In fact, I believe the 3560 model just lets you classify the type of traffic and maybe a bit rate whereas the 9300 uses the update MQC (Modular QoS CLI) model.
BTW, the 3560/3750 series QoS capabilities are a fairly large subset of a 9k switch, but the egress configuration is NOT, at all, like MQC. It more or less continues the MLS QoS configuration model of earlier switches.
What should be understood with switches, their QoS support is very much dependent on their hardware, and so, their QoS features vary considerably between switch platforms (especially between switch generations).
So, good news and bad news. The MLS QoS is confusing, but it "matches" the switch's capabilities. MQC is much more uniform in configuration but then you run into many restrictions of what QoS configuration statements are NOT supported on a particular switch platform. (Also, BTW, Cisco later switches, like 9k series, support more QoS features than earlier switches, and Cisco switch QoS support seems to be more uniform, across switches, then ever before. [NB: on the Catalyst 4500 series, QoS varied by supervisor. On the Catalyst 6500 series, QoS also varied by line card.]). The advantage of MQC, it works well moving to a newer generation because usually it provides backward compatibility, but (rarely) not always. (E.g. router QoS, same exact MQC syntax might work differently between pre-HQF and HQF.)
Above, I mentioned Cisco appears to be making QoS more uniform. However, I wonder at the cost of avoiding QoS innovation. (I'm a fanatic supporter of QoS. It's NOT a cure all for every network issue, but I believe it's often very much underappreciated. "Unfortunately", sellers of hardware and/or service providers benefit from selling you more bandwidth, then using bandwidth intelligently. Again, QoS isn't a cure all, but trying to solve all network issues by increasing bandwidth often isn't a cure all either.)
