linux-toradex.git/drivers/xen, branch v5.0

arm64/xen: fix xen-swiotlb cache flushing

2019-01-23T21:14:56+00:00

Xen-swiotlb hooks into the arm/arm64 arch code through a copy of the DMA
DMA mapping operations stored in the struct device arch data.

Switching arm64 to use the direct calls for the merged DMA direct /
swiotlb code broke this scheme.  Replace the indirect calls with
direct-calls in xen-swiotlb as well to fix this problem.

Fixes: 356da6d0cde3 ("dma-mapping: bypass indirect calls for dma-direct")
Reported-by: Julien Grall 
Signed-off-by: Christoph Hellwig 
Reviewed-by: Stefano Stabellini

Merge tag 'for-linus-5.0-rc3-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip

2019-01-18T17:53:41+00:00

Pull xen fixes from Juergen Gross:

 - Several fixes for the Xen pvcalls drivers (1 fix for the backend and
   8 for the frontend).

 - A fix for a rather longstanding bug in the Xen sched_clock()
   interface which led to weird time jumps when migrating the system.

 - A fix for avoiding accesses to x2apic MSRs in Xen PV guests.

* tag 'for-linus-5.0-rc3-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen: Fix x86 sched_clock() interface for xen
  pvcalls-front: fix potential null dereference
  always clear the X2APIC_ENABLE bit for PV guest
  pvcalls-front: Avoid get_free_pages(GFP_KERNEL) under spinlock
  xen/pvcalls: remove set but not used variable 'intf'
  pvcalls-back: set -ENOTCONN in pvcalls_conn_back_read
  pvcalls-front: don't return error when the ring is full
  pvcalls-front: properly allocate sk
  pvcalls-front: don't try to free unallocated rings
  pvcalls-front: read all data before closing the connection

xen: Fix x86 sched_clock() interface for xen

2019-01-16T18:06:05+00:00

Commit f94c8d11699759 ("sched/clock, x86/tsc: Rework the x86 'unstable'
sched_clock() interface") broke Xen guest time handling across
migration:

[  187.249951] Freezing user space processes ... (elapsed 0.001 seconds) done.
[  187.251137] OOM killer disabled.
[  187.251137] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.
[  187.252299] suspending xenstore...
[  187.266987] xen:grant_table: Grant tables using version 1 layout
[18446743811.706476] OOM killer enabled.
[18446743811.706478] Restarting tasks ... done.
[18446743811.720505] Setting capacity to 16777216

Fix that by setting xen_sched_clock_offset at resume time to ensure a
monotonic clock value.

[boris: replaced pr_info() with pr_info_once() in xen_callback_vector()
 to avoid printing with incorrect timestamp during resume (as we
 haven't re-adjusted the clock yet)]

Fixes: f94c8d11699759 ("sched/clock, x86/tsc: Rework the x86 'unstable' sched_clock() interface")
Cc:  # 4.11
Reported-by: Hans van Kranenburg 
Signed-off-by: Juergen Gross 
Tested-by: Hans van Kranenburg 
Signed-off-by: Boris Ostrovsky

pvcalls-front: fix potential null dereference

2019-01-15T21:13:47+00:00

 static checker warning:
    drivers/xen/pvcalls-front.c:373 alloc_active_ring()
    error: we previously assumed 'map->active.ring' could be null
           (see line 357)

drivers/xen/pvcalls-front.c
    351 static int alloc_active_ring(struct sock_mapping *map)
    352 {
    353     void *bytes;
    354
    355     map->active.ring = (struct pvcalls_data_intf *)
    356         get_zeroed_page(GFP_KERNEL);
    357     if (!map->active.ring)
                    ^^^^^^^^^^^^^^^^^
Check

    358         goto out;
    359
    360     map->active.ring->ring_order = PVCALLS_RING_ORDER;
    361     bytes = (void *)__get_free_pages(GFP_KERNEL | __GFP_ZERO,
    362                     PVCALLS_RING_ORDER);
    363     if (!bytes)
    364         goto out;
    365
    366     map->active.data.in = bytes;
    367     map->active.data.out = bytes +
    368         XEN_FLEX_RING_SIZE(PVCALLS_RING_ORDER);
    369
    370     return 0;
    371
    372 out:
--> 373     free_active_ring(map);
                                 ^^^
Add null check on map->active.ring before dereferencing it to avoid
any NULL pointer dereferences.

Fixes: 9f51c05dc41a ("pvcalls-front: Avoid get_free_pages(GFP_KERNEL) under spinlock")
Reported-by: Dan Carpenter 
Suggested-by: Boris Ostrovsky 
Signed-off-by: Wen Yang 
Reviewed-by: Boris Ostrovsky 
CC: Boris Ostrovsky 
CC: Juergen Gross 
CC: Stefano Stabellini 
CC: Dan Carpenter 
CC: xen-devel@lists.xenproject.org
CC: linux-kernel@vger.kernel.org
Signed-off-by: Boris Ostrovsky

pvcalls-front: Avoid get_free_pages(GFP_KERNEL) under spinlock

2019-01-10T14:19:36+00:00

The problem is that we call this with a spin lock held.
The call tree is:
pvcalls_front_accept() holds bedata->socket_lock.
    -> create_active()
        -> __get_free_pages() uses GFP_KERNEL

The create_active() function is only called from pvcalls_front_accept()
with a spin_lock held, The allocation is not allowed to sleep and
GFP_KERNEL is not sufficient.

This issue was detected by using the Coccinelle software.

v2: Add a function doing the allocations which is called
    outside the lock and passing the allocated data to
    create_active().

v3: Use the matching deallocators i.e., free_page()
    and free_pages(), respectively.

v4: It would be better to pre-populate map (struct sock_mapping),
    rather than introducing one more new struct.

v5: Since allocating the data outside of this call it should also
    be freed outside, when create_active() fails.
    Move kzalloc(sizeof(*map2), GFP_ATOMIC) outside spinlock and
    use GFP_KERNEL instead.

v6: Drop the superfluous calls.

Suggested-by: Juergen Gross 
Suggested-by: Boris Ostrovsky 
Suggested-by: Stefano Stabellini 
Signed-off-by: Wen Yang 
Acked-by: Stefano Stabellini 
CC: Julia Lawall 
CC: Boris Ostrovsky 
CC: Juergen Gross 
CC: Stefano Stabellini 
CC: xen-devel@lists.xenproject.org
CC: linux-kernel@vger.kernel.org
Signed-off-by: Boris Ostrovsky

xen/pvcalls: remove set but not used variable 'intf'

2019-01-04T15:55:27+00:00

Fixes gcc '-Wunused-but-set-variable' warning:

drivers/xen/pvcalls-back.c: In function 'pvcalls_sk_state_change':
drivers/xen/pvcalls-back.c:286:28: warning:
 variable 'intf' set but not used [-Wunused-but-set-variable]

It not used since e6587cdbd732 ("pvcalls-back: set -ENOTCONN in
pvcalls_conn_back_read")

Signed-off-by: YueHaibing 
Reviewed-by: Boris Ostrovsky 
Signed-off-by: Boris Ostrovsky

Remove 'type' argument from access_ok() function

2019-01-04T02:57:57+00:00

Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument
of the user address range verification function since we got rid of the
old racy i386-only code to walk page tables by hand.

It existed because the original 80386 would not honor the write protect
bit when in kernel mode, so you had to do COW by hand before doing any
user access.  But we haven't supported that in a long time, and these
days the 'type' argument is a purely historical artifact.

A discussion about extending 'user_access_begin()' to do the range
checking resulted this patch, because there is no way we're going to
move the old VERIFY_xyz interface to that model.  And it's best done at
the end of the merge window when I've done most of my merges, so let's
just get this done once and for all.

This patch was mostly done with a sed-script, with manual fix-ups for
the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form.

There were a couple of notable cases:

 - csky still had the old "verify_area()" name as an alias.

 - the iter_iov code had magical hardcoded knowledge of the actual
   values of VERIFY_{READ,WRITE} (not that they mattered, since nothing
   really used it)

 - microblaze used the type argument for a debug printout

but other than those oddities this should be a total no-op patch.

I tried to fix up all architectures, did fairly extensive grepping for
access_ok() uses, and the changes are trivial, but I may have missed
something.  Any missed conversion should be trivially fixable, though.

Signed-off-by: Linus Torvalds

pvcalls-back: set -ENOTCONN in pvcalls_conn_back_read

2019-01-02T16:00:38+00:00

When a connection is closing we receive on pvcalls_sk_state_change
notification. Instead of setting the connection as closed immediately
(-ENOTCONN), let's read one more time from it: pvcalls_conn_back_read
will set the connection as closed when necessary.

That way, we avoid races between pvcalls_sk_state_change and
pvcalls_back_ioworker.

Signed-off-by: Stefano Stabellini 
Reviewed-by: Boris Ostrovsky 
Signed-off-by: Boris Ostrovsky

pvcalls-front: don't return error when the ring is full

2019-01-02T16:00:31+00:00

When the ring is full, size == array_size. It is not an error condition,
so simply return 0 instead of an error.

Signed-off-by: Stefano Stabellini 
Reviewed-by: Boris Ostrovsky 
Signed-off-by: Boris Ostrovsky

pvcalls-front: properly allocate sk

2019-01-02T16:00:23+00:00

Don't use kzalloc: it ends up leaving sk->sk_prot not properly
initialized. Use sk_alloc instead and define our own trivial struct
proto.

Signed-off-by: Stefano Stabellini 
Reviewed-by: Boris Ostrovsky 
Signed-off-by: Boris Ostrovsky