That shouldn't happen. To use the x86 code as a basis, segkp_unlock() will call hat_unload() which will call either hat_kmap_unload() or hat_unload_callback().
Then either of those will call hat_pte_unmap(), which will eventually call hati_sync_pte_to_page() which syncs the ref and mod bits of the real PTE with the bits in the associated page_t. (SPARC follows the same general process, with hat_unload_callback() calling sfmmu_hblk_unload(), which then calls sfmmu_ttesync().) Therefore, by the time segkp_unlock() calls hat_ismod(), it will be perfectly legal to simply check the bits in the page_t. This message posted from opensolaris.org _______________________________________________ opensolaris-code mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
