Re-enable some Quantization UTs after Quantization flow updates (#108125)
**Summary**
This diff mainly has done 2 things:
- Re-enable the testcases skipped in commit: https://github.com/pytorch/pytorch/commit/9ae3d7ca90ebeaea1082f41f4c60ee3488954d78 due to the quantization flow update.
- Break down the original testcases into small testcases to make each testcase simpler.
**TestPlans**
```
python -m pytest test_mkldnn_pattern_matcher.py -k test_qconv2d
python -m pytest test_mkldnn_pattern_matcher.py -k test_qconv2d_relu
python -m pytest test_mkldnn_pattern_matcher.py -k test_qconv2d_add
python -m pytest test_mkldnn_pattern_matcher.py -k test_qconv2d_add_relu
python -m pytest test_mkldnn_pattern_matcher.py -k test_qconv2d_dequant_promotion
python -m pytest test_mkldnn_pattern_matcher.py -k test_qlinear
python -m pytest test_mkldnn_pattern_matcher.py -k test_qlinear_relu
python -m pytest test_mkldnn_pattern_matcher.py -k test_qlinear_dequant_promotion
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/108125
Approved by: https://github.com/jerryzh168, https://github.com/jgong5
