2020년 4월 3일 (금) 오전 4:42, Roman Gushchin <guro@xxxxxx>님이 작성:
> > In fact, I've tested this patch and your fixes for migration problem
> > and found that there is
> > still migration problem and failure rate is increased by this patch.
>
> Do you mind sharing any details? What kind of pages are those?
I don't investigate more since I had not enough time to do. If I
remember correctly,
it's the page used by journaling. I attach my test script below to
help you reproduce it.
My test setup is:
- virtual machine, 8 cpus and 1024 MB mem (256 MB cma mem)
- ubuntu 16.04 with custom kernel
- filesystem is ext4
> I'm using the following patch to dump failed pages:
>
> @@ -1455,6 +1455,9 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
> private, page, pass > 2, mode,
> reason);
>
> + if (rc && reason == MR_CONTIG_RANGE)
> + dump_page(page, "unmap_and_move");
> +
> switch(rc) {
> case -ENOMEM:
> /*
>
>
> > However, given that
> > there is no progress on this area for a long time, I think that
> > applying the change aggressively
> > is required to break the current situation.
>
> I totally agree!
>
> Btw, I've found that cma_release() grabs the cma->lock mutex,
> so it can't be called from the atomic context (I've got a lockdep warning).
>
> Of course, I can change the calling side, but I think it's better to change
> the cma code to make cma_release() more accepting. What do you think
> about the following patch?
For 2GB CMA area, we need to check 8192(?) bytes in worst case scenario and
I don't think it's small enough for spinlock. Even, there is no limit
on the size of
the cma area. If cma area is bigger, it takes more. So, I think that
spinlock() isn't
good here.
Anyway, below is the test script that I used.
Thanks.
-------------------------->8------------------------------------
RUNS=1
MAKE_CPUS=10
KERNEL_DIR=~~~~~~~~~~~~~~~
WORKING_DIR=`pwd`
RESULT_OUTPUT=$WORKING_DIR/log-cma-alloc.txt
BUILD_KERNEL=1
BUILD_KERNEL_PID=0
SHOW_CONSOLE=1
SHOW_LATENCY=1
CMA_AREA_NAME=cma_reserve
CMA_DEBUGFS_ROOT_DIR=/sys/kernel/debug/cma
CMA_DEBUGFS_AREA_DIR=$CMA_DEBUGFS_ROOT_DIR/cma-$CMA_AREA_NAME
CMA_AREA_COUNT=`sudo cat $CMA_DEBUGFS_AREA_DIR/count`
CMA_AREA_ORDERPERBIT=`sudo cat $CMA_DEBUGFS_AREA_DIR/order_per_bit`
CMA_AREA_PAGES=$(($CMA_AREA_COUNT * 2 ** $CMA_AREA_ORDERPERBIT))
CMA_ALLOC_DELAY=5
CMA_ALLOC_SPLIT=32
CMA_ALLOC_PAGES=$(($CMA_AREA_PAGES / $CMA_ALLOC_SPLIT))
function show_cma_info()
{
cat /proc/meminfo | grep -i cma
sudo cat $CMA_DEBUGFS_AREA_DIR/{count,used}
}
function time_begin()
{
echo $(date +%s.%N)
}
function time_elapsed()
{
tmp=$(date +%s.%N)
echo $tmp - $1 | bc -l
}
function time_sum()
{
echo $1 + $2 | bc -l
}
function time_avg()
{
echo $1 / $2 | bc -l
}
if [ "$1" == "show" ]; then
show_cma_info
exit 0
fi
if [ "$SHOW_CONSOLE" != "1" ]; then
exec 3>&1 4>&2 >$RESULT_OUTPUT 2>&1
fi
if [ "$BUILD_KERNEL" == "1" ]; then
pushd -
cd $KERNEL_DIR
make clean &> /dev/null;
make -j$MAKE_CPUS &> /dev/null &
BUILD_KERNEL_PID=$!
popd
echo "waiting until build kernel runs actively"
sleep 10
fi
echo "BUILD_KERNEL: $BUILD_KERNEL"
echo "BUILD_KERNEL_PID: $BUILD_KERNEL_PID"
echo "CMA_AREA_NAME: $CMA_AREA_NAME"
echo "CMA_AREA_PAGES: $CMA_AREA_PAGES"
echo "CMA_ALLOC_SPLIT: $CMA_ALLOC_SPLIT"
echo "CMA_ALLOC_PAGES: $CMA_ALLOC_PAGES"
for i in `seq $RUNS`;
do
echo "begin: $i"
show_cma_info
CMA_ALLOC_SUCC=0
T_BEGIN=`time_begin`
for j in `seq $CMA_ALLOC_SPLIT`;
do
sudo bash -c "echo $CMA_ALLOC_PAGES > $CMA_DEBUGFS_AREA_DIR/alloc" &> /dev/null
if [ "$?" == "0" ]; then
CMA_ALLOC_SUCC=$(($CMA_ALLOC_SUCC+1))
fi
done
T_ELAPSED=`time_elapsed $T_BEGIN`
sleep 5
echo "alloced: $CMA_ALLOC_SUCC"
show_cma_info
for j in `seq $CMA_ALLOC_SUCC`;
do
sudo bash -c "echo $CMA_ALLOC_PAGES > $CMA_DEBUGFS_AREA_DIR/free"
done
if [ "$SHOW_LATENCY" == "1" ]; then
T_AVG=`time_avg $T_ELAPSED $CMA_ALLOC_SPLIT`
echo "T_AVG: $T_AVG"
fi
sleep $CMA_ALLOC_DELAY
done
if [ "$BUILD_KERNEL_PID" != "0" ]; then
kill $BUILD_KERNEL_PID
fi
if [ "$SHOW_CONSOLE" != "1" ]; then
exec 1>&3 2>&4
fi
