$ Can $ $ a $ $ TLB $ $ hit $ $ lead $ $ to $ $ a $ $ page $ $ fault? $
My answer is yes, like saying even after a TLB hit if the page in memory is dirty and this will cause a page fault. Another case may be this, it is read-only and we want to write this page which will eventually cause a page fault.
Now, in this context, let me ask you a question that was asked during the GATE 2020 exam (one of the most prestigious exams in India for $ IIT & # 39; s $ for Masters and PhD). I know this is not a question and publication site, but to dispel my doubt, I have to, and the concept can be well understood through this question. The questions go like this ->
Consider a paging system that uses $ 1 – level $ page table residing in
main memory and a TLB for address translation. Each access to the main memory
take $ 100 ns $ and TLB research takes $ 20 ns $. Each page transfer
to / from disk takes $ 5,000 ns $. Suppose the TLB success rate is
95% $, the page defect rate is 10% $. Suppose that for $ 20% $ of the total
page fault, a dirty page must be rewritten on the disc before
the required page is read from the disc. The TLB update time is negligible. the
average memory access time in ns (round up to $ 1 $ decimal places) is
There is a lot of unrest around this question as the answer provided by the examining authority claims that it goes from $ 154.5ns $ to $ $ 155.5n, but many colleagues argue that it should be $ 725ns $ as there can still be page faults if there is a hit in the TLB. All of India is looking for your help.