Clinton isn't considered honest by a majority of any racial group, and only even earns a plurality among black voters. Among all other demographics, more voters distrust her than trust her. Even among black voters, around 31% say that she's dishonest. Overall, by pretending that only white voters distrust Clinton, Walsh is erasing the more than 40% of non-white respondents who don't trust her, either.
More to the point, however, political assessments of trust are hardly disinterested: they are significantly correlated with partisan affiliation, which is why Democrats will always tend to trust Democrats and Republicans will always tend to trust Republicans. To control for that, the better comparison would be between different racial groupings among self-identified Democrats:
Here, the trend is quite different. Clinton still enjoys disproportionate trust among black Democrats (though even a quarter of those respondents still distrust her), but she also enjoys plurality support among white Democrats as well. Her distrust numbers are higher among white voters than among black voters, but only by 9.7% - compare that with the 37% distrust gap between white and black voters when we don't control for partisanship. Meanwhile, the only group of Democrats who do tend to distrust Clinton are not whites, but Hispanics, by a slim 3% margin. Among other races, Clinton's distrust numbers are directly comparable to whites (at ~36%), while her trust numbers are appreciably lower (37% vs. 43.1%).
So when we control for partisanship, what we see is that distrust for Clinton is hardly driven by whiteness. The two most remarkable trends are, first, her outsized trust among black Democrats, and second, her plurality distrust among Hispanic Democrats. More white Democrats trust her than distrust her; that latter number, meanwhile, is comparable or less than her distrust among Hispanic and "other" Democrats, and only varies from her substantial distrust among black Democrats by around 9%.