From: Wen Gu <guwen@linux.alibaba.com>
To: Alexandra Winter <wintera@linux.ibm.com>,
wenjia@linux.ibm.com, hca@linux.ibm.com, gor@linux.ibm.com,
agordeev@linux.ibm.com, davem@davemloft.net, edumazet@google.com,
kuba@kernel.org, pabeni@redhat.com, kgraul@linux.ibm.com,
jaka@linux.ibm.com
Cc: borntraeger@linux.ibm.com, svens@linux.ibm.com,
alibuda@linux.alibaba.com, tonylu@linux.alibaba.com,
raspl@linux.ibm.com, schnelle@linux.ibm.com,
guangguan.wang@linux.alibaba.com, linux-s390@vger.kernel.org,
netdev@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH net-next v8 03/10] net/smc: unify the structs of accept or confirm message for v1 and v2
Date: Wed, 20 Dec 2023 20:16:50 +0800 [thread overview]
Message-ID: <38f06cfb-2d68-2b10-f82b-62a44c25b8f8@linux.alibaba.com> (raw)
In-Reply-To: <ab835e29-ad4a-4377-b80a-8ef6bb35ef7b@linux.ibm.com>
On 2023/12/20 19:37, Alexandra Winter wrote:
>
>
> On 19.12.23 15:26, Wen Gu wrote:
>> struct smc_clc_msg_accept_confirm { /* clc accept / confirm message */
>> - struct smc_clc_msg_hdr hdr;
>> - union {
>> - struct smcr_clc_msg_accept_confirm r0; /* SMC-R */
>> - struct { /* SMC-D */
>> - struct smcd_clc_msg_accept_confirm_common d0;
>> - u32 reserved5[3];
>> - };
>> - };
>> -} __packed; /* format defined in RFC7609 */
>> -
>> -struct smc_clc_msg_accept_confirm_v2 { /* clc accept / confirm message */
>> struct smc_clc_msg_hdr hdr;
>> union {
>> struct { /* SMC-R */
>> struct smcr_clc_msg_accept_confirm r0;
>> - u8 eid[SMC_MAX_EID_LEN];
>> - u8 reserved6[8];
>> - } r1;
>> + struct { /* v2 only */
>> + u8 eid[SMC_MAX_EID_LEN];
>> + u8 reserved6[8];
>> + } __packed r1;
>> + };
>> struct { /* SMC-D */
>> struct smcd_clc_msg_accept_confirm_common d0;
>> - __be16 chid;
>> - u8 eid[SMC_MAX_EID_LEN];
>> - u8 reserved5[8];
>> - } d1;
>> + struct { /* v2 only, but 12 bytes reserved in v1 */
>> + __be16 chid;
>> + u8 eid[SMC_MAX_EID_LEN];
>> + u8 reserved5[8];
>> + } __packed d1;
>> + };
>> };
>> };
>
>
> I still think the __packed at the outmost level is the safest place.
> Like you have it now the compiler could place unused memory between
> ro and r1 or between d0 and d1.
> Afaik compilers don't do that, if the blocks are word-aligned, but
> there is no guarantee.
>
> Up to you. My R-b still applies.
> Sandy
Thank you, Sandy.
IIUC, if only outmost level has __packed, it won't work for the inner block.
e.g.
If __packed is added at d1 and r1:
struct smc_clc_msg_accept_confirm { /* clc accept / confirm message */
struct smc_clc_msg_hdr hdr;
union {
struct { /* SMC-R */
struct smcr_clc_msg_accept_confirm r0;
struct { /* v2 only */
u8 eid[SMC_MAX_EID_LEN];
u8 reserved6[8];
} __packed r1;
};
struct { /* SMC-D */
struct smcd_clc_msg_accept_confirm_common d0;
struct { /* v2 only, but 12 bytes reserved in v1 */
__be16 chid;
u8 eid[SMC_MAX_EID_LEN];
u64 gid_ext;
} __packed d1;
};
};
};
According to pahole, it will be:
struct smc_clc_msg_accept_confirm {
struct smc_clc_msg_hdr hdr; /* 0 8 */
union {
struct {
struct smcr_clc_msg_accept_confirm r0; /* 8 56 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct {
u8 eid[32]; /* 64 32 */
u8 reserved6[8]; /* 96 8 */
} r1; /* 64 40 */
}; /* 8 96 */
struct {
struct smcd_clc_msg_accept_confirm_common d0; /* 8 24 */
struct {
__be16 chid; /* 32 2 */
u8 eid[32]; /* 34 32 */
/* --- cacheline 1 boundary (64 bytes) was 2 bytes ago --- */
u64 gid_ext; /* 66 8 */
} __attribute__((__packed__)) d1; /* 32 42 */
} __attribute__((__packed__)); /* 8 66 */
}; /* 8 96 */
/* size: 104, cachelines: 2, members: 2 */
/* last cacheline: 40 bytes */
};
If __packed is added at outmost level:
struct smc_clc_msg_accept_confirm { /* clc accept / confirm message */
struct smc_clc_msg_hdr hdr;
union {
struct { /* SMC-R */
struct smcr_clc_msg_accept_confirm r0;
struct { /* v2 only */
u8 eid[SMC_MAX_EID_LEN];
u8 reserved6[8];
} r1;
};
struct { /* SMC-D */
struct smcd_clc_msg_accept_confirm_common d0;
struct { /* v2 only, but 12 bytes reserved in v1 */
__be16 chid;
u8 eid[SMC_MAX_EID_LEN];
u64 gid_ext;
} d1;
};
};
} __packed;
According to pahole, it will be:
struct smc_clc_msg_accept_confirm {
struct smc_clc_msg_hdr hdr; /* 0 8 */
union {
struct {
struct smcr_clc_msg_accept_confirm r0; /* 8 56 */
/* --- cacheline 1 boundary (64 bytes) --- */
struct {
u8 eid[32]; /* 64 32 */
u8 reserved6[8]; /* 96 8 */
} r1; /* 64 40 */
}; /* 8 96 */
struct {
struct smcd_clc_msg_accept_confirm_common d0; /* 8 24 */
struct {
__be16 chid; /* 32 2 */
u8 eid[32]; /* 34 32 */
/* XXX 6 bytes hole, try to pack */
/* --- cacheline 1 boundary (64 bytes) was 8 bytes ago --- */
u64 gid_ext; /* 72 8 */
} d1; /* 32 48 */ <- doesn't work for inner d1.
}; /* 8 72 */
}; /* 8 96 */
/* size: 104, cachelines: 2, members: 2 */
/* last cacheline: 40 bytes */
};
I also considered add them all:
struct smc_clc_msg_accept_confirm { /* clc accept / confirm message */
struct smc_clc_msg_hdr hdr;
union {
struct { /* SMC-R */
struct smcr_clc_msg_accept_confirm r0;
struct { /* v2 only */
u8 eid[SMC_MAX_EID_LEN];
u8 reserved6[8];
} __packed r1;
} __packed;
struct { /* SMC-D */
struct smcd_clc_msg_accept_confirm_common d0;
struct { /* v2 only, but 12 bytes reserved in v1 */
__be16 chid;
u8 eid[SMC_MAX_EID_LEN];
u64 gid_ext;
} __packed d1;
} __packed;
};
} __packed;
but a little bit strange since for only d1 needs to packed, so I kept it as it is now.
Thanks,
Wen Gu
next prev parent reply other threads:[~2023-12-20 12:17 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-12-19 14:26 [PATCH net-next v8 00/10] net/smc: implement SMCv2.1 virtual ISM device support Wen Gu
2023-12-19 14:26 ` [PATCH net-next v8 01/10] net/smc: rename some 'fce' to 'fce_v2x' for clarity Wen Gu
2023-12-19 14:26 ` [PATCH net-next v8 02/10] net/smc: introduce sub-functions for smc_clc_send_confirm_accept() Wen Gu
2023-12-19 14:26 ` [PATCH net-next v8 03/10] net/smc: unify the structs of accept or confirm message for v1 and v2 Wen Gu
2023-12-20 10:27 ` Alexandra Winter
2023-12-20 11:37 ` Alexandra Winter
2023-12-20 12:16 ` Wen Gu [this message]
2023-12-19 14:26 ` [PATCH net-next v8 04/10] net/smc: support SMCv2.x supplemental features negotiation Wen Gu
2023-12-19 14:26 ` [PATCH net-next v8 05/10] net/smc: introduce virtual ISM device support feature Wen Gu
2023-12-19 14:26 ` [PATCH net-next v8 06/10] net/smc: define a reserved CHID range for virtual ISM devices Wen Gu
2023-12-19 14:26 ` [PATCH net-next v8 07/10] net/smc: compatible with 128-bits extended GID of virtual ISM device Wen Gu
2024-01-24 14:29 ` [REGRESSION] v6.8 SMC-D issues Alexandra Winter
2024-01-24 14:44 ` Alexandra Winter
2024-01-25 4:59 ` Wen Gu
2024-01-25 8:26 ` Alexandra Winter
2024-01-25 9:28 ` Wen Gu
2023-12-19 14:26 ` [PATCH net-next v8 08/10] net/smc: support extended GID in SMC-D lgr netlink attribute Wen Gu
2023-12-19 14:26 ` [PATCH net-next v8 09/10] net/smc: disable SEID on non-s390 archs where virtual ISM may be used Wen Gu
2023-12-19 14:26 ` [PATCH net-next v8 10/10] net/smc: manage system EID in SMC stack instead of ISM driver Wen Gu
2023-12-20 13:34 ` [PATCH net-next v8 00/10] net/smc: implement SMCv2.1 virtual ISM device support Wen Gu
2023-12-26 20:30 ` patchwork-bot+netdevbpf
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=38f06cfb-2d68-2b10-f82b-62a44c25b8f8@linux.alibaba.com \
--to=guwen@linux.alibaba.com \
--cc=agordeev@linux.ibm.com \
--cc=alibuda@linux.alibaba.com \
--cc=borntraeger@linux.ibm.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=gor@linux.ibm.com \
--cc=guangguan.wang@linux.alibaba.com \
--cc=hca@linux.ibm.com \
--cc=jaka@linux.ibm.com \
--cc=kgraul@linux.ibm.com \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-s390@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=raspl@linux.ibm.com \
--cc=schnelle@linux.ibm.com \
--cc=svens@linux.ibm.com \
--cc=tonylu@linux.alibaba.com \
--cc=wenjia@linux.ibm.com \
--cc=wintera@linux.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).