- 05 Feb, 2021 39 commits
-
-
Gustavo A. R. Silva authored
There is a regular need in the kernel to provide a way to declare having a dynamically sized set of trailing elements in a structure. Kernel code should always use “flexible array members”[1] for these cases. The older style of one-element or zero-length arrays should no longer be used[2]. Refactor the code according to the use of a flexible-array member in struct ice_res_tracker, instead of a one-element array and use the struct_size() helper to calculate the size for the allocations. Also, notice that the code below suggests that, currently, two too many bytes are being allocated with devm_kzalloc(), as the total number of entries (pf->irq_tracker->num_entries) for pf->irq_tracker->list[] is _vectors_ and sizeof(*pf->irq_tracker) also includes the size of the one-element array _list_ in struct ice_res_tracker. drivers/net/ethernet/intel/ice/ice_main.c:3511: 3511 /* populate SW interrupts pool with number of OS granted IRQs. */ 3512 pf->num_avail_sw_msix = (u16)vectors; 3513 pf->irq_tracker->num_entries = (u16)vectors; 3514 pf->irq_tracker->end = pf->irq_tracker->num_entries; With this change, the right amount of dynamic memory is now allocated because, contrary to one-element arrays which occupy at least as much space as a single object of the type, flexible-array members don't occupy such space in the containing structure. [1] https://en.wikipedia.org/wiki/Flexible_array_member [2] https://www.kernel.org/doc/html/v5.9-rc1/process/deprecated.html#zero-length-and-one-element-arraysBuilt-tested-by:
kernel test robot <lkp@intel.com> Signed-off-by:
Gustavo A. R. Silva <gustavoars@kernel.org> Tested-by:
Tony Brelinski <tonyx.brelinski@intel.com> Signed-off-by:
Tony Nguyen <anthony.l.nguyen@intel.com>
-
Jacob Keller authored
Just as we recently added support for other stored firmware flash versions, support display of the stored UNDI Option ROM version via devlink info. To do this, we need to introduce a new ice_get_inactive_orom_ver function. This is a little trickier than with other flash versions. The Option ROM version data was being read from a special "Boot Configuration" block of the NVM Preserved Field Area. This block only contains the *active* Option ROM version data. It is populated when the device firmware finishes updating the Option ROM. This method is ineffective at reading the stored Option ROM version data. Instead of reading from this section of the flash, replace this version extraction with one which locates the Combo Version information from within the Option ROM binary. This data is stored within the Option ROM at a 512 byte offset, in a simple structured format. The structure uses a simple modulo 256 checksum for integrity verification. Scan through the Option ROM to locate the CIVD data section, and extract the Combo Version. Refactor ice_get_orom_ver_info so that it takes the bank select enumeration parameter. Use this to implement ice_get_inactive_orom_ver. Although all ice devices have a Boot Configuration block in the NVM PFA, not all devices have a valid Option ROM. In this case, the old ice_get_orom_ver_info would "succeed" but report a version of all zeros. The new implementation would fail to locate the $CIV section in the Option ROM and report an error. Thus, we must ensure that ice_init_nvm does not fail if ice_get_orom_ver_info fails. Use the new ice_get_inactive_orom_ver to allow reporting the Option ROM versions for a pending update via devlink info. Signed-off-by:
Jacob Keller <jacob.e.keller@intel.com> Tested-by:
-
Jacob Keller authored
Add a function to read the inactive netlist bank for version information. To support this, refactor how we read the netlist version data. Instead of using the firmware AQ interface with a module ID, read from the flash as a flat NVM, using ice_read_flash_module. This change requires a slight adjustment to the offset values used, as reading from the flat NVM includes the type field (which was stripped by firmware previously). Cleanup the macro names and move them to ice_type.h. For clarity in how we calculate the offsets and so that programmers can easily map the offset value to the data sheet, use a wrapper macro to account for the offset adjustments. Use the newly added ice_get_inactive_netlist_ver function to extract the version data from the pending netlist module update. Add the stored variants of "fw.netlist", and "fw.netlist.build" to the info version map array. With this change, we now report the "fw.netlist" and "fw.netlist.build" versions into the stored section of the devlink info report. As with the main NVM module versions, if there is no pending update, we report the currently active values as stored. Signed-off-by:
-
Jacob Keller authored
The devlink info interface supports drivers reporting "stored" versions. These versions indicate the version of an update that has been downloaded to the device, but is not yet active. The code for extracting the NVM version recently changed to enable support for reading from either the active or the inactive bank. Use this to implement ice_get_inactive_nvm_ver, which will read the NVM version data from the inactive section of flash. When reporting the versions via devlink info, first read the device capabilities. Determine if there is a pending flash update, and if so, extract relevant version information from the inactive flash. Store these within the info context structure. When reporting "stored" firmware versions, devlink documentation indicates that we ought to always report a stored value, even if there is no pending update. In this common case, the stored version should match the running version. This means that each stored version should by default fallback to the same value as reported by the running handler. To support this, modify the version structure to have both a "getter" and a "fallback". Modify the control loop so that it will use the "fallback" function if the "getter" function does not report a version. To report versions for which we can read the stored value, use a new "stored()" macro. This macro will insert two entries into the version list. The first entry is the traditional running version. The second is the stored version, implemented with a fallback to the active version. This is a little tricky, but reduces the overall duplication of elements in the entry list, and ensures that running and stored values remain consistent. To avoid some duplication, add a combined() macro that will insert both the running and stored versions into the version entry list. Using this new support, add pending version reporter functions for "fw.psid.api" and "fw.bundle_id". This enables reporting the stored values for some of versions in the NVM module of the flash. Reporting management versions is not implemented by this patch. The active management version is reported to the driver via the AdminQ mailbox during load. Although the version must be in the firmware binary somewhere, accessing this from the inactive firmware is not trivial and has not been implemented in this change. Future changes will introduce support for reading the UNDI Option ROM version and the version associated with the Netlist module. Signed-off-by:
-
Jacob Keller authored
When reading from the flash memory of the device, the ice driver has two interfaces available to it. First, it can use a mediated interface via firmware that allows specifying a module ID. This allows reading from specific modules of the active flash bank. The second interface available is to perform flat reads. This allows complete access to the entire flash. However, using it requires the software to handle calculating module location and interpret pointer addresses. While most data required is accessible through the convenient first interface, certain flash contents are not. This includes the CSS header information associated with the Option ROM and NVM banks, as well as any access to the "inactive" banks used as scratch space for performing flash updates. In order to access all of the relevant flash contents, software must use the flat reads. Rather than forcing all flows to perform flat read calculations, introduce a new abstraction for reading from the flash: ice_read_flash_module. This function provides an abstraction for reading from either the active or inactive flash bank at the requested module. This interface is very similar to the abstraction provided via firmware, but allows access to additional modules, as well as providing a mechanism to request access to both flash banks. At first glance, it might make sense for this abstraction to allow specifying precisely which bank (1st or 2nd) the caller wishes to read. This is simpler to implement but more difficult to use. In practice, most callers only know whether they want the active bank, or the inactive bank. Rather than force callers to determine for themselves which bank to read from, implement ice_read_flash_module in terms of "active" vs "inactive". This significantly simplifies the implementation at the caller level and is a more useful abstraction over the flash contents. Make use of this new interface to refactor reading of the main NVM version information. Instead of using the firmware's mediated ShadowRAM function, use the ice_read_flash_module abstraction. To do this, notice that most reads of the NVM are going to be in 2-byte word chunks. To simplify using ice_read_flash_module for this case, ice_read_nvm_module is introduced. This is a simple wrapper around ice_read_flash_module which takes the correct pointer address for the NVM bank, and forces the 2-byte word format onto the caller. When reading the NVM versions, some fields are read from the Shadow RAM. The Shadow RAM is the first 64KB of flash memory, and is populated during device load. Most fields are copied from a section within the active NVM bank. In order to read this data from both the active and inactive NVM banks, we need to read not from the first 64KB of flash, but instead from the correct offset into the NVM bank. Introduce ice_read_nvm_sr_copy for this purpose. This function wraps around ice_read_nvm_module and has the same interface as the ice_read_sr_word, with the exception of allowing the caller to specify whether to read the active or inactive flash bank. With this change, it is now trivial to refactor ice_get_nvm_ver_info to read using the software mediated ice_read_flash_module interface instead of relying on the firmware mediated interface. This will be used in the following change to implement support for stored versions in the devlink info report. Additionally, the overall ice_read_flash_module interface will be used and extended to support all three major flash banks, and additionally to support reading the flash image security revision information. Signed-off-by:
-
Jacob Keller authored
The ice flash contains two copies of each of the NVM, Option ROM, and Netlist modules. Each bank has a pointer word and a size word. In order to correctly read from the active flash bank, the driver must calculate the offset manually. During NVM initialization, read the Shadow RAM control word and determine which bank is active for each NVM module. Additionally, cache the size and pointer values for use in calculating the correct offset. Signed-off-by:
-
Jacob Keller authored
The ice driver uses an array of structures which link an info name with a function that formats the associated version data into a string. All existing format functions simply format already captured static data from the driver hw structure. Future changes will introduce format functions for reporting the versions of flash sections stored but not yet applied. This type of version data is not stored as a member of the hw structure. This is because (a) it might not yet exist in the case there is no pending flash update, and (b) even if it does, it might change such as if an update is canceled or replaced by a new update before finalizing. We could simply have each format function gather its own data upon being called. However, in some cases the raw binary version data is a combination of multiple different reported fields. Additionally, the current interface doesn't have a way for the function to indicate that the version doesn't exist. Refactor this function interface to take a new ice_info_ctx structure instead of the buffer pointer and length. This context structure allows for future extensions to pre-gather version data that is stored within the context struct instead of the hw struct. Allocate this context structure initially at the start of ice_devlink_info_get. We use dynamic allocation instead of a local stack variable in order to avoid using too much kernel stack once we extend it with additional data structures. Modify the main loop that drives the info reporting so that the version buffer string is always cleared between each format. Explicitly check that the format function actually filled in a version string of non-zero length. If the string is not provided, simply skip this version without reporting an error. This allows for introducing format functions of versions which may or may not be present, such as the version of a pending update that has not yet been activated. Signed-off-by:
-
Jacob Keller authored
The ice_nvm_info structure has become somewhat of a dumping ground for all of the fields related to flash version. It holds the NVM version and EETRACK id, the OptionROM info structure, the flash size, the ShadowRAM size, and more. A future change is going to add the ability to read the NVM version and EETRACK ID from the inactive NVM bank. To make this simpler, it is useful to have these NVM version info fields extracted to their own structure. Rename ice_nvm_info into ice_flash_info, and create a separate ice_nvm_info structure that will contain the eetrack and NVM map version. Move the netlist_ver structure into ice_flash_info and rename it ice_netlist_info for consistency. Modify the static ice_get_orom_ver_info to take the option rom structure as a pointer. This makes it more obvious what portion of the hw struct is being modified. Do the same for ice_get_netlist_ver_info. Introduce a new ice_get_nvm_ver_info function, which will be similar to ice_get_orom_ver_info and ice_get_netlist_ver_info, used to keep the NVM version extraction code co-located. Signed-off-by:
-
Jacob Keller authored
When erasing, notify userspace of how long we will potentially take to erase a module. Doing so allows userspace to report the timeout, giving a clear indication of the upper time bound of the operation. Since we're re-using the erase timeout value, make it a macro rather than a magic number. Signed-off-by:
Shannon Nelson <snelson@pensando.io> Tested-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queueJakub Kicinski authored
Tony Nguyen says: ==================== 1GbE Intel Wired LAN Driver Updates 2021-02-03 This series contains updates to igc, igb, e1000e, and e1000 drivers. Sasha adds counting of good transmit packets and reporting of NVM version and gPHY version in ethtool firmware version. Replaces the use of strlcpy to the preferred strscpy. Fixes a typo that caused the wrong register to be output. He also removes an unused function pointer, some unneeded defines, and a non-applicable comment. All changes for igc. Gal Hammer fixes a typo which caused the RDBAL register values to be shown instead of TDBAL for igb. Nick Lowe enables RSS support for i211 devices for igb. Tom Rix fixes checkpatch warning by removing h from printk format specifier for igb. Kaixu Xia removes setting of a variable that is overwritten before next use for e1000e. Sudip Mukherjee removes an unneeded assignment for e1000. * '1GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue: e1000: drop unneeded assignment in e1000_set_itr() e1000e: remove the redundant value assignment in e1000_update_nvm_checksum_spt igb: remove h from printk format specifier igb: Enable RSS for Intel I211 Ethernet Controller igb: fix TDBAL register show incorrect value igc: Fix TDBAL register show incorrect value igc: Remove unused FUNC_1 mask igc: Remove unused local receiver mask igc: Prefer strscpy over strlcpy igc: Expose the gPHY firmware version igc: Expose the NVM version igc: Add Host Good Packets Transmitted Count igc: Remove MULR mask define igc: Remove igc_set_fw_version comment igc: Clean up nvm_operations structure ==================== Link: https://lore.kernel.org/r/20210204004259.3662059-1-anthony.l.nguyen@intel.comSigned-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Jakub Kicinski authored
Andrea Parri says: ==================== Amend "hv_netvsc: Copy packets sent by Hyper-V out of the receive buffer" Patch #2 also addresses the Smatch complaint reported here: https://lkml.kernel.org/r/YBp2oVIdMe+G%2FliJ@mwanda/ ==================== Link: https://lore.kernel.org/r/20210203113513.558864-1-parri.andrea@gmail.comSigned-off-by:
-
Andrea Parri (Microsoft) authored
Fix the typo. Signed-off-by:
Andrea Parri (Microsoft) <parri.andrea@gmail.com> Fixes: 0ba35fe9 ("hv_netvsc: Copy packets sent by Hyper-V out of the receive buffer") Signed-off-by:
-
Andrea Parri (Microsoft) authored
The recv_buf buffers are allocated in netvsc_device_add(). Later in netvsc_init_buf() the response to NVSP_MSG1_TYPE_SEND_RECV_BUF allows the host to set up a recv_section_size that could be bigger than the (default) value used for that allocation. The host-controlled value could be used by a malicious host to bypass the check on the packet's length in netvsc_receive() and hence to overflow the recv_buf buffer. Move the allocation of the recv_buf buffers into netvsc_init_but(). Reported-by:
Juan Vazquez <juvazq@microsoft.com> Signed-off-by:
-
Jakub Kicinski authored
Hayes Wang says: ==================== r8152: adjust flow for power cut The two patches are used to adjust the flow about resuming from the state of power cut. For the purpose, some functions have to be updated first. ==================== Link: https://lore.kernel.org/r/1394712342-15778-398-Taiwan-albertk@realtek.comSigned-off-by:
-
Hayes Wang authored
For runtime resuming, the RTL8153B may be resumed from the state of power cut, when enabling the feature of UPS. Then, the PHY would be reset, so it is necessary to be initailized again. Besides, the USB_U1U2_TIMER also has to be set again, so I move it from r8153b_init() to r8153b_hw_phy_cfg(). Signed-off-by:
Hayes Wang <hayeswang@realtek.com> Signed-off-by:
-
Hayes Wang authored
Replace r8153_patch_request() with rtl_phy_patch_request(). Replace r8153_pre_ram_code() with rtl_pre_ram_code(). Replace r8153_post_ram_code() with rtl_post_ram_code(). Add rtl_patch_key_set(). The new functions have an additional parameter. It is used to wait the patch request command finished. When the PHY is resumed from the state of power cut, the PHY is at a safe mode and the OCP_PHY_PATCH_STAT wouldn't be updated. For this situation, it is safe to set patch request command without waiting OCP_PHY_PATCH_STAT. Signed-off-by:
-
Tobias Waldekranz authored
On read, master should send 31 MSB of the register (only even values are ever used), followed by a 1 to indicate read. Then, reading two bytes, the device will output the register's value. On write, master sends 31 MSB of the register, followed by a 0 to indicate write, followed by two bytes containing the register value. Flexibilis' documentation (version 1.3) specifies the opposite polarity (#read/write), but the scope indicates that it is, in fact, read/#write. Signed-off-by:
Tobias Waldekranz <tobias@waldekranz.com> Reviewed-by:
George McCollister <george.mccollister@gmail.com> Link: https://lore.kernel.org/r/20210202191645.439-1-tobias@waldekranz.comSigned-off-by:
-
Vladimir Oltean authored
The flow steering struct ethtool_flow_ext::data field is __be32, so when the CFP code needs to check the VLAN egress tagging attribute in bit 0, it does this in CPU native endianness. So logically, the endianness conversion is set up the other way around, although in practice the same result is produced. Gets rid of build warning: warning: cast from restricted __be32 warning: incorrect type in argument 1 (different base types) expected unsigned int [usertype] val got restricted __be32 warning: cast from restricted __be32 warning: cast from restricted __be32 warning: cast from restricted __be32 warning: cast from restricted __be32 warning: restricted __be32 degrades to integer Signed-off-by:
Vladimir Oltean <olteanv@gmail.com> Acked-by:
Florian Fainelli <f.fainelli@gmail.com> Link: https://lore.kernel.org/r/20210203193918.2236994-1-olteanv@gmail.comSigned-off-by:
-
Bhaskar Chowdhury authored
s/initialsation/initialisation/ s/specifiing/specifying/ Signed-off-by:
Bhaskar Chowdhury <unixbhaskar@gmail.com> Acked-by:
Randy Dunlap <rdunlap@infradead.org> Link: https://lore.kernel.org/r/20210204031648.27300-1-unixbhaskar@gmail.comSigned-off-by:
-
Jakub Kicinski authored
Leon Romanovsky says: ==================== Fix W=1 compilation warnings in net/* folder Changelog: v2: * Patch 3: Added missing include file. v1: https://lore.kernel.org/lkml/20210203101612.4004322-1-leon@kernel.org * Removed Fixes lines. * Changed target from net to be net-next. * Patch 1: Moved function declaration to be outside config instead games with if/endif. * Patch 3: Moved declarations to new header file. v0: https://lore.kernel.org/lkml/20210202135544.3262383-1-leon@kernel.org ==================== Link: https://lore.kernel.org/r/20210203135112.4083711-1-leon@kernel.orgSigned-off-by:
-
Leon Romanovsky authored
Fix the following compilation warnings: net/netfilter/ipvs/ip_vs_proto_tcp.c:147:1: warning: no previous prototype for 'tcp_snat_handler' [-Wmissing-prototypes] 147 | tcp_snat_handler(struct sk_buff *skb, struct ip_vs_protocol *pp, | ^~~~~~~~~~~~~~~~ net/netfilter/ipvs/ip_vs_proto_udp.c:136:1: warning: no previous prototype for 'udp_snat_handler' [-Wmissing-prototypes] 136 | udp_snat_handler(struct sk_buff *skb, struct ip_vs_protocol *pp, | ^~~~~~~~~~~~~~~~ Signed-off-by:Leon Romanovsky <leonro@nvidia.com> Signed-off-by:
-
Leon Romanovsky authored
Fir the following compilation warnings: 1031 | INDIRECT_CALLABLE_SCOPE void udp_v6_early_demux(struct sk_buff *skb) net/ipv6/ip6_offload.c:182:41: warning: no previous prototype for ‘ipv6_gro_receive’ [-Wmissing-prototypes] 182 | INDIRECT_CALLABLE_SCOPE struct sk_buff *ipv6_gro_receive(struct list_head *head, | ^~~~~~~~~~~~~~~~ net/ipv6/ip6_offload.c:320:29: warning: no previous prototype for ‘ipv6_gro_complete’ [-Wmissing-prototypes] 320 | INDIRECT_CALLABLE_SCOPE int ipv6_gro_complete(struct sk_buff *skb, int nhoff) | ^~~~~~~~~~~~~~~~~ net/ipv6/ip6_offload.c:182:41: warning: no previous prototype for ‘ipv6_gro_receive’ [-Wmissing-prototypes] 182 | INDIRECT_CALLABLE_SCOPE struct sk_buff *ipv6_gro_receive(struct list_head *head, | ^~~~~~~~~~~~~~~~ net/ipv6/ip6_offload.c:320:29: warning: no previous prototype for ‘ipv6_gro_complete’ [-Wmissing-prototypes] 320 | INDIRECT_CALLABLE_SCOPE int ipv6_gro_complete(struct sk_buff *skb, int nhoff) Signed-off-by: -
Leon Romanovsky authored
Fix the following compilation warning: net/ipv6/udp.c:1031:30: warning: no previous prototype for 'udp_v6_early_demux' [-Wmissing-prototypes] 1031 | INDIRECT_CALLABLE_SCOPE void udp_v6_early_demux(struct sk_buff *skb) | ^~~~~~~~~~~~~~~~~~ net/ipv6/udp.c:1072:29: warning: no previous prototype for 'udpv6_rcv' [-Wmissing-prototypes] 1072 | INDIRECT_CALLABLE_SCOPE int udpv6_rcv(struct sk_buff *skb) | ^~~~~~~~~ Signed-off-by: -
Leon Romanovsky authored
The W=1 compilation of allmodconfig generates the following warning: net/ipv6/icmp.c:448:6: warning: no previous prototype for 'icmp6_send' [-Wmissing-prototypes] 448 | void icmp6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info, | ^~~~~~~~~~ Fix it by providing function declaration for builds with ipv6 as a module. Signed-off-by: -
Colin Ian King authored
The null check of filp->f_path.dentry->d_iname is redundant because it is an array of DNAME_INLINE_LEN chars and cannot be a null. Fix this by removing the null check. Addresses-Coverity: ("Array compared against 0") Fixes: 04987ca1 ("net: hns3: add debugfs support for tm nodes, priority and qset info") Signed-off-by:Colin Ian King <colin.king@canonical.com> Reviewed-by:
Jesse Brandeburg <jesse.brandeburg@intel.com> Link: https://lore.kernel.org/r/20210203131040.21656-1-colin.king@canonical.comSigned-off-by:
-
wengjianfeng authored
change 'piority' to 'priority' change 'succesfult' to 'successful' Signed-off-by:
wengjianfeng <wengjianfeng@yulong.com> Link: https://lore.kernel.org/r/20210203093842.11180-1-samirweng1979@163.comSigned-off-by:
-
Jakub Kicinski authored
Xin Long says: ==================== net: enable udp v6 sockets receiving v4 packets with UDP Currently, udp v6 socket can not process v4 packets with UDP GRO, as udp_encap_needed_key is not increased when udp_tunnel_encap_enable() is called for v6 socket. This patchset is to increase it and remove the unnecessary code in bareudp in Patch 1/2, and improve rxrpc encap_enable by calling udp_tunnel_encap_enable(). ==================== Link: https://lore.kernel.org/r/cover.1612342376.git.lucien.xin@gmail.comSigned-off-by:
-
Xin Long authored
When doing encap_enable/increasing encap_needed_key, up->encap_enabled is not set in rxrpc_open_socket(), and it will cause encap_needed_key not being decreased in udpv6_destroy_sock(). This patch is to improve it by just calling udp_tunnel_encap_enable() where it increases both UDP and UDPv6 encap_needed_key and sets up->encap_enabled. v4->v5: - add the missing '#include <net/udp_tunnel.h>', as David Howells noticed. Acked-and-tested-by:David Howells <dhowells@redhat.com> Signed-off-by:
Xin Long <lucien.xin@gmail.com> Signed-off-by:
-
Xin Long authored
When enabling encap for a ipv6 socket without udp_encap_needed_key increased, UDP GRO won't work for v4 mapped v6 address packets as sk will be NULL in udp4_gro_receive(). This patch is to enable it by increasing udp_encap_needed_key for v6 sockets in udp_tunnel_encap_enable(), and correspondingly decrease udp_encap_needed_key in udpv6_destroy_sock(). v1->v2: - add udp_encap_disable() and export it. v2->v3: - add the change for rxrpc and bareudp into one patch, as Alex suggested. v3->v4: - move rxrpc part to another patch. Acked-by:Willem de Bruijn <willemb@google.com> Signed-off-by:
-
Jian Yang authored
Traditionally loopback devices come up with initial state as DOWN for any new network-namespace. This would mean that anyone needing this device would have to bring this UP by issuing something like 'ip link set lo up'. This can be avoided if the initial state is set as UP. Signed-off-by:
Mahesh Bandewar <maheshb@google.com> Signed-off-by:
Jian Yang <jianyang@google.com> Link: https://lore.kernel.org/r/20210201233445.2044327-1-jianyang.kernel@gmail.comSigned-off-by:
-
Jakub Kicinski authored
Alexander Lobakin says: ==================== net: consolidate page_is_pfmemalloc() usage page_is_pfmemalloc() is used mostly by networking drivers to test if a page can be considered for reusing/recycling. It doesn't write anything to the struct page itself, so its sole argument can be constified, as well as the first argument of skb_propagate_pfmemalloc(). In Page Pool core code, it can be simply inlined instead. Most of the callers from NIC drivers were just doppelgangers of the same condition tests. Derive them into a new common function do deduplicate the code. Resend of v3 [2]: - it missed Patchwork and Netdev archives, probably due to server-side issues. Since v2 [1]: - use more intuitive name for the new inline function since there's nothing "reserved" in remote pages (Jakub Kicinski, John Hubbard); - fold likely() inside the helper itself to make driver code a bit fancier (Jakub Kicinski); - split function introduction and using into two separate commits; - collect some more tags (Jesse Brandeburg, David Rientjes). Since v1 [0]: - new: reduce code duplication by introducing a new common function to test if a page can be reused/recycled (David Rientjes); - collect autographs for Page Pool bits (Jesper Dangaard Brouer, Ilias Apalodimas). [0] https://lore.kernel.org/netdev/20210125164612.243838-1-alobakin@pm.me [1] https://lore.kernel.org/netdev/20210127201031.98544-1-alobakin@pm.me [2] https://lore.kernel.org/lkml/20210131120844.7529-1-alobakin@pm.me ==================== Link: https://lore.kernel.org/r/20210202133030.5760-1-alobakin@pm.meSigned-off-by:
-
Alexander Lobakin authored
pool_page_reusable() is a leftover from pre-NUMA-aware times. For now, this function is just a redundant wrapper over page_is_pfmemalloc(), so inline it into its sole call site. Signed-off-by:
Alexander Lobakin <alobakin@pm.me> Acked-by:
Jesper Dangaard Brouer <brouer@redhat.com> Reviewed-by:
Ilias Apalodimas <ilias.apalodimas@linaro.org> Reviewed-by:
David Rientjes <rientjes@google.com> Signed-off-by:
-
Alexander Lobakin authored
Now we can remove a bunch of identical functions from the drivers and make them use common dev_page_is_reusable(). All {,un}likely() checks are omitted since it's already present in this helper. Also update some comments near the call sites. Suggested-by: -
Alexander Lobakin authored
A bunch of drivers test the page before reusing/recycling for two common conditions: - if a page was allocated under memory pressure (pfmemalloc page); - if a page was allocated at a distant memory node (to exclude slowdowns). Introduce a new common inline for doing this, with likely() already folded inside to make driver code a bit simpler. Suggested-by:
-
Alexander Lobakin authored
The function doesn't write anything to the page struct itself, so this argument can be const. Misc: align second argument to the brace while at it. Signed-off-by:
-
Alexander Lobakin authored
The function only tests for page->index, so its argument should be const. Signed-off-by:
-
Brian Vazquez authored
This commit fixes the errores reported when building for powerpc: ERROR: modpost: "ip6_dst_check" [vmlinux] is a static EXPORT_SYMBOL ERROR: modpost: "ipv4_dst_check" [vmlinux] is a static EXPORT_SYMBOL ERROR: modpost: "ipv4_mtu" [vmlinux] is a static EXPORT_SYMBOL ERROR: modpost: "ip6_mtu" [vmlinux] is a static EXPORT_SYMBOL Fixes: f67fbeae ("net: use indirect call helpers for dst_mtu") Fixes: bbd807df ("net: indirect call helpers for ipv4/ipv6 dst_check functions") Reported-by:
Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by:
Brian Vazquez <brianvv@google.com> Link: https://lore.kernel.org/r/20210204181839.558951-2-brianvv@google.comSigned-off-by:
-
Brian Vazquez authored
When a static function is annotated with INDIRECT_CALLABLE_SCOPE and CONFIG_RETPOLINE is set, the static keyword is removed. Sometimes the function needs to be exported but EXPORT_SYMBOL can't be used because if CONFIG_RETPOLINE is not set, we will attempt to export a static symbol. This patch introduces a new indirect call wrapper: EXPORT_INDIRECT_CALLABLE. This basically does EXPORT_SYMBOL when CONFIG_RETPOLINE is set, but does nothing when it's not. Reported-by:
-
Marcelo Ricardo Leitner authored
Often userspace won't request the extack information, or they don't log it because of log level or so, and even when they do, sometimes it's not enough to know exactly what caused the error. Netlink extack is the standard way of reporting erros with descriptive error messages. With a trace point on it, we then can know exactly where the error happened, regardless of userspace app. Also, we can even see if the err msg was overwritten. The wrapper do_trace_netlink_extack() is because trace points shouldn't be called from .h files, as trace points are not that small, and the function call to do_trace_netlink_extack() on the macros is not protected by tracepoint_enabled() because the macros are called from modules, and this would require exporting some trace structs. As this is error path, it's better to export just the wrapper instead. v2: removed leftover tracepoint declaration Signed-off-by:
Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> Reviewed-by:
David Ahern <dsahern@kernel.org> Link: https://lore.kernel.org/r/4546b63e67b2989789d146498b13cc09e1fdc543.1612403190.git.marcelo.leitner@gmail.comSigned-off-by:
-
- 04 Feb, 2021 1 commit
-
-
Jiapeng Chong authored
Fix the following coccicheck warnings: ./drivers/net/xen-netfront.c:1816:52-54: WARNING !A || A && B is equivalent to !A || B. Reported-by:
Abaci Robot <abaci@linux.alibaba.com> Signed-off-by:
Jiapeng Chong <jiapeng.chong@linux.alibaba.com> Reviewed-by:
Juergen Gross <jgross@suse.com> Link: https://lore.kernel.org/r/1612261069-13315-1-git-send-email-jiapeng.chong@linux.alibaba.comSigned-off-by:
-
