skbuff: render the checksum comment to documentation

Long time ago Tom added a giant comment to skbuff.h explaining
checksums. Now that we have a place in Documentation for skbuff
docs we should render it. Sprinkle some markup while at it.

Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This commit is contained in:
Jakub Kicinski 2022-05-09 09:04:56 -07:00
parent 9ec7ea1462
commit 9facd94114
2 changed files with 119 additions and 84 deletions

View File

@ -29,3 +29,9 @@ dataref and headerless skbs
.. kernel-doc:: include/linux/skbuff.h .. kernel-doc:: include/linux/skbuff.h
:doc: dataref and headerless skbs :doc: dataref and headerless skbs
Checksum information
--------------------
.. kernel-doc:: include/linux/skbuff.h
:doc: skb checksums

View File

@ -43,98 +43,112 @@
#include <linux/netfilter/nf_conntrack_common.h> #include <linux/netfilter/nf_conntrack_common.h>
#endif #endif
/* The interface for checksum offload between the stack and networking drivers /**
* DOC: skb checksums
*
* The interface for checksum offload between the stack and networking drivers
* is as follows... * is as follows...
* *
* A. IP checksum related features * IP checksum related features
* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
* *
* Drivers advertise checksum offload capabilities in the features of a device. * Drivers advertise checksum offload capabilities in the features of a device.
* From the stack's point of view these are capabilities offered by the driver. * From the stack's point of view these are capabilities offered by the driver.
* A driver typically only advertises features that it is capable of offloading * A driver typically only advertises features that it is capable of offloading
* to its device. * to its device.
* *
* The checksum related features are: * .. flat-table:: Checksum related device features
* :widths: 1 10
* *
* NETIF_F_HW_CSUM - The driver (or its device) is able to compute one * * - %NETIF_F_HW_CSUM
* IP (one's complement) checksum for any combination * - The driver (or its device) is able to compute one
* of protocols or protocol layering. The checksum is * IP (one's complement) checksum for any combination
* computed and set in a packet per the CHECKSUM_PARTIAL * of protocols or protocol layering. The checksum is
* interface (see below). * computed and set in a packet per the CHECKSUM_PARTIAL
* interface (see below).
* *
* NETIF_F_IP_CSUM - Driver (device) is only able to checksum plain * * - %NETIF_F_IP_CSUM
* TCP or UDP packets over IPv4. These are specifically * - Driver (device) is only able to checksum plain
* unencapsulated packets of the form IPv4|TCP or * TCP or UDP packets over IPv4. These are specifically
* IPv4|UDP where the Protocol field in the IPv4 header * unencapsulated packets of the form IPv4|TCP or
* is TCP or UDP. The IPv4 header may contain IP options. * IPv4|UDP where the Protocol field in the IPv4 header
* This feature cannot be set in features for a device * is TCP or UDP. The IPv4 header may contain IP options.
* with NETIF_F_HW_CSUM also set. This feature is being * This feature cannot be set in features for a device
* DEPRECATED (see below). * with NETIF_F_HW_CSUM also set. This feature is being
* DEPRECATED (see below).
* *
* NETIF_F_IPV6_CSUM - Driver (device) is only able to checksum plain * * - %NETIF_F_IPV6_CSUM
* TCP or UDP packets over IPv6. These are specifically * - Driver (device) is only able to checksum plain
* unencapsulated packets of the form IPv6|TCP or * TCP or UDP packets over IPv6. These are specifically
* IPv6|UDP where the Next Header field in the IPv6 * unencapsulated packets of the form IPv6|TCP or
* header is either TCP or UDP. IPv6 extension headers * IPv6|UDP where the Next Header field in the IPv6
* are not supported with this feature. This feature * header is either TCP or UDP. IPv6 extension headers
* cannot be set in features for a device with * are not supported with this feature. This feature
* NETIF_F_HW_CSUM also set. This feature is being * cannot be set in features for a device with
* DEPRECATED (see below). * NETIF_F_HW_CSUM also set. This feature is being
* DEPRECATED (see below).
* *
* NETIF_F_RXCSUM - Driver (device) performs receive checksum offload. * * - %NETIF_F_RXCSUM
* This flag is only used to disable the RX checksum * - Driver (device) performs receive checksum offload.
* feature for a device. The stack will accept receive * This flag is only used to disable the RX checksum
* checksum indication in packets received on a device * feature for a device. The stack will accept receive
* regardless of whether NETIF_F_RXCSUM is set. * checksum indication in packets received on a device
* regardless of whether NETIF_F_RXCSUM is set.
* *
* B. Checksumming of received packets by device. Indication of checksum * Checksumming of received packets by device
* verification is set in skb->ip_summed. Possible values are: * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
* *
* CHECKSUM_NONE: * Indication of checksum verification is set in &sk_buff.ip_summed.
* Possible values are:
*
* - %CHECKSUM_NONE
* *
* Device did not checksum this packet e.g. due to lack of capabilities. * Device did not checksum this packet e.g. due to lack of capabilities.
* The packet contains full (though not verified) checksum in packet but * The packet contains full (though not verified) checksum in packet but
* not in skb->csum. Thus, skb->csum is undefined in this case. * not in skb->csum. Thus, skb->csum is undefined in this case.
* *
* CHECKSUM_UNNECESSARY: * - %CHECKSUM_UNNECESSARY
* *
* The hardware you're dealing with doesn't calculate the full checksum * The hardware you're dealing with doesn't calculate the full checksum
* (as in CHECKSUM_COMPLETE), but it does parse headers and verify checksums * (as in %CHECKSUM_COMPLETE), but it does parse headers and verify checksums
* for specific protocols. For such packets it will set CHECKSUM_UNNECESSARY * for specific protocols. For such packets it will set %CHECKSUM_UNNECESSARY
* if their checksums are okay. skb->csum is still undefined in this case * if their checksums are okay. &sk_buff.csum is still undefined in this case
* though. A driver or device must never modify the checksum field in the * though. A driver or device must never modify the checksum field in the
* packet even if checksum is verified. * packet even if checksum is verified.
* *
* CHECKSUM_UNNECESSARY is applicable to following protocols: * %CHECKSUM_UNNECESSARY is applicable to following protocols:
* TCP: IPv6 and IPv4. *
* UDP: IPv4 and IPv6. A device may apply CHECKSUM_UNNECESSARY to a * - TCP: IPv6 and IPv4.
* - UDP: IPv4 and IPv6. A device may apply CHECKSUM_UNNECESSARY to a
* zero UDP checksum for either IPv4 or IPv6, the networking stack * zero UDP checksum for either IPv4 or IPv6, the networking stack
* may perform further validation in this case. * may perform further validation in this case.
* GRE: only if the checksum is present in the header. * - GRE: only if the checksum is present in the header.
* SCTP: indicates the CRC in SCTP header has been validated. * - SCTP: indicates the CRC in SCTP header has been validated.
* FCOE: indicates the CRC in FC frame has been validated. * - FCOE: indicates the CRC in FC frame has been validated.
* *
* skb->csum_level indicates the number of consecutive checksums found in * &sk_buff.csum_level indicates the number of consecutive checksums found in
* the packet minus one that have been verified as CHECKSUM_UNNECESSARY. * the packet minus one that have been verified as %CHECKSUM_UNNECESSARY.
* For instance if a device receives an IPv6->UDP->GRE->IPv4->TCP packet * For instance if a device receives an IPv6->UDP->GRE->IPv4->TCP packet
* and a device is able to verify the checksums for UDP (possibly zero), * and a device is able to verify the checksums for UDP (possibly zero),
* GRE (checksum flag is set) and TCP, skb->csum_level would be set to * GRE (checksum flag is set) and TCP, &sk_buff.csum_level would be set to
* two. If the device were only able to verify the UDP checksum and not * two. If the device were only able to verify the UDP checksum and not
* GRE, either because it doesn't support GRE checksum or because GRE * GRE, either because it doesn't support GRE checksum or because GRE
* checksum is bad, skb->csum_level would be set to zero (TCP checksum is * checksum is bad, skb->csum_level would be set to zero (TCP checksum is
* not considered in this case). * not considered in this case).
* *
* CHECKSUM_COMPLETE: * - %CHECKSUM_COMPLETE
* *
* This is the most generic way. The device supplied checksum of the _whole_ * This is the most generic way. The device supplied checksum of the _whole_
* packet as seen by netif_rx() and fills in skb->csum. This means the * packet as seen by netif_rx() and fills in &sk_buff.csum. This means the
* hardware doesn't need to parse L3/L4 headers to implement this. * hardware doesn't need to parse L3/L4 headers to implement this.
* *
* Notes: * Notes:
*
* - Even if device supports only some protocols, but is able to produce * - Even if device supports only some protocols, but is able to produce
* skb->csum, it MUST use CHECKSUM_COMPLETE, not CHECKSUM_UNNECESSARY. * skb->csum, it MUST use CHECKSUM_COMPLETE, not CHECKSUM_UNNECESSARY.
* - CHECKSUM_COMPLETE is not applicable to SCTP and FCoE protocols. * - CHECKSUM_COMPLETE is not applicable to SCTP and FCoE protocols.
* *
* CHECKSUM_PARTIAL: * - %CHECKSUM_PARTIAL
* *
* A checksum is set up to be offloaded to a device as described in the * A checksum is set up to be offloaded to a device as described in the
* output description for CHECKSUM_PARTIAL. This may occur on a packet * output description for CHECKSUM_PARTIAL. This may occur on a packet
@ -146,14 +160,18 @@
* packet that are after the checksum being offloaded are not considered to * packet that are after the checksum being offloaded are not considered to
* be verified. * be verified.
* *
* C. Checksumming on transmit for non-GSO. The stack requests checksum offload * Checksumming on transmit for non-GSO
* in the skb->ip_summed for a packet. Values are: * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
* *
* CHECKSUM_PARTIAL: * The stack requests checksum offload in the &sk_buff.ip_summed for a packet.
* Values are:
*
* - %CHECKSUM_PARTIAL
* *
* The driver is required to checksum the packet as seen by hard_start_xmit() * The driver is required to checksum the packet as seen by hard_start_xmit()
* from skb->csum_start up to the end, and to record/write the checksum at * from &sk_buff.csum_start up to the end, and to record/write the checksum at
* offset skb->csum_start + skb->csum_offset. A driver may verify that the * offset &sk_buff.csum_start + &sk_buff.csum_offset.
* A driver may verify that the
* csum_start and csum_offset values are valid values given the length and * csum_start and csum_offset values are valid values given the length and
* offset of the packet, but it should not attempt to validate that the * offset of the packet, but it should not attempt to validate that the
* checksum refers to a legitimate transport layer checksum -- it is the * checksum refers to a legitimate transport layer checksum -- it is the
@ -165,55 +183,66 @@
* checksum calculation to the device, or call skb_checksum_help (in the case * checksum calculation to the device, or call skb_checksum_help (in the case
* that the device does not support offload for a particular checksum). * that the device does not support offload for a particular checksum).
* *
* NETIF_F_IP_CSUM and NETIF_F_IPV6_CSUM are being deprecated in favor of * %NETIF_F_IP_CSUM and %NETIF_F_IPV6_CSUM are being deprecated in favor of
* NETIF_F_HW_CSUM. New devices should use NETIF_F_HW_CSUM to indicate * %NETIF_F_HW_CSUM. New devices should use %NETIF_F_HW_CSUM to indicate
* checksum offload capability. * checksum offload capability.
* skb_csum_hwoffload_help() can be called to resolve CHECKSUM_PARTIAL based * skb_csum_hwoffload_help() can be called to resolve %CHECKSUM_PARTIAL based
* on network device checksumming capabilities: if a packet does not match * on network device checksumming capabilities: if a packet does not match
* them, skb_checksum_help or skb_crc32c_help (depending on the value of * them, skb_checksum_help() or skb_crc32c_help() (depending on the value of
* csum_not_inet, see item D.) is called to resolve the checksum. * &sk_buff.csum_not_inet, see :ref:`crc`)
* is called to resolve the checksum.
* *
* CHECKSUM_NONE: * - %CHECKSUM_NONE
* *
* The skb was already checksummed by the protocol, or a checksum is not * The skb was already checksummed by the protocol, or a checksum is not
* required. * required.
* *
* CHECKSUM_UNNECESSARY: * - %CHECKSUM_UNNECESSARY
* *
* This has the same meaning as CHECKSUM_NONE for checksum offload on * This has the same meaning as CHECKSUM_NONE for checksum offload on
* output. * output.
* *
* CHECKSUM_COMPLETE: * - %CHECKSUM_COMPLETE
*
* Not used in checksum output. If a driver observes a packet with this value * Not used in checksum output. If a driver observes a packet with this value
* set in skbuff, it should treat the packet as if CHECKSUM_NONE were set. * set in skbuff, it should treat the packet as if %CHECKSUM_NONE were set.
* *
* D. Non-IP checksum (CRC) offloads * .. _crc:
* *
* NETIF_F_SCTP_CRC - This feature indicates that a device is capable of * Non-IP checksum (CRC) offloads
* offloading the SCTP CRC in a packet. To perform this offload the stack * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
* will set csum_start and csum_offset accordingly, set ip_summed to
* CHECKSUM_PARTIAL and set csum_not_inet to 1, to provide an indication in
* the skbuff that the CHECKSUM_PARTIAL refers to CRC32c.
* A driver that supports both IP checksum offload and SCTP CRC32c offload
* must verify which offload is configured for a packet by testing the
* value of skb->csum_not_inet; skb_crc32c_csum_help is provided to resolve
* CHECKSUM_PARTIAL on skbs where csum_not_inet is set to 1.
* *
* NETIF_F_FCOE_CRC - This feature indicates that a device is capable of * .. flat-table::
* offloading the FCOE CRC in a packet. To perform this offload the stack * :widths: 1 10
* will set ip_summed to CHECKSUM_PARTIAL and set csum_start and csum_offset
* accordingly. Note that there is no indication in the skbuff that the
* CHECKSUM_PARTIAL refers to an FCOE checksum, so a driver that supports
* both IP checksum offload and FCOE CRC offload must verify which offload
* is configured for a packet, presumably by inspecting packet headers.
* *
* E. Checksumming on output with GSO. * * - %NETIF_F_SCTP_CRC
* - This feature indicates that a device is capable of
* offloading the SCTP CRC in a packet. To perform this offload the stack
* will set csum_start and csum_offset accordingly, set ip_summed to
* %CHECKSUM_PARTIAL and set csum_not_inet to 1, to provide an indication
* in the skbuff that the %CHECKSUM_PARTIAL refers to CRC32c.
* A driver that supports both IP checksum offload and SCTP CRC32c offload
* must verify which offload is configured for a packet by testing the
* value of &sk_buff.csum_not_inet; skb_crc32c_csum_help() is provided to
* resolve %CHECKSUM_PARTIAL on skbs where csum_not_inet is set to 1.
* *
* In the case of a GSO packet (skb_is_gso(skb) is true), checksum offload * * - %NETIF_F_FCOE_CRC
* - This feature indicates that a device is capable of offloading the FCOE
* CRC in a packet. To perform this offload the stack will set ip_summed
* to %CHECKSUM_PARTIAL and set csum_start and csum_offset
* accordingly. Note that there is no indication in the skbuff that the
* %CHECKSUM_PARTIAL refers to an FCOE checksum, so a driver that supports
* both IP checksum offload and FCOE CRC offload must verify which offload
* is configured for a packet, presumably by inspecting packet headers.
*
* Checksumming on output with GSO
* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*
* In the case of a GSO packet (skb_is_gso() is true), checksum offload
* is implied by the SKB_GSO_* flags in gso_type. Most obviously, if the * is implied by the SKB_GSO_* flags in gso_type. Most obviously, if the
* gso_type is SKB_GSO_TCPV4 or SKB_GSO_TCPV6, TCP checksum offload as * gso_type is %SKB_GSO_TCPV4 or %SKB_GSO_TCPV6, TCP checksum offload as
* part of the GSO operation is implied. If a checksum is being offloaded * part of the GSO operation is implied. If a checksum is being offloaded
* with GSO then ip_summed is CHECKSUM_PARTIAL, and both csum_start and * with GSO then ip_summed is %CHECKSUM_PARTIAL, and both csum_start and
* csum_offset are set to refer to the outermost checksum being offloaded * csum_offset are set to refer to the outermost checksum being offloaded
* (two offloaded checksums are possible with UDP encapsulation). * (two offloaded checksums are possible with UDP encapsulation).
*/ */