Bonferroni Correction


im doing a 2 way within subjects ANOVA and want to do post hoc tests, but SPSS obviously doesnt do post hoc tests for within subjects designs. So I have four conditions and 6 possible paired comparisons. I should therefore do 6 paired sample t tests to see if these are significant. is this correct? im aware however that i should do a Bonferroni correction on each but i have read 2 different ways to do this. I either divide .05 by the number of possible comparisons (i.e. 6) and then the p values that are below this value or seen as significant.
Someone else has told me tho that instead i multiply each p value from the tests by 6 then if they're still below .05 they are significant.

can anyone tell me which one is right?!

thanks, emily