Date: Wed, 5 May 1999 12:39:16 -0500 (CDT) From: Jonathan Lemon <jlemon@americantv.com> To: Doug Rabson <dfr@nlsystems.com>, CyberPsychotic <fygrave@tigerteam.net>, hackers@freebsd.org Subject: Re: io ports reading/writing Message-ID: <199905051739.MAA11985@free.pcs> In-Reply-To: <local.mail.freebsd-hackers/Pine.BSF.4.05.9905051016150.411-100000@herring.nlsystems.com> References: <local.mail.freebsd-hackers/Pine.GSO.4.05.9905051359030.632-100000@kyrnet.kg>
next in thread | previous in thread | raw e-mail | index | archive | help
In article <local.mail.freebsd-hackers/Pine.BSF.4.05.9905051016150.411-100000@herring.nlsystems.com> you write: >On Wed, 5 May 1999, CyberPsychotic wrote: > >> ~ >> ~ The access control for io ports is controlled by the file-system >> ~ permissions on /dev/io. In a standard setup, only root can access this >> ~ device. >> ~ >> >> yes. But I was refering to linux scheme, where you can set the port-range, >> so the code wouldn't make any unintentional damage. (like if you're working >> with cmos you could only permit 0x70/0x71 ports, so even if code goes nuts, >> your disks will be safe). This is basically programmer's problem of course, >> but the feature is very handy. > >I don't quite understand the i386 architecture at this level but I seem to >remember that this support would require significant changes in the way we >handle processes and there might have been some performance implications. >I don't think its a big problem in practice. It is supported via the i386_get_ioperm(2) system call, which requires "options VM86" in the kernel. But you are right, it is slower than just opening up "/dev/io". -- Jonathan To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199905051739.MAA11985>