Hello. Atsushi Nemoto wrote:
This fixes data corruption on some heavy load. Signed-off-by: Atsushi Nemoto <anemo@xxxxxxxxxxxxx>
Acked-by: Sergei Shtylyov <sshtylyov@xxxxxxxxxxxxx>
diff --git a/drivers/ide/tx4939ide.c b/drivers/ide/tx4939ide.c index bafb7d1..30d0d25 100644 --- a/drivers/ide/tx4939ide.c +++ b/drivers/ide/tx4939ide.c @@ -259,6 +259,12 @@ static int tx4939ide_build_dmatable(ide_drive_t *drive, struct request *rq) bcount = 0x10000 - (cur_addr & 0xffff); if (bcount > cur_len) bcount = cur_len; + /* + * This workaround for zero count seems required. + * (standard ide_build_dmatable do it too)
s/do/does/
+ */ + if ((bcount & 0xffff) == 0x0000)
Why not just bcount == 0x10000? MBR, Sergei
