value check for dma_set_mask_and_coherent()
  dmaengine: dw-axi-dmac: Fix uninitialized variable in axi_chan_block_xfer_start()

2 years agoMerge tag 'drm-fixes-2021-12-17-1' of git://anongit.freedesktop.org/drm/drm
Linus Torvalds [Fri, 17 Dec 2021 19:07:13 +0000 (11:07 -0800)]
Merge tag 'drm-fixes-2021-12-17-1' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "Mostly amdgpu fixes this week scattered around the driver, otherwise
  one i915, one ast, one simpledrm. There is a revert in the fb-helper
  for places userspace was using a string that we tried to change.

  i915:
   - Fix a bound check in the DMC fw load.

  ast:
   - NULL ptr deref fix

  simpledrm:
   - pixel clock units fix

  fb-helper:
   - userspace regression revert

  amdgpu:
   - Fix RLC register offset
   - GMC fix
   - Properly cache SMU FW version on Yellow Carp
   - Fix missing callback on DCN3.1
   - Reset DMCUB before HW init
   - Fix for GMC powergating on PCO
   - Fix a possible memory leak in GPU metrics table handling on RN"

* tag 'drm-fixes-2021-12-17-1' of git://anongit.freedesktop.org/drm/drm:
  drm/amd/pm: fix a potential gpu_metrics_table memory leak
  drm/amdgpu: correct the wrong cached state for GMC on PICASSO
  drm/amd/display: Reset DMCUB before HW init
  drm/amd/display: Set exit_optimized_pwr_state for DCN31
  drm/amd/pm: fix reading SMU FW version from amdgpu_firmware_info on YC
  drm/amdgpu: don't override default ECO_BITs setting
  drm/amdgpu: correct register access for RLC_JUMP_TABLE_RESTORE
  drm/i915/display: Fix an unsigned subtraction which can never be negative.
  drm/ast: potential dereference of null pointer
  drm: simpledrm: fix wrong unit with pixel clock
  Revert "drm/fb-helper: improve DRM fbdev emulation device names"

2 years agoRevert "usb: early: convert to readl_poll_timeout_atomic()"
Greg Kroah-Hartman [Fri, 17 Dec 2021 15:24:30 +0000 (16:24 +0100)]
Revert "usb: early: convert to readl_poll_timeout_atomic()"

This reverts commit 796eed4b2342c9d6b26c958e92af91253a2390e1.

This change causes boot lockups when using "arlyprintk=xdbc" because
ktime can not be used at this point in time in the boot process.  Also,
it is not needed for very small delays like this.

Reported-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Reported-by: Peter Zijlstra <peterz@infradead.org>
Cc: Jann Horn <jannh@google.com>
Cc: Chunfeng Yun <chunfeng.yun@mediatek.com>
Fixes: 796eed4b2342 ("usb: early: convert to readl_poll_timeout_atomic()")
Link: https://lore.kernel.org/r/c2b5c9bb-1b75-bf56-3754-b5b18812d65e@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2 years agoMerge tag 'usb-serial-5.16-rc6' of https://git.kernel.org/pub/scm/linux/kernel/git...
Greg Kroah-Hartman [Fri, 17 Dec 2021 09:37:02 +0000 (10:37 +0100)]
Merge tag 'usb-serial-5.16-rc6' of https://git./linux/kernel/git/johan/usb-serial into usb-linus

Johan writes:

USB-serial fixes for 5.16-rc6

Here's a fix for a reported problem in the cp210x gpio-registration code
and some more modem device ids.

All have been in linux-next with no reported issues.

* tag 'usb-serial-5.16-rc6' of https://git.kernel.org/pub/scm/linux/kernel/git/johan/usb-serial:
  USB: serial: option: add Telit FN990 compositions
  USB: serial: cp210x: fix CP2105 GPIO registration

2 years agoMAITAINERS: Change zonefs maintainer email address
Damien Le Moal [Fri, 17 Dec 2021 07:41:17 +0000 (16:41 +0900)]
MAITAINERS: Change zonefs maintainer email address

Update my email address from damien.lemoal@wdc.com to
damien.lemoal@opensource.wdc.com.

Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
2 years agozonefs: add MODULE_ALIAS_FS
Naohiro Aota [Fri, 17 Dec 2021 06:15:45 +0000 (15:15 +0900)]
zonefs: add MODULE_ALIAS_FS

Add MODULE_ALIAS_FS() to load the module automatically when you do "mount
-t zonefs".

Fixes: 8dcc1a9d90c1 ("fs: New zonefs file system")
Cc: stable <stable@vger.kernel.org> # 5.6+
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Reviewed-by: Johannes Thumshirn <jth@kernel.org>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
2 years agoriscv: dts: sifive unmatched: Link the tmp451 with its power supply
Vincent Pelletier [Tue, 16 Nov 2021 23:57:42 +0000 (23:57 +0000)]
riscv: dts: sifive unmatched: Link the tmp451 with its power supply

Fixes the following probe warning:
  lm90 0-004c: Looking up vcc-supply from device tree
  lm90 0-004c: Looking up vcc-supply property in node /soc/i2c@10030000/temperature-sensor@4c failed
  lm90 0-004c: supply vcc not found, using dummy regulator

Signed-off-by: Vincent Pelletier <plr.vincent@gmail.com>
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
2 years agoriscv: dts: sifive unmatched: Fix regulator for board rev3
Vincent Pelletier [Tue, 16 Nov 2021 23:57:41 +0000 (23:57 +0000)]
riscv: dts: sifive unmatched: Fix regulator for board rev3

The existing values are rejected by the da9063 regulator driver, as they
are unachievable with the declared chip setup (non-merged vcore and bmem
are unable to provide the declared curent).

Fix voltages to match rev3 schematics, which also matches their boot-up
configuration within the chip's available precision.
Declare bcore1/bcore2 and bmem/bio as merged.
Set ldo09 and ldo10 as always-on as their consumers are not declared but
exist.
Drop ldo current limits as there is no current limit feature for these
regulators in the DA9063. Fixes warnings like:
  DA9063_LDO3: Operation of current configuration missing

Signed-off-by: Vincent Pelletier <plr.vincent@gmail.com>
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
2 years agoriscv: dts: sifive unmatched: Expose the PMIC sub-functions
Vincent Pelletier [Tue, 16 Nov 2021 23:57:39 +0000 (23:57 +0000)]
riscv: dts: sifive unmatched: Expose the PMIC sub-functions

These sub-functions are available in the chip revision on this board, so
expose them.

Signed-off-by: Vincent Pelletier <plr.vincent@gmail.com>
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
2 years agoriscv: dts: sifive unmatched: Expose the board ID eeprom
Vincent Pelletier [Tue, 16 Nov 2021 23:57:38 +0000 (23:57 +0000)]
riscv: dts: sifive unmatched: Expose the board ID eeprom

Mark it as read-only as it is factory-programmed with identifying
information, and no executable nor configuration:
- eth MAC address
- board model (PCB version, BoM version)
- board serial number
Accidental modification would cause misidentification which could brick
the board, so marking read-only seem like both a safe and non-constraining
choice.

Signed-off-by: Vincent Pelletier <plr.vincent@gmail.com>
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
2 years agoriscv: dts: sifive unmatched: Name gpio lines
Vincent Pelletier [Tue, 16 Nov 2021 23:57:37 +0000 (23:57 +0000)]
riscv: dts: sifive unmatched: Name gpio lines

Follow the pin descriptions given in the version 3 of the board schematics.

Signed-off-by: Vincent Pelletier <plr.vincent@gmail.com>
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
2 years agoMerge tag 'amd-drm-fixes-5.16-2021-12-15' of https://gitlab.freedesktop.org/agd5f...
Dave Airlie [Fri, 17 Dec 2021 05:01:00 +0000 (15:01 +1000)]
Merge tag 'amd-drm-fixes-5.16-2021-12-15' of https://gitlab.freedesktop.org/agd5f/linux into drm-fixes

amd-drm-fixes-5.16-2021-12-15:

amdgpu:
- Fix RLC register offset
- GMC fix
- Properly cache SMU FW version on Yellow Carp
- Fix missing callback on DCN3.1
- Reset DMCUB before HW init
- Fix for GMC powergating on PCO
- Fix a possible memory leak in GPU metrics table handling on RN

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20211216035239.5787-1-alexander.deucher@amd.com
2 years agoMerge tag 'drm-misc-fixes-2021-12-16-1' of ssh://git.freedesktop.org/git/drm/drm...
Dave Airlie [Fri, 17 Dec 2021 04:16:57 +0000 (14:16 +1000)]
Merge tag 'drm-misc-fixes-2021-12-16-1' of ssh://git.freedesktop.org/git/drm/drm-misc into drm-fixes

One null pointer dereference fix for ast, a pixel clock unit fix for
simpledrm and a user-space regression revert for fb-helper

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Maxime Ripard <maxime@cerno.tech>
Link: https://patchwork.freedesktop.org/patch/msgid/20211216082603.pm6yzlckmxvwnqyv@houat
2 years agolibata: if T_LENGTH is zero, dma direction should be DMA_NONE
George Kennedy [Tue, 14 Dec 2021 14:45:10 +0000 (09:45 -0500)]
libata: if T_LENGTH is zero, dma direction should be DMA_NONE

Avoid data corruption by rejecting pass-through commands where
T_LENGTH is zero (No data is transferred) and the dma direction
is not DMA_NONE.

Cc: <stable@vger.kernel.org>
Reported-by: syzkaller<syzkaller@googlegroups.com>
Signed-off-by: George Kennedy<george.kennedy@oracle.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
2 years agoMerge tag 'audit-pr-20211216' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoor...
Linus Torvalds [Thu, 16 Dec 2021 23:24:46 +0000 (15:24 -0800)]
Merge tag 'audit-pr-20211216' of git://git./linux/kernel/git/pcmoore/audit

Pull audit fix from Paul Moore:
 "A single patch to fix a problem where the audit queue could grow
  unbounded when the audit daemon is forcibly stopped"

* tag 'audit-pr-20211216' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/audit:
  audit: improve robustness of the audit queue handling

2 years agoMerge tag 'net-5.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Thu, 16 Dec 2021 23:02:14 +0000 (15:02 -0800)]
Merge tag 'net-5.16-rc6' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Networking fixes, including fixes from mac80211, wifi, bpf.

  Relatively large batches of fixes from BPF and the WiFi stack, calm in
  general networking.

  Current release - regressions:

   - dpaa2-eth: fix buffer overrun when reporting ethtool statistics

  Current release - new code bugs:

   - bpf: fix incorrect state pruning for <8B spill/fill

   - iavf:
       - add missing unlocks in iavf_watchdog_task()
       - do not override the adapter state in the watchdog task (again)

   - mlxsw: spectrum_router: consolidate MAC profiles when possible

  Previous releases - regressions:

   - mac80211 fixes:
       - rate control, avoid driver crash for retransmitted frames
       - regression in SSN handling of addba tx
       - a memory leak where sta_info is not freed
       - marking TX-during-stop for TX in in_reconfig, prevent stall

   - cfg80211: acquire wiphy mutex on regulatory work

   - wifi drivers: fix build regressions and LED config dependency

   - virtio_net: fix rx_drops stat for small pkts

   - dsa: mv88e6xxx: unforce speed & duplex in mac_link_down()

  Previous releases - always broken:

   - bpf fixes:
       - kernel address leakage in atomic fetch
       - kernel address leakage in atomic cmpxchg's r0 aux reg
       - signed bounds propagation after mov32
       - extable fixup offset
       - extable address check

   - mac80211:
       - fix the size used for building probe request
       - send ADDBA requests using the tid/queue of the aggregation
         session
       - agg-tx: don't schedule_and_wake_txq() under sta->lock, avoid
         deadlocks
       - validate extended element ID is present

   - mptcp:
       - never allow the PM to close a listener subflow (null-defer)
       - clear 'kern' flag from fallback sockets, prevent crash
       - fix deadlock in __mptcp_push_pending()

   - inet_diag: fix kernel-infoleak for UDP sockets

   - xsk: do not sleep in poll() when need_wakeup set

   - smc: avoid very long waits in smc_release()

   - sch_ets: don't remove idle classes from the round-robin list

   - netdevsim:
       - zero-initialize memory for bpf map's value, prevent info leak
       - don't let user space overwrite read only (max) ethtool parms

   - ixgbe: set X550 MDIO speed before talking to PHY

   - stmmac:
       - fix null-deref in flower deletion w/ VLAN prio Rx steering
       - dwmac-rk: fix oob read in rk_gmac_setup

   - ice: time stamping fixes

   - systemport: add global locking for descriptor life cycle"

* tag 'net-5.16-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (89 commits)
  bpf, selftests: Fix racing issue in btf_skc_cls_ingress test
  selftest/bpf: Add a test that reads various addresses.
  bpf: Fix extable address check.
  bpf: Fix extable fixup offset.
  bpf, selftests: Add test case trying to taint map value pointer
  bpf: Make 32->64 bounds propagation slightly more robust
  bpf: Fix signed bounds propagation after mov32
  sit: do not call ipip6_dev_free() from sit_init_net()
  net: systemport: Add global locking for descriptor lifecycle
  net/smc: Prevent smc_release() from long blocking
  net: Fix double 0x prefix print in SKB dump
  virtio_net: fix rx_drops stat for small pkts
  dsa: mv88e6xxx: fix debug print for SPEED_UNFORCED
  sfc_ef100: potential dereference of null pointer
  net: stmmac: dwmac-rk: fix oob read in rk_gmac_setup
  net: usb: lan78xx: add Allied Telesis AT29M2-AF
  net/packet: rx_owner_map depends on pg_vec
  netdevsim: Zero-initialize memory for new map's value in function nsim_bpf_map_alloc
  dpaa2-eth: fix ethtool statistics
  ixgbe: set X550 MDIO speed before talking to PHY
  ...

2 years agoMerge tag 'soc-fixes-5.16-3' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Linus Torvalds [Thu, 16 Dec 2021 22:48:57 +0000 (14:48 -0800)]
Merge tag 'soc-fixes-5.16-3' of git://git./linux/kernel/git/soc/soc

Pull ARM SoC fixes from Arnd Bergmann:
 "There are a number of DT fixes, mostly for mistakes found through
  static checking of the dts files again, as well as a couple of minor
  changes to address incorrect DT settings.

  For i.MX, there is yet another series of devitree changes to update
  RGMII delay settings for ethernet, which is an ongoing problem after
  some driver changes.

  For SoC specific device drivers, a number of smaller fixes came up:

   - i.MX SoC identification was incorrectly registered non-i.MX
     machines when the driver is built-in

   - One fix on imx8m-blk-ctrl driver to get i.MX8MM MIPI reset work
     properly

   - a few compile fixes for warnings that get in the way of -Werror

   - a string overflow in the scpi firmware driver

   - a boot failure with FORTIFY_SOURCE on Rockchips machines

   - broken error handling in the AMD TEE driver

   - a revert for a tegra reset driver commit that broke HDA"

* tag 'soc-fixes-5.16-3' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (25 commits)
  soc/tegra: fuse: Fix bitwise vs. logical OR warning
  firmware: arm_scpi: Fix string overflow in SCPI genpd driver
  soc: imx: Register SoC device only on i.MX boards
  soc: imx: imx8m-blk-ctrl: Fix imx8mm mipi reset
  ARM: dts: imx6ull-pinfunc: Fix CSI_DATA07__ESAI_TX0 pad name
  arm64: dts: imx8mq: remove interconnect property from lcdif
  ARM: socfpga: dts: fix qspi node compatible
  arm64: dts: apple: add #interrupt-cells property to pinctrl nodes
  dt-bindings: i2c: apple,i2c: allow multiple compatibles
  arm64: meson: remove COMMON_CLK
  arm64: meson: fix dts for JetHub D1
  tee: amdtee: fix an IS_ERR() vs NULL bug
  arm64: dts: apple: change ethernet0 device type to ethernet
  arm64: dts: ten64: remove redundant interrupt declaration for gpio-keys
  arm64: dts: rockchip: fix poweroff on helios64
  arm64: dts: rockchip: fix audio-supply for Rock Pi 4
  arm64: dts: rockchip: fix rk3399-leez-p710 vcc3v3-lan supply
  arm64: dts: rockchip: fix rk3308-roc-cc vcc-sd supply
  arm64: dts: rockchip: remove mmc-hs400-enhanced-strobe from rk3399-khadas-edge
  ARM: rockchip: Use memcpy_toio instead of memcpy on smp bring-up
  ...

2 years agoselinux: fix sleeping function called from invalid context
Scott Mayhew [Wed, 15 Dec 2021 21:28:40 +0000 (16:28 -0500)]
selinux: fix sleeping function called from invalid context

selinux_sb_mnt_opts_compat() is called via sget_fc() under the sb_lock
spinlock, so it can't use GFP_KERNEL allocations:

[  868.565200] BUG: sleeping function called from invalid context at
               include/linux/sched/mm.h:230
[  868.568246] in_atomic(): 1, irqs_disabled(): 0,
               non_block: 0, pid: 4914, name: mount.nfs
[  868.569626] preempt_count: 1, expected: 0
[  868.570215] RCU nest depth: 0, expected: 0
[  868.570809] Preemption disabled at:
[  868.570810] [<0000000000000000>] 0x0
[  868.571848] CPU: 1 PID: 4914 Comm: mount.nfs Kdump: loaded
               Tainted: G        W         5.16.0-rc5.2585cf9dfa #1
[  868.573273] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009),
               BIOS 1.14.0-4.fc34 04/01/2014
[  868.574478] Call Trace:
[  868.574844]  <TASK>
[  868.575156]  dump_stack_lvl+0x34/0x44
[  868.575692]  __might_resched.cold+0xd6/0x10f
[  868.576308]  slab_pre_alloc_hook.constprop.0+0x89/0xf0
[  868.577046]  __kmalloc_track_caller+0x72/0x420
[  868.577684]  ? security_context_to_sid_core+0x48/0x2b0
[  868.578569]  kmemdup_nul+0x22/0x50
[  868.579108]  security_context_to_sid_core+0x48/0x2b0
[  868.579854]  ? _nfs4_proc_pathconf+0xff/0x110 [nfsv4]
[  868.580742]  ? nfs_reconfigure+0x80/0x80 [nfs]
[  868.581355]  security_context_str_to_sid+0x36/0x40
[  868.581960]  selinux_sb_mnt_opts_compat+0xb5/0x1e0
[  868.582550]  ? nfs_reconfigure+0x80/0x80 [nfs]
[  868.583098]  security_sb_mnt_opts_compat+0x2a/0x40
[  868.583676]  nfs_compare_super+0x113/0x220 [nfs]
[  868.584249]  ? nfs_try_mount_request+0x210/0x210 [nfs]
[  868.584879]  sget_fc+0xb5/0x2f0
[  868.585267]  nfs_get_tree_common+0x91/0x4a0 [nfs]
[  868.585834]  vfs_get_tree+0x25/0xb0
[  868.586241]  fc_mount+0xe/0x30
[  868.586605]  do_nfs4_mount+0x130/0x380 [nfsv4]
[  868.587160]  nfs4_try_get_tree+0x47/0xb0 [nfsv4]
[  868.587724]  vfs_get_tree+0x25/0xb0
[  868.588193]  do_new_mount+0x176/0x310
[  868.588782]  __x64_sys_mount+0x103/0x140
[  868.589388]  do_syscall_64+0x3b/0x90
[  868.589935]  entry_SYSCALL_64_after_hwframe+0x44/0xae
[  868.590699] RIP: 0033:0x7f2b371c6c4e
[  868.591239] Code: 48 8b 0d dd 71 0e 00 f7 d8 64 89 01 48 83 c8 ff c3 66 2e
                     0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa 49 89 ca b8 a5 00
                     00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d aa 71
                     0e 00 f7 d8 64 89 01 48
[  868.593810] RSP: 002b:00007ffc83775d88 EFLAGS: 00000246
               ORIG_RAX: 00000000000000a5
[  868.594691] RAX: ffffffffffffffda RBX: 00007ffc83775f10 RCX: 00007f2b371c6c4e
[  868.595504] RDX: 0000555d517247a0 RSI: 0000555d51724700 RDI: 0000555d51724540
[  868.596317] RBP: 00007ffc83775f10 R08: 0000555d51726890 R09: 0000555d51726890
[  868.597162] R10: 0000000000000000 R11: 0000000000000246 R12: 0000555d51726890
[  868.598005] R13: 0000000000000003 R14: 0000555d517246e0 R15: 0000555d511ac925
[  868.598826]  </TASK>

Cc: stable@vger.kernel.org
Fixes: 69c4a42d72eb ("lsm,selinux: add new hook to compare new mount to an existing mount")
Signed-off-by: Scott Mayhew <smayhew@redhat.com>
[PM: cleanup/line-wrap the backtrace]
Signed-off-by: Paul Moore <paul@paul-moore.com>
2 years agoMerge https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
Jakub Kicinski [Thu, 16 Dec 2021 21:06:49 +0000 (13:06 -0800)]
Merge https://git./linux/kernel/git/bpf/bpf

Daniel Borkmann says:

====================
pull-request: bpf 2021-12-16

We've added 15 non-merge commits during the last 7 day(s) which contain
a total of 12 files changed, 434 insertions(+), 30 deletions(-).

The main changes are:

1) Fix incorrect verifier state pruning behavior for <8B register spill/fill,
   from Paul Chaignon.

2) Fix x86-64 JIT's extable handling for fentry/fexit when return pointer
   is an ERR_PTR(), from Alexei Starovoitov.

3) Fix 3 different possibilities that BPF verifier missed where unprivileged
   could leak kernel addresses, from Daniel Borkmann.

4) Fix xsk's poll behavior under need_wakeup flag, from Magnus Karlsson.

5) Fix an oob-write in test_verifier due to a missed MAX_NR_MAPS bump,
   from Kumar Kartikeya Dwivedi.

6) Fix a race in test_btf_skc_cls_ingress selftest, from Martin KaFai Lau.

* https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
  bpf, selftests: Fix racing issue in btf_skc_cls_ingress test
  selftest/bpf: Add a test that reads various addresses.
  bpf: Fix extable address check.
  bpf: Fix extable fixup offset.
  bpf, selftests: Add test case trying to taint map value pointer
  bpf: Make 32->64 bounds propagation slightly more robust
  bpf: Fix signed bounds propagation after mov32
  bpf, selftests: Update test case for atomic cmpxchg on r0 with pointer
  bpf: Fix kernel address leakage in atomic cmpxchg's r0 aux reg
  bpf, selftests: Add test case for atomic fetch on spilled pointer
  bpf: Fix kernel address leakage in atomic fetch
  selftests/bpf: Fix OOB write in test_verifier
  xsk: Do not sleep in poll() when need_wakeup set
  selftests/bpf: Tests for state pruning with u32 spill/fill
  bpf: Fix incorrect state pruning for <8B spill/fill
====================

Link: https://lore.kernel.org/r/20211216210005.13815-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agobpf, selftests: Fix racing issue in btf_skc_cls_ingress test
Martin KaFai Lau [Thu, 16 Dec 2021 19:16:30 +0000 (11:16 -0800)]
bpf, selftests: Fix racing issue in btf_skc_cls_ingress test

The libbpf CI reported occasional failure in btf_skc_cls_ingress:

  test_syncookie:FAIL:Unexpected syncookie states gen_cookie:80326634 recv_cookie:0
  bpf prog error at line 97

"error at line 97" means the bpf prog cannot find the listening socket
when the final ack is received.  It then skipped processing
the syncookie in the final ack which then led to "recv_cookie:0".

The problem is the userspace program did not do accept() and went
ahead to close(listen_fd) before the kernel (and the bpf prog) had
a chance to process the final ack.

The fix is to add accept() call so that the userspace will wait for
the kernel to finish processing the final ack first before close()-ing
everything.

Fixes: 9a856cae2217 ("bpf: selftest: Add test_btf_skc_cls_ingress")
Reported-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20211216191630.466151-1-kafai@fb.com
2 years agoselftest/bpf: Add a test that reads various addresses.
Alexei Starovoitov [Wed, 15 Dec 2021 20:35:34 +0000 (12:35 -0800)]
selftest/bpf: Add a test that reads various addresses.

Add a function to bpf_testmod that returns invalid kernel and user addresses.
Then attach an fexit program to that function that tries to read
memory through these addresses.

This logic checks that bpf_probe_read_kernel and BPF_PROBE_MEM logic is sane.

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2 years agobpf: Fix extable address check.
Alexei Starovoitov [Wed, 15 Dec 2021 03:25:13 +0000 (19:25 -0800)]
bpf: Fix extable address check.

The verifier checks that PTR_TO_BTF_ID pointer is either valid or NULL,
but it cannot distinguish IS_ERR pointer from valid one.

When offset is added to IS_ERR pointer it may become small positive
value which is a user address that is not handled by extable logic
and has to be checked for at the runtime.

Tighten BPF_PROBE_MEM pointer check code to prevent this case.

Fixes: 4c5de127598e ("bpf: Emit explicit NULL pointer checks for PROBE_LDX instructions.")
Reported-by: Lorenzo Fontana <lorenzo.fontana@elastic.co>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2 years agobpf: Fix extable fixup offset.
Alexei Starovoitov [Thu, 16 Dec 2021 02:38:30 +0000 (18:38 -0800)]
bpf: Fix extable fixup offset.

The prog - start_of_ldx is the offset before the faulting ldx to the location
after it, so this will be used to adjust pt_regs->ip for jumping over it and
continuing, and with old temp it would have been fixed up to the wrong offset,
causing crash.

Fixes: 4c5de127598e ("bpf: Emit explicit NULL pointer checks for PROBE_LDX instructions.")
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2 years agoMerge tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Thu, 16 Dec 2021 19:48:59 +0000 (11:48 -0800)]
Merge tag 'clk-fixes-for-linus' of git://git./linux/kernel/git/clk/linux

Pull clk fix from Stephen Boyd:
 "A single fix for the clk framework that needed some more bake time in
  linux-next.

  The problem is that two clks being registered at the same time can
  lead to a busted clk tree if the parent isn't fully registered by the
  time the child finds the parent. We rejigger the place where we mark
  the parent as fully registered so that the child can't find the parent
  until things are proper"

* tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux:
  clk: Don't parent clks until the parent is fully registered

2 years agobpf, selftests: Add test case trying to taint map value pointer
Daniel Borkmann [Wed, 15 Dec 2021 23:48:54 +0000 (23:48 +0000)]
bpf, selftests: Add test case trying to taint map value pointer

Add a test case which tries to taint map value pointer arithmetic into a
unknown scalar with subsequent export through the map.

Before fix:

  # ./test_verifier 1186
  #1186/u map access: trying to leak tained dst reg FAIL
  Unexpected success to load!
  verification time 24 usec
  stack depth 8
  processed 15 insns (limit 1000000) max_states_per_insn 0 total_states 1 peak_states 1 mark_read 1
  #1186/p map access: trying to leak tained dst reg FAIL
  Unexpected success to load!
  verification time 8 usec
  stack depth 8
  processed 15 insns (limit 1000000) max_states_per_insn 0 total_states 1 peak_states 1 mark_read 1
  Summary: 0 PASSED, 0 SKIPPED, 2 FAILED

After fix:

  # ./test_verifier 1186
  #1186/u map access: trying to leak tained dst reg OK
  #1186/p map access: trying to leak tained dst reg OK
  Summary: 2 PASSED, 0 SKIPPED, 0 FAILED

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
2 years agobpf: Make 32->64 bounds propagation slightly more robust
Daniel Borkmann [Wed, 15 Dec 2021 22:28:48 +0000 (22:28 +0000)]
bpf: Make 32->64 bounds propagation slightly more robust

Make the bounds propagation in __reg_assign_32_into_64() slightly more
robust and readable by aligning it similarly as we did back in the
__reg_combine_64_into_32() counterpart. Meaning, only propagate or
pessimize them as a smin/smax pair.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
2 years agobpf: Fix signed bounds propagation after mov32
Daniel Borkmann [Wed, 15 Dec 2021 22:02:19 +0000 (22:02 +0000)]
bpf: Fix signed bounds propagation after mov32

For the case where both s32_{min,max}_value bounds are positive, the
__reg_assign_32_into_64() directly propagates them to their 64 bit
counterparts, otherwise it pessimises them into [0,u32_max] universe and
tries to refine them later on by learning through the tnum as per comment
in mentioned function. However, that does not always happen, for example,
in mov32 operation we call zext_32_to_64(dst_reg) which invokes the
__reg_assign_32_into_64() as is without subsequent bounds update as
elsewhere thus no refinement based on tnum takes place.

Thus, not calling into the __update_reg_bounds() / __reg_deduce_bounds() /
__reg_bound_offset() triplet as we do, for example, in case of ALU ops via
adjust_scalar_min_max_vals(), will lead to more pessimistic bounds when
dumping the full register state:

Before fix:

  0: (b4) w0 = -1
  1: R0_w=invP4294967295
     (id=0,imm=ffffffff,
      smin_value=4294967295,smax_value=4294967295,
      umin_value=4294967295,umax_value=4294967295,
      var_off=(0xffffffff; 0x0),
      s32_min_value=-1,s32_max_value=-1,
      u32_min_value=-1,u32_max_value=-1)

  1: (bc) w0 = w0
  2: R0_w=invP4294967295
     (id=0,imm=ffffffff,
      smin_value=0,smax_value=4294967295,
      umin_value=4294967295,umax_value=4294967295,
      var_off=(0xffffffff; 0x0),
      s32_min_value=-1,s32_max_value=-1,
      u32_min_value=-1,u32_max_value=-1)

Technically, the smin_value=0 and smax_value=4294967295 bounds are not
incorrect, but given the register is still a constant, they break assumptions
about const scalars that smin_value == smax_value and umin_value == umax_value.

After fix:

  0: (b4) w0 = -1
  1: R0_w=invP4294967295
     (id=0,imm=ffffffff,
      smin_value=4294967295,smax_value=4294967295,
      umin_value=4294967295,umax_value=4294967295,
      var_off=(0xffffffff; 0x0),
      s32_min_value=-1,s32_max_value=-1,
      u32_min_value=-1,u32_max_value=-1)

  1: (bc) w0 = w0
  2: R0_w=invP4294967295
     (id=0,imm=ffffffff,
      smin_value=4294967295,smax_value=4294967295,
      umin_value=4294967295,umax_value=4294967295,
      var_off=(0xffffffff; 0x0),
      s32_min_value=-1,s32_max_value=-1,
      u32_min_value=-1,u32_max_value=-1)

Without the smin_value == smax_value and umin_value == umax_value invariant
being intact for const scalars, it is possible to leak out kernel pointers
from unprivileged user space if the latter is enabled. For example, when such
registers are involved in pointer arithmtics, then adjust_ptr_min_max_vals()
will taint the destination register into an unknown scalar, and the latter
can be exported and stored e.g. into a BPF map value.

Fixes: 3f50f132d840 ("bpf: Verifier, do explicit ALU32 bounds tracking")
Reported-by: Kuee K1r0a <liulin063@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
2 years agoMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Linus Torvalds [Thu, 16 Dec 2021 18:44:20 +0000 (10:44 -0800)]
Merge tag 'arm64-fixes' of git://git./linux/kernel/git/arm64/linux

Pull arm64 fix from Catalin Marinas:
 "Fix missing error code on kexec failure path"

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: kexec: Fix missing error code 'ret' warning in load_other_segments()

2 years agoMerge tag 'for-5.16/dm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/device...
Linus Torvalds [Thu, 16 Dec 2021 18:05:49 +0000 (10:05 -0800)]
Merge tag 'for-5.16/dm-fixes' of git://git./linux/kernel/git/device-mapper/linux-dm

Pull device mapper fixes from Mike Snitzer:

 - Fix use after free in DM btree remove's rebalance_children()

 - Fix DM integrity data corruption, introduced during 5.16 merge, due
   to improper use of bvec_kmap_local()

* tag 'for-5.16/dm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm integrity: fix data corruption due to improper use of bvec_kmap_local
  dm btree remove: fix use after free in rebalance_children()

2 years agoarm64: kexec: Fix missing error code 'ret' warning in load_other_segments()
Lakshmi Ramasubramanian [Fri, 10 Dec 2021 01:01:21 +0000 (17:01 -0800)]
arm64: kexec: Fix missing error code 'ret' warning in load_other_segments()

Since commit ac10be5cdbfa ("arm64: Use common
of_kexec_alloc_and_setup_fdt()"), smatch reports the following warning:

  arch/arm64/kernel/machine_kexec_file.c:152 load_other_segments()
  warn: missing error code 'ret'

Return code is not set to an error code in load_other_segments() when
of_kexec_alloc_and_setup_fdt() call returns a NULL dtb. This results
in status success (return code set to 0) being returned from
load_other_segments().

Set return code to -EINVAL if of_kexec_alloc_and_setup_fdt() returns
NULL dtb.

Signed-off-by: Lakshmi Ramasubramanian <nramas@linux.microsoft.com>
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Fixes: ac10be5cdbfa ("arm64: Use common of_kexec_alloc_and_setup_fdt()")
Link: https://lore.kernel.org/r/20211210010121.101823-1-nramas@linux.microsoft.com
Signed-off-by: Will Deacon <will@kernel.org>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2 years agoafs: Fix mmap
David Howells [Tue, 14 Dec 2021 09:22:12 +0000 (09:22 +0000)]
afs: Fix mmap

Fix afs_add_open_map() to check that the vnode isn't already on the list
when it adds it.  It's possible that afs_drop_open_mmap() decremented
the cb_nr_mmap counter, but hadn't yet got into the locked section to
remove it.

Also vnode->cb_mmap_link should be initialised, so fix that too.

Fixes: 6e0e99d58a65 ("afs: Fix mmap coherency vs 3rd-party changes")
Reported-by: kafs-testing+fedora34_64checkkafs-build-300@auristor.com
Suggested-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Tested-by: kafs-testing+fedora34_64checkkafs-build-300@auristor.com
cc: linux-afs@lists.infradead.org
Link: https://lore.kernel.org/r/686465.1639435380@warthog.procyon.org.uk/
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agosit: do not call ipip6_dev_free() from sit_init_net()
Eric Dumazet [Thu, 16 Dec 2021 11:17:41 +0000 (03:17 -0800)]
sit: do not call ipip6_dev_free() from sit_init_net()

ipip6_dev_free is sit dev->priv_destructor, already called
by register_netdevice() if something goes wrong.

Alternative would be to make ipip6_dev_free() robust against
multiple invocations, but other drivers do not implement this
strategy.

syzbot reported:

dst_release underflow
WARNING: CPU: 0 PID: 5059 at net/core/dst.c:173 dst_release+0xd8/0xe0 net/core/dst.c:173
Modules linked in:
CPU: 1 PID: 5059 Comm: syz-executor.4 Not tainted 5.16.0-rc5-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:dst_release+0xd8/0xe0 net/core/dst.c:173
Code: 4c 89 f2 89 d9 31 c0 5b 41 5e 5d e9 da d5 44 f9 e8 1d 90 5f f9 c6 05 87 48 c6 05 01 48 c7 c7 80 44 99 8b 31 c0 e8 e8 67 29 f9 <0f> 0b eb 85 0f 1f 40 00 53 48 89 fb e8 f7 8f 5f f9 48 83 c3 a8 48
RSP: 0018:ffffc9000aa5faa0 EFLAGS: 00010246
RAX: d6894a925dd15a00 RBX: 00000000ffffffff RCX: 0000000000040000
RDX: ffffc90005e19000 RSI: 000000000003ffff RDI: 0000000000040000
RBP: 0000000000000000 R08: ffffffff816a1f42 R09: ffffed1017344f2c
R10: ffffed1017344f2c R11: 0000000000000000 R12: 0000607f462b1358
R13: 1ffffffff1bfd305 R14: ffffe8ffffcb1358 R15: dffffc0000000000
FS:  00007f66c71a2700(0000) GS:ffff8880b9a00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f88aaed5058 CR3: 0000000023e0f000 CR4: 00000000003506f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 <TASK>
 dst_cache_destroy+0x107/0x1e0 net/core/dst_cache.c:160
 ipip6_dev_free net/ipv6/sit.c:1414 [inline]
 sit_init_net+0x229/0x550 net/ipv6/sit.c:1936
 ops_init+0x313/0x430 net/core/net_namespace.c:140
 setup_net+0x35b/0x9d0 net/core/net_namespace.c:326
 copy_net_ns+0x359/0x5c0 net/core/net_namespace.c:470
 create_new_namespaces+0x4ce/0xa00 kernel/nsproxy.c:110
 unshare_nsproxy_namespaces+0x11e/0x180 kernel/nsproxy.c:226
 ksys_unshare+0x57d/0xb50 kernel/fork.c:3075
 __do_sys_unshare kernel/fork.c:3146 [inline]
 __se_sys_unshare kernel/fork.c:3144 [inline]
 __x64_sys_unshare+0x34/0x40 kernel/fork.c:3144
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x44/0xd0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae
RIP: 0033:0x7f66c882ce99
Code: ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 bc ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007f66c71a2168 EFLAGS: 00000246 ORIG_RAX: 0000000000000110
RAX: ffffffffffffffda RBX: 00007f66c893ff60 RCX: 00007f66c882ce99
RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000048040200
RBP: 00007f66c8886ff1 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
R13: 00007fff6634832f R14: 00007f66c71a2300 R15: 0000000000022000
 </TASK>

Fixes: cf124db566e6 ("net: Fix inconsistent teardown and release of private netdev state.")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Link: https://lore.kernel.org/r/20211216111741.1387540-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: systemport: Add global locking for descriptor lifecycle
Florian Fainelli [Wed, 15 Dec 2021 20:24:49 +0000 (12:24 -0800)]
net: systemport: Add global locking for descriptor lifecycle

The descriptor list is a shared resource across all of the transmit queues, and
the locking mechanism used today only protects concurrency across a given
transmit queue between the transmit and reclaiming. This creates an opportunity
for the SYSTEMPORT hardware to work on corrupted descriptors if we have
multiple producers at once which is the case when using multiple transmit
queues.

This was particularly noticeable when using multiple flows/transmit queues and
it showed up in interesting ways in that UDP packets would get a correct UDP
header checksum being calculated over an incorrect packet length. Similarly TCP
packets would get an equally correct checksum computed by the hardware over an
incorrect packet length.

The SYSTEMPORT hardware maintains an internal descriptor list that it re-arranges
when the driver produces a new descriptor anytime it writes to the
WRITE_PORT_{HI,LO} registers, there is however some delay in the hardware to
re-organize its descriptors and it is possible that concurrent TX queues
eventually break this internal allocation scheme to the point where the
length/status part of the descriptor gets used for an incorrect data buffer.

The fix is to impose a global serialization for all TX queues in the short
section where we are writing to the WRITE_PORT_{HI,LO} registers which solves
the corruption even with multiple concurrent TX queues being used.

Fixes: 80105befdb4b ("net: systemport: add Broadcom SYSTEMPORT Ethernet MAC driver")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20211215202450.4086240-1-f.fainelli@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet/smc: Prevent smc_release() from long blocking
D. Wythe [Wed, 15 Dec 2021 12:29:21 +0000 (20:29 +0800)]
net/smc: Prevent smc_release() from long blocking

In nginx/wrk benchmark, there's a hung problem with high probability
on case likes that: (client will last several minutes to exit)

server: smc_run nginx

client: smc_run wrk -c 10000 -t 1 http://server

Client hangs with the following backtrace:

0 [ffffa7ce8Of3bbf8] __schedule at ffffffff9f9eOd5f
1 [ffffa7ce8Of3bc88] schedule at ffffffff9f9eløe6
2 [ffffa7ce8Of3bcaO] schedule_timeout at ffffffff9f9e3f3c
3 [ffffa7ce8Of3bd2O] wait_for_common at ffffffff9f9el9de
4 [ffffa7ce8Of3bd8O] __flush_work at ffffffff9fOfeOl3
5 [ffffa7ce8øf3bdfO] smc_release at ffffffffcO697d24 [smc]
6 [ffffa7ce8Of3be2O] __sock_release at ffffffff9f8O2e2d
7 [ffffa7ce8Of3be4ø] sock_close at ffffffff9f8ø2ebl
8 [ffffa7ce8øf3be48] __fput at ffffffff9f334f93
9 [ffffa7ce8Of3be78] task_work_run at ffffffff9flOlff5
10 [ffffa7ce8Of3beaO] do_exit at ffffffff9fOe5Ol2
11 [ffffa7ce8Of3bflO] do_group_exit at ffffffff9fOe592a
12 [ffffa7ce8Of3bf38] __x64_sys_exit_group at ffffffff9fOe5994
13 [ffffa7ce8Of3bf4O] do_syscall_64 at ffffffff9f9d4373
14 [ffffa7ce8Of3bfsO] entry_SYSCALL_64_after_hwframe at ffffffff9fa0007c

This issue dues to flush_work(), which is used to wait for
smc_connect_work() to finish in smc_release(). Once lots of
smc_connect_work() was pending or all executing work dangling,
smc_release() has to block until one worker comes to free, which
is equivalent to wait another smc_connnect_work() to finish.

In order to fix this, There are two changes:

1. For those idle smc_connect_work(), cancel it from the workqueue; for
   executing smc_connect_work(), waiting for it to finish. For that
   purpose, replace flush_work() with cancel_work_sync().

2. Since smc_connect() hold a reference for passive closing, if
   smc_connect_work() has been cancelled, release the reference.

Fixes: 24ac3a08e658 ("net/smc: rebuild nonblocking connect")
Reported-by: Tony Lu <tonylu@linux.alibaba.com>
Tested-by: Dust Li <dust.li@linux.alibaba.com>
Reviewed-by: Dust Li <dust.li@linux.alibaba.com>
Reviewed-by: Tony Lu <tonylu@linux.alibaba.com>
Signed-off-by: D. Wythe <alibuda@linux.alibaba.com>
Acked-by: Karsten Graul <kgraul@linux.ibm.com>
Link: https://lore.kernel.org/r/1639571361-101128-1-git-send-email-alibuda@linux.alibaba.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge tag 'tegra-for-5.16-soc-fixes' of git://git.kernel.org/pub/scm/linux/kernel...
Arnd Bergmann [Thu, 16 Dec 2021 14:02:25 +0000 (15:02 +0100)]
Merge tag 'tegra-for-5.16-soc-fixes' of git://git./linux/kernel/git/tegra/linux into arm/fixes

soc/tegra: Fixes for v5.16-rc6

This contains a single build fix without which ARM allmodconfig builds
are broken if -Werror is enabled.

* tag 'tegra-for-5.16-soc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux:
  soc/tegra: fuse: Fix bitwise vs. logical OR warning

Link: https://lore.kernel.org/r/20211215162618.3568474-1-thierry.reding@gmail.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
2 years agonet: Fix double 0x prefix print in SKB dump
Gal Pressman [Thu, 16 Dec 2021 09:28:25 +0000 (11:28 +0200)]
net: Fix double 0x prefix print in SKB dump

When printing netdev features %pNF already takes care of the 0x prefix,
remove the explicit one.

Fixes: 6413139dfc64 ("skbuff: increase verbosity when dumping skb data")
Signed-off-by: Gal Pressman <gal@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agovirtio_net: fix rx_drops stat for small pkts
Wenliang Wang [Thu, 16 Dec 2021 03:11:35 +0000 (11:11 +0800)]
virtio_net: fix rx_drops stat for small pkts

We found the stat of rx drops for small pkts does not increment when
build_skb fail, it's not coherent with other mode's rx drops stat.

Signed-off-by: Wenliang Wang <wangwenliang.1995@bytedance.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodsa: mv88e6xxx: fix debug print for SPEED_UNFORCED
Andrey Eremeev [Wed, 15 Dec 2021 17:30:32 +0000 (20:30 +0300)]
dsa: mv88e6xxx: fix debug print for SPEED_UNFORCED

Debug print uses invalid check to detect if speed is unforced:
(speed != SPEED_UNFORCED) should be used instead of (!speed).

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Signed-off-by: Andrey Eremeev <Axtone4all@yandex.ru>
Fixes: 96a2b40c7bd3 ("net: dsa: mv88e6xxx: add port's MAC speed setter")
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agosfc_ef100: potential dereference of null pointer
Jiasheng Jiang [Wed, 15 Dec 2021 14:37:31 +0000 (22:37 +0800)]
sfc_ef100: potential dereference of null pointer

The return value of kmalloc() needs to be checked.
To avoid use in efx_nic_update_stats() in case of the failure of alloc.

Fixes: b593b6f1b492 ("sfc_ef100: statistics gathering")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: stmmac: dwmac-rk: fix oob read in rk_gmac_setup
John Keeping [Tue, 14 Dec 2021 19:10:09 +0000 (19:10 +0000)]
net: stmmac: dwmac-rk: fix oob read in rk_gmac_setup

KASAN reports an out-of-bounds read in rk_gmac_setup on the line:

while (ops->regs[i]) {

This happens for most platforms since the regs flexible array member is
empty, so the memory after the ops structure is being read here.  It
seems that mostly this happens to contain zero anyway, so we get lucky
and everything still works.

To avoid adding redundant data to nearly all the ops structures, add a
new flag to indicate whether the regs field is valid and avoid this loop
when it is not.

Fixes: 3bb3d6b1c195 ("net: stmmac: Add RK3566/RK3568 SoC support")
Signed-off-by: John Keeping <john@metanate.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch '1GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue
David S. Miller [Thu, 16 Dec 2021 10:27:12 +0000 (10:27 +0000)]
Merge branch '1GbE' of git://git./linux/kernel/git/tnguy/net-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2021-12-15

This series contains updates to igb, igbvf, igc and ixgbe drivers.

Karen moves checks for invalid VF MAC filters to occur earlier for
igb.

Letu Ren fixes a double free issue in igbvf probe.

Sasha fixes incorrect min value being used when calculating for max for
igc.

Robert Schlabbach adds documentation on enabling NBASE-T support for
ixgbe.

Cyril Novikov adds missing initialization of MDIO bus speed for ixgbe.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: usb: lan78xx: add Allied Telesis AT29M2-AF
Greg Jesionowski [Tue, 14 Dec 2021 22:10:27 +0000 (15:10 -0700)]
net: usb: lan78xx: add Allied Telesis AT29M2-AF

This adds the vendor and product IDs for the AT29M2-AF which is a
lan7801-based device.

Signed-off-by: Greg Jesionowski <jesionowskigreg@gmail.com>
Link: https://lore.kernel.org/r/20211214221027.305784-1-jesionowskigreg@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet/packet: rx_owner_map depends on pg_vec
Willem de Bruijn [Wed, 15 Dec 2021 14:39:37 +0000 (09:39 -0500)]
net/packet: rx_owner_map depends on pg_vec

Packet sockets may switch ring versions. Avoid misinterpreting state
between versions, whose fields share a union. rx_owner_map is only
allocated with a packet ring (pg_vec) and both are swapped together.
If pg_vec is NULL, meaning no packet ring was allocated, then neither
was rx_owner_map. And the field may be old state from a tpacket_v3.

Fixes: 61fad6816fc1 ("net/packet: tpacket_rcv: avoid a producer race condition")
Reported-by: Syzbot <syzbot+1ac0994a0a0c55151121@syzkaller.appspotmail.com>
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20211215143937.106178-1-willemdebruijn.kernel@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonetdevsim: Zero-initialize memory for new map's value in function nsim_bpf_map_alloc
Haimin Zhang [Wed, 15 Dec 2021 11:15:30 +0000 (19:15 +0800)]
netdevsim: Zero-initialize memory for new map's value in function nsim_bpf_map_alloc

Zero-initialize memory for new map's value in function nsim_bpf_map_alloc
since it may cause a potential kernel information leak issue, as follows:
1. nsim_bpf_map_alloc calls nsim_map_alloc_elem to allocate elements for
a new map.
2. nsim_map_alloc_elem uses kmalloc to allocate map's value, but doesn't
zero it.
3. A user application can use IOCTL BPF_MAP_LOOKUP_ELEM to get specific
element's information in the map.
4. The kernel function map_lookup_elem will call bpf_map_copy_value to get
the information allocated at step-2, then use copy_to_user to copy to the
user buffer.
This can only leak information for an array map.

Fixes: 395cacb5f1a0 ("netdevsim: bpf: support fake map offload")
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Haimin Zhang <tcs.kernel@gmail.com>
Link: https://lore.kernel.org/r/20211215111530.72103-1-tcs.kernel@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>