Is there any norm or general practice about the papers in the reference list that has been published by the same author (self citations or by a different author)? Is there a maximum limit as a percentage of total references?
There should be no limitation. A reference is added because of its importance for the material that you are addressing in your paper. So if an author contributed including self citation that is relevant from your paper's requirements, then it should be justifiable.
Unless you declare in your title that you are writing about one author, you must have at least two references, otherwise there can be no controversy between scholars that you reviewed!
In fact, this is not a matter of quantity/number but this should be a matter of quality of these citations, if there are several scientifically sound citations (articles or other publications) linked to the same author and there are scientific reasons to use these citations in your research paper, you accordingly can state all of them in the text and in references list without limitation in number (however, in some cases peer-reviewers or editors of journals may propose to omit some citations that seem to be redundant). Overall, it is highly suggested to use in a research paper, in addition to citations related to the same author, other potential citations related to several other authors.
I think, this is up to the author but usually there is some conditions to follow, primary references is more important than secondary references and new up-to date references is more acceptable. Thus, more related and direct references to the work by either author or other are recommended. However, too many references and citations are not welcomed by several journals and usually is mentioned in the instructions to author (in summary, new and closely related references to the work are recommended) .
There is no limit imposed on the number of references you cite in your paper, even though some journals may impose a limit on the number of pages a paper can have. However, if a particular author is relevant to you research topic, then they can be cited as many times as required. Here, I guess, only the content structure of your argumentation and the type of the paper you are writing determine the number of times an author is cited. Perhaps more citations in a qualitative type article, and less in a quantitative type article since qualitative articles often require more text and tend to be longer articles.
To complete the statement of Dr. Nebi Caka: Thomson Reuters considers self-citation beyond 20% as suspect of abuse AS THE JOURNAL’s IF IS HIGHER AND PRESTIGE ENHANCED BY SELF-CITATION (see this in the article through the link above). As far as I know this is especially true for the case of editorial or review articles written by editors or editorial-board members of journals who use self-citation to increase the IF value of these latter.
Assuming the self referenced works are relevant for the paper, what does having more than 25% self citations tell the editor or referee? It indicates either (1) you did not perform a full literature review when writing the paper, or (2) the area of research is so specialized, that it is only of interest to the self citing author. Neither possibility bodes favourably for your work. Also, keep in mind the average number of citations per paper for articles in the journal you select to submit to. If you have twenty citations, and the average for the journal is forty, perhaps you should perform further literature review to decrease you self citation ratio. It also suggests you should consider writing an article which summarizes the area of self citation, and cite only this review article in the future.