Markus reported that BTS is sporadically missing the tail of the trace
in the perf_event data buffer: [decode error (1): instruction overflow]
shown in GDB; and bisected it to the conversion of debug_store to PTI.

A little "optimization" crept into alloc_bts_buffer(), which mistakenly
placed bts_interrupt_threshold away from the 24-byte record boundary.
Intel SDM Vol 3B 17.4.9 says "This address must point to an offset from
the BTS buffer base that is a multiple of the BTS record size."

Revert "max" from a byte count to a record count, to calculate the
bts_interrupt_threshold correctly: which turns out to fix problem seen.

Fixes: c1961a4631da ("x86/events/intel/ds: Map debug buffers in cpu_entry_area")
Signed-off-by: Hugh Dickins <hu...@google.com>
Reported-and-tested-by: Markus T Metzger <markus.t.metz...@intel.com>
Cc: <sta...@vger.kernel.org> # v4.14+
---
Sorry for the spam: I missed out [PATCH] and x...@kernel.org yesterday.

 arch/x86/events/intel/ds.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

--- 4.18-rc4/arch/x86/events/intel/ds.c 2018-06-03 14:15:21.000000000 -0700
+++ linux/arch/x86/events/intel/ds.c    2018-07-12 17:38:28.471378616 -0700
@@ -408,9 +408,11 @@ static int alloc_bts_buffer(int cpu)
        ds->bts_buffer_base = (unsigned long) cea;
        ds_update_cea(cea, buffer, BTS_BUFFER_SIZE, PAGE_KERNEL);
        ds->bts_index = ds->bts_buffer_base;
-       max = BTS_RECORD_SIZE * (BTS_BUFFER_SIZE / BTS_RECORD_SIZE);
-       ds->bts_absolute_maximum = ds->bts_buffer_base + max;
-       ds->bts_interrupt_threshold = ds->bts_absolute_maximum - (max / 16);
+       max = BTS_BUFFER_SIZE / BTS_RECORD_SIZE;
+       ds->bts_absolute_maximum = ds->bts_buffer_base +
+                                       max * BTS_RECORD_SIZE;
+       ds->bts_interrupt_threshold = ds->bts_absolute_maximum -
+                                       (max / 16) * BTS_RECORD_SIZE;
        return 0;
 }
 

Reply via email to