- 10 Oct, 2014 27 commits
-
-
David S. Miller authored
Eric Dumazet says: ==================== net: fix races accessing page->_count This is illegal to use atomic_set(&page->_count, ...) even if we 'own' the page. Other entities in the kernel need to use get_page_unless_zero() to get a reference to the page before testing page properties, so we could loose a refcount increment. The only case it is valid is when page->_count is 0, we can use this in __netdev_alloc_frag() Note that I never seen crashes caused by these races, the issue was reported by Andres Lagar-Cavilla and Hugh Dickins. ==================== Signed-off-by:David S. Miller <davem@davemloft.net>
-
Eric Dumazet authored
This is illegal to use atomic_set(&page->_count, ...) even if we 'own' the page. Other entities in the kernel need to use get_page_unless_zero() to get a reference to the page before testing page properties, so we could loose a refcount increment. The only case it is valid is when page->_count is 0 Fixes: 540eb7bf ("net: Update alloc frag to reduce get/put page usage and recycle pages") Signed-off-by:
Eric Dumaze <edumazet@google.com> Signed-off-by:
-
Eric Dumazet authored
This is illegal to use atomic_set(&page->_count, ...) even if we 'own' the page. Other entities in the kernel need to use get_page_unless_zero() to get a reference to the page before testing page properties, so we could loose a refcount increment. Signed-off-by:
-
Eric Dumazet authored
This is illegal to use atomic_set(&page->_count, 2) even if we 'own' the page. Other entities in the kernel need to use get_page_unless_zero() to get a reference to the page before testing page properties, so we could loose a refcount increment. Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by:
-
Eric Dumazet authored
This is illegal to use atomic_set(&page->_count, 2) even if we 'own' the page. Other entities in the kernel need to use get_page_unless_zero() to get a reference to the page before testing page properties, so we could loose a refcount increment. Signed-off-by:
-
Eric Dumazet authored
This is illegal to use atomic_set(&page->_count, 2) even if we 'own' the page. Other entities in the kernel need to use get_page_unless_zero() to get a reference to the page before testing page properties, so we could loose a refcount increment. Signed-off-by:
-
Sascha Hauer authored
The KSZ8021 and KSZ8031 support RMII reference input clocks of 25MHz and 50MHz. Both PHYs differ in the default frequency they expect after reset. If this differs from the actual input clock, then register 0x1f bit 7 must be changed. Signed-off-by:
Sascha Hauer <s.hauer@pengutronix.de> Reviewed-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
Alexander Duyck authored
This patch addresses a kernel unaligned access bug seen on a sparc64 system with an igb adapter. Specifically the __skb_flow_get_ports was returning a be32 pointer which was then having the value directly returned. In order to prevent this it is actually easier to simply not populate the ports or address values when an skb is not present. In this case the assumption is that the data isn't needed and rather than slow down the faster aligned accesses by making them have to assume the unaligned path on architectures that don't support efficent unaligned access it makes more sense to simply switch off the bits that were copying the source and destination address/port for the case where we only care about the protocol types and lengths which are normally 16 bit fields anyway. Reported-by:
Alexander Duyck <alexander.h.duyck@redhat.com> Signed-off-by:
-
Li RongQing authored
1. sk_run_filter has been renamed, sk_filter() is using SK_RUN_FILTER. 2. Remove wrong comments about storing intermediate value. 3. replace sk_run_filter with __bpf_prog_run for check_load_and_stores's comments Cc: Alexei Starovoitov <ast@plumgrid.com> Signed-off-by:
Li RongQing <roy.qing.li@gmail.com> Acked-by:
Alexei Starovoitov <ast@plumgrid.com> Signed-off-by:
-
Li RongQing authored
__sk_run_filter has been renamed as __bpf_prog_run, so replace them in comments Signed-off-by:
-
David S. Miller authored
Jason Baron says: ==================== macvlan: optimize receive path So after porting this optimization to net-next, I found that the netperf results of TCP_RR regress right at the maximum peak of transactions/sec. That is as I increase the number of threads via the first argument to super_netperf, the number of transactions/sec keep increasing, peak, and then start decreasing. It is right at the peak, that I see a small regression with this patch (see results in patch 2/2). Without the patch, the ksoftirqd threads are the top cpu consumers threads on the system, since the extra 'netif_rx()', is queuing more softirq work, whereas with the patch, the ksoftirqd threads are below all of the 'netserver' threads in terms of their cpu usage. So there appears to be some interaction between how softirqs are serviced at the peak here and this patch. I think the test results are still supportive of this approach, but I wanted to be clear on my findings. ==================== Signed-off-by: -
jbaron@akamai.com authored
The netif_rx() call on the fast path of macvlan_handle_frame() appears to be there to ensure that we properly throttle incoming packets. However, it would appear as though the proper throttling is already in place for all possible ingress paths, and that the call is redundant. If packets are arriving from the physical NIC, we've already throttled them by this point. Otherwise, if they are coming via macvlan_queue_xmit(), it calls either 'dev_forward_skb()', which ends up calling netif_rx_internal(), or else in the broadcast case, we are throttling via macvlan_broadcast_enqueue(). The test results below are from off the box to an lxc instance running macvlan. Once the tranactions/sec stop increasing, the cpu idle time has gone to 0. Results are from a quad core Intel E3-1270 V2@3.50GHz box with bnx2x 10G card. for i in {10,100,200,300,400,500}; do super_netperf $i -H $ip -t TCP_RR; done Average of 5 runs. trans/sec trans/sec (3.17-rc7-net-next) (3.17-rc7-net-next + this patch) ---------- ---------- 208101 211534 (+1.6%) 839493 850162 (+1.3%) 845071 844053 (-.12%) 816330 819623 (+.4%) 778700 789938 (+1.4%) 735984 754408 (+2.5%) Signed-off-by:Jason Baron <jbaron@akamai.com> Signed-off-by:
-
jbaron@akamai.com authored
Pass last argument to macvlan_count_rx() as the correct bool type. Signed-off-by:
-
David S. Miller authored
Iyappan Subramanian says: ==================== Add 10GbE support to APM X-Gene SoC ethernet driver Adding 10GbE support to APM X-Gene SoC ethernet driver. v4: Address comments from v3 * dtb: resolved merge conflict for the net tree v3: Address comments from v2 * dtb: changed to use all-zeros for the mac address v2: Address comments from v1 * created preparatory patch to review before adding new functionality * dtb: updated to use tabs consistently v1: * Initial version ==================== Signed-off-by: -
Iyappan Subramanian authored
Signed-off-by:
Iyappan Subramanian <isubramanian@apm.com> Signed-off-by:
Keyur Chudgar <kchudgar@apm.com> Signed-off-by:
-
Iyappan Subramanian authored
- Added 10GbE support - Removed unused macros/variables - Moved mac_init call to the end of hardware init Signed-off-by:
-
Iyappan Subramanian authored
- Rearranged code to pave the way for adding 10GbE support - Added mac_ops structure containing function pointers for mac specific functions - Added port_ops structure containing function pointers for port specific functions Signed-off-by:
-
Iyappan Subramanian authored
Added 10GbE interface and clock nodes. Signed-off-by:
-
Iyappan Subramanian authored
Signed-off-by:
-
Iyappan Subramanian authored
Updated APM X-Gene ethernet driver maintainers list. Signed-off-by:
-
Alexei Starovoitov authored
minimal configurations where EPOLL, PERF_EVENTS, etc are disabled, but NET is enabled, are failing to build with link error: kernel/built-in.o: In function `bpf_prog_load': syscall.c:(.text+0x3b728): undefined reference to `anon_inode_getfd' fix it by selecting ANON_INODES when NET is enabled Reported-by:
Michal Sojka <sojkam1@fel.cvut.cz> Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-nextDavid S. Miller authored
Pablo Neira Ayuso says: ==================== Netfilter fixes for net-next This batch contains two fixes for what you have in your net-next, they are: 1) Remove nf_send_reset6() from header file. This function now resides in the nf_reject_ipv6 module. Reported by Eric Dumazet. 2) Fix wrong NFT_REJECT_ICMPX_MAX definition and adjust code to fix errors reported by Dan Carpenter's static analysis tools. ==================== Signed-off-by: -
David S. Miller authored
Merge tag 'master-2014-10-08' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-next John W. Linville says: ==================== pull request: wireless-next 2014-10-09 Please pull this batch of fixes intended for the 3.18 stream! Andrea Merello makes rtl818x_pci use a more reasonable transmission rate for HW generated frames. Fabian Frederick tweaks some kernel-doc bits to avoid warnings. Larry Finger corrects a possible unaligned access in the rtlwifi code. Marek Puzyniak avoids a kernel panic in ath9k_hw_reset. Sujith Manoharan goes for the hat trick -- he fixes a smatch warning in the shared ath code, he fixes a crash in ath9k, and he corrects a sequence number assignment problem in ath9k too. For ease of merging, I pulled the last bits of the wireless tree as well... Please let me know if there are problems! ==================== Signed-off-by:
-
Vince Bridgers authored
Testing revealed that the local variable mc_filter was dimensioned incorrectly for all possible configurations and get_mac_addr should have been set_mac_addr (a typo). Make sure mc_filter is dimensioned to 8 32-bit unsigned longs - the largest size of the Synopsys multicast filter register set. Signed-off-by:
Vince Bridgers <vbridger@opensource.altera.com> Acked-by:
Giuseppe Cavallaro <peppe.cavallaro@st.com> Signed-off-by:
-
Geert Uytterhoeven authored
If NO_DMA=y: drivers/built-in.o: In function `rxq_deinit': pxa168_eth.c:(.text+0x2a2f2e): undefined reference to `dma_free_coherent' drivers/built-in.o: In function `txq_reclaim': pxa168_eth.c:(.text+0x2a3044): undefined reference to `dma_unmap_single' drivers/built-in.o: In function `txq_deinit': pxa168_eth.c:(.text+0x2a310a): undefined reference to `dma_free_coherent' drivers/built-in.o: In function `txq_init': pxa168_eth.c:(.text+0x2a3226): undefined reference to `dma_alloc_coherent' drivers/built-in.o: In function `rxq_init': pxa168_eth.c:(.text+0x2a32d4): undefined reference to `dma_alloc_coherent' drivers/built-in.o: In function `init_hash_table': pxa168_eth.c:(.text+0x2a3354): undefined reference to `dma_alloc_coherent' drivers/built-in.o: In function `rxq_refill': pxa168_eth.c:(.text+0x2a345a): undefined reference to `dma_map_single' drivers/built-in.o: In function `rxq_process': pxa168_eth.c:(.text+0x2a39cc): undefined reference to `dma_unmap_single' drivers/built-in.o: In function `pxa168_eth_remove': pxa168_eth.c:(.text+0x2a3b84): undefined reference to `dma_free_coherent' drivers/built-in.o: In function `pxa168_eth_start_xmit': pxa168_eth.c:(.text+0x2a3e8a): undefined reference to `dma_map_single' Signed-off-by:
Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by:
-
Pranith Kumar authored
The latest linus git tip (3.18-rc1) fails with the following build failure. Fix this by making PTP support explicit for fm10k driver. rivers/built-in.o: In function `fm10k_ptp_register': (.text+0x12e760): undefined reference to `ptp_clock_registER' drivers/built-in.o: In function `fm10k_ptp_unregister': (.text+0x12e7dc): undefined reference to `ptp_clock_unregister' Makefile:930: recipe for target 'vmlinux' failed Signed-off-by:
Pranith Kumar <bobby.prani@gmail.com> Signed-off-by:
-
LEROY Christophe authored
[linux-devel:devel-hourly-2014100909 3763/3915] drivers/net/ethernet/freescale/fs_enet/mac-scc.c:119:32: error: 'SCCE_ENET_TXF' undeclared Due to patch d43a396a net: fs_enet: Add NAPI TX, it appears that some target compilations are broken. This is due to the fact that unlike the FEC, the SCC and FCC don't have a TXF event (complete Frame transmitted) but only TXB (buffer transmitted). Signed-off-by:
Christophe Leroy <christophe.leroy@c-s.fr> Signed-off-by:
-
- 09 Oct, 2014 13 commits
-
-
Jesper Dangaard Brouer authored
Restore the quota fairness between qdisc's, that we broke with commit 5772e9a3 ("qdisc: bulk dequeue support for qdiscs with TCQ_F_ONETXQUEUE"). Before that commit, the quota in __qdisc_run() were in packets as dequeue_skb() would only dequeue a single packet, that assumption broke with bulk dequeue. We choose not to account for the number of packets inside the TSO/GSO packets (accessable via "skb_gso_segs"). As the previous fairness also had this "defect". Thus, GSO/TSO packets counts as a single packet. Further more, we choose to slack on accuracy, by allowing a bulk dequeue try_bulk_dequeue_skb() to exceed the "packets" limit, only limited by the BQL bytelimit. This is done because BQL prefers to get its full budget for appropriate feedback from TX completion. In future, we might consider reworking this further and, if it allows, switch to a time-based model, as suggested by Eric. Right now, we only restore old semantics. Joint work with Eric, Hannes, Daniel and Jesper. Hannes wrote the first patch in cooperation with Daniel and Jesper. Eric rewrote the patch. Fixes: 5772e9a3 ("qdisc: bulk dequeue support for qdiscs with TCQ_F_ONETXQUEUE") Signed-off-by:
Jesper Dangaard Brouer <brouer@redhat.com> Signed-off-by:
Hannes Frederic Sowa <hannes@stressinduktion.org> Signed-off-by:
Daniel Borkmann <dborkman@redhat.com> Signed-off-by:
-
David S. Miller authored
Hayes Wang says: ==================== r8152: use mutex for hw settings v2: Make sure the autoresume wouldn't occur inside the mutex, otherwise the dead lock would happen. For the purpose, adjust some code about autosuspend/autoresume. v1: Use mutex to avoid that the serial hw settings would be interrupted by other settings. Although there is no problem now, it makes the driver more safe. ==================== Signed-off-by: -
hayeswang authored
Use the mutex to avoid the settings are interrupted by other ones. Signed-off-by:
Hayes Wang <hayeswang@realtek.com> Signed-off-by:
-
hayeswang authored
Add usb_autopm_xxx for rtl8152_get_settings() ,and remove usb_autopm_xxx from read_mii_word() and write_mii_word(). Signed-off-by:
-
hayeswang authored
Resume the device before setting the feature. Signed-off-by:
-
Masanari Iida authored
This patch fix following warning. Warning(.//net/core/skbuff.c:4142): No description found for parameter 'header_len' Warning(.//net/core/skbuff.c:4142): No description found for parameter 'data_len' Warning(.//net/core/skbuff.c:4142): No description found for parameter 'max_page_order' Warning(.//net/core/skbuff.c:4142): No description found for parameter 'errcode' Warning(.//net/core/skbuff.c:4142): No description found for parameter 'gfp_mask' Acutually the descriptions exist, but missing "@" in front. This problem start to happen when following commit was merged into Linus's tree during 3.18-rc1 merge period. commit 2e4e4410 net: add alloc_skb_with_frags() helper Signed-off-by:
Masanari Iida <standby24x7@gmail.com> Signed-off-by:
-
David S. Miller authored
Hariprasad Shenai says: ==================== cxgb4/cxgb4vf: Misc fixes and 40G support for cxgb4vf This patch series adds 40G support for cxgb4vf driver. Update the LSO length for cxgb4vf, fix macro. Wait for device to get ready before reading PL_WHOAMI register. The patches series is created against 'net-next' tree. And includes patches on cxgb4 and cxgb4vf driver. We have included all the maintainers of respective drivers. Kindly review the change and let us know in case of any review comments. ==================== Signed-off-by: -
Hariprasad Shenai authored
Call t4_wait_dev_ready() before attempting to read the PL_WHOAMI register (to determine which function we have been attached to). This prevents us from failing on that read if it comes right after a RESET. Signed-off-by:
Hariprasad Shenai <hariprasad@chelsio.com> Signed-off-by:
-
Hariprasad Shenai authored
Add 40G support for cxgb4vf driver. ethtool speed values are just numbers of megabits and there is no SPEED_40000 in ethtool speed values. To be consistent, use integer constants directly for all speeds. Use is_x_10g_port()("is 10Gb/s or higher") in cfg_queues() instead of is_10g_port() ("is exactly 10Gb/s"). Else we will end up using a single "Queue Set" on 40Gb/s adapters. Signed-off-by: -
Hariprasad Shenai authored
Update the lso length for T5 adapter and fix PIDX_T5 macro Signed-off-by:
-
David S. Miller authored
Claudiu Manoil says: ==================== gianfar: ARM port driver updates (1/2) This is the first round of driver portability fixes and clean-up with the main purpose to make gianfar portable on ARM, for the ARM based SoC that integrates the eTSEC ethernet controller - "ls1021a". The patches primarily address compile time errors, when compiling gianfar on ARM. They replace PPC specific functions and macros with architecture independent ones, solve arch specific header inclusions, guard code that relates to PPC only, and even address some simple endianess issues (see MAC address setup patch). The patches addressing the bulk of remaining endianess issues, like handling DMA fields (BD and FCB), will follow with the second round. ==================== Reviewed-by:
Kim Phillips <kim.phillips@freescale.com> Signed-off-by:
-
Claudiu Manoil authored
Replace PPC specific eieio() with arch independent wmb() for other architectures, i.e. ARM. The eieio() macro is not defined on ARM and generates build error. Signed-off-by:
Claudiu Manoil <claudiu.manoil@freescale.com> Signed-off-by:
-
Claudiu Manoil authored
Use arch independent code to replace the powerpc dependent spin_event_timeout() from gfar_halt_nodisable(). Added GRS/GTS read accessors to clean-up the implementation of gfar_halt_nodisable(). Signed-off-by:
-
