linux.git/drivers/staging/android/ashmem.c, branch v5.10

staging: android: ashmem: Fix lockdep warning for write operation

2020-07-31T12:41:43+00:00

syzbot report [1] describes a deadlock when write operation against an
ashmem fd executed at the time when ashmem is shrinking its cache results
in the following lock sequence:

Possible unsafe locking scenario:

        CPU0                    CPU1
        ----                    ----
   lock(fs_reclaim);
                                lock(&sb->s_type->i_mutex_key#13);
                                lock(fs_reclaim);
   lock(&sb->s_type->i_mutex_key#13);

kswapd takes fs_reclaim and then inode_lock while generic_perform_write
takes inode_lock and then fs_reclaim. However ashmem does not support
writing into backing shmem with a write syscall. The only way to change
its content is to mmap it and operate on mapped memory. Therefore the race
that lockdep is warning about is not valid. Resolve this by introducing a
separate lockdep class for the backing shmem inodes.

[1]: https://lkml.kernel.org/lkml/0000000000000b5f9d059aa2037f@google.com/

Reported-by: syzbot+7a0d9d0b26efefe61780@syzkaller.appspotmail.com
Signed-off-by: Suren Baghdasaryan 
Cc: stable 
Reviewed-by: Joel Fernandes (Google) 
Link: https://lore.kernel.org/r/20200730192632.3088194-1-surenb@google.com
Signed-off-by: Greg Kroah-Hartman

mmap locking API: convert mmap_sem comments

2020-06-09T16:39:14+00:00

Convert comments that reference mmap_sem to reference mmap_lock instead.

[akpm@linux-foundation.org: fix up linux-next leftovers]
[akpm@linux-foundation.org: s/lockaphore/lock/, per Vlastimil]
[akpm@linux-foundation.org: more linux-next fixups, per Michel]

Signed-off-by: Michel Lespinasse 
Signed-off-by: Andrew Morton 
Reviewed-by: Vlastimil Babka 
Reviewed-by: Daniel Jordan 
Cc: Davidlohr Bueso 
Cc: David Rientjes 
Cc: Hugh Dickins 
Cc: Jason Gunthorpe 
Cc: Jerome Glisse 
Cc: John Hubbard 
Cc: Laurent Dufour 
Cc: Liam Howlett 
Cc: Matthew Wilcox 
Cc: Peter Zijlstra 
Cc: Ying Han 
Link: http://lkml.kernel.org/r/20200520052908.204642-13-walken@google.com
Signed-off-by: Linus Torvalds

staging: android: ashmem: Disallow ashmem memory from being remapped

2020-02-10T16:49:11+00:00

When ashmem file is mmapped, the resulting vma->vm_file points to the
backing shmem file with the generic fops that do not check ashmem
permissions like fops of ashmem do. If an mremap is done on the ashmem
region, then the permission checks will be skipped. Fix that by disallowing
mapping operation on the backing shmem file.

Reported-by: Jann Horn 
Signed-off-by: Suren Baghdasaryan 
Cc: stable  # 4.4,4.9,4.14,4.18,5.4
Signed-off-by: Todd Kjos 
Reviewed-by: Joel Fernandes (Google) 
Link: https://lore.kernel.org/r/20200127235616.48920-1-tkjos@google.com
Signed-off-by: Greg Kroah-Hartman

staging: android: ashmem: Replace strcpy with strscpy

2020-01-14T15:17:11+00:00

Replaced strcpy call with safer strscpy call with given length.
This elimates the need to manually null-terminate the given string,
since strscpy will null terminate the destination anyway.:

Signed-off-by: Ole Wiedemann 
Co-developed-by: Sebastian Scherbel 
Signed-off-by: Sebastian Scherbel 
Acked-by: Christian Brauner 
Acked-by: Joel Fernandes (Google) 
Acked-by: Todd Kjos 
Link: https://lore.kernel.org/r/20191213131032.22579-1-ole.wiedemann@fau.de
Signed-off-by: Greg Kroah-Hartman

staging: android: ashmem: Avoid range_alloc() allocation with ashmem_mutex held.

2019-02-26T10:50:17+00:00

ashmem_pin() is calling range_shrink() without checking whether
range_alloc() succeeded. Also, doing memory allocation with ashmem_mutex
held should be avoided because ashmem_shrink_scan() tries to hold it.

Therefore, move memory allocation for range_alloc() to ashmem_pin_unpin()
and make range_alloc() not to fail.

This patch is mostly meant for backporting purpose for fuzz testing on
stable/distributor kernels, for there is a plan to remove this code in
near future.

Signed-off-by: Tetsuo Handa 
Cc: stable@vger.kernel.org
Reviewed-by: Joel Fernandes 
Signed-off-by: Greg Kroah-Hartman

staging: android: ashmem: Don't call fallocate() with ashmem_mutex held.

2019-02-19T10:28:43+00:00

syzbot is hitting lockdep warnings [1][2][3]. This patch tries to fix
the warning by eliminating ashmem_shrink_scan() => {shmem|vfs}_fallocate()
sequence.

[1] https://syzkaller.appspot.com/bug?id=87c399f6fa6955006080b24142e2ce7680295ad4
[2] https://syzkaller.appspot.com/bug?id=7ebea492de7521048355fc84210220e1038a7908
[3] https://syzkaller.appspot.com/bug?id=e02419c12131c24e2a957ea050c2ab6dcbbc3270

Reported-by: syzbot 
Reported-by: syzbot 
Reported-by: syzbot 
Signed-off-by: Tetsuo Handa 
Cc: stable@vger.kernel.org
Acked-by: Joel Fernandes (Google) 
Signed-off-by: Greg Kroah-Hartman

staging: android: formatting change in ashmem.c

2019-02-08T09:31:14+00:00

Resolved "Line over 80 characters" warning from checkpatch

Signed-off-by: K Hodges 
Signed-off-by: Greg Kroah-Hartman

staging: android: ashmem: doc: Fix spelling

2018-12-19T09:40:51+00:00

Fixed spelling in comment section.

Signed-off-by: Daniel Bovensiepen 
Acked-by: Joel Fernandes (Google) 
Signed-off-by: Greg Kroah-Hartman

Merge tag 'staging-4.19-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging

2018-08-18T18:00:00+00:00

Pull staging and IIO updates from Greg KH:
 "Here are the big staging/iio patches for 4.19-rc1.

  Lots of churn here, with tons of cleanups happening in staging
  drivers, a removal of an old crypto driver that no one was using
  (skein), and the addition of some new IIO drivers. Also added was a
  "gasket" driver from Google that needs loads of work and the erofs
  filesystem.

  Even with adding all of the new drivers and a new filesystem, we are
  only adding about 1000 lines overall to the kernel linecount, which
  shows just how much cleanup happened, and how big the unused crypto
  driver was.

  All of these have been in the linux-next tree for a while now with no
  reported issues"

* tag 'staging-4.19-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging: (903 commits)
  staging:rtl8192u: Remove unused macro definitions - Style
  staging:rtl8192u: Add spaces around '+' operator - Style
  staging:rtl8192u: Remove stale comment - Style
  staging: rtl8188eu: remove unused mp_custom_oid.h
  staging: fbtft: Add spaces around / - Style
  staging: fbtft: Erases some repetitive usage of function name - Style
  staging: fbtft: Adjust some empty-line problems - Style
  staging: fbtft: Removes one nesting level to help readability - Style
  staging: fbtft: Changes gamma table to define.
  staging: fbtft: A bit more information on dev_err.
  staging: fbtft: Fixes some alignment issues - Style
  staging: fbtft: Puts macro arguments in parenthesis to avoid precedence issues - Style
  staging: rtl8188eu: remove unused array dB_Invert_Table
  staging: rtl8188eu: remove whitespace, add missing blank line
  staging: rtl8188eu: use is_multicast_ether_addr in rtw_sta_mgt.c
  staging: rtl8188eu: remove whitespace - style
  staging: rtl8188eu: cleanup block comment - style
  staging: rtl8188eu: use is_multicast_ether_addr in rtl8188eu_xmit.c
  staging: rtl8188eu: use is_multicast_ether_addr in recv_linux.c
  staging: rtlwifi: refactor rtl_get_tcb_desc
  ...

staging: ashmem: Fix SIGBUS crash when traversing mmaped ashmem pages

2018-08-01T17:08:57+00:00

Amit Pundir and Youling in parallel reported crashes with recent
mainline kernels running Android:

  F DEBUG   : *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
  F DEBUG   : Build fingerprint: 'Android/db410c32_only/db410c32_only:Q/OC-MR1/102:userdebug/test-key
  F DEBUG   : Revision: '0'
  F DEBUG   : ABI: 'arm'
  F DEBUG   : pid: 2261, tid: 2261, name: zygote  >>> zygote <<<
  F DEBUG   : signal 7 (SIGBUS), code 2 (BUS_ADRERR), fault addr 0xec00008
  ...  ...
  F DEBUG   : backtrace:
  F DEBUG   :     #00 pc 00001c04  /system/lib/libc.so (memset+48)
  F DEBUG   :     #01 pc 0010c513  /system/lib/libart.so (create_mspace_with_base+82)
  F DEBUG   :     #02 pc 0015c601  /system/lib/libart.so (art::gc::space::DlMallocSpace::CreateMspace(void*, unsigned int, unsigned int)+40)
  F DEBUG   :     #03 pc 0015c3ed  /system/lib/libart.so (art::gc::space::DlMallocSpace::CreateFromMemMap(art::MemMap*, std::__1::basic_string, std::__1::allocator> const&, unsigned int, unsigned int, unsigned int, unsigned int, bool)+36)
  ...

This was bisected back to commit bfd40eaff5ab ("mm: fix
vma_is_anonymous() false-positives").

create_mspace_with_base() in the trace above, utilizes ashmem, and with
ashmem, for shared mappings we use shmem_zero_setup(), which sets the
vma->vm_ops to &shmem_vm_ops.  But for private ashmem mappings nothing
sets the vma->vm_ops.

Looking at the problematic patch, it seems to add a requirement that one
call vma_set_anonymous() on a vma, otherwise the dummy_vm_ops will be
used.  Using the dummy_vm_ops seem to triggger SIGBUS when traversing
unmapped pages.

Thus, this patch adds a call to vma_set_anonymous() for ashmem private
mappings and seems to avoid the reported problem.

Fixes: bfd40eaff5ab ("mm: fix vma_is_anonymous() false-positives")
Cc: Kirill Shutemov 
Cc: Andrew Morton 
Cc: Dmitry Vyukov 
Cc: Oleg Nesterov 
Cc: Andrea Arcangeli 
Cc: Greg Kroah-Hartman 
Cc: Hugh Dickins 
Cc: Joel Fernandes 
Cc: Colin Cross 
Cc: Matthew Wilcox 
Reported-by: Amit Pundir 
Reported-by: Youling 257 
Signed-off-by: John Stultz 
Signed-off-by: Linus Torvalds