March 2018 - Linux-stable-mirror

Patch "PCI: hv: Properly handle PCI bus remove" has been added to the 4.9-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled PCI: hv: Properly handle PCI bus remove to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: pci-hv-properly-handle-pci-bus-remove.patch and it can be found in the queue-4.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From foo@baz Sun Mar 18 16:55:33 CET 2018 From: Long Li <longli(a)microsoft.com> Date: Thu, 23 Mar 2017 14:58:10 -0700 Subject: PCI: hv: Properly handle PCI bus remove From: Long Li <longli(a)microsoft.com> [ Upstream commit d3a78d8bf759d8848339dcc367c4c1678b57a08b ] hv_pci_devices_present() is called in hv_pci_remove() when we remove a PCI device from the host, e.g., by disabling SR-IOV on a device. In hv_pci_remove(), the bus is already removed before the call, so we don't need to rescan the bus in the workqueue scheduled from hv_pci_devices_present(). By introducing bus state hv_pcibus_removed, we can avoid this situation. Reported-by: Xiaofeng Wang <xiaofwan(a)redhat.com> Signed-off-by: Long Li <longli(a)microsoft.com> Signed-off-by: Bjorn Helgaas <bhelgaas(a)google.com> Acked-by: K. Y. Srinivasan <kys(a)microsoft.com> Signed-off-by: Sasha Levin <alexander.levin(a)microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/pci/host/pci-hyperv.c | 20 +++++++++++++++++--- 1 file changed, 17 insertions(+), 3 deletions(-) --- a/drivers/pci/host/pci-hyperv.c +++ b/drivers/pci/host/pci-hyperv.c @@ -351,6 +351,7 @@ enum hv_pcibus_state { hv_pcibus_init = 0, hv_pcibus_probed, hv_pcibus_installed, + hv_pcibus_removed, hv_pcibus_maximum }; @@ -1489,13 +1490,24 @@ static void pci_devices_present_work(str put_pcichild(hpdev, hv_pcidev_ref_initial); } - /* Tell the core to rescan bus because there may have been changes. */ - if (hbus->state == hv_pcibus_installed) { + switch(hbus->state) { + case hv_pcibus_installed: + /* + * Tell the core to rescan bus + * because there may have been changes. + */ pci_lock_rescan_remove(); pci_scan_child_bus(hbus->pci_bus); pci_unlock_rescan_remove(); - } else { + break; + + case hv_pcibus_init: + case hv_pcibus_probed: survey_child_resources(hbus); + break; + + default: + break; } up(&hbus->enum_sem); @@ -2170,6 +2182,7 @@ static int hv_pci_probe(struct hv_device hbus = kzalloc(sizeof(*hbus), GFP_KERNEL); if (!hbus) return -ENOMEM; + hbus->state = hv_pcibus_init; /* * The PCI bus "domain" is what is called "segment" in ACPI and @@ -2312,6 +2325,7 @@ static int hv_pci_remove(struct hv_devic pci_stop_root_bus(hbus->pci_bus); pci_remove_root_bus(hbus->pci_bus); pci_unlock_rescan_remove(); + hbus->state = hv_pcibus_removed; } ret = hv_send_resources_released(hdev); Patches currently in stable-queue which might be from longli(a)microsoft.com are queue-4.9/pci-hv-lock-pci-bus-on-device-eject.patch queue-4.9/pci-hv-properly-handle-pci-bus-remove.patch

6 years, 10 months

1
0
0 0

Patch "PCI: hv: Lock PCI bus on device eject" has been added to the 4.9-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled PCI: hv: Lock PCI bus on device eject to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: pci-hv-lock-pci-bus-on-device-eject.patch and it can be found in the queue-4.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From foo@baz Sun Mar 18 16:55:33 CET 2018 From: Long Li <longli(a)microsoft.com> Date: Thu, 23 Mar 2017 14:58:32 -0700 Subject: PCI: hv: Lock PCI bus on device eject From: Long Li <longli(a)microsoft.com> [ Upstream commit 414428c5da1c71986727c2fa5cdf1ed071e398d7 ] A PCI_EJECT message can arrive at the same time we are calling pci_scan_child_bus() in the workqueue for the previous PCI_BUS_RELATIONS message or in create_root_hv_pci_bus(). In this case we could potentially modify the bus from multiple places. Properly lock the bus access. Thanks Dexuan Cui <decui(a)microsoft.com> for pointing out the race condition in create_root_hv_pci_bus(). Reported-by: Xiaofeng Wang <xiaofwan(a)redhat.com> Signed-off-by: Long Li <longli(a)microsoft.com> Signed-off-by: Bjorn Helgaas <bhelgaas(a)google.com> Acked-by: K. Y. Srinivasan <kys(a)microsoft.com> Signed-off-by: Sasha Levin <alexander.levin(a)microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/pci/host/pci-hyperv.c | 4 ++++ 1 file changed, 4 insertions(+) --- a/drivers/pci/host/pci-hyperv.c +++ b/drivers/pci/host/pci-hyperv.c @@ -1206,9 +1206,11 @@ static int create_root_hv_pci_bus(struct hbus->pci_bus->msi = &hbus->msi_chip; hbus->pci_bus->msi->dev = &hbus->hdev->device; + pci_lock_rescan_remove(); pci_scan_child_bus(hbus->pci_bus); pci_bus_assign_resources(hbus->pci_bus); pci_bus_add_devices(hbus->pci_bus); + pci_unlock_rescan_remove(); hbus->state = hv_pcibus_installed; return 0; } @@ -1597,8 +1599,10 @@ static void hv_eject_device_work(struct pdev = pci_get_domain_bus_and_slot(hpdev->hbus->sysdata.domain, 0, wslot); if (pdev) { + pci_lock_rescan_remove(); pci_stop_and_remove_bus_device(pdev); pci_dev_put(pdev); + pci_unlock_rescan_remove(); } memset(&ctxt, 0, sizeof(ctxt)); Patches currently in stable-queue which might be from longli(a)microsoft.com are queue-4.9/pci-hv-lock-pci-bus-on-device-eject.patch queue-4.9/pci-hv-properly-handle-pci-bus-remove.patch

6 years, 10 months

1
0
0 0

Patch "PCI: Apply Cavium ACS quirk only to CN81xx/CN83xx/CN88xx devices" has been added to the 4.9-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled PCI: Apply Cavium ACS quirk only to CN81xx/CN83xx/CN88xx devices to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: pci-apply-cavium-acs-quirk-only-to-cn81xx-cn83xx-cn88xx-devices.patch and it can be found in the queue-4.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From foo@baz Sun Mar 18 16:55:33 CET 2018 From: Manish Jaggi <mjaggi(a)caviumnetworks.com> Date: Thu, 30 Mar 2017 18:47:14 -0500 Subject: PCI: Apply Cavium ACS quirk only to CN81xx/CN83xx/CN88xx devices From: Manish Jaggi <mjaggi(a)caviumnetworks.com> [ Upstream commit b77d537d00d08fcf0bf641cd3491dd7df0ad1475 ] Only apply the Cavium ACS quirk to devices with ID in the range 0xa000-0xa0ff. These are the on-chip PCI devices for CN81xx/CN83xx/CN88xx. Fixes: b404bcfbf035 ("PCI: Add ACS quirk for all Cavium devices") Reported-by: Alex Williamson <alex.williamson(a)redhat.com> Signed-off-by: Manish Jaggi <mjaggi(a)cavium.com> Acked-by: David Daney <david.daney(a)cavium.com> Acked-by: Alex Williamson <alex.williamson(a)redhat.com> Signed-off-by: Sasha Levin <alexander.levin(a)microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/pci/quirks.c | 3 +++ 1 file changed, 3 insertions(+) --- a/drivers/pci/quirks.c +++ b/drivers/pci/quirks.c @@ -4104,6 +4104,9 @@ static int pci_quirk_cavium_acs(struct p */ acs_flags &= ~(PCI_ACS_RR | PCI_ACS_CR | PCI_ACS_SV | PCI_ACS_UF); + if (!((dev->device >= 0xa000) && (dev->device <= 0xa0ff))) + return -ENOTTY; + return acs_flags ? 0 : 1; } Patches currently in stable-queue which might be from mjaggi(a)caviumnetworks.com are queue-4.9/pci-apply-cavium-acs-quirk-only-to-cn81xx-cn83xx-cn88xx-devices.patch

6 years, 10 months

1
0
0 0

Patch "omapfb: dss: Handle return errors in dss_init_ports()" has been added to the 4.9-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled omapfb: dss: Handle return errors in dss_init_ports() to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: omapfb-dss-handle-return-errors-in-dss_init_ports.patch and it can be found in the queue-4.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From foo@baz Sun Mar 18 16:55:33 CET 2018 From: Arvind Yadav <arvind.yadav.cs(a)gmail.com> Date: Mon, 20 Mar 2017 18:30:59 +0100 Subject: omapfb: dss: Handle return errors in dss_init_ports() From: Arvind Yadav <arvind.yadav.cs(a)gmail.com> [ Upstream commit 0348aaa34412e24ebe622a2b1b013e68d6ae5412 ] dss_init_ports() is not handling return errors from dpi_init_port() and sdi_init_port(). It is also always returning 0 currently which results in part of error handling code in dss_bind() being unused. Fix dss_init_ports() to handle return errors from dpi_init_port() and sdi_init_port(). Signed-off-by: Arvind Yadav <arvind.yadav.cs(a)gmail.com> Cc: tomi.valkeinen(a)ti.com [b.zolnierkie: fail early on errors, minor fixups] Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie(a)samsung.com> Signed-off-by: Sasha Levin <alexander.levin(a)microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/video/fbdev/omap2/omapfb/dss/dss.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) --- a/drivers/video/fbdev/omap2/omapfb/dss/dss.c +++ b/drivers/video/fbdev/omap2/omapfb/dss/dss.c @@ -941,11 +941,13 @@ static int dss_init_features(struct plat return 0; } +static void dss_uninit_ports(struct platform_device *pdev); + static int dss_init_ports(struct platform_device *pdev) { struct device_node *parent = pdev->dev.of_node; struct device_node *port; - int r; + int r, ret = 0; if (parent == NULL) return 0; @@ -972,17 +974,21 @@ static int dss_init_ports(struct platfor switch (port_type) { case OMAP_DISPLAY_TYPE_DPI: - dpi_init_port(pdev, port); + ret = dpi_init_port(pdev, port); break; case OMAP_DISPLAY_TYPE_SDI: - sdi_init_port(pdev, port); + ret = sdi_init_port(pdev, port); break; default: break; } - } while ((port = omapdss_of_get_next_port(parent, port)) != NULL); + } while (!ret && + (port = omapdss_of_get_next_port(parent, port)) != NULL); - return 0; + if (ret) + dss_uninit_ports(pdev); + + return ret; } static void dss_uninit_ports(struct platform_device *pdev) Patches currently in stable-queue which might be from arvind.yadav.cs(a)gmail.com are queue-4.9/omapfb-dss-handle-return-errors-in-dss_init_ports.patch

6 years, 10 months

1
0
0 0

Patch "of: fix of_device_get_modalias returned length when truncating buffers" has been added to the 4.9-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled of: fix of_device_get_modalias returned length when truncating buffers to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: of-fix-of_device_get_modalias-returned-length-when-truncating-buffers.patch and it can be found in the queue-4.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From foo@baz Sun Mar 18 16:55:33 CET 2018 From: Rob Herring <robh(a)kernel.org> Date: Mon, 16 Jan 2017 14:28:39 -0600 Subject: of: fix of_device_get_modalias returned length when truncating buffers From: Rob Herring <robh(a)kernel.org> [ Upstream commit bcf54d5385abaea9c8026aae6f4eeb348671a52d ] If the length of the modalias is greater than the buffer size, then the modalias is truncated. However the untruncated length is returned which will cause an error. Fix this to return the truncated length. If an error in the case was desired, then then we should just return -ENOMEM. The reality is no device will ever have 4KB of compatible strings to hit this case. Signed-off-by: Rob Herring <robh(a)kernel.org> Cc: Frank Rowand <frowand.list(a)gmail.com> Signed-off-by: Sasha Levin <alexander.levin(a)microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/of/device.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) --- a/drivers/of/device.c +++ b/drivers/of/device.c @@ -223,7 +223,7 @@ ssize_t of_device_get_modalias(struct de str[i] = '_'; } - return tsize; + return repend; } EXPORT_SYMBOL_GPL(of_device_get_modalias); Patches currently in stable-queue which might be from robh(a)kernel.org are queue-4.9/of-fix-of_device_get_modalias-returned-length-when-truncating-buffers.patch

6 years, 10 months

1
0
0 0

Patch "NFC: pn533: change order of free_irq and dev unregistration" has been added to the 4.9-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled NFC: pn533: change order of free_irq and dev unregistration to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: nfc-pn533-change-order-of-free_irq-and-dev-unregistration.patch and it can be found in the queue-4.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From foo@baz Sun Mar 18 16:55:33 CET 2018 From: Andrey Rusalin <arusalin(a)dev.rtsoft.ru> Date: Wed, 28 Dec 2016 20:10:57 +0300 Subject: NFC: pn533: change order of free_irq and dev unregistration From: Andrey Rusalin <arusalin(a)dev.rtsoft.ru> [ Upstream commit 068a496c4525c638ffab56449d905b88ef97fe32 ] Change order of free_irq and dev unregistration. It fixes situation when device already unregistered and an interrupt happens and nobody can handle it. Signed-off-by: Andrey Rusalin <arusalin(a)dev.rtsoft.ru> Signed-off-by: Samuel Ortiz <sameo(a)linux.intel.com> Signed-off-by: Sasha Levin <alexander.levin(a)microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/nfc/pn533/i2c.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) --- a/drivers/nfc/pn533/i2c.c +++ b/drivers/nfc/pn533/i2c.c @@ -242,10 +242,10 @@ static int pn533_i2c_remove(struct i2c_c dev_dbg(&client->dev, "%s\n", __func__); - pn533_unregister_device(phy->priv); - free_irq(client->irq, phy); + pn533_unregister_device(phy->priv); + return 0; } Patches currently in stable-queue which might be from arusalin(a)dev.rtsoft.ru are queue-4.9/nfc-pn533-change-order-of-free_irq-and-dev-unregistration.patch

6 years, 10 months

1
0
0 0

Patch "NFC: nfcmrvl: Include unaligned.h instead of access_ok.h" has been added to the 4.9-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled NFC: nfcmrvl: Include unaligned.h instead of access_ok.h to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: nfc-nfcmrvl-include-unaligned.h-instead-of-access_ok.h.patch and it can be found in the queue-4.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From foo@baz Sun Mar 18 16:55:33 CET 2018 From: Tobias Klauser <tklauser(a)distanz.ch> Date: Wed, 26 Oct 2016 11:00:12 +0200 Subject: NFC: nfcmrvl: Include unaligned.h instead of access_ok.h From: Tobias Klauser <tklauser(a)distanz.ch> [ Upstream commit d916d923724d59cde99ee588f15eec59dd863bbd ] Including linux/unaligned/access_ok.h causes the allmodconfig build on ia64 (and maybe others) to fail with the following warnings: include/linux/unaligned/access_ok.h:7:19: error: redefinition of 'get_unaligned_le16' include/linux/unaligned/access_ok.h:12:19: error: redefinition of 'get_unaligned_le32' include/linux/unaligned/access_ok.h:17:19: error: redefinition of 'get_unaligned_le64' include/linux/unaligned/access_ok.h:22:19: error: redefinition of 'get_unaligned_be16' include/linux/unaligned/access_ok.h:27:19: error: redefinition of 'get_unaligned_be32' include/linux/unaligned/access_ok.h:32:19: error: redefinition of 'get_unaligned_be64' include/linux/unaligned/access_ok.h:37:20: error: redefinition of 'put_unaligned_le16' include/linux/unaligned/access_ok.h:42:20: error: redefinition of 'put_unaligned_le32' include/linux/unaligned/access_ok.h:42:20: error: redefinition of 'put_unaligned_le64' include/linux/unaligned/access_ok.h:42:20: error: redefinition of 'put_unaligned_be16' include/linux/unaligned/access_ok.h:42:20: error: redefinition of 'put_unaligned_be32' include/linux/unaligned/access_ok.h:42:20: error: redefinition of 'put_unaligned_be64' Fix these by including asm/unaligned.h instead and leave it up to the architecture to decide how to implement unaligned accesses. Fixes: 3194c6870158 ("NFC: nfcmrvl: add firmware download support") Reported-by: kbuild test robot <fengguang.wu(a)intel.com> Link: https://lkml.org/lkml/2016/10/22/247 Cc: Vincent Cuissard <cuissard(a)marvell.com> Signed-off-by: Tobias Klauser <tklauser(a)distanz.ch> Signed-off-by: Samuel Ortiz <sameo(a)linux.intel.com> Signed-off-by: Sasha Levin <alexander.levin(a)microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/nfc/nfcmrvl/fw_dnld.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) --- a/drivers/nfc/nfcmrvl/fw_dnld.c +++ b/drivers/nfc/nfcmrvl/fw_dnld.c @@ -17,7 +17,7 @@ */ #include <linux/module.h> -#include <linux/unaligned/access_ok.h> +#include <asm/unaligned.h> #include <linux/firmware.h> #include <linux/nfc.h> #include <net/nfc/nci.h> Patches currently in stable-queue which might be from tklauser(a)distanz.ch are queue-4.9/nfc-nfcmrvl-include-unaligned.h-instead-of-access_ok.h.patch

6 years, 10 months

1
0
0 0

Patch "NFC: nfcmrvl: double free on error path" has been added to the 4.9-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled NFC: nfcmrvl: double free on error path to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: nfc-nfcmrvl-double-free-on-error-path.patch and it can be found in the queue-4.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From foo@baz Sun Mar 18 16:55:33 CET 2018 From: Dan Carpenter <dan.carpenter(a)oracle.com> Date: Wed, 8 Mar 2017 08:22:37 +0300 Subject: NFC: nfcmrvl: double free on error path From: Dan Carpenter <dan.carpenter(a)oracle.com> [ Upstream commit ca42fb9e52d155547e6cf18cf26bce3e1a6af4ea ] The nci_spi_send() function calls kfree_skb(skb) on both error and success so this extra kfree_skb() is a double free. Fixes: caf6e49bf6d0 ("NFC: nfcmrvl: add spi driver") Signed-off-by: Dan Carpenter <dan.carpenter(a)oracle.com> Signed-off-by: Samuel Ortiz <sameo(a)linux.intel.com> Signed-off-by: Sasha Levin <alexander.levin(a)microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/nfc/nfcmrvl/spi.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) --- a/drivers/nfc/nfcmrvl/spi.c +++ b/drivers/nfc/nfcmrvl/spi.c @@ -96,10 +96,9 @@ static int nfcmrvl_spi_nci_send(struct n /* Send the SPI packet */ err = nci_spi_send(drv_data->nci_spi, &drv_data->handshake_completion, skb); - if (err != 0) { + if (err) nfc_err(priv->dev, "spi_send failed %d", err); - kfree_skb(skb); - } + return err; } Patches currently in stable-queue which might be from dan.carpenter(a)oracle.com are queue-4.9/media-cpia2-fix-a-couple-off-by-one-bugs.patch queue-4.9/nfc-nfcmrvl-double-free-on-error-path.patch queue-4.9/asoc-nuc900-fix-a-loop-timeout-test.patch

6 years, 10 months

1
0
0 0

Patch "netem: apply correct delay when rate throttling" has been added to the 4.9-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled netem: apply correct delay when rate throttling to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: netem-apply-correct-delay-when-rate-throttling.patch and it can be found in the queue-4.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From foo@baz Sun Mar 18 16:55:33 CET 2018 From: Nik Unger <njunger(a)uwaterloo.ca> Date: Mon, 13 Mar 2017 10:16:58 -0700 Subject: netem: apply correct delay when rate throttling From: Nik Unger <njunger(a)uwaterloo.ca> [ Upstream commit 5080f39e8c72e01cf37e8359023e7018e2a4901e ] I recently reported on the netem list that iperf network benchmarks show unexpected results when a bandwidth throttling rate has been configured for netem. Specifically: 1) The measured link bandwidth *increases* when a higher delay is added 2) The measured link bandwidth appears higher than the specified limit 3) The measured link bandwidth for the same very slow settings varies significantly across machines The issue can be reproduced by using tc to configure netem with a 512kbit rate and various (none, 1us, 50ms, 100ms, 200ms) delays on a veth pair between network namespaces, and then using iperf (or any other network benchmarking tool) to test throughput. Complete detailed instructions are in the original email chain here: https://lists.linuxfoundation.org/pipermail/netem/2017-February/001672.html There appear to be two underlying bugs causing these effects: - The first issue causes long delays when the rate is slow and no delay is configured (e.g., "rate 512kbit"). This is because SKBs are not orphaned when no delay is configured, so orphaning does not occur until *after* the rate-induced delay has been applied. For this reason, adding a tiny delay (e.g., "rate 512kbit delay 1us") dramatically increases the measured bandwidth. - The second issue is that rate-induced delays are not correctly applied, allowing SKB delays to occur in parallel. The indended approach is to compute the delay for an SKB and to add this delay to the end of the current queue. However, the code does not detect existing SKBs in the queue due to improperly testing sch->q.qlen, which is nonzero even when packets exist only in the rbtree. Consequently, new SKBs do not wait for the current queue to empty. When packet delays vary significantly (e.g., if packet sizes are different), then this also causes unintended reordering. I modified the code to expect a delay (and orphan the SKB) when a rate is configured. I also added some defensive tests that correctly find the latest scheduled delivery time, even if it is (unexpectedly) for a packet in sch->q. I have tested these changes on the latest kernel (4.11.0-rc1+) and the iperf / ping test results are as expected. Signed-off-by: Nik Unger <njunger(a)uwaterloo.ca> Signed-off-by: Stephen Hemminger <stephen(a)networkplumber.org> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Sasha Levin <alexander.levin(a)microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- net/sched/sch_netem.c | 26 ++++++++++++++++++-------- 1 file changed, 18 insertions(+), 8 deletions(-) --- a/net/sched/sch_netem.c +++ b/net/sched/sch_netem.c @@ -462,7 +462,7 @@ static int netem_enqueue(struct sk_buff /* If a delay is expected, orphan the skb. (orphaning usually takes * place at TX completion time, so _before_ the link transit delay) */ - if (q->latency || q->jitter) + if (q->latency || q->jitter || q->rate) skb_orphan_partial(skb); /* @@ -530,21 +530,31 @@ static int netem_enqueue(struct sk_buff now = psched_get_time(); if (q->rate) { - struct sk_buff *last; + struct netem_skb_cb *last = NULL; + + if (sch->q.tail) + last = netem_skb_cb(sch->q.tail); + if (q->t_root.rb_node) { + struct sk_buff *t_skb; + struct netem_skb_cb *t_last; + + t_skb = netem_rb_to_skb(rb_last(&q->t_root)); + t_last = netem_skb_cb(t_skb); + if (!last || + t_last->time_to_send > last->time_to_send) { + last = t_last; + } + } - if (sch->q.qlen) - last = sch->q.tail; - else - last = netem_rb_to_skb(rb_last(&q->t_root)); if (last) { /* * Last packet in queue is reference point (now), * calculate this time bonus and subtract * from delay. */ - delay -= netem_skb_cb(last)->time_to_send - now; + delay -= last->time_to_send - now; delay = max_t(psched_tdiff_t, 0, delay); - now = netem_skb_cb(last)->time_to_send; + now = last->time_to_send; } delay += packet_len_2_sched_time(qdisc_pkt_len(skb), q); Patches currently in stable-queue which might be from njunger(a)uwaterloo.ca are queue-4.9/netem-apply-correct-delay-when-rate-throttling.patch

6 years, 10 months

1
0
0 0

Patch "net: xfrm: allow clearing socket xfrm policies." has been added to the 4.9-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled net: xfrm: allow clearing socket xfrm policies. to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: net-xfrm-allow-clearing-socket-xfrm-policies.patch and it can be found in the queue-4.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From foo@baz Sun Mar 18 16:55:33 CET 2018 From: Lorenzo Colitti <lorenzo(a)google.com> Date: Mon, 20 Nov 2017 19:26:02 +0900 Subject: net: xfrm: allow clearing socket xfrm policies. From: Lorenzo Colitti <lorenzo(a)google.com> [ Upstream commit be8f8284cd897af2482d4e54fbc2bdfc15557259 ] Currently it is possible to add or update socket policies, but not clear them. Therefore, once a socket policy has been applied, the socket cannot be used for unencrypted traffic. This patch allows (privileged) users to clear socket policies by passing in a NULL pointer and zero length argument to the {IP,IPV6}_{IPSEC,XFRM}_POLICY setsockopts. This results in both the incoming and outgoing policies being cleared. The simple approach taken in this patch cannot clear socket policies in only one direction. If desired this could be added in the future, for example by continuing to pass in a length of zero (which currently is guaranteed to return EMSGSIZE) and making the policy be a pointer to an integer that contains one of the XFRM_POLICY_{IN,OUT} enum values. An alternative would have been to interpret the length as a signed integer and use XFRM_POLICY_IN (i.e., 0) to clear the input policy and -XFRM_POLICY_OUT (i.e., -1) to clear the output policy. Tested: https://android-review.googlesource.com/539816 Signed-off-by: Lorenzo Colitti <lorenzo(a)google.com> Signed-off-by: Steffen Klassert <steffen.klassert(a)secunet.com> Signed-off-by: Sasha Levin <alexander.levin(a)microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- net/xfrm/xfrm_policy.c | 2 +- net/xfrm/xfrm_state.c | 7 +++++++ 2 files changed, 8 insertions(+), 1 deletion(-) --- a/net/xfrm/xfrm_policy.c +++ b/net/xfrm/xfrm_policy.c @@ -1346,7 +1346,7 @@ EXPORT_SYMBOL(xfrm_policy_delete); int xfrm_sk_policy_insert(struct sock *sk, int dir, struct xfrm_policy *pol) { - struct net *net = xp_net(pol); + struct net *net = sock_net(sk); struct xfrm_policy *old_pol; #ifdef CONFIG_XFRM_SUB_POLICY --- a/net/xfrm/xfrm_state.c +++ b/net/xfrm/xfrm_state.c @@ -1883,6 +1883,13 @@ int xfrm_user_policy(struct sock *sk, in struct xfrm_mgr *km; struct xfrm_policy *pol = NULL; + if (!optval && !optlen) { + xfrm_sk_policy_insert(sk, XFRM_POLICY_IN, NULL); + xfrm_sk_policy_insert(sk, XFRM_POLICY_OUT, NULL); + __sk_dst_reset(sk); + return 0; + } + if (optlen <= 0 || optlen > PAGE_SIZE) return -EMSGSIZE; Patches currently in stable-queue which might be from lorenzo(a)google.com are queue-4.9/net-xfrm-allow-clearing-socket-xfrm-policies.patch

6 years, 10 months

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror March 2018