From: John Garry <john.garry@huawei.com>
To: Guenter Roeck <linux@roeck-us.net>
Cc: <bhelgaas@google.com>, <rafael@kernel.org>, <arnd@arndb.de>,
<lorenzo.pieralisi@arm.com>, <linux-kernel@vger.kernel.org>,
<linux-pci@vger.kernel.org>, <bp@suse.de>,
<wangkefeng.wang@huawei.com>, <linuxarm@huawei.com>,
<andy.shevchenko@gmail.com>,
<linux-arm-kernel@lists.infradead.org>, <catalin.marinas@arm.com>,
<will.deacon@arm.com>
Subject: Re: [PATCH v3 3/4] lib: logic_pio: Reject accesses to unregistered CPU MMIO regions
Date: Thu, 4 Apr 2019 17:52:35 +0100 [thread overview]
Message-ID: <24cc8006-0f0d-6b20-a466-e4a32a0bb656@huawei.com> (raw)
In-Reply-To: <20190404164130.GA12203@roeck-us.net>
On 04/04/2019 17:41, Guenter Roeck wrote:
> On Fri, Apr 05, 2019 at 12:00:01AM +0800, John Garry wrote:
>> Currently when accessing logical indirect PIO addresses in
>> logic_{in, out}{,s}, we first ensure that the region is registered.
>>
>> However, no such check exists for CPU MMIO regions. The CPU MMIO regions
>> would be registered by the PCI host - when PCI_IOBASE is defined - in
>> pci_register_io_range().
>>
>> We have seen scenarios when systems which don't have a PCI host or, they
>> do, and the PCI host probe fails, that certain devices attempts to still
>> attempt to access PCI IO ports; examples are in [1] and [2].
>>
>> And even though we should protect against this by ensuring the driver
>> calls request_{muxed_}region(), some don't do this:
>>
>> root@(none)$ insmod hwmon/f71805f.ko
>> Unable to handle kernel paging request at virtual address ffff7dfffee0002e
>> Mem abort info:
>> ESR = 0x96000046
>> Exception class = DABT (current EL), IL = 32 bits
>> SET = 0, FnV = 0
>> EA = 0, S1PTW = 0
>> Data abort info:
>> ISV = 0, ISS = 0x00000046
>> CM = 0, WnR = 1
>> swapper pgtable: 4k pages, 48-bit VAs, pgdp = (____ptrval____)
>> [ffff7dfffee0002e] pgd=000000000141c003, pud=000000000141d003, pmd=0000000000000000
>> Internal error: Oops: 96000046 [#1] PREEMPT SMP
>> Modules linked in: f71805f(+)
>> CPU: 20 PID: 2736 Comm: insmod Not tainted 5.1.0-rc1-00003-g6f1bfec2a620-dirty #99
>> Hardware name: Huawei Taishan 2280 /D05, BIOS Hisilicon D05 IT21 Nemo 2.0 RC0 04/18/2018
>> pstate: 80000005 (Nzcv daif -PAN -UAO)
>> pc : logic_outb+0x54/0xb8
>> lr : f71805f_find+0x2c/0x1b8 [f71805f]
>> sp : ffff000025fbba90
>> x29: ffff000025fbba90 x28: ffff000008b944d0
>> x27: ffff000025fbbdf0 x26: 0000000000000100
>> x25: ffff801f8c270580 x24: ffff000011420000
>> x23: ffff000025fbbb3e x22: ffff000025fbbb40
>> x21: ffff000008b991b8 x20: 0000000000000087
>> x19: 000000000000002e x18: ffffffffffffffff
>> x17: 0000000000000000 x16: 0000000000000000
>> x15: ffff00001127d6c8 x14: 0000000000000000
>> x13: 0000000000000000 x12: 0000000000000000
>> x11: 0000000000010820 x10: 0000841fdac40000
>> x9 : 0000000000000001 x8 : 0000000040000000
>> x7 : 0000000000210d00 x6 : 0000000000000000
>> x5 : ffff801fb6a46040 x4 : ffff841febeaeda0
>> x3 : 0000000000ffbffe x2 : ffff000025fbbb40
>> x1 : ffff7dfffee0002e x0 : ffff7dfffee00000
>> Process insmod (pid: 2736, stack limit = 0x(____ptrval____))
>> Call trace:
>> logic_outb+0x54/0xb8
>> f71805f_find+0x2c/0x1b8 [f71805f]
>> f71805f_init+0x38/0xe48 [f71805f]
>> do_one_initcall+0x5c/0x198
>> do_init_module+0x54/0x1b0
>> load_module+0x1dc4/0x2158
>> __se_sys_init_module+0x14c/0x1e8
>> __arm64_sys_init_module+0x18/0x20
>> el0_svc_common+0x5c/0x100
>> el0_svc_handler+0x2c/0x80
>> el0_svc+0x8/0xc
>> Code: d2bfdc00 f2cfbfe0 f2ffffe0 8b000021 (39000034)
>> ---[ end trace 10ea80bde051bbfc ]---
>> root@(none)$
>>
>> Note that the f71805f driver does not call request_{muxed_}region(), as it
>> should.
>>
Hi Guenter,
> ... which is the real problem, one that is not solved by this patch. This may
> result in parallel and descructive accesses if there is another device on the
> LPC bus, and another driver accessing that device. Personally I'd rather have
> request_muxed_region() added to the f71805f driver.
Right, we should and will still fix f71805f. If you recall, I did have
the f71805f fix in the v1 series, but you committed that it was
orthogonal, so I decided to take it out of this work for now.
And even if we fix up f71805f and other known drivers which don't call
request_muxed_region(), we still need to police against these rogue
accesses, which is what this patch attempts to do.
Thanks,
John
>
> Guenter
>
>> This patch adds a check to ensure that the CPU MMIO region is registered
>> prior to accessing the PCI IO ports.
>>
>> [1] https://lore.kernel.org/linux-pci/56F209A9.4040304@huawei.com
>> [2] https://lore.kernel.org/linux-arm-kernel/e6995b4a-184a-d8d4-f4d4-9ce75d8f47c0@huawei.com/
>>
>> This patch includes some other tidy-up.
>>
>> Signed-off-by: John Garry <john.garry@huawei.com>
>> ---
>> lib/logic_pio.c | 103 +++++++++++++++++++++++++++++++++++-------------
>> 1 file changed, 75 insertions(+), 28 deletions(-)
>>
>> diff --git a/lib/logic_pio.c b/lib/logic_pio.c
>> index 431cd8d99236..3d8d986e9dcb 100644
>> --- a/lib/logic_pio.c
>> +++ b/lib/logic_pio.c
>> @@ -193,95 +193,135 @@ unsigned long logic_pio_trans_cpuaddr(resource_size_t addr)
>>
>> #if defined(PCI_IOBASE)
>> #if defined(CONFIG_INDIRECT_PIO)
>> +#define INVALID_RANGE(range) \
>> + (!(range) || ((range)->flags == LOGIC_PIO_INDIRECT && !(range)->ops))
>> +
>> #define BUILD_LOGIC_IO(bw, type) \
>> type logic_in##bw(unsigned long addr) \
>> { \
>> type ret = (type)~0; \
>> + struct logic_pio_hwaddr *range = find_io_range(addr); \
>> + \
>> + if (INVALID_RANGE(range)) { \
>> + WARN_ON_ONCE(1); \
>> + return ret; \
>> + } \
>> \
>> if (addr < MMIO_UPPER_LIMIT) { \
>> ret = read##bw(PCI_IOBASE + addr); \
>> } else if (addr >= MMIO_UPPER_LIMIT && addr < IO_SPACE_LIMIT) { \
>> - struct logic_pio_hwaddr *entry = find_io_range(addr); \
>> size_t sz = sizeof(type); \
>> + void *hostdata = range->hostdata; \
>> \
>> - if (entry && entry->ops) \
>> - ret = entry->ops->in(entry->hostdata, addr, sz);\
>> - else \
>> - WARN_ON_ONCE(1); \
>> + if (range->ops->in) \
>> + ret = range->ops->in(hostdata, addr, sz); \
>> } \
>> return ret; \
>> } \
>> \
>> -void logic_out##bw(type value, unsigned long addr) \
>> +void logic_out##bw(type val, unsigned long addr) \
>> { \
>> + struct logic_pio_hwaddr *range = find_io_range(addr); \
>> + \
>> + if (INVALID_RANGE(range)) { \
>> + WARN_ON_ONCE(1); \
>> + return; \
>> + } \
>> + \
>> if (addr < MMIO_UPPER_LIMIT) { \
>> - write##bw(value, PCI_IOBASE + addr); \
>> + write##bw(val, PCI_IOBASE + addr); \
>> } else if (addr >= MMIO_UPPER_LIMIT && addr < IO_SPACE_LIMIT) { \
>> - struct logic_pio_hwaddr *entry = find_io_range(addr); \
>> size_t sz = sizeof(type); \
>> + void *hostdata = range->hostdata; \
>> \
>> - if (entry && entry->ops) \
>> - entry->ops->out(entry->hostdata, \
>> - addr, value, sz); \
>> - else \
>> - WARN_ON_ONCE(1); \
>> + if (range->ops->out) \
>> + range->ops->out(hostdata, addr, val, sz); \
>> } \
>> } \
>> \
>> void logic_ins##bw(unsigned long addr, void *buf, unsigned int cnt) \
>> { \
>> + struct logic_pio_hwaddr *range = find_io_range(addr); \
>> + \
>> + if (INVALID_RANGE(range)) { \
>> + WARN_ON_ONCE(1); \
>> + return; \
>> + } \
>> + \
>> if (addr < MMIO_UPPER_LIMIT) { \
>> reads##bw(PCI_IOBASE + addr, buf, cnt); \
>> } else if (addr >= MMIO_UPPER_LIMIT && addr < IO_SPACE_LIMIT) { \
>> - struct logic_pio_hwaddr *entry = find_io_range(addr); \
>> size_t sz = sizeof(type); \
>> + void *hostdata = range->hostdata; \
>> \
>> - if (entry && entry->ops) \
>> - entry->ops->ins(entry->hostdata, \
>> - addr, buf, sz, cnt); \
>> - else \
>> - WARN_ON_ONCE(1); \
>> + if (range->ops->ins) \
>> + range->ops->ins(hostdata, addr, buf, sz, cnt); \
>> } \
>> - \
>> } \
>> \
>> void logic_outs##bw(unsigned long addr, const void *buf, \
>> unsigned int cnt) \
>> { \
>> + struct logic_pio_hwaddr *range = find_io_range(addr); \
>> + \
>> + if (INVALID_RANGE(range)) { \
>> + WARN_ON_ONCE(1); \
>> + return; \
>> + } \
>> + \
>> if (addr < MMIO_UPPER_LIMIT) { \
>> writes##bw(PCI_IOBASE + addr, buf, cnt); \
>> } else if (addr >= MMIO_UPPER_LIMIT && addr < IO_SPACE_LIMIT) { \
>> - struct logic_pio_hwaddr *entry = find_io_range(addr); \
>> size_t sz = sizeof(type); \
>> + void *hostdata = range->hostdata; \
>> \
>> - if (entry && entry->ops) \
>> - entry->ops->outs(entry->hostdata, \
>> - addr, buf, sz, cnt); \
>> - else \
>> - WARN_ON_ONCE(1); \
>> + if (range->ops->outs) \
>> + range->ops->outs(hostdata, addr, buf, sz, cnt); \
>> } \
>> }
>>
>> #else /* CONFIG_INDIRECT_PIO */
>>
>> +#define INVALID_RANGE(range) (!(range))
>> +
>> #define BUILD_LOGIC_IO(bw, type) \
>> type logic_in##bw(unsigned long addr) \
>> { \
>> type ret = (type)~0; \
>> + struct logic_pio_hwaddr *range = find_io_range(addr); \
>> + \
>> + if (INVALID_RANGE(range)) { \
>> + WARN_ON_ONCE(1); \
>> + return ret; \
>> + } \
>> \
>> if (addr < MMIO_UPPER_LIMIT) \
>> ret = read##bw(PCI_IOBASE + addr); \
>> return ret; \
>> } \
>> \
>> -void logic_out##bw(type value, unsigned long addr) \
>> +void logic_out##bw(type val, unsigned long addr) \
>> { \
>> + struct logic_pio_hwaddr *range = find_io_range(addr); \
>> + \
>> + if (INVALID_RANGE(range)) { \
>> + WARN_ON_ONCE(1); \
>> + return; \
>> + } \
>> + \
>> if (addr < MMIO_UPPER_LIMIT) \
>> - write##bw(value, PCI_IOBASE + addr); \
>> + write##bw(val, PCI_IOBASE + addr); \
>> } \
>> \
>> void logic_ins##bw(unsigned long addr, void *buf, unsigned int cnt) \
>> { \
>> + struct logic_pio_hwaddr *range = find_io_range(addr); \
>> + \
>> + if (INVALID_RANGE(range)) { \
>> + WARN_ON_ONCE(1); \
>> + return; \
>> + } \
>> + \
>> if (addr < MMIO_UPPER_LIMIT) \
>> reads##bw(PCI_IOBASE + addr, buf, cnt); \
>> } \
>> @@ -289,6 +329,13 @@ void logic_ins##bw(unsigned long addr, void *buf, unsigned int cnt) \
>> void logic_outs##bw(unsigned long addr, const void *buf, \
>> unsigned int cnt) \
>> { \
>> + struct logic_pio_hwaddr *range = find_io_range(addr); \
>> + \
>> + if (INVALID_RANGE(range)) { \
>> + WARN_ON_ONCE(1); \
>> + return; \
>> + } \
>> + \
>> if (addr < MMIO_UPPER_LIMIT) \
>> writes##bw(PCI_IOBASE + addr, buf, cnt); \
>> }
>> --
>> 2.17.1
>>
>
> .
>
next prev parent reply other threads:[~2019-04-04 16:52 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-04-04 15:59 [PATCH v3 0/4] Fix system crash for accessing unmapped IO port regions John Garry
2019-04-04 15:59 ` [RFC PATCH v3 1/4] resource: Request IO port regions from children of ioport_resource John Garry
2019-04-04 16:00 ` [PATCH v3 2/4] lib: logic_pio: Use logical PIO low-level accessors for !CONFIG_INDIRECT_PIO John Garry
2019-04-04 16:00 ` [PATCH v3 3/4] lib: logic_pio: Reject accesses to unregistered CPU MMIO regions John Garry
2019-04-04 16:41 ` Guenter Roeck
2019-04-04 16:52 ` John Garry [this message]
2019-04-04 17:43 ` Guenter Roeck
2019-04-04 18:58 ` Bjorn Helgaas
2019-04-05 8:10 ` John Garry
2019-04-05 18:06 ` Bjorn Helgaas
2019-04-05 18:29 ` Guenter Roeck
2019-04-08 8:19 ` John Garry
2019-04-08 13:47 ` Guenter Roeck
2019-04-08 16:35 ` John Garry
2019-04-08 16:50 ` Will Deacon
2019-04-09 10:38 ` John Garry
2019-04-08 8:01 ` John Garry
2019-04-04 16:00 ` [PATCH v3 4/4] lib: logic_pio: Fix up some prints John Garry
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=24cc8006-0f0d-6b20-a466-e4a32a0bb656@huawei.com \
--to=john.garry@huawei.com \
--cc=andy.shevchenko@gmail.com \
--cc=arnd@arndb.de \
--cc=bhelgaas@google.com \
--cc=bp@suse.de \
--cc=catalin.marinas@arm.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=linux@roeck-us.net \
--cc=linuxarm@huawei.com \
--cc=lorenzo.pieralisi@arm.com \
--cc=rafael@kernel.org \
--cc=wangkefeng.wang@huawei.com \
--cc=will.deacon@arm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Powered by JetHome