Models of disk galaxy formation commonly predict the existence of an extended reservoir of accreted hot gas surrounding massive spirals at low redshift. As a test of these models, we use X-ray and H-alpha data of the two massive, quiescent edge-on spirals NGC 5746 and NGC 5170 to investigate the amount and origin of any hot gas in their halos. Contrary to our earlier claim, the Chandra analysis of NGC 5746, employing more recent calibration data, does not reveal any significant evidence for diffuse X-ray emission outside the optical disk, with a 3-sigma upper limit to the halo X-ray luminosity of 4e39 erg/s. An identical study of the less massive NGC 5170 also fails to detect any extraplanar X-ray emission. By extracting hot halo properties of disk galaxies formed in cosmological hydrodynamical simulations, we compare these results to expectations for cosmological accretion of hot gas by spirals. For Milky Way-sized galaxies, these high-resolution simulations predict hot halo X-ray luminosities which are lower by a factor of ~2 compared to our earlier results reported by Toft et al. (2002). We find the new simulation predictions to be consistent with our observational constraints for both NGC 5746 and NGC 5170, while also confirming that the hot gas detected so far around more actively star-forming spirals is in general probably associated with stellar activity in the disk. Observational results on quiescent disk galaxies at the high-mass end are nevertheless providing powerful constraints on theoretical predictions, and hence on the assumed input physics in numerical studies of disk galaxy formation and evolution.