Today, vfio mappings for a single GPU (BAR 0) may consume more than 30 umem entries (e.g. configured by qemu with vhost-net), so bump default to allow running with a few GPUs without module parameter adjustment - it is harmless, since umem rarely used and iotlb already have larger limit.
Signed-off-by: Nikolay Kuratov <[email protected]> --- drivers/vhost/vhost.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c index 2f2c45d20883..3ac2ba272716 100644 --- a/drivers/vhost/vhost.c +++ b/drivers/vhost/vhost.c @@ -34,10 +34,10 @@ #include "vhost.h" -static ushort max_mem_regions = 64; +static ushort max_mem_regions = 256; module_param(max_mem_regions, ushort, 0444); MODULE_PARM_DESC(max_mem_regions, - "Maximum number of memory regions in memory map. (default: 64)"); + "Maximum number of memory regions in memory map. (default: 256)"); static int max_iotlb_entries = 2048; module_param(max_iotlb_entries, int, 0444); MODULE_PARM_DESC(max_iotlb_entries, -- 2.34.1
