From: Yongseok Koh <yskoh@mellanox.com>
To: Andy Green <andy@warmcat.com>
Cc: Hemant Agrawal <hemant.agrawal@nxp.com>, dpdk stable <stable@dpdk.org>
Subject: [dpdk-stable] patch 'bus/dpaa: fix inconsistent struct alignment' has been queued to LTS release 17.11.5
Date: Thu, 3 Jan 2019 00:13:33 -0800 [thread overview]
Message-ID: <20190103081400.14191-10-yskoh@mellanox.com> (raw)
In-Reply-To: <20190103081400.14191-1-yskoh@mellanox.com>
Hi,
FYI, your patch has been queued to LTS release 17.11.5
Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 01/04/19. So please
shout if anyone has objections.
Also note that after the patch there's a diff of the upstream commit vs the patch applied
to the branch. If the code is different (ie: not only metadata diffs), due for example to
a change in context or macro names, please double check it.
Thanks.
Yongseok
---
>From 84f4ff2891bab8ddca99b73983f7c3108b532528 Mon Sep 17 00:00:00 2001
From: Andy Green <andy@warmcat.com>
Date: Mon, 14 May 2018 13:00:12 +0800
Subject: [PATCH] bus/dpaa: fix inconsistent struct alignment
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
[ upstream commit dd6f8d712e927397da27f2a3d90a759eeac4f694 ]
The actual descriptor for qm_mr_entry is 64-byte aligned.
But the original code plays a trick, and puts a u8 common
to the three descriptor subtypes in the union afterwards
outside their structure definitions.
Unfortunately since they compose a struct qm_fd with
alignment 8, this trick destroys the ability of the compiler
to understand what has happened, resulting in this kind of
problem:
drivers/bus/dpaa/include/fsl_qman.h:354:3: error:
alignment 1 of ‘struct <anonymous>’ is less than 8 [-Werror=packed-not-aligned]
} __packed dcern;
on gcc 8 / Fedora 28 out of the box.
This patch moves the u8 verb into the structure definitions
composed into the union, so the alignment of the parent struct
containing the alignment 8 object can also be seen to be
alignment 8 by the compiler. Uses of .verb are fixed up to use
.ern.verb (the same offset of +0 inside all the structs in
the union).
The final struct layout should be unchanged.
Fixes: c47ff048b99a ("bus/dpaa: add QMAN driver core routines")
Fixes: f6fadc3e6310 ("bus/dpaa: add QMAN interface driver")
Signed-off-by: Andy Green <andy@warmcat.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Tested-by: Hemant Agrawal <hemant.agrawal@nxp.com>
---
drivers/bus/dpaa/base/qbman/qman.c | 14 +++++++-------
drivers/bus/dpaa/include/fsl_qman.h | 24 +++++++++++++-----------
2 files changed, 20 insertions(+), 18 deletions(-)
diff --git a/drivers/bus/dpaa/base/qbman/qman.c b/drivers/bus/dpaa/base/qbman/qman.c
index b8511103c..84f0cb0be 100644
--- a/drivers/bus/dpaa/base/qbman/qman.c
+++ b/drivers/bus/dpaa/base/qbman/qman.c
@@ -344,9 +344,9 @@ loop:
if (!msg)
return 0;
}
- if ((msg->verb & QM_MR_VERB_TYPE_MASK) != QM_MR_VERB_FQRNI) {
+ if ((msg->ern.verb & QM_MR_VERB_TYPE_MASK) != QM_MR_VERB_FQRNI) {
/* We aren't draining anything but FQRNIs */
- pr_err("Found verb 0x%x in MR\n", msg->verb);
+ pr_err("Found verb 0x%x in MR\n", msg->ern.verb);
return -1;
}
qm_mr_next(p);
@@ -513,7 +513,7 @@ static inline void qm_mr_pvb_update(struct qm_portal *portal)
/* when accessing 'verb', use __raw_readb() to ensure that compiler
* inlining doesn't try to optimise out "excess reads".
*/
- if ((__raw_readb(&res->verb) & QM_MR_VERB_VBIT) == mr->vbit) {
+ if ((__raw_readb(&res->ern.verb) & QM_MR_VERB_VBIT) == mr->vbit) {
mr->pi = (mr->pi + 1) & (QM_MR_SIZE - 1);
if (!mr->pi)
mr->vbit ^= QM_MR_VERB_VBIT;
@@ -810,7 +810,7 @@ mr_loop:
goto mr_done;
swapped_msg = *msg;
hw_fd_to_cpu(&swapped_msg.ern.fd);
- verb = msg->verb & QM_MR_VERB_TYPE_MASK;
+ verb = msg->ern.verb & QM_MR_VERB_TYPE_MASK;
/* The message is a software ERN iff the 0x20 bit is set */
if (verb & 0x20) {
switch (verb) {
@@ -1504,7 +1504,7 @@ int qman_retire_fq(struct qman_fq *fq, u32 *flags)
*/
struct qm_mr_entry msg;
- msg.verb = QM_MR_VERB_FQRNI;
+ msg.ern.verb = QM_MR_VERB_FQRNI;
msg.fq.fqs = mcr->alterfq.fqs;
msg.fq.fqid = fq->fqid;
#ifdef CONFIG_FSL_QMAN_FQ_LOOKUP
@@ -2389,7 +2389,7 @@ int qman_shutdown_fq(u32 fqid)
qm_mr_pvb_update(low_p);
msg = qm_mr_current(low_p);
while (msg) {
- if ((msg->verb &
+ if ((msg->ern.verb &
QM_MR_VERB_TYPE_MASK)
== QM_MR_VERB_FQRN)
found_fqrn = 1;
@@ -2457,7 +2457,7 @@ int qman_shutdown_fq(u32 fqid)
qm_mr_pvb_update(low_p);
msg = qm_mr_current(low_p);
while (msg) {
- if ((msg->verb & QM_MR_VERB_TYPE_MASK) ==
+ if ((msg->ern.verb & QM_MR_VERB_TYPE_MASK) ==
QM_MR_VERB_FQRL)
orl_empty = 1;
qm_mr_next(low_p);
diff --git a/drivers/bus/dpaa/include/fsl_qman.h b/drivers/bus/dpaa/include/fsl_qman.h
index 72556dc12..5d00bc73a 100644
--- a/drivers/bus/dpaa/include/fsl_qman.h
+++ b/drivers/bus/dpaa/include/fsl_qman.h
@@ -316,20 +316,20 @@ static inline dma_addr_t qm_sg_addr(const struct qm_sg_entry *sg)
} while (0)
/* See 1.5.8.1: "Enqueue Command" */
-struct qm_eqcr_entry {
+struct __rte_aligned(8) qm_eqcr_entry {
u8 __dont_write_directly__verb;
u8 dca;
u16 seqnum;
u32 orp; /* 24-bit */
u32 fqid; /* 24-bit */
u32 tag;
- struct qm_fd fd;
+ struct qm_fd fd; /* this has alignment 8 */
u8 __reserved3[32];
} __packed;
/* "Frame Dequeue Response" */
-struct qm_dqrr_entry {
+struct __rte_aligned(8) qm_dqrr_entry {
u8 verb;
u8 stat;
u16 seqnum; /* 15-bit */
@@ -337,7 +337,7 @@ struct qm_dqrr_entry {
u8 __reserved2[3];
u32 fqid; /* 24-bit */
u32 contextB;
- struct qm_fd fd;
+ struct qm_fd fd; /* this has alignment 8 */
u8 __reserved4[32];
};
@@ -355,18 +355,19 @@ struct qm_dqrr_entry {
/* "ERN Message Response" */
/* "FQ State Change Notification" */
struct qm_mr_entry {
- u8 verb;
union {
struct {
+ u8 verb;
u8 dca;
u16 seqnum;
u8 rc; /* Rejection Code */
u32 orp:24;
u32 fqid; /* 24-bit */
u32 tag;
- struct qm_fd fd;
- } __packed ern;
+ struct qm_fd fd; /* this has alignment 8 */
+ } __packed __rte_aligned(8) ern;
struct {
+ u8 verb;
#if __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
u8 colour:2; /* See QM_MR_DCERN_COLOUR_* */
u8 __reserved1:4;
@@ -381,18 +382,19 @@ struct qm_mr_entry {
u32 __reserved3:24;
u32 fqid; /* 24-bit */
u32 tag;
- struct qm_fd fd;
- } __packed dcern;
+ struct qm_fd fd; /* this has alignment 8 */
+ } __packed __rte_aligned(8) dcern;
struct {
+ u8 verb;
u8 fqs; /* Frame Queue Status */
u8 __reserved1[6];
u32 fqid; /* 24-bit */
u32 contextB;
u8 __reserved2[16];
- } __packed fq; /* FQRN/FQRNI/FQRL/FQPN */
+ } __packed __rte_aligned(8) fq; /* FQRN/FQRNI/FQRL/FQPN */
};
u8 __reserved2[32];
-} __packed;
+} __packed __rte_aligned(8);
#define QM_MR_VERB_VBIT 0x80
/*
* ERNs originating from direct-connect portals ("dcern") use 0x20 as a verb
--
2.11.0
---
Diff of the applied patch vs upstream commit (please double-check if non-empty:
---
--- - 2019-01-02 23:59:12.646228426 -0800
+++ 0010-bus-dpaa-fix-inconsistent-struct-alignment.patch 2019-01-02 23:59:12.036815000 -0800
@@ -1,4 +1,4 @@
-From dd6f8d712e927397da27f2a3d90a759eeac4f694 Mon Sep 17 00:00:00 2001
+From 84f4ff2891bab8ddca99b73983f7c3108b532528 Mon Sep 17 00:00:00 2001
From: Andy Green <andy@warmcat.com>
Date: Mon, 14 May 2018 13:00:12 +0800
Subject: [PATCH] bus/dpaa: fix inconsistent struct alignment
@@ -6,6 +6,8 @@
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
+[ upstream commit dd6f8d712e927397da27f2a3d90a759eeac4f694 ]
+
The actual descriptor for qm_mr_entry is 64-byte aligned.
But the original code plays a trick, and puts a u8 common
@@ -34,7 +36,6 @@
Fixes: c47ff048b99a ("bus/dpaa: add QMAN driver core routines")
Fixes: f6fadc3e6310 ("bus/dpaa: add QMAN interface driver")
-Cc: stable@dpdk.org
Signed-off-by: Andy Green <andy@warmcat.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
@@ -45,10 +46,10 @@
2 files changed, 20 insertions(+), 18 deletions(-)
diff --git a/drivers/bus/dpaa/base/qbman/qman.c b/drivers/bus/dpaa/base/qbman/qman.c
-index 2810fdd26..27d98cc10 100644
+index b8511103c..84f0cb0be 100644
--- a/drivers/bus/dpaa/base/qbman/qman.c
+++ b/drivers/bus/dpaa/base/qbman/qman.c
-@@ -314,9 +314,9 @@ loop:
+@@ -344,9 +344,9 @@ loop:
if (!msg)
return 0;
}
@@ -60,7 +61,7 @@
return -1;
}
qm_mr_next(p);
-@@ -483,7 +483,7 @@ static inline void qm_mr_pvb_update(struct qm_portal *portal)
+@@ -513,7 +513,7 @@ static inline void qm_mr_pvb_update(struct qm_portal *portal)
/* when accessing 'verb', use __raw_readb() to ensure that compiler
* inlining doesn't try to optimise out "excess reads".
*/
@@ -69,7 +70,7 @@
mr->pi = (mr->pi + 1) & (QM_MR_SIZE - 1);
if (!mr->pi)
mr->vbit ^= QM_MR_VERB_VBIT;
-@@ -832,7 +832,7 @@ mr_loop:
+@@ -810,7 +810,7 @@ mr_loop:
goto mr_done;
swapped_msg = *msg;
hw_fd_to_cpu(&swapped_msg.ern.fd);
@@ -78,7 +79,7 @@
/* The message is a software ERN iff the 0x20 bit is set */
if (verb & 0x20) {
switch (verb) {
-@@ -1666,7 +1666,7 @@ int qman_retire_fq(struct qman_fq *fq, u32 *flags)
+@@ -1504,7 +1504,7 @@ int qman_retire_fq(struct qman_fq *fq, u32 *flags)
*/
struct qm_mr_entry msg;
@@ -87,7 +88,7 @@
msg.fq.fqs = mcr->alterfq.fqs;
msg.fq.fqid = fq->fqid;
#ifdef CONFIG_FSL_QMAN_FQ_LOOKUP
-@@ -2643,7 +2643,7 @@ int qman_shutdown_fq(u32 fqid)
+@@ -2389,7 +2389,7 @@ int qman_shutdown_fq(u32 fqid)
qm_mr_pvb_update(low_p);
msg = qm_mr_current(low_p);
while (msg) {
@@ -96,7 +97,7 @@
QM_MR_VERB_TYPE_MASK)
== QM_MR_VERB_FQRN)
found_fqrn = 1;
-@@ -2711,7 +2711,7 @@ int qman_shutdown_fq(u32 fqid)
+@@ -2457,7 +2457,7 @@ int qman_shutdown_fq(u32 fqid)
qm_mr_pvb_update(low_p);
msg = qm_mr_current(low_p);
while (msg) {
@@ -106,10 +107,10 @@
orl_empty = 1;
qm_mr_next(low_p);
diff --git a/drivers/bus/dpaa/include/fsl_qman.h b/drivers/bus/dpaa/include/fsl_qman.h
-index e9793f30d..e4ad7ae48 100644
+index 72556dc12..5d00bc73a 100644
--- a/drivers/bus/dpaa/include/fsl_qman.h
+++ b/drivers/bus/dpaa/include/fsl_qman.h
-@@ -284,20 +284,20 @@ static inline dma_addr_t qm_sg_addr(const struct qm_sg_entry *sg)
+@@ -316,20 +316,20 @@ static inline dma_addr_t qm_sg_addr(const struct qm_sg_entry *sg)
} while (0)
/* See 1.5.8.1: "Enqueue Command" */
@@ -133,7 +134,7 @@
u8 verb;
u8 stat;
u16 seqnum; /* 15-bit */
-@@ -305,7 +305,7 @@ struct qm_dqrr_entry {
+@@ -337,7 +337,7 @@ struct qm_dqrr_entry {
u8 __reserved2[3];
u32 fqid; /* 24-bit */
u32 contextB;
@@ -142,7 +143,7 @@
u8 __reserved4[32];
};
-@@ -323,18 +323,19 @@ struct qm_dqrr_entry {
+@@ -355,18 +355,19 @@ struct qm_dqrr_entry {
/* "ERN Message Response" */
/* "FQ State Change Notification" */
struct qm_mr_entry {
@@ -165,7 +166,7 @@
#if __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
u8 colour:2; /* See QM_MR_DCERN_COLOUR_* */
u8 __reserved1:4;
-@@ -349,18 +350,19 @@ struct qm_mr_entry {
+@@ -381,18 +382,19 @@ struct qm_mr_entry {
u32 __reserved3:24;
u32 fqid; /* 24-bit */
u32 tag;
next prev parent reply other threads:[~2019-01-03 8:14 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-01-03 8:13 [dpdk-stable] patch 'eal: explicit cast of builtin for bsf32' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'eal: explicit cast of core id when getting index' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'eal: declare trace buffer at top of own block' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'spinlock/x86: move stack declaration before code' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'net: move stack variable at top of VLAN strip function' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'ethdev: explicit cast of buffered Tx number' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'hash: move stack declaration at top of CRC32c function' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'hash: explicit casts for truncation in CRC32c' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'bus/pci: replace strncpy by strlcpy' " Yongseok Koh
2019-01-03 8:13 ` Yongseok Koh [this message]
2019-01-03 8:13 ` [dpdk-stable] patch 'net/nfp: fix memcpy out of source range' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'net/qede: replace strncpy by strlcpy' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'net/qede: fix strncpy' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'net/sfc: make sure that stats name is nul-terminated' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'app/procinfo: fix sprintf overrun' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'devtools: provide more generic grep in git check' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'eal: explicit cast of strlcpy return' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'eal: fix casts in random functions' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'eal: explicit cast in constant byte swap' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'ring: remove useless variables' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'ring: remove signed type flip-flopping' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'mbuf: fix reference counter integer promotion' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'mbuf: explicit casts of reference counter' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'mbuf: explicit cast of headroom on reset' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'mbuf: explicit cast of size on detach' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'net: explicit cast of multicast bit clearing' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'net: explicit cast of IP checksum to 16-bit' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'net: explicit cast of protocol in IPv6 checksum' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'ethdev: explicit cast of queue count return' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'eal/x86: fix type of variable in memcpy function' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'eal: explicit cast in rwlock functions' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'net: explicit cast in L4 checksum' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'mbuf: fix type of private size in detach' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'mbuf: fix type of variables in linearize function' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'mbuf: avoid implicit demotion in 64-bit arithmetic' " Yongseok Koh
2019-01-03 8:13 ` [dpdk-stable] patch 'mbuf: avoid integer promotion in prepend/adj/chain' " Yongseok Koh
2019-01-03 8:14 ` [dpdk-stable] patch 'ethdev: fix type and scope of variables in Rx burst' " Yongseok Koh
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=20190103081400.14191-10-yskoh@mellanox.com \
--to=yskoh@mellanox.com \
--cc=andy@warmcat.com \
--cc=hemant.agrawal@nxp.com \
--cc=stable@dpdk.org \
/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).