[PATCH] swiotlb: ensure io_tlb_default_mem spinlock always initialised

2022-07-10 Thread Ben Dooks
If the io_tlb_default_mem is used by a device that then gets used
by the swiotlb code, the spinlock warning is triggered causing a
lot of stack-trace.

Fix this by making the structure's lock initialised at build time.

Avoids the following BUG trigger:

[3.046401] BUG: spinlock bad magic on CPU#3, kworker/u8:0/7
[3.046689]  lock: io_tlb_default_mem+0x30/0x60, .magic: , .owner: 
/-1, .owner_cpu: 0
[3.047217] CPU: 3 PID: 7 Comm: kworker/u8:0 Not tainted 
5.17.0-00056-g1e9bac738084-dirty #310
[3.048363] Workqueue: events_unbound deferred_probe_work_func
[3.048892] Call Trace:
[3.049224] [] dump_backtrace+0x1c/0x24
[3.049576] [] show_stack+0x2c/0x38
[3.049898] [] dump_stack_lvl+0x40/0x58
[3.050216] [] dump_stack+0x14/0x1c
[3.050460] [] spin_dump+0x62/0x6e
[3.050681] [] do_raw_spin_lock+0xb0/0xd0
[3.050934] [] _raw_spin_lock_irqsave+0x20/0x2c
[3.051157] [] swiotlb_tbl_map_single+0xce/0x3da
[3.051372] [] swiotlb_map+0x3a/0x15c
[3.051668] [] dma_map_page_attrs+0x76/0x162
[3.051975] [] dw_pcie_host_init+0x326/0x3f2

Signed-off-by: Ben Dooks 
---
 kernel/dma/swiotlb.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index cb50f8d38360..a707a944c39a 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -65,7 +65,7 @@
 static bool swiotlb_force_bounce;
 static bool swiotlb_force_disable;
 
-struct io_tlb_mem io_tlb_default_mem;
+struct io_tlb_mem io_tlb_default_mem = { .lock = 
__SPIN_LOCK_UNLOCKED(io_tlb_default_mem.lock) } ;
 
 phys_addr_t swiotlb_unencrypted_base;
 
-- 
2.35.1

___
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu


Re: [PATCH] swiotlb: ensure io_tlb_default_mem spinlock always initialised

2022-07-10 Thread Robin Murphy

On 2022-07-08 18:08, Ben Dooks wrote:

If the io_tlb_default_mem is used by a device that then gets used
by the swiotlb code, the spinlock warning is triggered causing a
lot of stack-trace.


Hang on, how have we got as far as trying to allocate out of an 
uninitialised SWIOTLB at all? That seems like either something's gone 
more fundamentally wrong or we're missing a proper check somewhere. I 
don't think papering over it like this is right.


Thanks,
Robin.


Fix this by making the structure's lock initialised at build time.

Avoids the following BUG trigger:

[3.046401] BUG: spinlock bad magic on CPU#3, kworker/u8:0/7
[3.046689]  lock: io_tlb_default_mem+0x30/0x60, .magic: , .owner: 
/-1, .owner_cpu: 0
[3.047217] CPU: 3 PID: 7 Comm: kworker/u8:0 Not tainted 
5.17.0-00056-g1e9bac738084-dirty #310
[3.048363] Workqueue: events_unbound deferred_probe_work_func
[3.048892] Call Trace:
[3.049224] [] dump_backtrace+0x1c/0x24
[3.049576] [] show_stack+0x2c/0x38
[3.049898] [] dump_stack_lvl+0x40/0x58
[3.050216] [] dump_stack+0x14/0x1c
[3.050460] [] spin_dump+0x62/0x6e
[3.050681] [] do_raw_spin_lock+0xb0/0xd0
[3.050934] [] _raw_spin_lock_irqsave+0x20/0x2c
[3.051157] [] swiotlb_tbl_map_single+0xce/0x3da
[3.051372] [] swiotlb_map+0x3a/0x15c
[3.051668] [] dma_map_page_attrs+0x76/0x162
[3.051975] [] dw_pcie_host_init+0x326/0x3f2

Signed-off-by: Ben Dooks 
---
  kernel/dma/swiotlb.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
index cb50f8d38360..a707a944c39a 100644
--- a/kernel/dma/swiotlb.c
+++ b/kernel/dma/swiotlb.c
@@ -65,7 +65,7 @@
  static bool swiotlb_force_bounce;
  static bool swiotlb_force_disable;
  
-struct io_tlb_mem io_tlb_default_mem;

+struct io_tlb_mem io_tlb_default_mem = { .lock = 
__SPIN_LOCK_UNLOCKED(io_tlb_default_mem.lock) } ;
  
  phys_addr_t swiotlb_unencrypted_base;
  

___
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu