linux-kernel - Re: [PATCH] soc/tegra: refactor soc_is

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]
Date:   Thu, 22 Nov 2018 11:49:27 +0100
From:   Arnd Bergmann <arnd@...db.de>
To:     tiny.windzz@...il.com
Cc:     Thierry Reding <thierry.reding@...il.com>,
        Jonathan Hunter <jonathanh@...dia.com>,
        "open list:TEGRA ARCHITECTURE SUPPORT" <linux-tegra@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Anders Roxell <anders.roxell@...aro.org>
Subject: Re: [PATCH] soc/tegra: refactor soc_is_tegra()

On Thu, Nov 22, 2018 at 11:45 AM Arnd Bergmann <arnd@...db.de> wrote:
>
> On Wed, Nov 21, 2018 at 3:40 PM Yangtao Li <tiny.windzz@...il.com> wrote:
> >
> > of_find_node_by_path() acquires a reference to the node returned by
> > it and that reference needs to be dropped by its caller.soc_is_tegra()
> > doesn't do that, so fix it.Call of_machine_is_compatible() to refactor
> > soc_is_tegra() whcih automatically manages the reference count.
> >
> > Signed-off-by: Yangtao Li <tiny.windzz@...il.com>
> > ---

Anders ran into a crash after this patch, on a non-tegra qemu platform:

 [    0.055907] ASID allocator initialised with 32768 entries
[    0.063381] rcu: Hierarchical SRCU implementation.
[    0.137238] Unable to handle kernel paging request at virtual
address ffff00000935f0c0
[    0.137274] Mem abort info:
[    0.137291]   ESR = 0x96000007
[    0.137320]   Exception class = DABT (current EL), IL = 32 bits
[    0.137337]   SET = 0, FnV = 0
[    0.137352]   EA = 0, S1PTW = 0
[    0.137370] Data abort info:
[    0.137387]   ISV = 0, ISS = 0x00000007
[    0.137401]   CM = 0, WnR = 0
[    0.137456] swapper pgtable: 4k pages, 48-bit VAs, pgdp = (____ptrval____)
[    0.137479] [ffff00000935f0c0] pgd=00000000bdfff003,
pud=00000000bdffe003, pmd=00000000bdffa003, pte=0000000000000000
[    0.137644] Internal error: Oops: 96000007 [#1] PREEMPT SMP
[    0.137766] Modules linked in:
[    0.137950] CPU: 0 PID: 1 Comm: swapper/0 Not tainted
4.20.0-rc3-next-20181122-00006-g38d8a1f80349-dirty #2
[    0.137975] Hardware name: linux,dummy-virt (DT)
[    0.138071] pstate: 00000085 (nzcv daIf -PAN -UAO)
[    0.138110] pc : __of_device_is_compatible+0x30/0x138
[    0.138132] lr : of_device_is_compatible+0x40/0x68
[    0.138147] sp : ffff00000804bc80
[    0.138167] x29: ffff00000804bc80 x28: 0000000000000000
[    0.138207] x27: 0000000000000000 x26: 0000000000000000
[    0.138224] x25: ffff80007dfed2a8 x24: ffff000009301000
[    0.138239] x23: 0000000000000000 x22: 0000000000000000
[    0.138254] x21: ffff00000935f0c0 x20: ffff00000935f0c0
[    0.138269] x19: 0000000000000000 x18: 0000000000000400
[    0.138284] x17: 0000000000000000 x16: 0000000000000000
[    0.138298] x15: 0000000000000400 x14: 0000000000000400
[    0.138313] x13: 0000000000000000 x12: 0000000000000020
[    0.138327] x11: 0000000000000008 x10: 0101010101010101
[    0.138357] x9 : 6862726efffefeff x8 : 7f7f7f7f7f7f7f7f
[    0.138373] x7 : fefefefefefeff2e x6 : 0080808080808080
[    0.138387] x5 : 0000000000000002 x4 : 0000000000000001
[    0.138402] x3 : 0000000000000000 x2 : 0000000000000000
[    0.138416] x1 : ffff00000935f0c0 x0 : ffff80007dfed2a8
[    0.138475] Process swapper/0 (pid: 1, stack limit = 0x(____ptrval____))
[    0.138540] Call trace:
[    0.138607]  __of_device_is_compatible+0x30/0x138
[    0.138632]  of_device_is_compatible+0x40/0x68
[    0.138654]  of_machine_is_compatible+0x34/0x68
[    0.138672]  soc_is_tegra+0x2c/0x40
[    0.138689]  tegra_flowctrl_init+0x28/0x108
[    0.138706]  do_one_initcall+0x5c/0x178
[    0.138722]  kernel_init_freeable+0xd0/0x240
[    0.138741]  kernel_init+0x10/0x108
[    0.138755]  ret_from_fork+0x10/0x18
[    0.138913] Code: b4000861 f90013b5 aa0103f5 52800013 (39400021)
[    0.139229] ---[ end trace d4d0fc77e9b04fa6 ]---
[    0.139448] note: swapper/0[1] exited with preempt_count 1
[    0.140598] Kernel panic - not syncing: Attempted to kill init!
exitcode=0x0000000b
[    0.140767] ---[ end Kernel panic - not syncing: Attempted to kill
init! exitcode=0x0000000b ]---

I'm not completely sure what's wrong with the patch, but I assume
it never worked on non-tegra machines.

> > diff --git a/drivers/soc/tegra/common.c b/drivers/soc/tegra/common.c
> > index cd8f41351add..0b40700b672a 100644
> > --- a/drivers/soc/tegra/common.c
> > +++ b/drivers/soc/tegra/common.c
> > @@ -22,11 +22,13 @@ static const struct of_device_id tegra_machine_match[] = {
> >
> >  bool soc_is_tegra(void)
> >  {
> > -       struct device_node *root;
> > +       struct of_device_id *match = tegra_machine_match;
> >
> > -       root = of_find_node_by_path("/");
> > -       if (!root)
> > -               return false;
> > +       while(match->compatible){
> > +               if(of_machine_is_compatible(match->compatible))
> > +                       return true;
> > +               match++;
> > +       }
> >
> > -       return of_match_node(tegra_machine_match, root) != NULL;
> > +       return false;

I would also note that this is a rather inefficient way to check
for a particular platform, as we to do the string search through
the DT for each entry in the table now instead of doing it once.

       Arnd