Date: Thu, 19 Apr 2007 20:28:35 +0300 From: Diomidis Spinellis <dds@aueb.gr> To: Robert Watson <rwatson@FreeBSD.org> Cc: arch@FreeBSD.org, re@FreeBSD.org Subject: Re: Accounting changes Message-ID: <4627A6C3.2070409@aueb.gr> In-Reply-To: <20070419101815.Y2913@fledge.watson.org> References: <461958CC.4040804@aueb.gr> <20070414170218.M76326@fledge.watson.org> <4621E826.6050306@aueb.gr> <20070415105157.J84174@fledge.watson.org> <46231C64.9010707@aueb.gr> <20070419101815.Y2913@fledge.watson.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Robert Watson wrote:
> If we're willing to assume architectures can only read their own
> accounting files (the status quo), the above argument doesn't really
> make sense. You end up with a series of versions of "struct acct", and
> that code is architecture-neutral. Thinking about it more, I'm not sure
> a per file header is even required or desired (as I had previously
> suggested), simply a per-record versioning scheme, allowing a reboot
> onto a new kernel to continue to write to the existing accounting data.
> Read the first 16 bytes, if the first byte is non-0 then it's the
> original "struct acct" layout, and otherwise the second byte is the
> version number to use. Or in the interests of forward compatibility,
> include a length parameter in another 16 bytes so you can skip over
> records if necessary in order to allow the kernel to move back and
> forward across file versions if there's a problem after the upgrade.
I will work on developing a mechanism for backwards compatibility.
Adding a version and length field in each record as you now propose is I
think the way to go. However, it is not trivial, because lastcomm(1)
often reads the file backwards. One approach would be to add a flag in
ac_flags indicating a new version, and second copy of the length field
before ac_flags. Thus, the new structure would be something like:
#define AC_COMM_LEN 16
struct acct {
char ac_zero; /* zero identifies new version */
char ac_version; /* record version number */
u_int16_t ac_len; /* record length */
/* Business as usual: */
char ac_comm[AC_COMM_LEN]; /* command name */
comp_t ac_utime; /* user time */
comp_t ac_stime; /* system time */
comp_t ac_etime; /* elapsed time */
time_t ac_btime; /* starting time */
uid_t ac_uid; /* user id */
gid_t ac_gid; /* group id */
u_int16_t ac_mem; /* average memory usage */
comp_t ac_io; /* count of IO blocks */
__dev_t ac_tty; /* controlling tty */
/* Changes here: */
u_int16_t ac_len2; /* record length */
#define AFORK 0x01 /* forked but not exec'ed */
/* ASU is no longer supported */
#define ASU 0x02 /* used super-user permissions */
#define ACOMPAT 0x04 /* used compatibility mode */
#define ACORE 0x08 /* dumped core */
#define AXSIG 0x10 /* killed by a signal */
#define ANVER 0x20 /* new record version */
u_int8_t ac_flag; /* accounting flags */
};
This structure allows the file to be read in both directions, and also
provides backwards compatibility.
Diomidis
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4627A6C3.2070409>