[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <54648EA202000078000471CF@mail.emea.novell.com>
Date: Thu, 13 Nov 2014 09:57:38 +0000
From: "Jan Beulich" <JBeulich@...e.com>
To: "David Vrabel" <david.vrabel@...rix.com>
Cc: <x86@...nel.org>, "Thomas Gleixner" <tglx@...utronix.de>,
"Greg Kroah-Hartman" <gregkh@...uxfoundation.org>,
<xen-devel@...ts.xenproject.org>,
"Boris Ostrovsky" <boris.ostrovsky@...cle.com>,
"Ingo Molnar" <mingo@...hat.com>, <linux-kernel@...r.kernel.org>,
"H. Peter Anvin" <hpa@...or.com>
Subject: Re: [Xen-devel] [PATCH 3/3] x86/xen: use the maximum MFN to
calculate the required DMA mask
>>> On 12.11.14 at 16:55, <JBeulich@...e.com> wrote:
>>>> On 12.11.14 at 16:25, <david.vrabel@...rix.com> wrote:
>> +u64
>> +xen_swiotlb_get_required_mask(struct device *dev)
>> +{
>> + u64 max_mfn;
>> +
>> + max_mfn = HYPERVISOR_memory_op(XENMEM_maximum_ram_page, NULL);
>> +
>> + return DMA_BIT_MASK(fls64(max_mfn << PAGE_SHIFT) + 1);
>> +}
>
> The value the hypercall returns is exclusive and an unsigned long.
The latter has further relevance: For one, HYPERVISOR_memory_op()
returns "int" currently - this needs to become "long" for the purpose
here (also for eventual future uses of XENMEM_maximum_gpfn and
XENMEM_{current,maximum}_reservation).
And then, using u64 rather than unsigned long for max_mfn would
still cause problems for 32-bit kernels on systems with memory
extending beyond the 8Tb boundary (since the long -> u64
conversion really goes long -> s64 -> u64).
Jan
--
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/
Powered by blists - more mailing lists