- 24 May, 2023 40 commits
-
-
Jakub Kicinski authored
Russell King says: ==================== net: pcs: xpcs: cleanups for clause 73 support This series cleans up xpcs code, moving much of the clause 73 code out of the driver into places where others can make use of it. Specifically, we add a helper to convert a clause 73 advertisement to ethtool link modes to mdio.h, and a helper to resolve the clause 73 negotiation state to phylink, which includes the pause modes. In doing this cleanup, several issues were identified with the original xpcs implementation: 1) it masks the link partner advertisement with its own advertisement so userspace can't see what the full link partner advertisement was. 2) it was always setting pause modes irrespective of the advertisements on either end of the link. 3) it was reading the STAT1 registers multiple times. Reading STAT1 has the side effect of unlatching the link-down status, so multiple reads should be avoided. This patch series addresses the first two first by addressing the issues, and then by moving over to the new helpers. The third issue is solved by restructuring the xpcs code. ==================== Link: https://lore.kernel.org/r/ZGyR/jDyYTYzRklg@shell.armlinux.org.ukSigned-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Russell King (Oracle) authored
Avoid reading the STAT1 registers more than once while getting the PCS state, as this register contains latching-low bits that are lost after the first read. Signed-off-by:
Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Reviewed-by:
Andrew Lunn <andrew@lunn.ch> Signed-off-by:
-
Russell King (Oracle) authored
Use phylink_resolve_c73() to resolve the clause 73 autonegotiation result. Signed-off-by:
-
Russell King (Oracle) authored
xpcs was indicating symmetric pause should be enabled regardless of the advertisements by either party. Fix this to use linkmode_resolve_pause() now that we're no longer obliterating the link partner's advertisement by logically anding it with our own. This is transitional, the function will be entirely replaced with phylink_resolve_c73() in the following patch. Reviewed-by:
-
Russell King (Oracle) authored
lp_advertising is supposed to reflect the link partner's advertisement unmodified by the local advertisement, but xpcs bitwise ands it with the local advertisement prior to calculating the resolution of the negotiation. Fix this by moving the bitwise and to xpcs_resolve_lpa_c73() so it can place the results in a temporary bitmap before passing that to ixpcs_get_max_usxgmii_speed(). Reviewed-by:
-
Russell King (Oracle) authored
Convert xpcs clause 73 reading to use the newly introduced mii_c73_to_linkmode() helper to translate the link partner advertisement to an ethtool bitmap. Reviewed-by:
-
Russell King (Oracle) authored
Read the clause 73 link partner advertisement in a loop and then translate to the ethtool modes. Reviewed-by:
-
Russell King (Oracle) authored
Add a function to resolve clause 73 negotiation according to the priority resolution function described in clause 73.3.6. Signed-off-by:
-
Russell King (Oracle) authored
Phylink had two chunks of code virtually the same for resolving the negotiated pause modes. Factor this down to one function. Reviewed-by:
-
Russell King (Oracle) authored
Add a helper to convert a clause 73 advertisement to an ethtool bitmap. Reviewed-by:
-
David S. Miller authored
Jiri Pirko says: ==================== devlink: small port_new/del() cleanup This patchset cleans up couple of leftovers after recent devlink locking changes. Previously, both port_new/dev() commands were called without holding instance lock. Currently all devlink commands are called with instance lock held. The first patch just removes redundant port notification. The second one removes couple of outdated comments. The last patch changes port_dev() to have devlink_port pointer as an arg instead of port_index, which makes it similar to the rest of port related ops. ==================== Reviewed-by:
David S. Miller <davem@davemloft.net>
-
Jiri Pirko authored
Historically there was a reason why port_dev() along with for example port_split() did get port_index instead of the devlink_port pointer. With the locking changes that were done which ensured devlink instance mutex is hold for every command, the port ops could get devlink_port pointer directly. Change the forgotten port_dev() op to be as others and pass devlink_port pointer instead of port_index. Signed-off-by:
Jiri Pirko <jiri@nvidia.com> Reviewed-by:
Simon Horman <simon.horman@corigine.com> Signed-off-by:
-
Jiri Pirko authored
All commands are called holding instance lock. Remove the outdated comment that says otherwise. Signed-off-by:
-
Jiri Pirko authored
The notification about created port is send from devl_port_register() function called from ops->port_new(). No need to send it again here, so remove the call and the helper function. Signed-off-by:
-
David S. Miller authored
Donald Hunter says: ==================== tools: ynl: Add byte-order support for struct members This patchset adds support to ynl for handling byte-order in struct members. The first patch is a refactor to use predefined Struct() objects instead of generating byte-order specific formats on the fly. The second patch adds byte-order handling for struct members. ==================== Reviewed-by:
-
Donald Hunter authored
Add support for byte-order in struct members in the genetlink-legacy spec. Signed-off-by:
Donald Hunter <donald.hunter@gmail.com> Acked-by:
-
Donald Hunter authored
Use a dict of predefined Struct() objects to decode scalar types in native, big or little endian format. This removes the repetitive code for the scalar variants and ensures all the signed variants are supported. Signed-off-by:
-
Russell King (Oracle) authored
When taking a network interface down (or removing a SFP module) after the PHY has encountered an error, phy_stop() complains incorrectly that it was called from HALTED state. The reason this is incorrect is that the network driver will have called phy_start() when the interface was brought up, and the fact that the PHY has a problem bears no relationship to the administrative state of the interface. Taking the interface administratively down (which calls phy_stop()) is always the right thing to do after a successful phy_start() call, whether or not the PHY has encountered an error. Signed-off-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
David S. Miller authored
Guillaume Nault says: ==================== ipv4: Remove RTO_ONLINK from udp, ping and raw sockets. udp_sendmsg(), ping_v4_sendmsg() and raw_sendmsg() use similar patterns for restricting their route lookup to on-link hosts. Although they use slightly different code, they all use RTO_ONLINK to override the least significant bit of their tos value. RTO_ONLINK is used to restrict the route scope even when the scope is set to RT_SCOPE_UNIVERSE. Therefore it isn't necessary: we can properly set the scope to RT_SCOPE_LINK instead. Removing RTO_ONLINK will allow to convert .flowi4_tos to dscp_t in the future, thus allowing to properly separate the DSCP from the ECN bits in the networking stack. This patch series defines a common helper to figure out what's the scope of the route lookup. This unifies the way udp, ping and raw sockets get their routing scope and removes their dependency on RTO_ONLINK. ==================== Signed-off-by: -
Guillaume Nault authored
Use ip_sendmsg_scope() to properly initialise the scope in flowi4_init_output(), instead of overriding tos with the RTO_ONLINK flag. The objective is to eventually remove RTO_ONLINK, which will allow converting .flowi4_tos to dscp_t. Now that the scope is determined by ip_sendmsg_scope(), we need to check its result to set the 'connected' variable. Signed-off-by:
Guillaume Nault <gnault@redhat.com> Reviewed-by:
David Ahern <dsahern@kernel.org> Signed-off-by:
-
Guillaume Nault authored
Use ip_sendmsg_scope() to properly initialise the scope in flowi4_init_output(), instead of overriding tos with the RTO_ONLINK flag. The objective is to eventually remove RTO_ONLINK, which will allow converting .flowi4_tos to dscp_t. The MSG_DONTROUTE and SOCK_LOCALROUTE cases were already handled by raw_sendmsg() (SOCK_LOCALROUTE was handled by the RT_CONN_FLAGS*() macros called by get_rtconn_flags()). However, opt.is_strictroute wasn't taken into account. Therefore, a side effect of this patch is to now honour opt.is_strictroute, and thus align raw_sendmsg() with ping_v4_sendmsg() and udp_sendmsg(). Since raw_sendmsg() was the only user of get_rtconn_flags(), we can now remove this function. Signed-off-by:
-
Guillaume Nault authored
Define a new helper to figure out the correct route scope to use on TX, depending on socket configuration, ancillary data and send flags. Use this new helper to properly initialise the scope in flowi4_init_output(), instead of overriding tos with the RTO_ONLINK flag. The objective is to eventually remove RTO_ONLINK, which will allow converting .flowi4_tos to dscp_t. Signed-off-by:
-
Coco Li authored
Add support for using IPv6 Big TCP on DQ which can handle large TSO/GRO packets. See https://lwn.net/Articles/895398/. This can improve the throughput and CPU usage. Perf test result: ip -d link show $DEV gso_max_size 185000 gso_max_segs 65535 tso_max_size 262143 tso_max_segs 65535 gro_max_size 185000 For performance, tested with neper using 9k MTU on hardware that supports 200Gb/s line rate. In single streams when line rate is not saturated, we expect throughput improvements. When the networking is performing at line rate, we expect cpu usage improvements. Tcp_stream (unidirectional stream test, T=thread, F=flow): skb=180kb, T=1, F=1, no zerocopy: throughput average=64576.88 Mb/s, sender stime=8.3, receiver stime=10.68 skb=64kb, T=1, F=1, no zerocopy: throughput average=64862.54 Mb/s, sender stime=9.96, receiver stime=12.67 skb=180kb, T=1, F=1, yes zerocopy: throughput average=146604.97 Mb/s, sender stime=10.61, receiver stime=5.52 skb=64kb, T=1, F=1, yes zerocopy: throughput average=131357.78 Mb/s, sender stime=12.11, receiver stime=12.25 skb=180kb, T=20, F=100, no zerocopy: throughput average=182411.37 Mb/s, sender stime=41.62, receiver stime=79.4 skb=64kb, T=20, F=100, no zerocopy: throughput average=182892.02 Mb/s, sender stime=57.39, receiver stime=72.69 skb=180kb, T=20, F=100, yes zerocopy: throughput average=182337.65 Mb/s, sender stime=27.94, receiver stime=39.7 skb=64kb, T=20, F=100, yes zerocopy: throughput average=182144.20 Mb/s, sender stime=47.06, receiver stime=39.01 Signed-off-by:
Ziwei Xiao <ziweixiao@google.com> Signed-off-by:
Coco Li <lixiaoyan@google.com> Reviewed-by:
-
Jakub Kicinski authored
David Howells says: ==================== splice, net: Replace sendpage with sendmsg(MSG_SPLICE_PAGES), part 1 Here's the first tranche of patches towards providing a MSG_SPLICE_PAGES internal sendmsg flag that is intended to replace the ->sendpage() op with calls to sendmsg(). MSG_SPLICE_PAGES is a hint that tells the protocol that it should splice the pages supplied if it can and copy them if not. This will allow splice to pass multiple pages in a single call and allow certain parts of higher protocols (e.g. sunrpc, iwarp) to pass an entire message in one go rather than having to send them piecemeal. This should also make it easier to handle the splicing of multipage folios. A helper, skb_splice_from_iter() is provided to do the work of splicing or copying data from an iterator. If a page is determined to be unspliceable (such as being in the slab), then the helper will give an error. Note that this facility is not made available to userspace and does not provide any sort of callback. This set consists of the following parts: (1) Define the MSG_SPLICE_PAGES flag and prevent sys_sendmsg() from being able to set it. (2) Add an extra argument to skb_append_pagefrags() so that something other than MAX_SKB_FRAGS can be used (sysctl_max_skb_frags for example). (3) Add the skb_splice_from_iter() helper to handle splicing pages into skbuffs for MSG_SPLICE_PAGES that can be shared by TCP, IP/UDP and AF_UNIX. (4) Implement MSG_SPLICE_PAGES support in TCP. (5) Make do_tcp_sendpages() just wrap sendmsg() and then fold it in to its various callers. (6) Implement MSG_SPLICE_PAGES support in IP and make udp_sendpage() just a wrapper around sendmsg(). (7) Implement MSG_SPLICE_PAGES support in IP6/UDP6. (8) Implement MSG_SPLICE_PAGES support in AF_UNIX. (9) Make AF_UNIX copy unspliceable pages. Link: https://lore.kernel.org/r/20230316152618.711970-1-dhowells@redhat.com/ # v1 Link: https://lore.kernel.org/r/20230329141354.516864-1-dhowells@redhat.com/ # v2 Link: https://lore.kernel.org/r/20230331160914.1608208-1-dhowells@redhat.com/ # v3 Link: https://lore.kernel.org/r/20230405165339.3468808-1-dhowells@redhat.com/ # v4 Link: https://lore.kernel.org/r/20230406094245.3633290-1-dhowells@redhat.com/ # v5 Link: https://lore.kernel.org/r/20230411160902.4134381-1-dhowells@redhat.com/ # v6 Link: https://lore.kernel.org/r/20230515093345.396978-1-dhowells@redhat.com/ # v7 Link: https://lore.kernel.org/r/20230518113453.1350757-1-dhowells@redhat.com/ # v8 ==================== Link: https://lore.kernel.org/r/20230522121125.2595254-1-dhowells@redhat.comSigned-off-by: -
David Howells authored
Convert unix_stream_sendpage() to use sendmsg() with MSG_SPLICE_PAGES rather than directly splicing in the pages itself. This allows ->sendpage() to be replaced by something that can handle multiple multipage folios in a single transaction. Signed-off-by:
David Howells <dhowells@redhat.com> cc: Kuniyuki Iwashima <kuniyu@amazon.com> cc: Jens Axboe <axboe@kernel.dk> cc: Matthew Wilcox <willy@infradead.org> Signed-off-by:
-
David Howells authored
Make AF_UNIX sendmsg() support MSG_SPLICE_PAGES, splicing in pages from the source iterator if possible and copying the data in otherwise. This allows ->sendpage() to be replaced by something that can handle multiple multipage folios in a single transaction. Signed-off-by:
-
David Howells authored
ip_append_page() is no longer used with the removal of udp_sendpage(), so remove it. Signed-off-by:
-
David Howells authored
Convert udp_sendpage() to use sendmsg() with MSG_SPLICE_PAGES rather than directly splicing in the pages itself. This allows ->sendpage() to be replaced by something that can handle multiple multipage folios in a single transaction. Signed-off-by:
-
David Howells authored
Make IP6/UDP6 sendmsg() support MSG_SPLICE_PAGES. This causes pages to be spliced from the source iterator if possible, copying the data if not. This allows ->sendpage() to be replaced by something that can handle multiple multipage folios in a single transaction. Signed-off-by:
-
David Howells authored
Make IP/UDP sendmsg() support MSG_SPLICE_PAGES. This causes pages to be spliced from the source iterator. This allows ->sendpage() to be replaced by something that can handle multiple multipage folios in a single transaction. Signed-off-by:
-
David Howells authored
Fold do_tcp_sendpages() into its last remaining caller, tcp_sendpage_locked(). Signed-off-by:
-
David Howells authored
do_tcp_sendpages() is now just a small wrapper around tcp_sendmsg_locked(), so inline it, allowing do_tcp_sendpages() to be removed. This is part of replacing ->sendpage() with a call to sendmsg() with MSG_SPLICE_PAGES set. Signed-off-by:
Bernard Metzler <bmt@zurich.ibm.com> Reviewed-by:
Tom Talpey <tom@talpey.com> cc: Jason Gunthorpe <jgg@ziepe.ca> cc: Leon Romanovsky <leon@kernel.org> cc: Jens Axboe <axboe@kernel.dk> cc: Matthew Wilcox <willy@infradead.org> Signed-off-by:
-
David Howells authored
do_tcp_sendpages() is now just a small wrapper around tcp_sendmsg_locked(), so inline it, allowing do_tcp_sendpages() to be removed. This is part of replacing ->sendpage() with a call to sendmsg() with MSG_SPLICE_PAGES set. Signed-off-by:
-
David Howells authored
do_tcp_sendpages() is now just a small wrapper around tcp_sendmsg_locked(), so inline it, allowing do_tcp_sendpages() to be removed. This is part of replacing ->sendpage() with a call to sendmsg() with MSG_SPLICE_PAGES set. Signed-off-by:
-
David Howells authored
do_tcp_sendpages() is now just a small wrapper around tcp_sendmsg_locked(), so inline it. This is part of replacing ->sendpage() with a call to sendmsg() with MSG_SPLICE_PAGES set. Signed-off-by:
-
David Howells authored
Convert do_tcp_sendpages() to use sendmsg() with MSG_SPLICE_PAGES rather than directly splicing in the pages itself. do_tcp_sendpages() can then be inlined in subsequent patches into its callers. This allows ->sendpage() to be replaced by something that can handle multiple multipage folios in a single transaction. Signed-off-by:
-
David Howells authored
Make TCP's sendmsg() support MSG_SPLICE_PAGES. This causes pages to be spliced or copied (if it cannot be spliced) from the source iterator. This allows ->sendpage() to be replaced by something that can handle multiple multipage folios in a single transaction. Signed-off-by:
-
David Howells authored
Add a function to handle MSG_SPLICE_PAGES being passed internally to sendmsg(). Pages are spliced into the given socket buffer if possible and copied in if not (e.g. they're slab pages or have a zero refcount). Signed-off-by:
-
David Howells authored
Pass the maximum number of fragments into skb_append_pagefrags() rather than using MAX_SKB_FRAGS so that it can be used from code that wants to specify sysctl_max_skb_frags. Signed-off-by:
-
David Howells authored
Declare MSG_SPLICE_PAGES, an internal sendmsg() flag, that hints to a network protocol that it should splice pages from the source iterator rather than copying the data if it can. This flag is added to a list that is cleared by sendmsg syscalls on entry. This is intended as a replacement for the ->sendpage() op, allowing a way to splice in several multipage folios in one go. Signed-off-by:
Willem de Bruijn <willemb@google.com> cc: Jens Axboe <axboe@kernel.dk> cc: Matthew Wilcox <willy@infradead.org> Signed-off-by:
-
