From: Kishon Vijay Abraham I <kishon@ti.com>
To: Alan Mikhak <alan.mikhak@sifive.com>,
<linux-kernel@vger.kernel.org>, <linux-pci@vger.kernel.org>,
<lorenzo.pieralisi@arm.com>, Bjorn Helgaas <bhelgaas@google.com>,
Palmer Dabbelt <palmer@sifive.com>,
Paul Walmsley <paul.walmsley@sifive.com>
Subject: Re: [PATCH] PCI: endpoint: cast the page number to phys_addr_t
Date: Wed, 9 Oct 2019 14:33:26 +0530 [thread overview]
Message-ID: <69ec3cdf-a7e8-d926-ccba-a1edbb92348d@ti.com> (raw)
In-Reply-To: <CABEDWGx5MzsdcKzNzCtt3DxXAEWK69Bm-QBK0248rGAvWaU22w@mail.gmail.com>
Hi Alan,
On 07/10/19 11:14 PM, Alan Mikhak wrote:
> On Fri, Oct 4, 2019 at 6:49 PM Alan Mikhak <alan.mikhak@sifive.com> wrote:
>>
>> From: Alan Mikhak <alan.mikhak@sifive.com>
>>
>> Modify pci_epc_mem_alloc_addr() to cast the variable 'pageno'
>> from type 'int' to 'phys_addr_t' before shifting left. This
>> cast is needed to avoid treating bit 31 of 'pageno' as the
>> sign bit which would otherwise get sign-extended to produce
>> a negative value. When added to the base address of PCI memory
>> space, the negative value would produce an invalid physical
>> address which falls before the start of the PCI memory space.
>>
>> Signed-off-by: Alan Mikhak <alan.mikhak@sifive.com>
Thanks for the patch.
The change-log title should start with "capitalized verb"
linux-pci follows certain guidelines listed here
https://lore.kernel.org/r/20171026223701.GA25649@bhelgaas-glaptop.roam.corp.google.com/
Once that gets fixed
Acked-by: Kishon Vijay Abraham I <kishon@ti.com>
>> ---
>> drivers/pci/endpoint/pci-epc-mem.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
>> index 2bf8bd1f0563..d2b174ce15de 100644
>> --- a/drivers/pci/endpoint/pci-epc-mem.c
>> +++ b/drivers/pci/endpoint/pci-epc-mem.c
>> @@ -134,7 +134,7 @@ void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
>> if (pageno < 0)
>> return NULL;
>>
>> - *phys_addr = mem->phys_base + (pageno << page_shift);
>> + *phys_addr = mem->phys_base + ((phys_addr_t)pageno << page_shift);
>> virt_addr = ioremap(*phys_addr, size);
>> if (!virt_addr)
>> bitmap_release_region(mem->bitmap, pageno, order);
>> --
>> 2.7.4
>>
>
> Hi Kishon,
>
> This issue was observed when requesting pci_epc_mem_alloc_addr()
> to allocate a region of size 0x40010000ULL (1GB + 64KB) from a
> 128GB PCI address space with page sizes being 64KB. This resulted
> in 'pageno' value of '0x8000' as the first available page in a
> contiguous region for the requested size due to other smaller
> regions having been allocated earlier. With 64KB page sizes,
> the variable 'page_shift' holds a value of 0x10. Shifting 'pageno'
> 16 bits to the left results in an 'int' value whose bit 31 is set.
>
> [ 10.565256] __pci_epc_mem_init: mem size 0x2000000000 page_size 0x10000
> [ 10.571613] __pci_epc_mem_init: mem pages 0x200000 bitmap_size
> 0x40000 page_shift 0x10
>
> PCI memory base 0x2000000000
> PCI memory size 128M 0x2000000000
> page_size 64K 0x10000
> page_shift 16 0x10
> pages 2M 0x200000
> bitmap_size 256K 0x40000
>
> [ 702.050299] pci_epc_mem_alloc_addr: size 0x10000 order 0x0 pageno
> 0x4 virt_add 0xffffffd0047b0000 phys_addr 0x2000040000
> [ 702.061424] pci_epc_mem_alloc_addr: size 0x10000 order 0x0 pageno
> 0x5 virt_add 0xffffffd0047d0000 phys_addr 0x2000050000
> [ 702.203933] pci_epc_mem_alloc_addr: size 0x40010000 order 0xf
> pageno 0x8000 virt_add 0xffffffd004800000 phys_addr 0x1f80000000
> [ 702.216547] Oops - store (or AMO) access fault [#1]
> :::
> [ 702.310198] sstatus: 0000000200000120 sbadaddr: ffffffd004804000
> scause: 0000000000000007
Thank you Alan for testing this and sending a patch to fix it.
Cheers
Kishon
prev parent reply other threads:[~2019-10-09 9:04 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-10-05 1:49 [PATCH] PCI: endpoint: cast the page number to phys_addr_t Alan Mikhak
2019-10-07 17:44 ` Alan Mikhak
2019-10-07 17:52 ` Alan Mikhak
2019-10-09 9:03 ` Kishon Vijay Abraham I [this message]
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=69ec3cdf-a7e8-d926-ccba-a1edbb92348d@ti.com \
--to=kishon@ti.com \
--cc=alan.mikhak@sifive.com \
--cc=bhelgaas@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=lorenzo.pieralisi@arm.com \
--cc=palmer@sifive.com \
--cc=paul.walmsley@sifive.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;
as well as URLs for NNTP newsgroup(s).