public inbox for linux-s390@vger.kernel.org
 help / color / mirror / Atom feed
From: Pierre Morel <pmorel@linux.ibm.com>
To: Janosch Frank <frankja@linux.ibm.com>, kvm@vger.kernel.org
Cc: linux-s390@vger.kernel.org, david@redhat.com, thuth@redhat.com,
	cohuck@redhat.com, imbrenda@linux.ibm.com, drjones@redhat.com,
	pbonzini@redhat.com
Subject: Re: [kvm-unit-tests PATCH v5 2/3] s390x: define UV compatible I/O allocation
Date: Fri, 22 Jan 2021 16:38:31 +0100	[thread overview]
Message-ID: <a692e401-874a-5d20-7f78-d0024ca57583@linux.ibm.com> (raw)
In-Reply-To: <fcb7e165-d1a6-9d0b-a8e8-e91f39e2e0f7@linux.ibm.com>



On 1/22/21 2:57 PM, Janosch Frank wrote:
> On 1/22/21 2:27 PM, Pierre Morel wrote:
>> To centralize the memory allocation for I/O we define
>> the alloc_io_mem/free_io_mem functions which share the I/O
>> memory with the host in case the guest runs with
>> protected virtualization.
>>
>> These functions allocate on a page integral granularity to
>> ensure a dedicated sharing of the allocated objects.
>>
>> Signed-off-by: Pierre Morel <pmorel@linux.ibm.com>
> 
> Acked-by: Janosch Frank <frankja@de.ibm.com>
> 
>> ---
>>   lib/s390x/malloc_io.c | 71 +++++++++++++++++++++++++++++++++++++++++++
>>   lib/s390x/malloc_io.h | 45 +++++++++++++++++++++++++++
>>   s390x/Makefile        |  1 +
>>   3 files changed, 117 insertions(+)
>>   create mode 100644 lib/s390x/malloc_io.c
>>   create mode 100644 lib/s390x/malloc_io.h
>>
>> diff --git a/lib/s390x/malloc_io.c b/lib/s390x/malloc_io.c
>> new file mode 100644
>> index 0000000..b01222e
>> --- /dev/null
>> +++ b/lib/s390x/malloc_io.c
>> @@ -0,0 +1,71 @@
>> +/* SPDX-License-Identifier: GPL-2.0-only */
>> +/*
>> + * I/O page allocation
>> + *
>> + * Copyright (c) 2021 IBM Corp
>> + *
>> + * Authors:
>> + *  Pierre Morel <pmorel@linux.ibm.com>
>> + *
>> + * Using this interface provide host access to the allocated pages in
>> + * case the guest is a secure guest.
> 
> s/secure/protected/

OK

> 
>> + * This is needed for I/O buffers.
>> + *
>> + */
>> +#include <libcflat.h>
>> +#include <asm/page.h>
>> +#include <asm/uv.h>
>> +#include <malloc_io.h>
>> +#include <alloc_page.h>
>> +#include <asm/facility.h>
>> +#include <bitops.h>
>> +
>> +static int share_pages(void *p, int count)
>> +{
>> +	int i = 0;
>> +
>> +	for (i = 0; i < count; i++, p += PAGE_SIZE)
>> +		if (uv_set_shared((unsigned long)p))
>> +			break;
>> +	return i;
>> +}
>> +
>> +static void unshare_pages(void *p, int count)
>> +{
>> +	int i;
>> +
>> +	for (i = count; i > 0; i--, p += PAGE_SIZE)
>> +		uv_remove_shared((unsigned long)p);
>> +}
>> +
>> +void *alloc_io_mem(int size, int flags)
>> +{
>> +	int order = get_order(size >> PAGE_SHIFT);
>> +	void *p;
>> +	int n;
>> +
>> +	assert(size);
>> +
>> +	p = alloc_pages_flags(order, AREA_DMA31 | flags);
> 
> Pardon my lack of IO knowledge but do we also need to restrict the data
> to below 2g or only the IO control structures?
> 
> Not that it currently matters much for a unit test with 256mb of memory...

it depends, if you know how work most scatter-gather it is easy:

INSTR-> ORB -> CCW -> DATA

first level metadata: yes
- Intruction point with 31bit to ORB
- ORB point with 31b to CCW
- CCW point with 31b to data

first level of data, the one pointed by the CCW: yes

This is all what is used in this patch.

To be developed in the future:

INSTR-> ORB -> CCW -> [M]IDA -> DATA

second level metadata:
they are pointed to by CCW...
- IDA : yes
- MIDA: yes

second level data (indirect buffering)
- IDA format 1: 31b address: yes
- IDA format 2: 64b address: no
- MIDA        : 64b address: no

Or to sum up only data accessed with indirect addressing by IDA format 2 
and MIDA can be anywhere.

> 
>> +	if (!p || !test_facility(158))
>> +		return p;
>> +
>> +	n = share_pages(p, 1 << order);
>> +	if (n == 1 << order)
>> +		return p;
>> +
>> +	unshare_pages(p, n);
>> +	free_pages(p);
>> +	return NULL;
>> +}
>> +
>> +void free_io_mem(void *p, int size)
>> +{
>> +	int order = get_order(size >> PAGE_SHIFT);
>> +
>> +	assert(IS_ALIGNED((uintptr_t)p, PAGE_SIZE));
>> +
>> +	if (test_facility(158))
>> +		unshare_pages(p, 1 << order);
>> +	free_pages(p);
>> +}
>> diff --git a/lib/s390x/malloc_io.h b/lib/s390x/malloc_io.h
>> new file mode 100644
>> index 0000000..cc5fad7
>> --- /dev/null
>> +++ b/lib/s390x/malloc_io.h
>> @@ -0,0 +1,45 @@
>> +/* SPDX-License-Identifier: GPL-2.0-only */
>> +/*
>> + * I/O allocations
>> + *
>> + * Copyright (c) 2021 IBM Corp
>> + *
>> + * Authors:
>> + *  Pierre Morel <pmorel@linux.ibm.com>
>> + *
>> + */
>> +#ifndef _S390X_MALLOC_IO_H_
>> +#define _S390X_MALLOC_IO_H_
>> +
>> +/*
>> + * Allocates a page aligned page bound range of contiguous real or
>> + * absolute memory in the DMA31 region large enough to contain size
>> + * bytes.
>> + * If Protected Virtualisation facility is present, shares the pages
> 
> s/Virtualisation/Virtualization/
> 
> But I can fix this and the nitpick above up when picking.

:) Thomas made me change this in the last round.

As far as I know "z" is used in the uS ans "s" in England (and France).

regards,
Pierre

-- 
Pierre Morel
IBM Lab Boeblingen

  reply	other threads:[~2021-01-22 15:39 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-01-22 13:27 [kvm-unit-tests PATCH v5 0/3] s390x: css: pv: css test adaptation for PV Pierre Morel
2021-01-22 13:27 ` [kvm-unit-tests PATCH v5 1/3] s390x: pv: implement routine to share/unshare memory Pierre Morel
2021-01-22 13:33   ` Janosch Frank
2021-01-22 15:44     ` Pierre Morel
2021-01-22 13:27 ` [kvm-unit-tests PATCH v5 2/3] s390x: define UV compatible I/O allocation Pierre Morel
2021-01-22 13:57   ` Janosch Frank
2021-01-22 15:38     ` Pierre Morel [this message]
2021-01-22 14:12   ` Thomas Huth
2021-01-22 15:43     ` Pierre Morel
2021-01-27 10:47   ` Cornelia Huck
2021-01-27 12:40     ` Pierre Morel
2021-01-22 13:27 ` [kvm-unit-tests PATCH v5 3/3] s390x: css: pv: css test adaptation for PV Pierre Morel
2021-01-27 12:37 ` [kvm-unit-tests PATCH v5 0/3] " Janosch Frank
2021-01-27 12:40   ` Pierre Morel

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=a692e401-874a-5d20-7f78-d0024ca57583@linux.ibm.com \
    --to=pmorel@linux.ibm.com \
    --cc=cohuck@redhat.com \
    --cc=david@redhat.com \
    --cc=drjones@redhat.com \
    --cc=frankja@linux.ibm.com \
    --cc=imbrenda@linux.ibm.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-s390@vger.kernel.org \
    --cc=pbonzini@redhat.com \
    --cc=thuth@redhat.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