linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: Alexey Kardashevskiy <aik@ozlabs.ru>
To: Gavin Shan <gwshan@linux.vnet.ibm.com>
Cc: Wei Yang <weiyang@linux.vnet.ibm.com>,
	linux-kernel@vger.kernel.org,
	Alex Williamson <alex.williamson@redhat.com>,
	Paul Mackerras <paulus@samba.org>,
	linuxppc-dev@lists.ozlabs.org,
	David Gibson <david@gibson.dropbear.id.au>
Subject: Re: [PATCH kernel v10 15/34] powerpc/powernv/ioda/ioda2: Rework TCE invalidation in tce_build()/tce_free()
Date: Thu, 14 May 2015 13:19:38 +1000	[thread overview]
Message-ID: <5554144A.50701@ozlabs.ru> (raw)
In-Reply-To: <20150514004824.GA18437@gwshan>

On 05/14/2015 10:48 AM, Gavin Shan wrote:
> On Tue, May 12, 2015 at 01:39:04AM +1000, Alexey Kardashevskiy wrote:
>> The pnv_pci_ioda_tce_invalidate() helper invalidates TCE cache. It is
>> supposed to be called on IODA1/2 and not called on p5ioc2. It receives
>> start and end host addresses of TCE table.
>>
>> IODA2 actually needs PCI addresses to invalidate the cache. Those
>> can be calculated from host addresses but since we are going
>> to implement multi-level TCE tables, calculating PCI address from
>> a host address might get either tricky or ugly as TCE table remains flat
>> on PCI bus but not in RAM.
>>
>> This moves pnv_pci_ioda_tce_invalidate() from generic pnv_tce_build/
>> pnt_tce_free and defines IODA1/2-specific callbacks which call generic
>> ones and do PHB-model-specific TCE cache invalidation. P5IOC2 keeps
>> using generic callbacks as before.
>>
>> This changes pnv_pci_ioda2_tce_invalidate() to receives TCE index and
>> number of pages which are PCI addresses shifted by IOMMU page shift.
>>
>> No change in behaviour is expected.
>>
>> Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
>> Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
>
> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
>
>> ---
>> Changes:
>> v10:
>> * moved before "Switch from iommu_table to new iommu_table_group" as it adds
>> list of groups to iommu_table and tce invalidation depends on it
>>
>> v9:
>> * removed confusing comment from commit log about unintentional calling of
>> pnv_pci_ioda_tce_invalidate()
>> * moved mechanical changes away to "powerpc/iommu: Move tce_xxx callbacks from ppc_md to iommu_table"
>> * fixed bug with broken invalidation in pnv_pci_ioda2_tce_invalidate -
>> @index includes @tbl->it_offset but old code added it anyway which later broke
>> DDW
>> ---
>> arch/powerpc/platforms/powernv/pci-ioda.c | 81 ++++++++++++++++++++++---------
>> arch/powerpc/platforms/powernv/pci.c      | 17 ++-----
>> 2 files changed, 61 insertions(+), 37 deletions(-)
>>
>> diff --git a/arch/powerpc/platforms/powernv/pci-ioda.c b/arch/powerpc/platforms/powernv/pci-ioda.c
>> index 2924abe..1b43e25 100644
>> --- a/arch/powerpc/platforms/powernv/pci-ioda.c
>> +++ b/arch/powerpc/platforms/powernv/pci-ioda.c
>> @@ -1678,18 +1678,19 @@ static void pnv_ioda_setup_bus_dma(struct pnv_ioda_pe *pe,
>> 	}
>> }
>>
>> -static void pnv_pci_ioda1_tce_invalidate(struct pnv_ioda_pe *pe,
>> -					 struct iommu_table *tbl,
>> -					 __be64 *startp, __be64 *endp, bool rm)
>> +static void pnv_pci_ioda1_tce_invalidate(struct iommu_table *tbl,
>> +		unsigned long index, unsigned long npages, bool rm)
>> {
>> +	struct pnv_ioda_pe *pe = tbl->data;
>> 	__be64 __iomem *invalidate = rm ?
>> 		(__be64 __iomem *)pe->tce_inval_reg_phys :
>> 		(__be64 __iomem *)tbl->it_index;
>> 	unsigned long start, end, inc;
>> 	const unsigned shift = tbl->it_page_shift;
>>
>> -	start = __pa(startp);
>> -	end = __pa(endp);
>> +	start = __pa(((__be64 *)tbl->it_base) + index - tbl->it_offset);
>> +	end = __pa(((__be64 *)tbl->it_base) + index - tbl->it_offset +
>> +			npages - 1);
>
> Platform is the only one knowing the TCE table layout and iommu_table_ops->get()
> helps to retrieve TCE entry for the given index. If iommu_table_ops->get() had
> returned the address of the TCE entry, not the content. Here, iommu_table_ops->get()
> can be reused and we hide the platform specific TCE table layout in iommu_table_ops->get()
> backend. However, it's not a big deal and it probably introduces more changes
> than expected. You judge it's worthy to do it or improve it later :-)


This will require a separate patch to convert tce_get() from returning 
value to returning address. I could do that, yes. Furthermore there is even 
pnv_tce() helper added later in this patchset which can be used for this in 
this patch (if I moved that patch earlier). But this would be a bigger 
change which is not very much related to what the patch does - 
cut-n-pasting invalidate() bits. May be later. The patchset is way too big 
already :(


>>
>> 	/* BML uses this case for p6/p7/galaxy2: Shift addr and put in node */
>> 	if (tbl->it_busno) {
>> @@ -1725,16 +1726,39 @@ static void pnv_pci_ioda1_tce_invalidate(struct pnv_ioda_pe *pe,
>> 	 */
>> }
>>
>> +static int pnv_ioda1_tce_build(struct iommu_table *tbl, long index,
>> +		long npages, unsigned long uaddr,
>> +		enum dma_data_direction direction,
>> +		struct dma_attrs *attrs)
>> +{
>> +	long ret = pnv_tce_build(tbl, index, npages, uaddr, direction,
>> +			attrs);
>
> The return value from pnv_tce_build() is "int" :-)

Oops.


>> +
>> +	if (!ret && (tbl->it_type & TCE_PCI_SWINV_CREATE))
>> +		pnv_pci_ioda1_tce_invalidate(tbl, index, npages, false);
>> +
>> +	return ret;
>> +}
>> +
>> +static void pnv_ioda1_tce_free(struct iommu_table *tbl, long index,
>> +		long npages)
>> +{
>> +	pnv_tce_free(tbl, index, npages);
>> +
>> +	if (tbl->it_type & TCE_PCI_SWINV_FREE)
>> +		pnv_pci_ioda1_tce_invalidate(tbl, index, npages, false);
>> +}
>> +
>> static struct iommu_table_ops pnv_ioda1_iommu_ops = {
>> -	.set = pnv_tce_build,
>> -	.clear = pnv_tce_free,
>> +	.set = pnv_ioda1_tce_build,
>> +	.clear = pnv_ioda1_tce_free,
>> 	.get = pnv_tce_get,
>> };
>>
>> -static void pnv_pci_ioda2_tce_invalidate(struct pnv_ioda_pe *pe,
>> -					 struct iommu_table *tbl,
>> -					 __be64 *startp, __be64 *endp, bool rm)
>> +static void pnv_pci_ioda2_tce_invalidate(struct iommu_table *tbl,
>> +		unsigned long index, unsigned long npages, bool rm)
>> {
>> +	struct pnv_ioda_pe *pe = tbl->data;
>> 	unsigned long start, end, inc;
>> 	__be64 __iomem *invalidate = rm ?
>> 		(__be64 __iomem *)pe->tce_inval_reg_phys :
>> @@ -1747,10 +1771,8 @@ static void pnv_pci_ioda2_tce_invalidate(struct pnv_ioda_pe *pe,
>> 	end = start;
>>
>> 	/* Figure out the start, end and step */
>> -	inc = tbl->it_offset + (((u64)startp - tbl->it_base) / sizeof(u64));
>> -	start |= (inc << shift);
>> -	inc = tbl->it_offset + (((u64)endp - tbl->it_base) / sizeof(u64));
>> -	end |= (inc << shift);
>> +	start |= (index << shift);
>> +	end |= ((index + npages - 1) << shift);
>> 	inc = (0x1ull << shift);
>> 	mb();
>>
>> @@ -1763,21 +1785,32 @@ static void pnv_pci_ioda2_tce_invalidate(struct pnv_ioda_pe *pe,
>> 	}
>> }
>>
>> -void pnv_pci_ioda_tce_invalidate(struct iommu_table *tbl,
>> -				 __be64 *startp, __be64 *endp, bool rm)
>> +static int pnv_ioda2_tce_build(struct iommu_table *tbl, long index,
>> +		long npages, unsigned long uaddr,
>> +		enum dma_data_direction direction,
>> +		struct dma_attrs *attrs)
>> {
>> -	struct pnv_ioda_pe *pe = tbl->data;
>> -	struct pnv_phb *phb = pe->phb;
>> +	long ret = pnv_tce_build(tbl, index, npages, uaddr, direction,
>> +			attrs);
>
> s/long/int


I better make them all long.

>
>>
>> -	if (phb->type == PNV_PHB_IODA1)
>> -		pnv_pci_ioda1_tce_invalidate(pe, tbl, startp, endp, rm);
>> -	else
>> -		pnv_pci_ioda2_tce_invalidate(pe, tbl, startp, endp, rm);
>> +	if (!ret && (tbl->it_type & TCE_PCI_SWINV_CREATE))
>> +		pnv_pci_ioda2_tce_invalidate(tbl, index, npages, false);
>> +
>> +	return ret;
>> +}
>> +
>> +static void pnv_ioda2_tce_free(struct iommu_table *tbl, long index,
>> +		long npages)
>> +{
>> +	pnv_tce_free(tbl, index, npages);
>> +
>> +	if (tbl->it_type & TCE_PCI_SWINV_FREE)
>> +		pnv_pci_ioda2_tce_invalidate(tbl, index, npages, false);
>> }
>>
>> static struct iommu_table_ops pnv_ioda2_iommu_ops = {
>> -	.set = pnv_tce_build,
>> -	.clear = pnv_tce_free,
>> +	.set = pnv_ioda2_tce_build,
>> +	.clear = pnv_ioda2_tce_free,
>> 	.get = pnv_tce_get,
>> };
>>
>> diff --git a/arch/powerpc/platforms/powernv/pci.c b/arch/powerpc/platforms/powernv/pci.c
>> index 4c3bbb1..84b4ea4 100644
>> --- a/arch/powerpc/platforms/powernv/pci.c
>> +++ b/arch/powerpc/platforms/powernv/pci.c
>> @@ -577,37 +577,28 @@ int pnv_tce_build(struct iommu_table *tbl, long index, long npages,
>> 		struct dma_attrs *attrs)
>> {
>> 	u64 proto_tce = iommu_direction_to_tce_perm(direction);
>> -	__be64 *tcep, *tces;
>> +	__be64 *tcep;
>> 	u64 rpn;
>>
>> -	tces = tcep = ((__be64 *)tbl->it_base) + index - tbl->it_offset;
>> +	tcep = ((__be64 *)tbl->it_base) + index - tbl->it_offset;
>> 	rpn = __pa(uaddr) >> tbl->it_page_shift;
>>
>> 	while (npages--)
>> 		*(tcep++) = cpu_to_be64(proto_tce |
>> 				(rpn++ << tbl->it_page_shift));
>>
>> -	/* Some implementations won't cache invalid TCEs and thus may not
>> -	 * need that flush. We'll probably turn it_type into a bit mask
>> -	 * of flags if that becomes the case
>> -	 */
>> -	if (tbl->it_type & TCE_PCI_SWINV_CREATE)
>> -		pnv_pci_ioda_tce_invalidate(tbl, tces, tcep - 1, false);
>>
>> 	return 0;
>> }
>>
>> void pnv_tce_free(struct iommu_table *tbl, long index, long npages)
>> {
>> -	__be64 *tcep, *tces;
>> +	__be64 *tcep;
>>
>> -	tces = tcep = ((__be64 *)tbl->it_base) + index - tbl->it_offset;
>> +	tcep = ((__be64 *)tbl->it_base) + index - tbl->it_offset;
>>
>> 	while (npages--)
>> 		*(tcep++) = cpu_to_be64(0);
>> -
>> -	if (tbl->it_type & TCE_PCI_SWINV_FREE)
>> -		pnv_pci_ioda_tce_invalidate(tbl, tces, tcep - 1, false);
>> }
>>
>> unsigned long pnv_tce_get(struct iommu_table *tbl, long index)
>> --
>> 2.4.0.rc3.8.gfb3e7d5
>>
>


-- 
Alexey

  reply	other threads:[~2015-05-14  3:19 UTC|newest]

Thread overview: 82+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-05-11 15:38 [PATCH kernel v10 00/34] powerpc/iommu/vfio: Enable Dynamic DMA windows Alexey Kardashevskiy
2015-05-11 15:38 ` [PATCH kernel v10 01/34] powerpc/eeh/ioda2: Use device::iommu_group to check IOMMU group Alexey Kardashevskiy
2015-05-12  1:51   ` Gavin Shan
2015-05-11 15:38 ` [PATCH kernel v10 02/34] powerpc/iommu/powernv: Get rid of set_iommu_table_base_and_group Alexey Kardashevskiy
2015-05-13  5:18   ` Gavin Shan
2015-05-13  7:26     ` Alexey Kardashevskiy
2015-05-11 15:38 ` [PATCH kernel v10 03/34] powerpc/powernv/ioda: Clean up IOMMU group registration Alexey Kardashevskiy
2015-05-13  5:21   ` Gavin Shan
2015-05-11 15:38 ` [PATCH kernel v10 04/34] powerpc/iommu: Put IOMMU group explicitly Alexey Kardashevskiy
2015-05-13  5:27   ` Gavin Shan
2015-05-11 15:38 ` [PATCH kernel v10 05/34] powerpc/iommu: Always release iommu_table in iommu_free_table() Alexey Kardashevskiy
2015-05-13  5:33   ` Gavin Shan
2015-05-13  6:30     ` Alexey Kardashevskiy
2015-05-13 12:51       ` Thomas Huth
2015-05-13 23:27         ` Gavin Shan
2015-05-14  2:34           ` Alexey Kardashevskiy
2015-05-14  2:53             ` Alex Williamson
2015-05-14  6:29               ` Alexey Kardashevskiy
2015-05-11 15:38 ` [PATCH kernel v10 06/34] vfio: powerpc/spapr: Move page pinning from arch code to VFIO IOMMU driver Alexey Kardashevskiy
2015-05-13  5:58   ` Gavin Shan
2015-05-13  6:32     ` Alexey Kardashevskiy
2015-05-11 15:38 ` [PATCH kernel v10 07/34] vfio: powerpc/spapr: Check that IOMMU page is fully contained by system page Alexey Kardashevskiy
2015-05-13  6:06   ` Gavin Shan
2015-05-11 15:38 ` [PATCH kernel v10 08/34] vfio: powerpc/spapr: Use it_page_size Alexey Kardashevskiy
2015-05-13  6:12   ` Gavin Shan
2015-05-11 15:38 ` [PATCH kernel v10 09/34] vfio: powerpc/spapr: Move locked_vm accounting to helpers Alexey Kardashevskiy
2015-05-13  6:18   ` Gavin Shan
2015-05-11 15:38 ` [PATCH kernel v10 10/34] vfio: powerpc/spapr: Disable DMA mappings on disabled container Alexey Kardashevskiy
2015-05-13  6:20   ` Gavin Shan
2015-05-11 15:39 ` [PATCH kernel v10 11/34] vfio: powerpc/spapr: Moving pinning/unpinning to helpers Alexey Kardashevskiy
2015-05-13  6:32   ` Gavin Shan
2015-05-13  7:30     ` Alexey Kardashevskiy
2015-05-11 15:39 ` [PATCH kernel v10 12/34] vfio: powerpc/spapr: Rework groups attaching Alexey Kardashevskiy
2015-05-13 23:35   ` Gavin Shan
2015-05-11 15:39 ` [PATCH kernel v10 13/34] powerpc/powernv: Do not set "read" flag if direction==DMA_NONE Alexey Kardashevskiy
2015-05-14  0:00   ` Gavin Shan
2015-05-14  2:51     ` Alexey Kardashevskiy
2015-05-11 15:39 ` [PATCH kernel v10 14/34] powerpc/iommu: Move tce_xxx callbacks from ppc_md to iommu_table Alexey Kardashevskiy
2015-05-14  0:23   ` Gavin Shan
2015-05-14  3:07     ` Alexey Kardashevskiy
2015-05-11 15:39 ` [PATCH kernel v10 15/34] powerpc/powernv/ioda/ioda2: Rework TCE invalidation in tce_build()/tce_free() Alexey Kardashevskiy
2015-05-14  0:48   ` Gavin Shan
2015-05-14  3:19     ` Alexey Kardashevskiy [this message]
2015-05-11 15:39 ` [PATCH kernel v10 16/34] powerpc/spapr: vfio: Replace iommu_table with iommu_table_group Alexey Kardashevskiy
2015-05-13 21:30   ` Alex Williamson
2015-05-14  1:21   ` Gavin Shan
2015-05-14  3:31     ` Alexey Kardashevskiy
2015-05-11 15:39 ` [PATCH kernel v10 17/34] powerpc/spapr: vfio: Switch from iommu_table to new iommu_table_group Alexey Kardashevskiy
2015-05-14  1:52   ` Gavin Shan
2015-05-11 15:39 ` [PATCH kernel v10 18/34] vfio: powerpc/spapr/iommu/powernv/ioda2: Rework IOMMU ownership control Alexey Kardashevskiy
2015-05-14  2:01   ` Gavin Shan
2015-05-11 15:39 ` [PATCH kernel v10 19/34] powerpc/iommu: Fix IOMMU ownership control functions Alexey Kardashevskiy
2015-05-14  3:36   ` Gavin Shan
2015-05-11 15:39 ` [PATCH kernel v10 20/34] powerpc/powernv/ioda2: Move TCE kill register address to PE Alexey Kardashevskiy
2015-05-14  2:10   ` Gavin Shan
2015-05-14  3:39     ` Alexey Kardashevskiy
2015-05-11 15:39 ` [PATCH kernel v10 21/34] powerpc/powernv/ioda2: Add TCE invalidation for all attached groups Alexey Kardashevskiy
2015-05-14  2:22   ` Gavin Shan
2015-05-14  3:50     ` Alexey Kardashevskiy
2015-05-11 15:39 ` [PATCH kernel v10 22/34] powerpc/powernv: Implement accessor to TCE entry Alexey Kardashevskiy
2015-05-14  2:34   ` Gavin Shan
2015-05-11 15:39 ` [PATCH kernel v10 23/34] powerpc/iommu/powernv: Release replaced TCE Alexey Kardashevskiy
2015-05-13 15:00   ` Thomas Huth
2015-05-14  3:53     ` Alexey Kardashevskiy
2015-05-15  8:09       ` Thomas Huth
2015-05-11 15:39 ` [PATCH kernel v10 24/34] powerpc/powernv/ioda2: Rework iommu_table creation Alexey Kardashevskiy
2015-05-14  4:14   ` Gavin Shan
2015-05-11 15:39 ` [PATCH kernel v10 25/34] powerpc/powernv/ioda2: Introduce helpers to allocate TCE pages Alexey Kardashevskiy
2015-05-14  4:31   ` Gavin Shan
2015-05-11 15:39 ` [PATCH kernel v10 26/34] powerpc/powernv/ioda2: Introduce pnv_pci_ioda2_set_window Alexey Kardashevskiy
2015-05-14  5:01   ` Gavin Shan
2015-05-11 15:39 ` [PATCH kernel v10 27/34] powerpc/powernv: Implement multilevel TCE tables Alexey Kardashevskiy
2015-05-11 15:39 ` [PATCH kernel v10 28/34] vfio: powerpc/spapr: powerpc/powernv/ioda: Define and implement DMA windows API Alexey Kardashevskiy
2015-05-13 21:30   ` Alex Williamson
2015-05-11 15:39 ` [PATCH kernel v10 29/34] powerpc/powernv/ioda2: Use new helpers to do proper cleanup on PE release Alexey Kardashevskiy
2015-05-11 15:39 ` [PATCH kernel v10 30/34] powerpc/iommu/ioda2: Add get_table_size() to calculate the size of future table Alexey Kardashevskiy
2015-05-11 15:39 ` [PATCH kernel v10 31/34] vfio: powerpc/spapr: powerpc/powernv/ioda2: Use DMA windows API in ownership control Alexey Kardashevskiy
2015-05-11 15:39 ` [PATCH kernel v10 32/34] powerpc/mmu: Add userspace-to-physical addresses translation cache Alexey Kardashevskiy
2015-05-11 15:39 ` [PATCH kernel v10 33/34] vfio: powerpc/spapr: Register memory and define IOMMU v2 Alexey Kardashevskiy
2015-05-13 21:30   ` Alex Williamson
2015-05-14  6:08     ` Alexey Kardashevskiy
2015-05-11 15:39 ` [PATCH kernel v10 34/34] vfio: powerpc/spapr: Support Dynamic DMA windows Alexey Kardashevskiy

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=5554144A.50701@ozlabs.ru \
    --to=aik@ozlabs.ru \
    --cc=alex.williamson@redhat.com \
    --cc=david@gibson.dropbear.id.au \
    --cc=gwshan@linux.vnet.ibm.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linuxppc-dev@lists.ozlabs.org \
    --cc=paulus@samba.org \
    --cc=weiyang@linux.vnet.ibm.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).