> blk_init_allocated_queue() allocates 1 request for flush and > 4 requests > for a memory pool. The memory pool requests only get used under memory > pressure. That is why the error doesn't come up straight away. This seems correct, I can "trivially" trigger the bug with a while-malloc loop + firefox.
> Reported-by: Seraphime Kirkovski <[email protected]> > Fixes: 304419d8a7e92 ("mmc: core: Allocate per-request data using the block > layer core") > Signed-off-by: Adrian Hunter <[email protected]> As I said, this fixes it for me, you can add Tested-By: Seraphime Kirkovski <[email protected]> Although I'm not sure this covers the same bug Pavel encountered. My kernel doesn't panic, it makes KASAN scream + #GP eventually followed by a lockup. Anyway, thanks for the fix, Seraphime
