linux-kernel - Re: [PATCH 12/14] x86 ioapic: Simplify probe_nr_irqs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <m1pr2mv01y.fsf@fess.ebiederm.org>
Date:	Mon, 29 Mar 2010 22:41:13 -0700
From:	ebiederm@...ssion.com (Eric W. Biederman)
To:	Yinghai Lu <yinghai@...nel.org>
Cc:	"H. Peter Anvin" <hpa@...or.com>,
	Thomas Gleixner <tglx@...utronix.de>,
	Ingo Molnar <mingo@...e.hu>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Jesse Barnes <jbarnes@...tuousgeek.org>,
	linux-kernel@...r.kernel.org, Thomas Renninger <trenn@...e.de>,
	Suresh Siddha <suresh.b.siddha@...el.com>, len.brown@...el.com,
	Tony Luck <tony.luck@...el.com>,
	Fenghua Yu <fenghua.yu@...el.com>, linux-acpi@...r.kernel.org,
	Iranna D Ankad <iranna.ankad@...ibm.com>,
	Gary Hade <garyhade@...ibm.com>,
	Natalie Protasevich <protasnb@...il.com>
Subject: Re: [PATCH 12/14] x86 ioapic: Simplify probe_nr_irqs_gsi.

Yinghai Lu <yinghai@...nel.org> writes:

> On 03/29/2010 09:43 PM, Eric W. Biederman wrote:
>> Yinghai Lu <yinghai@...nel.org> writes:
>> 
>>> On Mon, Mar 29, 2010 at 4:20 PM, Eric W. Biederman
>>> <ebiederm@...ssion.com> wrote:
>>>> From: Eric W. Biederman <ebiederm@...ssion.com>
>>>>
>>>> Use the global gsi_end value now that all ioapics have
>>>> valid gsi numbers instead of a combination of acpi_probe_gsi
>>>> and walking all of the ioapics and couting their number of
>>>> entries by hand if acpi_probe_gsi gave us an answer we did
>>>> not like.
>>>>
>>>> This fixes a small bug in probe_nr_irqs_gsi.  Previously
>>>> acpi_probe_gsi unnecessarily added 1 to the maximum
>>>> gsi_end value.  gsi_end is already one past the end of
>>>> the number of gsi's so the additional increment was
>>>> superfluous.
>>>>
>>>> Signed-off-by: Eric W. Biederman <ebiederm@...ssion.com>
>>>> ---
>>>>  arch/x86/include/asm/mpspec.h  |    6 ------
>>>>  arch/x86/kernel/acpi/boot.c    |   23 -----------------------
>>>>  arch/x86/kernel/apic/io_apic.c |   17 +++--------------
>>>>  3 files changed, 3 insertions(+), 43 deletions(-)
>>>>
>>>> diff --git a/arch/x86/include/asm/mpspec.h b/arch/x86/include/asm/mpspec.h
>>>> index 29994f0..c82868e 100644
>>>> --- a/arch/x86/include/asm/mpspec.h
>>>> +++ b/arch/x86/include/asm/mpspec.h
>>>> @@ -105,12 +105,6 @@ extern void mp_config_acpi_legacy_irqs(void);
>>>>  struct device;
>>>>  extern int mp_register_gsi(struct device *dev, u32 gsi, int edge_level,
>>>>                                 int active_high_low);
>>>> -extern int acpi_probe_gsi(void);
>>>> -#else /* !CONFIG_ACPI: */
>>>> -static inline int acpi_probe_gsi(void)
>>>> -{
>>>> -       return 0;
>>>> -}
>>>>  #endif /* CONFIG_ACPI */
>>>>
>>>>  #define PHYSID_ARRAY_SIZE      BITS_TO_LONGS(MAX_APICS)
>>>> diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
>>>> index 9c48e99..0e514a1 100644
>>>> --- a/arch/x86/kernel/acpi/boot.c
>>>> +++ b/arch/x86/kernel/acpi/boot.c
>>>> @@ -875,29 +875,6 @@ static int __init acpi_parse_madt_lapic_entries(void)
>>>>  extern int es7000_plat;
>>>>  #endif
>>>>
>>>> -int __init acpi_probe_gsi(void)
>>>> -{
>>>> -       int idx;
>>>> -       int gsi;
>>>> -       int max_gsi = 0;
>>>> -
>>>> -       if (acpi_disabled)
>>>> -               return 0;
>>>> -
>>>> -       if (!acpi_ioapic)
>>>> -               return 0;
>>>> -
>>>> -       max_gsi = 0;
>>>> -       for (idx = 0; idx < nr_ioapics; idx++) {
>>>> -               gsi = mp_gsi_routing[idx].gsi_end;
>>>> -
>>>> -               if (gsi > max_gsi)
>>>> -                       max_gsi = gsi;
>>>> -       }
>>>> -
>>>> -       return max_gsi + 1;
>>>> -}
>>>> -
>>>>  static void assign_to_mp_irq(struct mpc_intsrc *m,
>>>>                                    struct mpc_intsrc *mp_irq)
>>>>  {
>>>> diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
>>>> index 996cf8f..b57b7a3 100644
>>>> --- a/arch/x86/kernel/apic/io_apic.c
>>>> +++ b/arch/x86/kernel/apic/io_apic.c
>>>> @@ -3837,22 +3837,11 @@ int __init io_apic_get_redir_entries (int ioapic)
>>>>
>>>>  void __init probe_nr_irqs_gsi(void)
>>>>  {
>>>> -       int nr = 0;
>>>> +       int nr;
>>>>
>>>> -       nr = acpi_probe_gsi();
>>>> -       if (nr > nr_irqs_gsi) {
>>>> +       nr = gsi_end;
>>>
>>> you may need +1 here
>> 
>> As documented in my comment that extra +1 has every appearance of a
>> bug.  Nothing is at gsi_end.  gsi_end is already at 1 past the last in
>> use gsi.  Therefore an extra +1 puts us two past the end for no
>> apparent reason.
>
> io_apic_get_redir_entries(), and io apic register readingout will return 23 if the total entries is 24.

Good catch.  I don't know if I have ever seen a function with a name to function
correspondence.  Apparently sfi.c also got this wrong when it was written.
and I missed the +1 in other places.

I guess this calls for another patch way at the beginning of my series that fixes
this brain damage.

Eric
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/