[dpdk-dev] [PATCH 0/4] fix mempool creation with Xen Dom0

[dpdk-dev] [PATCH 0/4] fix mempool creation with Xen Dom0

[Olivier Matz]

Since the recent mempool rework [1], Xen Dom0 is broken.
This series aims at fixing it. I think it should be integrated
in 16.07.

As I don't have a full testing platform, any help in validating
this patchset would be appreciated.

[1] http://www.dpdk.org/ml/archives/dev/2016-May/039229.html

Olivier Matz (4):
  eal: fix typo in Xen Dom0 specific code
  mbuf: set errno on pool creation error
  eal: fix retrieval of phys address with Xen Dom0
  mempool: fix creation with Xen Dom0

 lib/librte_eal/linuxapp/eal/eal_memory.c     | 23 +++++++++++++++++++++++
 lib/librte_eal/linuxapp/eal/eal_xen_memory.c |  4 ++--
 lib/librte_mbuf/rte_mbuf.c                   |  5 ++++-
 lib/librte_mempool/rte_mempool.c             |  6 +++++-
 4 files changed, 34 insertions(+), 4 deletions(-)

-- 
2.8.1

[dpdk-dev] [PATCH 1/4] eal: fix typo in Xen Dom0 specific code

[Olivier Matz]

Fix the compilation with CONFIG_RTE_LIBRTE_XEN_DOM0=y, by correcting the
typo in variable names.

Fixes: 8dab48370129 ("xen: return machine address without knowing memseg id")

Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
---
 lib/librte_eal/linuxapp/eal/eal_xen_memory.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/librte_eal/linuxapp/eal/eal_xen_memory.c b/lib/librte_eal/linuxapp/eal/eal_xen_memory.c
index 0b612bb..bddbdb0 100644
--- a/lib/librte_eal/linuxapp/eal/eal_xen_memory.c
+++ b/lib/librte_eal/linuxapp/eal/eal_xen_memory.c
@@ -167,8 +167,8 @@ rte_xen_mem_phy2mch(int32_t memseg_id, const phys_addr_t phy_addr)
 	if (memseg_id == -1) {
 		for (i = 0; i < RTE_MAX_MEMSEG; i++) {
 			if ((phy_addr >= memseg[i].phys_addr) &&
-					(phys_addr < memseg[i].phys_addr +
-						memseg[i].size)) {
+					(phy_addr < memseg[i].phys_addr +
+						memseg[i].len)) {
 				memseg_id = i;
 				break;
 			}
-- 
2.8.1

[dpdk-dev] [PATCH 2/4] mbuf: set errno on pool creation error

[Olivier Matz]

In rte_pktmbuf_pool_create(), the rte_errno variable was not always
set on failure.

Fixes: 152ca517900b ("mbuf: use default mempool handler from config")

Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
---
 lib/librte_mbuf/rte_mbuf.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/lib/librte_mbuf/rte_mbuf.c b/lib/librte_mbuf/rte_mbuf.c
index 601e528..4846b89 100644
--- a/lib/librte_mbuf/rte_mbuf.c
+++ b/lib/librte_mbuf/rte_mbuf.c
@@ -156,6 +156,7 @@ rte_pktmbuf_pool_create(const char *name, unsigned n,
 	struct rte_mempool *mp;
 	struct rte_pktmbuf_pool_private mbp_priv;
 	unsigned elt_size;
+	int ret;
 
 	if (RTE_ALIGN(priv_size, RTE_MBUF_PRIV_ALIGN) != priv_size) {
 		RTE_LOG(ERR, MBUF, "mbuf priv_size=%u is not aligned\n",
@@ -181,8 +182,10 @@ rte_pktmbuf_pool_create(const char *name, unsigned n,
 	}
 	rte_pktmbuf_pool_init(mp, &mbp_priv);
 
-	if (rte_mempool_populate_default(mp) < 0) {
+	ret = rte_mempool_populate_default(mp);
+	if (ret < 0) {
 		rte_mempool_free(mp);
+		rte_errno = -ret;
 		return NULL;
 	}
 
-- 
2.8.1

[dpdk-dev] [PATCH 3/4] eal: fix retrieval of phys address with Xen Dom0

[Olivier Matz]

When using Xen Dom0, it looks that /proc/self/pagemap returns 0.
This breaks the creation of mbufs pool.

We can workaround this in rte_mem_virt2phy() by browsing the dpdk memory
segments. This only works for dpdk memory, but it's enough to fix the
mempool creation.

Fixes: c042ba20674a ("mempool: rework support of Xen dom0")
Fixes: 3097de6e6bfb ("mem: get physical address of any pointer")

Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
---
 lib/librte_eal/linuxapp/eal/eal_memory.c | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

diff --git a/lib/librte_eal/linuxapp/eal/eal_memory.c b/lib/librte_eal/linuxapp/eal/eal_memory.c
index b663244..42a29fa 100644
--- a/lib/librte_eal/linuxapp/eal/eal_memory.c
+++ b/lib/librte_eal/linuxapp/eal/eal_memory.c
@@ -164,6 +164,29 @@ rte_mem_virt2phy(const void *virtaddr)
 	int page_size;
 	off_t offset;
 
+	/* when using dom0, /proc/self/pagemap always returns 0, check in
+	 * dpdk memory by browsing the memsegs */
+	if (rte_xen_dom0_supported()) {
+		struct rte_mem_config *mcfg;
+		struct rte_memseg *memseg;
+		unsigned i;
+
+		mcfg = rte_eal_get_configuration()->mem_config;
+		for (i = 0; i < RTE_MAX_MEMSEG; i++) {
+			memseg = &mcfg->memseg[i];
+			if (memseg->addr == NULL)
+				break;
+			if (virtaddr > memseg->addr &&
+					virtaddr < RTE_PTR_ADD(memseg->addr,
+						memseg->len)) {
+				return memseg->phys_addr +
+					RTE_PTR_DIFF(virtaddr, memseg->addr);
+			}
+		}
+
+		return RTE_BAD_PHYS_ADDR;
+	}
+
 	/* Cannot parse /proc/self/pagemap, no need to log errors everywhere */
 	if (!proc_pagemap_readable)
 		return RTE_BAD_PHYS_ADDR;
-- 
2.8.1

[dpdk-dev] [PATCH 4/4] mempool: fix creation with Xen Dom0

[Olivier Matz]

Restore the use of 2M hugepages when using Xen Dom0 that was
dropped during mempool rework.

Fixes: c042ba20674a ("mempool: rework support of Xen dom0")

Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
---
 lib/librte_mempool/rte_mempool.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/lib/librte_mempool/rte_mempool.c b/lib/librte_mempool/rte_mempool.c
index d78d02b..6ec0906 100644
--- a/lib/librte_mempool/rte_mempool.c
+++ b/lib/librte_mempool/rte_mempool.c
@@ -524,7 +524,11 @@ rte_mempool_populate_default(struct rte_mempool *mp)
 	if (mp->nb_mem_chunks != 0)
 		return -EEXIST;
 
-	if (rte_eal_has_hugepages()) {
+	if (rte_xen_dom0_supported()) {
+		pg_sz = RTE_PGSIZE_2M;
+		pg_shift = rte_bsf32(pg_sz);
+		align = pg_sz;
+	} else if (rte_eal_has_hugepages()) {
 		pg_shift = 0; /* not needed, zone is physically contiguous */
 		pg_sz = 0;
 		align = RTE_CACHE_LINE_SIZE;
-- 
2.8.1

Re: [dpdk-dev] [PATCH 0/4] fix mempool creation with Xen Dom0

[Olivier Matz]

On 07/11/2016 12:20 PM, Olivier Matz wrote:
> Since the recent mempool rework [1], Xen Dom0 is broken.
> This series aims at fixing it. I think it should be integrated
> in 16.07.
> 
> As I don't have a full testing platform, any help in validating
> this patchset would be appreciated.
> 
> [1] http://www.dpdk.org/ml/archives/dev/2016-May/039229.html

I forgot to mention:

Reported-by: Xu, HuilongX <huilongx.xu@intel.com>

Thanks

Re: [dpdk-dev] [PATCH 0/4] fix mempool creation with Xen Dom0

[Thomas Monjalon]

2016-07-11 12:20, Olivier Matz:
> Since the recent mempool rework [1], Xen Dom0 is broken.
> This series aims at fixing it. I think it should be integrated
> in 16.07.
> 
> As I don't have a full testing platform, any help in validating
> this patchset would be appreciated.
> 
> [1] http://www.dpdk.org/ml/archives/dev/2016-May/039229.html
> 
> Olivier Matz (4):
>   eal: fix typo in Xen Dom0 specific code
>   mbuf: set errno on pool creation error
>   eal: fix retrieval of phys address with Xen Dom0
>   mempool: fix creation with Xen Dom0

Applied with an additional fix:

    xen: fix build as shared library
    
    When building as shared library, the compiler complains for
        undefined reference to `rte_xen_mem_phy2mch'
    
    The symbol rte_xen_mem_phy2mch was introduced in DPDK 2.2
    and has been called in mempool recently via rte_mem_phy2mch.
    
    Fixes: c042ba20674a ("mempool: rework support of Xen dom0")

Re: [dpdk-dev] [PATCH 0/4] fix mempool creation with Xen Dom0

[Chen, WeichunX]

Add huilong

-----Original Message-----
From: Thomas Monjalon [mailto:thomas.monjalon@6wind.com] 
Sent: Tuesday, July 12, 2016 1:10 AM
To: Olivier Matz <olivier.matz@6wind.com>
Cc: dev@dpdk.org; Xu, HuilongX <huilongx.xu@intel.com>; Cao, Waterman <waterman.cao@intel.com>; Liu, Yuanhan <yuanhan.liu@intel.com>; Chen, WeichunX <weichunx.chen@intel.com>; Liu, Yu Y <yu.y.liu@intel.com>
Subject: Re: [PATCH 0/4] fix mempool creation with Xen Dom0

2016-07-11 12:20, Olivier Matz:
> Since the recent mempool rework [1], Xen Dom0 is broken.
> This series aims at fixing it. I think it should be integrated in 
> 16.07.
> 
> As I don't have a full testing platform, any help in validating this 
> patchset would be appreciated.
> 
> [1] http://www.dpdk.org/ml/archives/dev/2016-May/039229.html
> 
> Olivier Matz (4):
>   eal: fix typo in Xen Dom0 specific code
>   mbuf: set errno on pool creation error
>   eal: fix retrieval of phys address with Xen Dom0
>   mempool: fix creation with Xen Dom0

Applied with an additional fix:

    xen: fix build as shared library
    
    When building as shared library, the compiler complains for
        undefined reference to `rte_xen_mem_phy2mch'
    
    The symbol rte_xen_mem_phy2mch was introduced in DPDK 2.2
    and has been called in mempool recently via rte_mem_phy2mch.
    
    Fixes: c042ba20674a ("mempool: rework support of Xen dom0")