On Tue, Jun 21, 2022 at 09:29:13AM +0200, Sebastien Marie wrote:
> On Tue, Jun 21, 2022 at 02:58:46PM +1000, Jonathan Gray wrote:
> > Intel CPUs used to have strings like
> > cpu0: Intel(R) Pentium(R) M processor 1200MHz ("GenuineIntel" 686-class)
> > 1.20 GHz
> > cpu0: Intel(R) Core(TM) i7-5600U CPU @ 2.60GHz, 2494.61 MHz, 06-3d-04
> > recent CPUs use
> > cpu0: 11th Gen Intel(R) Core(TM) i5-1130G7 @ 1.10GHz, 30009.37 MHz, 06-8c-01
> > cpu0: 12th Gen Intel(R) Core(TM) i5-12400, 4390.71 MHz, 06-97-02
> > cpu0: 12th Gen Intel(R) Core(TM) i7-1260P, 1995.55 MHz, 06-9a-03
> >
> > Index: patterns.c
> > ===================================================================
> > RCS file: /cvs/src/usr.sbin/fw_update/patterns.c,v
> > retrieving revision 1.3
> > diff -u -p -r1.3 patterns.c
> > --- patterns.c 10 Mar 2022 07:12:13 -0000 1.3
> > +++ patterns.c 21 Jun 2022 04:31:24 -0000
> > @@ -94,7 +94,7 @@ main(void)
> > printf("%s\n", "bwfm");
> > printf("%s\n", "bwi");
> > printf("%s\n", "intel");
> > - printf("%s\n", "intel ^cpu0: Intel");
> > + printf("%s\n", "intel ^cpu0:*Intel(R)");
> > printf("%s\n", "inteldrm");
> > print_devices("inteldrm", i915_devices, nitems(i915_devices));
> > printf("%s\n", "ipw");
> >
>
> If I properly understood the way patterns file is used in fw_update, having a
> '*' in the middle of the searched string could be hasardous and will result
> in
> possible false positive (installing a firmware whereas not need).
>
> With the pattern "intel ^cpu0:*Intel(R)", the search is done using:
>
> _d="intel"
> _m="^cpu0:*Intel(R)"
> _nl=$(echo) # newline: \n
>
> [ "$_m" ] || _m="${_nl}${_d}[0-9] at " # no change
> [ "$_m" = "${_m#^}" ] || _m="${_nl}${_m#^}" # changed to
> _m="\ncpu0:*Intel(R)"
>
> if [[ ${_dmesgtail} = *$_m* ]] ; then
> echo "$_d"
> fi
>
> The final searched string (glob) is "*\ncpu0:*Intel(R)*".
>
> It means that if the dmesg contains "\ncpu0:" and "Intel(R)" (in this order,
> but
> not necessary on the same line) it will match.
>
> On dmesglog, I found only few candidates for "Intel(R)" string not on a
> "cpuX:"
> line:
>
> bios0: Intel(R) Client Systems NUC8i5BEH
> ugen0 at uhub2 port 2 "Intel Corporation Intel(R) Centrino(R)
> Wireless-N + WiMAX 6150" rev 2.00/0.00 addr 3
> ugen0 at uhub2 port 6 "Intel(R) Corporation Intel(R) Centrino(R)
> Advanced-N + WiMAX 6250" rev 2.00/0.00 addr 4
> uhidev0 at uhub2 port 1 configuration 1 interface 0 "Intel Corporation
> Intel(R) Sensor Solution" rev 2.00/0.01 addr 3
>
> the bios0 entry is *before* cpu0, so it will not match. But the others
> entries
> will match (assuming the dmesg contains a cpu0: line, which somehow expected).
>
> Below an alternative diff to match the Intel cpu without using '*':
But then even more cases are required
cpu0: Genuine Intel(R) CPU @ 600MHz, 600.10 MHz
cpu0: Genuine Intel(R) CPU @ 1.00GHz, 1000.13 MHz, 06-26-01
cpu0: Genuine Intel(R) CPU L2400 @ 1.66GHz ("GenuineIntel" 686-class) 1.67 GHz,
06-0e-08
(we don't have microcode for these)
cpu0: Mobile Intel(R) Pentium(R) III CPU - M 1000MHz ("GenuineIntel" 686-class)
1 GHz
cpu0: Mobile Intel(R) Pentium(R) 4 - M CPU 2.00GHz ("GenuineIntel" 686-class) 2
GHz
cpu0: Mobile Genuine Intel(R) processor 1400MHz ("GenuineIntel" 686-class)
test chips seem to use "Genuine Intel(R) CPU 0000"
>
> diff 5177244f5c04ec0ff6b9b724b540740d62d2dcfa /home/semarie/repos/openbsd/src
> blob - d7cbab70c1654d8dc6ca067cea55b440fd70d426
> file + usr.sbin/fw_update/patterns.c
> --- usr.sbin/fw_update/patterns.c
> +++ usr.sbin/fw_update/patterns.c
> @@ -95,6 +95,7 @@ main(void)
> printf("%s\n", "bwi");
> printf("%s\n", "intel");
> printf("%s\n", "intel ^cpu0: Intel");
> + printf("%s\n", "intel ^cpu0: [0-9][0-9]th Gen Intel");
> printf("%s\n", "inteldrm");
> print_devices("inteldrm", i915_devices, nitems(i915_devices));
> printf("%s\n", "ipw");
>
> Comments ?
> --
> Sebastien Marie
>
>
