EDAC/i10nm: Add Intel Emerald Rapids server support
BugLink: https://bugs.launchpad.net/bugs/2028746
The Emerald Rapids CPU model uses similar memory controller registers
as Sapphire Rapids server. Add Emerald Rapids CPU model number ID for
EDAC support.
Tested-by: Li Zhang <li4.zhang@intel.com> Signed-off-by: Qiuxu Zhuo <qiuxu.zhuo@intel.com> Signed-off-by: Tony Luck <tony.luck@intel.com> Link: https://lore.kernel.org/all/20230113032802.41752-1-qiuxu.zhuo@intel.com
(cherry picked from commit e4b2bc6616e21f4a7ce4e7452f716e3db8fe66b6) Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
EDAC/skx_common: Delete duplicated and unreachable code
BugLink: https://bugs.launchpad.net/bugs/2028746
skx_mce_check_error() returns early if the error isn't from memory.
So when skx_mce_output_error() is invoked from skx_mce_check_error(),
it doesn't need to re-check whether the error is from memory. Delete
the duplicated and unreachable code from skx_mce_output_error().
Signed-off-by: Qiuxu Zhuo <qiuxu.zhuo@intel.com> Signed-off-by: Tony Luck <tony.luck@intel.com> Link: https://lore.kernel.org/all/20230113032802.41752-1-qiuxu.zhuo@intel.com
(cherry picked from commit d2415e2e5330fc11f4c688fa518751bdc90259f5) Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
EDAC/skx_common: Enable EDAC support for the "near" memory
BugLink: https://bugs.launchpad.net/bugs/2028746
The current {skx,i10nm}_edac miss the EDAC support to decode errors from
the 1st level memory (the fast "near" memory as cache) of the 2-level
memory system. Introduce a helper function skx_error_in_mem() to check
whether errors are from memory at the beginning of skx_mce_check_error().
As long as the errors are from memory (either the 1-level memory system
or the 2-level memory system), decode the errors.
Reported-and-tested-by: Youquan Song <youquan.song@intel.com> Signed-off-by: Qiuxu Zhuo <qiuxu.zhuo@intel.com> Signed-off-by: Tony Luck <tony.luck@intel.com> Link: https://lore.kernel.org/all/20230113032802.41752-1-qiuxu.zhuo@intel.com
(cherry picked from commit 6e8746cb735166eaf3ceb086b31bb0431f5e3532) Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
video/aperture: Provide a VGA helper for gma500 and internal use
BugLink: https://bugs.launchpad.net/bugs/2028749
The hardware for gma500 is different from the rest, as it uses stolen
framebuffer memory that is not available via PCI BAR. The regular PCI
removal helper cannot detect the framebuffer, while the non-PCI helper
misses possible conflicting VGA devices (i.e., a framebuffer or text
console).
Gma500 therefore calls both helpers to catch all cases. It's confusing
as it implies that there's something about the PCI device that requires
ownership management. The relationship between the PCI device and the
VGA devices is non-obvious. At worst, readers might assume that calling
two functions for clearing aperture ownership is a bug in the driver.
Hence, move the PCI removal helper's code for VGA functionality into
a separate function and call this function from gma500. Documents the
purpose of each call to aperture helpers. The change contains comments
and example code form the discussion at [1].
Daniel Vetter [Wed, 26 Jul 2023 08:19:00 +0000 (10:19 +0200)]
fbdev: Simplify fb_is_primary_device for x86
BugLink: https://bugs.launchpad.net/bugs/2028749
vga_default_device really is supposed to cover all corners, at least
for x86. Additionally checking for rom shadowing should be redundant,
because the bios/fw only does that for the boot vga device.
If this turns out to be wrong, then most likely that's a special case
which should be added to the vgaarb code, not replicated all over.
Patch motived by changes to the aperture helpers, which also have this
open code in a bunch of places, and which are all removed in a
clean-up series. This function here is just for selecting the default
fbdev device for fbcon, but I figured for consistency it might be good
to throw this patch in on top.
Note that the shadow rom check predates vgaarb, which was only wired
up in commit 88674088d10c ("x86: Use vga_default_device() when
determining whether an fb is primary"). That patch doesn't explain why
we still fall back to the shadow rom check.
Daniel Vetter [Wed, 26 Jul 2023 08:19:00 +0000 (10:19 +0200)]
video/aperture: Only remove sysfb on the default vga pci device
BugLink: https://bugs.launchpad.net/bugs/2028749
Instead of calling aperture_remove_conflicting_devices() to remove the
conflicting devices, just call to aperture_detach_devices() to detach
the device that matches the same PCI BAR / aperture range. Since the
former is just a wrapper of the latter plus a sysfb_disable() call,
and now that's done in this function but only for the primary devices.
This fixes a regression introduced by commit ee7a69aa38d8 ("fbdev:
Disable sysfb device registration when removing conflicting FBs"),
where we remove the sysfb when loading a driver for an unrelated pci
device, resulting in the user losing their efifb console or similar.
Note that in practice this only is a problem with the nvidia blob,
because that's the only gpu driver people might install which does not
come with an fbdev driver of it's own. For everyone else the real gpu
driver will restore a working console.
Also note that in the referenced bug there's confusion that this same
bug also happens on amdgpu. But that was just another amdgpu specific
regression, which just happened to happen at roughly the same time and
with the same user-observable symptoms. That bug is fixed now, see
https://bugzilla.kernel.org/show_bug.cgi?id=216331#c15
Note that we should not have any such issues on non-pci multi-gpu
issues, because I could only find two such cases:
- SoC with some external panel over spi or similar. These panel
drivers do not use drm_aperture_remove_conflicting_framebuffers(),
so no problem.
- vga+mga, which is a direct console driver and entirely bypasses all
this.
For the above reasons the cc: stable is just notionally, this patch
will need a backport and that's up to nvidia if they care enough.
v2:
- Explain a bit better why other multi-gpu that aren't pci shouldn't
have any issues with making all this fully pci specific.
Daniel Vetter [Wed, 26 Jul 2023 08:19:00 +0000 (10:19 +0200)]
video/aperture: Drop primary argument
BugLink: https://bugs.launchpad.net/bugs/2028749
With the preceding patches it's become defunct. Also I'm about to add
a different boolean argument, so it's better to keep the confusion
down to the absolute minimum.
v2: Since the hypervfb patch got droppped (it's only a pci device for
gen1 vm, not for gen2) there is one leftover user in an actual driver
left to touch.
- It's really the only one where this matters. I tried looking around,
and I didn't find any non-pci vga-compatible controllers for x86
(since that's the only platform where we had this until a few
patches ago), where a driver participating in the aperture claim
dance would interfere.
- I also don't expect that any future bus anytime soon will
not just look like pci towards the OS, that's been the case for like
25+ years by now for practically everything (even non non-x86).
- Also it's a bit funny if we have one part of the vga removal in the
pci function, and the other in the generic one.
v2: Rebase.
v4:
- fix Daniel's S-o-b address
v5:
- add back an S-o-b tag with Daniel's Intel address
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: Javier Martinez Canillas <javierm@redhat.com> Cc: Helge Deller <deller@gmx.de> Cc: linux-fbdev@vger.kernel.org Reviewed-by: Javier Martinez Canillas <javierm@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20230406132109.32050-6-tzimmermann@suse.de
(cherry picked from commit f1d599d315fb7b7343cddaf365e671aaa8453aca) Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Daniel Vetter [Wed, 26 Jul 2023 08:19:00 +0000 (10:19 +0200)]
video/aperture: Only kick vgacon when the pdev is decoding vga
BugLink: https://bugs.launchpad.net/bugs/2028749
Otherwise it's a bit silly, and we might throw out the driver for the
screen the user is actually looking at. I haven't found a bug report
for this case yet, but we did get bug reports for the analog case
where we're throwing out the efifb driver.
v2: Flip the check around to make it clear it's a special case for
kicking out the vgacon driver only (Thomas)
Daniel Vetter [Wed, 26 Jul 2023 08:19:00 +0000 (10:19 +0200)]
drm/aperture: Remove primary argument
BugLink: https://bugs.launchpad.net/bugs/2028749
Only really pci devices have a business setting this - it's for
figuring out whether the legacy vga stuff should be nuked too. And
with the preceding two patches those are all using the pci version of
this.
Which means for all other callers primary == false and we can remove
it now.
v2:
- Reorder to avoid compile fail (Thomas)
- Include gma500, which retained it's called to the non-pci version.
v4:
- fix Daniel's S-o-b address
v5:
- add back an S-o-b tag with Daniel's Intel address
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: Javier Martinez Canillas <javierm@redhat.com> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Maxime Ripard <mripard@kernel.org> Cc: Deepak Rawat <drawat.floss@gmail.com> Cc: Neil Armstrong <neil.armstrong@linaro.org> Cc: Kevin Hilman <khilman@baylibre.com> Cc: Jerome Brunet <jbrunet@baylibre.com> Cc: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Cc: Thierry Reding <thierry.reding@gmail.com> Cc: Jonathan Hunter <jonathanh@nvidia.com> Cc: Emma Anholt <emma@anholt.net> Cc: Helge Deller <deller@gmx.de> Cc: David Airlie <airlied@gmail.com> Cc: Daniel Vetter <daniel@ffwll.ch> Cc: linux-hyperv@vger.kernel.org Cc: linux-amlogic@lists.infradead.org Cc: linux-arm-kernel@lists.infradead.org Cc: linux-tegra@vger.kernel.org Cc: linux-fbdev@vger.kernel.org Acked-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Acked-by: Thierry Reding <treding@nvidia.com> Reviewed-by: Javier Martinez Canillas <javierm@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20230406132109.32050-4-tzimmermann@suse.de
(backported from commit 62aeaeaa1b267c5149abee6b45967a5df3feed58)
[khfeng: msm_fbdev.c already uses different routine, include drivers] Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Daniel Vetter [Wed, 26 Jul 2023 08:19:00 +0000 (10:19 +0200)]
video/aperture: use generic code to figure out the vga default device
BugLink: https://bugs.launchpad.net/bugs/2028749
Since vgaarb has been promoted to be a core piece of the pci subsystem
we don't have to open code random guesses anymore, we actually know
this in a platform agnostic way, and there's no need for an x86
specific hack. See also commit 1d38fe6ee6a8 ("PCI/VGA: Move vgaarb to
drivers/pci")
This should not result in any functional change, and the non-x86
multi-gpu pci systems are probably rare enough to not matter (I don't
know of any tbh). But it's a nice cleanup, so let's do it.
There's been a few questions on previous iterations on dri-devel and
irc:
- fb_is_primary_device() seems to be yet another implementation of
this theme, and at least on x86 it checks for both
vga_default_device OR rom shadowing. There shouldn't ever be a case
where rom shadowing gives any additional hints about the boot vga
device, but if there is then the default vga selection in vgaarb
should probably be fixed. And not special-case checks replicated all
over.
- Thomas also brought up that on most !x86 systems
fb_is_primary_device() returns 0, except on sparc/parisc. But these
2 special cases are about platform specific devices and not pci, so
shouldn't have any interactions.
- Furthermore fb_is_primary_device() is a bit a red herring since it's
only used to select the right fbdev driver for fbcon, and not for
the fw handover dance which the aperture helpers handle. At least
for x86 we might want to look into unifying them, but that's a
separate thing.
v2: Extend commit message trying to summarize various discussions.
v4:
- make the test for the primary device easier to read (Javier)
- fix commit message style (i.e., commit 1234 ("..."))
- fix Daniel's S-o-b address
v5:
- add back an S-o-b tag with Daniel's Intel address
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: Javier Martinez Canillas <javierm@redhat.com> Cc: Helge Deller <deller@gmx.de> Cc: linux-fbdev@vger.kernel.org Cc: Bjorn Helgaas <bhelgaas@google.com> Cc: linux-pci@vger.kernel.org Reviewed-by: Javier Martinez Canillas <javierm@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20230406132109.32050-3-tzimmermann@suse.de
(cherry picked from commit db082219569e3392cd38c9e322151450c855eed4) Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Daniel Vetter [Wed, 26 Jul 2023 08:19:00 +0000 (10:19 +0200)]
drm/gma500: Use drm_aperture_remove_conflicting_pci_framebuffers
BugLink: https://bugs.launchpad.net/bugs/2028749
This one nukes all framebuffers, which is a bit much. In reality
gma500 is igpu and never shipped with anything discrete, so there should
not be any difference.
v2: Unfortunately the framebuffer sits outside of the pci bars for
gma500, and so only using the pci helpers won't be enough. Otoh if we
only use non-pci helper, then we don't get the vga handling, and
subsequent refactoring to untangle these special cases won't work.
It's not pretty, but the simplest fix (since gma500 really is the only
quirky pci driver like this we have) is to just have both calls.
v4:
- fix Daniel's S-o-b address
v5:
- add back an S-o-b tag with Daniel's Intel address
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de> Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com> Cc: Thomas Zimmermann <tzimmermann@suse.de> Cc: Javier Martinez Canillas <javierm@redhat.com> Reviewed-by: Javier Martinez Canillas <javierm@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20230406132109.32050-2-tzimmermann@suse.de
(cherry picked from commit 80e993988b97fe794f3ec2be6db05fe30f9353c3) Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Hans de Goede [Wed, 28 Jun 2023 02:42:00 +0000 (04:42 +0200)]
ACPI: video: Stop trying to use vendor backlight control on laptops from after ~2012
BugLink: https://bugs.launchpad.net/bugs/2023638
There have been 2 separate reports now about a non working
"dell_backlight" device getting registered under /sys/class/backlight
1 report for a Raptor Lake based Dell and 1 report for a Meteor Lake
(development) platform.
On hw from the last 10 years dell-laptop will not register "dell_backlight"
because acpi_video_get_backlight_type() will return acpi_backlight_video
there if called before the GPU/kms driver loads. So it does not matter if
the GPU driver's native backlight is registered after dell-laptop loads.
But it seems that on the latest generation laptops the ACPI tables
no longer contain acpi_video backlight control support which causes
acpi_video_get_backlight_type() to return acpi_backlight_vendor causing
"dell_backlight" to get registered if the dell-laptop module is loaded
before the GPU/kms driver.
Vendor specific backlight control like the "dell_backlight" device is
only necessary on quite old hw (from before acpi_video backlight control
was introduced). Work around "dell_backlight" registering on very new
hw (where acpi_video backlight control seems to be no more) by making
acpi_video_get_backlight_type() return acpi_backlight_none instead
of acpi_backlight_vendor as final fallback when the ACPI tables have
support for Windows 8 or later (laptops from after ~2012).
Suggested-by: Matthew Garrett <mjg59@srcf.ucam.org> Reported-by: AceLan Kao <acelan.kao@canonical.com> Closes: https://lore.kernel.org/platform-driver-x86/20230607034331.576623-1-acelan.kao@canonical.com/ Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
(cherry picked from commit aa8a950a5d6b2094830aff834198777371ff91ff) Signed-off-by: Chia-Lin Kao (AceLan) <acelan.kao@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Kai-Heng Feng [Thu, 25 May 2023 07:18:00 +0000 (09:18 +0200)]
net: wwan: t7xx: Ensure init is completed before system sleep
BugLink: https://launchpad.net/bugs/2020743
When the system attempts to sleep while mtk_t7xx is not ready, the driver
cannot put the device to sleep:
[ 12.472918] mtk_t7xx 0000:57:00.0: [PM] Exiting suspend, modem in invalid state
[ 12.472936] mtk_t7xx 0000:57:00.0: PM: pci_pm_suspend(): t7xx_pci_pm_suspend+0x0/0x20 [mtk_t7xx] returns -14
[ 12.473678] mtk_t7xx 0000:57:00.0: PM: dpm_run_callback(): pci_pm_suspend+0x0/0x1b0 returns -14
[ 12.473711] mtk_t7xx 0000:57:00.0: PM: failed to suspend async: error -14
[ 12.764776] PM: Some devices failed to suspend, or early wake event detected
Mediatek confirmed the device can take a rather long time to complete
its initialization, so wait for up to 20 seconds until init is done.
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com> Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit ab87603b251134441a67385ecc9d3371be17b7a7) Signed-off-by: Chia-Lin Kao (AceLan) <acelan.kao@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Bjørn Mork [Mon, 22 May 2023 07:57:00 +0000 (09:57 +0200)]
r8152: add USB device driver for config selection
BugLink: https://launchpad.net/bugs/2020295
Subclassing the generic USB device driver to override the
default configuration selection regardless of matching interface
drivers.
The r815x family devices expose a vendor specific function which
the r8152 interface driver wants to handle. This is the preferred
device mode. Additionally one or more USB class functions are
usually supported for hosts lacking a vendor specific driver. The
choice is USB configuration based, with one alternate function per
configuration.
Example device with both NCM and ECM alternate cfgs:
A problem with this is that Linux will prefer class functions over
vendor specific functions. Using the above example, Linux defaults
to cfg #2, running the device in a sub-optimal NCM mode.
Previously we've attempted to work around the problem by
blacklisting the devices in the ECM class driver "cdc_ether", and
matching on the ECM class function in the vendor specific interface
driver. The latter has been used to switch back to the vendor
specific configuration when the driver is probed for a class
function.
This workaround has several issues;
- class driver blacklists is additional maintanence cruft in an
unrelated driver
- class driver blacklists prevents users from optionally running
the devices in class mode
- each device needs double match entries in the vendor driver
- the initial probing as a class function slows down device
discovery
Now these issues have become even worse with the introduction of
firmware supporting both NCM and ECM, where NCM ends up as the
default mode in Linux. To use the same workaround, we now have
to blacklist the devices in to two different class drivers and
add yet another match entry to the vendor specific driver.
This patch implements an alternative workaround strategy -
independent of the interface drivers. It avoids adding a
blacklist to the cdc_ncm driver and will let us remove the
existing blacklist from the cdc_ether driver.
As an additional bonus, removing the blacklists allow users to
select one of the other device modes if wanted.
Signed-off-by: Bjørn Mork <bjorn@mork.no> Signed-off-by: David S. Miller <davem@davemloft.net>
(backported from commit ec51fbd1b8a2bca2948dede99c14ec63dc57ff6b)
[AceLan: conflicts with below commit which introduce a ID to the same table
in Jammy kernel. Fixed by preserved the ID and changed to the same new format. 2edbc1459c1b ("r8152: add vendor/device ID pair for Microsoft Devkit")] Signed-off-by: Chia-Lin Kao (AceLan) <acelan.kao@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Acked-by: Roxana Nicolescu <roxana.nicolescu@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Add a fix for the Zen2 VZEROUPPER data corruption bug where under
certain circumstances executing VZEROUPPER can cause register
corruption or leak data.
The optimal fix is through microcode but in the case the proper
microcode revision has not been applied, enable a fallback fix using
a chicken bit.
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
CVE-2023-20593
(backported from commit 522b1d69219d8f083173819fde04f994aa051a98)
[cascardo: small conflict due to missing commit 8cc68c9c9e92dbaae51a711454c66eb668045508] Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Acked-by: Cengiz Can <cengiz.can@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
x86/cpu/amd: Move the errata checking functionality up
Avoid new and remove old forward declarations.
No functional changes.
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
CVE-2023-20593
(backported from commit 8b6f687743dacce83dbb0c7cfacf88bab00f808a)
[cascardo: small context conflict at set_dr_addr_mask] Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com> Acked-by: Cengiz Can <cengiz.can@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
netfilter: nft_set_pipapo: fix improper element removal
end key should be equal to start unless NFT_SET_EXT_KEY_END is present.
Its possible to add elements that only have a start key
("{ 1.0.0.0 . 2.0.0.0 }") without an internval end.
Insertion treats this via:
if (nft_set_ext_exists(ext, NFT_SET_EXT_KEY_END))
end = (const u8 *)nft_set_ext_key_end(ext)->data;
else
end = start;
but removal side always uses nft_set_ext_key_end().
This is wrong and leads to garbage remaining in the set after removal
next lookup/insert attempt will give:
BUG: KASAN: slab-use-after-free in pipapo_get+0x8eb/0xb90
Read of size 1 at addr ffff888100d50586 by task nft-pipapo_uaf_/1399
Call Trace:
kasan_report+0x105/0x140
pipapo_get+0x8eb/0xb90
nft_pipapo_insert+0x1dc/0x1710
nf_tables_newsetelem+0x31f5/0x4e00
..
Fixes: 3c4287f62044 ("nf_tables: Add set type for arbitrary concatenation of ranges") Reported-by: lonial con <kongln9170@gmail.com> Reviewed-by: Stefano Brivio <sbrivio@redhat.com> Signed-off-by: Florian Westphal <fw@strlen.de>
CVE-2023-4004
(cherry picked from commit 87b5a5c209405cb6b57424cdfa226a6dbd349232) Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com> Acked-by: Cengiz Can <cengiz.can@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Pedro Tammela [Thu, 27 Jul 2023 23:22:00 +0000 (01:22 +0200)]
net/sched: sch_qfq: account for stab overhead in qfq_enqueue
Lion says:
-------
In the QFQ scheduler a similar issue to CVE-2023-31436
persists.
Consider the following code in net/sched/sch_qfq.c:
static int qfq_enqueue(struct sk_buff *skb, struct Qdisc *sch,
struct sk_buff **to_free)
{
unsigned int len = qdisc_pkt_len(skb), gso_segs;
// ...
if (unlikely(cl->agg->lmax < len)) {
pr_debug("qfq: increasing maxpkt from %u to %u for class %u",
cl->agg->lmax, len, cl->common.classid);
err = qfq_change_agg(sch, cl, cl->agg->class_weight, len);
if (err) {
cl->qstats.drops++;
return qdisc_drop(skb, sch, to_free);
}
// ...
}
Similarly to CVE-2023-31436, "lmax" is increased without any bounds
checks according to the packet length "len". Usually this would not
impose a problem because packet sizes are naturally limited.
This is however not the actual packet length, rather the
"qdisc_pkt_len(skb)" which might apply size transformations according to
"struct qdisc_size_table" as created by "qdisc_get_stab()" in
net/sched/sch_api.c if the TCA_STAB option was set when modifying the qdisc.
A user may choose virtually any size using such a table.
As a result the same issue as in CVE-2023-31436 can occur, allowing heap
out-of-bounds read / writes in the kmalloc-8192 cache.
-------
We can create the issue with the following commands:
tc qdisc add dev $DEV root handle 1: stab mtu 2048 tsize 512 mpu 0 \
overhead 999999999 linklayer ethernet qfq
tc class add dev $DEV parent 1: classid 1:1 htb rate 6mbit burst 15k
tc filter add dev $DEV parent 1: matchall classid 1:1
ping -I $DEV 1.1.1.2
This is caused by incorrectly assuming that qdisc_pkt_len() returns a
length within the QFQ_MIN_LMAX < len < QFQ_MAX_LMAX.
Fixes: 462dbc9101ac ("pkt_sched: QFQ Plus: fair-queueing service at DRR cost") Reported-by: Lion <nnamrec@gmail.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: Pedro Tammela <pctammela@mojatatu.com> Reviewed-by: Simon Horman <simon.horman@corigine.com> Signed-off-by: Paolo Abeni <pabeni@redhat.com>
CVE-2023-3611
(cherry picked from commit 3e337087c3b5805fe0b8a46ba622a962880b5d64) Signed-off-by: Cengiz Can <cengiz.can@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Acked-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Pedro Tammela [Thu, 27 Jul 2023 23:22:00 +0000 (01:22 +0200)]
net/sched: sch_qfq: refactor parsing of netlink parameters
Two parameters can be transformed into netlink policies and
validated while parsing the netlink message.
Reviewed-by: Simon Horman <simon.horman@corigine.com> Acked-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: Pedro Tammela <pctammela@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
CVE-2023-3611
(cherry picked from commit 25369891fcef373540f8b4e0b3bccf77a04490d5)
[cengizcan: prerequisite commit] Signed-off-by: Cengiz Can <cengiz.can@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Acked-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Add bound flag to rule and chain transactions as in 6a0a8d10a366
("netfilter: nf_tables: use-after-free in failing rule with bound set")
to skip them in case that the chain is already bound from the abort
path.
This patch fixes an imbalance in the chain use refcnt that triggers a
WARN_ON on the table and chain destroy path.
This patch also disallows nested chain bindings, which is not
supported from userspace.
The logic to deal with chain binding in nft_data_hold() and
nft_data_release() is not correct. The NFT_TRANS_PREPARE state needs a
special handling in case a chain is bound but next expressions in the
same rule fail to initialize as described by 1240eb93f061 ("netfilter:
nf_tables: incorrect error path handling with NFT_MSG_NEWRULE").
The chain is left bound if rule construction fails, so the objects
stored in this chain (and the chain itself) are released by the
transaction records from the abort path, follow up patch ("netfilter:
nf_tables: add NFT_TRANS_PREPARE_ERROR to deal with bound set/chain")
completes this error handling.
When deleting an existing rule, chain bound flag is set off so the
rule expression .destroy path releases the objects.
Fixes: d0e2c7de92c7 ("netfilter: nf_tables: add NFT_CHAIN_BINDING") Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
CVE-2023-3610
(cherry picked from commit 4bedf9eee016286c835e3d8fa981ddece5338795) Signed-off-by: Cengiz Can <cengiz.can@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Chao Yu [Tue, 1 Aug 2023 22:08:00 +0000 (00:08 +0200)]
f2fs: fix to avoid NULL pointer dereference f2fs_write_end_io()
butt3rflyh4ck reports a bug as below:
When a thread always calls F2FS_IOC_RESIZE_FS to resize fs, if resize fs is
failed, f2fs kernel thread would invoke callback function to update f2fs io
info, it would call f2fs_write_end_io and may trigger null-ptr-deref in
NODE_MAPPING.
The root cause is below race case can cause leaving dirty metadata
in f2fs after filesystem is remount as ro:
Thread A Thread B
- f2fs_ioc_resize_fs
- f2fs_readonly --- return false
- f2fs_resize_fs
- f2fs_remount
- write_checkpoint
- set f2fs as ro
- free_segment_range
- update meta_inode's data
Then, if f2fs_put_super() fails to write_checkpoint due to readonly
status, and meta_inode's dirty data will be writebacked after node_inode
is put, finally, f2fs_write_end_io will access NULL pointer on
sbi->node_inode.
Chris Chiu [Mon, 24 Jul 2023 15:26:00 +0000 (17:26 +0200)]
ALSA: hda/realtek: Enable 4 amplifiers instead of 2 on a HP platform
BugLink: https://bugs.launchpad.net/bugs/2023197
In the commit 7bb62340951a ("ALSA: hda/realtek: fix speaker, mute/micmute
LEDs not work on a HP platform"), speakers and LEDs are fixed but only 2
CS35L41 amplifiers on SPI bus connected to Realtek codec are enabled. Need
the ALC245_FIXUP_CS35L41_SPI_4_HP_GPIO_LED to get all amplifiers working.
Signed-off-by: Chris Chiu <chris.chiu@canonical.com> Fixes: 7bb62340951a ("ALSA: hda/realtek: fix speaker, mute/micmute LEDs not work on a HP platform") Cc: <stable@vger.kernel.org> Link: https://lore.kernel.org/r/20230606145747.135966-1-chris.chiu@canonical.com Signed-off-by: Takashi Iwai <tiwai@suse.de>
(cherry picked from commit b752a385b584d385683c65cb76a1298f1379a88c) Signed-off-by: Chris Chiu <chris.chiu@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Acked-by: Roxana Nicolescu <roxana.nicolescu@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
BugLink: https://bugs.launchpad.net/bugs/2016398
Opened files reported in /proc/pid/map_files can be shows with the wrong
mount point using overlayfs with filesystem namspaces.
This incorrect behavior is fixed:
UBUNTU: SAUCE: overlayfs: fix incorrect mnt_id of files opened from map_files
However, the fix introduced a new regression, the reference to the
original file stored in vma->vm_prfile is not properly released when
vma->vm_prfile is replaced with a new file.
This can cause a reference counter unbalance, leading errors such as
"target is busy" when trying to unmount overlayfs, even if the
filesystem has not active reference.
Fix by properly releasing the original file stored in vm_prfile.
Fixes: 508fdae3f62dd ("UBUNTU: SAUCE: overlayfs: fix incorrect mnt_id of files opened from map_files") Signed-off-by: Andrea Righi <andrea.righi@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Acked-by: Roxana Nicolescu <roxana.nicolescu@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Ville Syrjälä [Wed, 19 Jul 2023 13:11:25 +0000 (21:11 +0800)]
drm/i915: Allow arbitrary refresh rates with VRR eDP panels
BugLink: https://bugs.launchpad.net/bugs/2024273
If the panel supports VRR it must be capable of accepting
timings with arbitrary vblank length, within the valid VRR
range. Use that fact to allow the user to request any refresh
rate they like. We simply pick the next highest fixed mode
from our list, and adjust the vblank to get the desired refresh
rate in the end.
Of course currently everything to do with the vrefresh is
using 1Hz precision, so might not be exact. But we can improve
that in the future by just upping our vrefresh precision.
drm/amd: Align SMU11 SMU_MSG_OverridePcieParameters implementation with SMU13
BugLink: https://bugs.launchpad.net/bugs/2027957
SMU13 overrides dynamic PCIe lane width and dynamic speed by when on
certain hosts. commit 38e4ced80479 ("drm/amd/pm: conditionally disable
pcie lane switching for some sienna_cichlid SKUs") worked around this
issue by setting up certain SKUs to set up certain limits, but the same
fundamental problem with those hosts affects all SMU11 implmentations
as well, so align the SMU11 and SMU13 driver handling.
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Reviewed-by: Evan Quan <evan.quan@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org # 6.1.x
(backported from commit e701156ccc6c7a5f104a968dda74cd6434178712
linux-next) Signed-off-by: Koba Ko <koba.ko@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Signed-off-by: Roxana Nicolescu <roxana.nicolescu@canonical.com>
drm/amd: Move helper for dynamic speed switch check out of smu13
BugLink: https://bugs.launchpad.net/bugs/2027957
This helper is used for checking if the connected host supports
the feature, it can be moved into generic code to be used by other
smu implementations as well.
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Reviewed-by: Evan Quan <evan.quan@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org # 6.1.x
(cherry picked from commit 188623076d0f1a500583d392b6187056bf7cc71a
linux-next) Signed-off-by: Koba Ko <koba.ko@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Signed-off-by: Roxana Nicolescu <roxana.nicolescu@canonical.com>
drm/amd/pm: conditionally disable pcie lane/speed switching for SMU13
BugLink: https://bugs.launchpad.net/bugs/2027957
Intel platforms such as Sapphire Rapids and Raptor Lake don't support
dynamic pcie lane or speed switching.
This limitation seems to carry over from one generation to another.
To be safer, disable dynamic pcie lane width and speed switching when
running on an Intel platform.
BugLink: https://bugs.launchpad.net/bugs/2027773
Shift operation of 'exp' and 'shift' variables exceeds the maximum number
of shift values in the u32 range leading to UBSAN shift-out-of-bounds.
BugLink: https://bugs.launchpad.net/bugs/2027773
As float32 is also used in other places as a data type, it is necessary
to rename the float32 variable in order to avoid confusion.
Cc: stable@vger.kernel.org Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com> Signed-off-by: Basavaraj Natikar <Basavaraj.Natikar@amd.com> Signed-off-by: Akshata MukundShetty <akshata.mukundshetty@amd.com> Link: https://lore.kernel.org/r/20230707065722.9036-2-Basavaraj.Natikar@amd.com Signed-off-by: Benjamin Tissoires <bentiss@kernel.org>
(cherry picked from commit c1685a862a4bea863537f06abaa37a123aef493c) Signed-off-by: You-Sheng Yang <vicamo.yang@canonical.com> Acked-by: John Cabaj <john.cabaj@canonical.com> Acked-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Roxana Nicolescu <roxana.nicolescu@canonical.com>
cifs: fix mid leak during reconnection after timeout threshold
BugLink: https://bugs.launchpad.net/bugs/2029138
When the number of responses with status of STATUS_IO_TIMEOUT
exceeds a specified threshold (NUM_STATUS_IO_TIMEOUT), we reconnect
the connection. But we do not return the mid, or the credits
returned for the mid, or reduce the number of in-flight requests.
This bug could result in the server->in_flight count to go bad,
and also cause a leak in the mids.
This change moves the check to a few lines below where the
response is decrypted, even of the response is read from the
transform header. This way, the code for returning the mids
can be reused.
Also, the cifs_reconnect was reconnecting just the transport
connection before. In case of multi-channel, this may not be
what we want to do after several timeouts. Changed that to
reconnect the session and the tree too.
Also renamed NUM_STATUS_IO_TIMEOUT to a more appropriate name
MAX_STATUS_IO_TIMEOUT.
Fixes: 8e670f77c4a5 ("Handle STATUS_IO_TIMEOUT gracefully") Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
(cherry picked from commit 69cba9d3c1284e0838ae408830a02c4a063104bc) Signed-off-by: Tim Gardner <tim.gardner@canonical.com> Acked-by: Cengiz Can <cengiz.can@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Signed-off-by: Roxana Nicolescu <roxana.nicolescu@canonical.com>
Kamal Mostafa [Fri, 28 Jul 2023 20:19:29 +0000 (13:19 -0700)]
UBUNTU: Upstream stable to v6.1.32, v6.3.6
BugLink: https://bugs.launchpad.net/bugs/2028979 Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
commit b11d31ae01e6 ("blk-wbt: remove unnecessary check in
wbt_enable_default()") removes the checking of CONFIG_BLK_WBT_MQ by
mistake, which is used to control enable or disable wbt by default.
Fix the problem by adding back the checking. This patch also do a litter
cleanup to make related code more readable.
In devm_cxl_add_port() the port creation may fail and its associated
pointer does not contain a valid address. During error message
generation this invalid port address is used. Fix that wrong address
access.
Fixes: f3cd264c4ec1 ("cxl: Unify debug messages when calling devm_cxl_add_port()") Signed-off-by: Robert Richter <rrichter@amd.com> Reviewed-by: Dave Jiang <dave.jiang@intel.com> Link: https://lore.kernel.org/r/20230519215436.3394532-1-rrichter@amd.com Signed-off-by: Dan Williams <dan.j.williams@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Two dma_wmb() are added in the XDP TX path to ensure proper ordering of
descriptor and buffer updates:
1. A dma_wmb() is added after updating the last BD to make sure
the updates to rest of the descriptor are visible before
transferring ownership to FEC.
2. A dma_wmb() is also added after updating the bdp to ensure these
updates are visible before updating txq->bd.cur.
3. Start the xmit of the frame immediately right after configuring the
tx descriptor.
Fixes: 6d6b39f180b8 ("net: fec: add initial XDP support") Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com> Reviewed-by: Wei Fang <wei.fang@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
If static allocation and dynamic allocation GPIOs are present,
dynamic allocation pollutes the numberspace for static allocation,
causing static allocation to fail.
Enforce dynamic allocation above GPIO_DYNAMIC_BASE.
Seen on a GTA04 when omap-gpio (static) and twl-gpio (dynamic)
raced:
[some successful registrations of omap_gpio instances]
[ 2.553833] twl4030_gpio twl4030-gpio: gpio (irq 145) chaining IRQs 161..178
[ 2.561401] gpiochip_find_base: found new base at 160
[ 2.564392] gpio gpiochip5: (twl4030): added GPIO chardev (254:5)
[ 2.564544] gpio gpiochip5: registered GPIOs 160 to 177 on twl4030
[...]
[ 2.692169] omap-gpmc 6e000000.gpmc: GPMC revision 5.0
[ 2.697357] gpmc_mem_init: disabling cs 0 mapped at 0x0-0x1000000
[ 2.703643] gpiochip_find_base: found new base at 178
[ 2.704376] gpio gpiochip6: (omap-gpmc): added GPIO chardev (254:6)
[ 2.704589] gpio gpiochip6: registered GPIOs 178 to 181 on omap-gpmc
[...]
[ 2.840393] gpio gpiochip7: Static allocation of GPIO base is deprecated, use dynamic allocation.
[ 2.849365] gpio gpiochip7: (gpio-160-191): GPIO integer space overlap, cannot add chip
[ 2.857513] gpiochip_add_data_with_key: GPIOs 160..191 (gpio-160-191) failed to register, -16
[ 2.866149] omap_gpio 48310000.gpio: error -EBUSY: Could not register gpio chip
On that device it is fixed invasively by
commit 92bf78b33b0b4 ("gpio: omap: use dynamic allocation of base")
but let's also fix that for devices where there is still
a mixture of static and dynamic allocation.
Fixes: 7b61212f2a07 ("gpiolib: Get rid of ARCH_NR_GPIOS") Signed-off-by: Andreas Kemnade <andreas@kemnade.info> Reviewed-by: <christophe.leroy@csgroup.eu> Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Warning: Kernel ABI header at 'tools/include/uapi/linux/in.h' differs from latest version at 'include/uapi/linux/in.h'
diff -u tools/include/uapi/linux/in.h include/uapi/linux/in.h
Signed-off-by: Yanteng Si <siyanteng@loongson.cn> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: loongson-kernel@lists.loongnix.cn Link: https://lore.kernel.org/r/23aabc69956ac94fbf388b05c8be08a64e8c7ccc.1683712945.git.siyanteng@loongson.cn Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Currently, offloaded conntrack entries (flows) can only be deleted
after they are removed from offload, which is either by timeout,
tcp state change or tc ct rule deletion. This can cause issues for
users wishing to manually delete or flush existing entries.
Support deletion of offloaded conntrack entries.
Example usage:
# Delete all offloaded (and non offloaded) conntrack entries
# whose source address is 1.2.3.4
$ conntrack -D -s 1.2.3.4
# Delete all entries
$ conntrack -F
Signed-off-by: Paul Blakey <paulb@nvidia.com> Reviewed-by: Simon Horman <simon.horman@corigine.com> Acked-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Florian Westphal <fw@strlen.de> Cc: Demi Marie Obenour <demi@invisiblethingslab.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Schedutil normally calls the adjust_perf callback for drivers with
adjust_perf callback available and fast_switch_possible flag set.
However, when frequency invariance is disabled and schedutil tries to
invoke fast_switch. So, there is a chance of kernel crash if this
function pointer is not set. To protect against this scenario add
fast_switch callback to amd_pstate driver.
Fixes: 1d215f0319c2 ("cpufreq: amd-pstate: Add fast switch function for AMD P-State") Signed-off-by: Gautham R. Shenoy <gautham.shenoy@amd.com> Signed-off-by: Wyes Karny <wyes.karny@amd.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Driver should update policy->cur after updating the frequency.
Currently amd_pstate doesn't update policy->cur when `adjust_perf`
is used. Which causes /proc/cpuinfo to show wrong cpu frequency.
Fix this by updating policy->cur with correct frequency value in
adjust_perf function callback.
- Before the fix: (setting min freq to 1.5 MHz)
[root@amd]# cat /proc/cpuinfo | grep "cpu MHz" | sort | uniq --count
1 cpu MHz : 1777.016
1 cpu MHz : 1797.160
1 cpu MHz : 1797.270
189 cpu MHz : 400.000
- After the fix: (setting min freq to 1.5 MHz)
[root@amd]# cat /proc/cpuinfo | grep "cpu MHz" | sort | uniq --count
1 cpu MHz : 1753.353
1 cpu MHz : 1756.838
1 cpu MHz : 1776.466
1 cpu MHz : 1776.873
1 cpu MHz : 1777.308
1 cpu MHz : 1779.900
183 cpu MHz : 1805.231
1 cpu MHz : 1956.815
1 cpu MHz : 2246.203
1 cpu MHz : 2259.984
Fixes: 1d215f0319c2 ("cpufreq: amd-pstate: Add fast switch function for AMD P-State") Signed-off-by: Wyes Karny <wyes.karny@amd.com>
[ rjw: Subject edits ] Cc: 5.17+ <stable@vger.kernel.org> # 5.17+ Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
commit <8af870aa5b847> ("block: enable bio caching use for passthru IO")
introduced bio-cache for passthru IO. In case when nr_vecs are greater
than BIO_INLINE_VECS, bio and bvecs are allocated from mempool (instead
of percpu cache) and REQ_ALLOC_CACHE is cleared. This causes the side
effect of not freeing bio/bvecs into mempool on completion.
This patch lets the passthru IO fallback to allocation using bio_kmalloc
when nr_vecs are greater than BIO_INLINE_VECS. The corresponding bio
is freed during call to blk_mq_map_bio_put during completion.
Cc: stable@vger.kernel.org # 6.1
fixes <8af870aa5b847> ("block: enable bio caching use for passthru IO")
Commit a71f388045ed ("thermal/drivers/mellanox: Use generic
thermal_zone_get_trip() function") was backported as a dependency of the
fix in upstream commit 6d206b1ea9f4 ("mlxsw: core_thermal: Fix fan speed
in maximum cooling state"). However, it is dependent on changes in the
thermal core that were merged in v6.3. Without them, the mlxsw driver is
unable to register its thermal zone:
mlxsw_spectrum 0000:03:00.0: Failed to register thermal zone
mlxsw_spectrum 0000:03:00.0: cannot register bus device
mlxsw_spectrum: probe of 0000:03:00.0 failed with error -22
Fix this by reverting this commit and instead fix the small conflict
with the above mentioned fix. Tested using the test case mentioned in
the change log of the fix:
Previously, channel open messages were always sent to monitors on the first
ioctl() call for unbound HCI sockets, even if the command and arguments
were completely invalid. This can leave an exploitable hole with the abuse
of invalid ioctl calls.
This commit hardens the ioctl processing logic by first checking if the
command is valid, and immediately returning with an ENOIOCTLCMD error code
if it is not. This ensures that ioctl calls with invalid commands are free
of side effects, and increases the difficulty of further exploitation by
forcing exploitation to find a way to pass a valid command first.
Signed-off-by: Ruihan Li <lrh2000@pku.edu.cn> Co-developed-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Dragos-Marian Panait <dragos.panait@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
APUs before Raven didn't support s0ix. As we just relieved some
of the safety checks for s0ix to improve power consumption on
APUs that support it but that are missing BIOS support a new
blind spot was introduced that a user could "try" to run s0ix.
Plug this hole so that if users try to run s0ix on anything older
than Raven it will just skip suspend of the GPU.
Fixes: cf488dcd0ab7 ("drm/amd: Allow s0ix without BIOS support") Suggested-by: Alexander Deucher <Alexander.Deucher@amd.com> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
As part of version 2 we don't need to have wake_packet_bufsize
and wake_packet_length. The first one is already calculated by the driver,
the latter is sent as part of the wake packet notification.
By default the VSC8501 and VSC8502 RGMII/GMII/MII RX_CLK output is
disabled. To allow packet forwarding towards the MAC it needs to be
enabled.
For other PHYs supported by this driver the clock output is enabled
by default.
Fixes: d3169863310d ("net: phy: mscc: add support for VSC8502") Signed-off-by: David Epping <david.epping@missinglinkelectronics.com> Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Reviewed-by: Vladimir Oltean <olteanv@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
page_pool_ring_[un]lock() use in_softirq() to decide which
spin lock variant to use, and when they are called in the
context with in_softirq() being false, spin_lock_bh() is
called in page_pool_ring_lock() while spin_unlock() is
called in page_pool_ring_unlock(), because spin_lock_bh()
has disabled the softirq in page_pool_ring_lock(), which
causes inconsistency for spin lock pair calling.
This patch fixes it by returning in_softirq state from
page_pool_producer_lock(), and use it to decide which
spin lock variant to use in page_pool_producer_unlock().
As pool->ring has both producer and consumer lock, so
rename it to page_pool_producer_[un]lock() to reflect
the actual usage. Also move them to page_pool.c as they
are only used there, and remove the 'inline' as the
compiler may have better idea to do inlining or not.
Fixes: 7886244736a4 ("net: page_pool: Add bulk support for ptr_ring") Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com> Acked-by: Jesper Dangaard Brouer <brouer@redhat.com> Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org> Link: https://lore.kernel.org/r/20230522031714.5089-1-linyunsheng@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
We use BH context only for synchronization, so we don't care if it's
actually serving softirq or not.
As a side node, in case of threaded NAPI, in_serving_softirq() will
return false because it's in process context with BH off, making
page_pool_recycle_in_cache() unreachable.
Signed-off-by: Qingfang DENG <qingfang.deng@siflower.com.cn> Tested-by: Felix Fietkau <nbd@nbd.name> Signed-off-by: David S. Miller <davem@davemloft.net>
Stable-dep-of: 368d3cb406cd ("page_pool: fix inconsistency for page_pool_ring_[un]lock()") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Check physical PFN is valid before converting the PFN to a struct page
pointer to be returned to caller of vfio_pin_pages().
vfio_pin_pages() pins user pages with contiguous IOVA.
If the IOVA of a user page to be pinned belongs to vma of vm_flags
VM_PFNMAP, pin_user_pages_remote() will return -EFAULT without returning
struct page address for this PFN. This is because usually this kind of PFN
(e.g. MMIO PFN) has no valid struct page address associated.
Upon this error, vaddr_get_pfns() will obtain the physical PFN directly.
While previously vfio_pin_pages() returns to caller PFN arrays directly,
after commit 34a255e67615 ("vfio: Replace phys_pfn with pages for vfio_pin_pages()"),
PFNs will be converted to "struct page *" unconditionally and therefore
the returned "struct page *" array may contain invalid struct page
addresses.
Given current in-tree users of vfio_pin_pages() only expect "struct page *
returned, check PFN validity and return -EINVAL to let the caller be
aware of IOVAs to be pinned containing PFN not able to be returned in
"struct page *" array. So that, the caller will not consume the returned
pointer (e.g. test PageReserved()) and avoid error like "supervisor read
access in kernel mode".
Fixes: 34a255e67615 ("vfio: Replace phys_pfn with pages for vfio_pin_pages()") Cc: Sean Christopherson <seanjc@google.com> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com> Signed-off-by: Yan Zhao <yan.y.zhao@intel.com> Reviewed-by: Sean Christopherson <seanjc@google.com> Link: https://lore.kernel.org/r/20230519065843.10653-1-yan.y.zhao@intel.com Signed-off-by: Alex Williamson <alex.williamson@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
If multiple CPUs are sharing the same hardware queue, it can
cause leak in the active queue counter tracking when __blk_mq_tag_busy()
is executed simultaneously.
Fixes: ee78ec1077d3 ("blk-mq: blk_mq_tag_busy is no need to return a value") Signed-off-by: Tian Lan <tian.lan@twosigma.com> Reviewed-by: Ming Lei <ming.lei@redhat.com> Reviewed-by: Damien Le Moal <dlemoal@kernel.org> Reviewed-by: John Garry <john.g.garry@oracle.com> Link: https://lore.kernel.org/r/20230522210555.794134-1-tilan7663@gmail.com Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
The read_skb() logic is incrementing the tcp->copied_seq which is used for
among other things calculating how many outstanding bytes can be read by
the application. This results in application errors, if the application
does an ioctl(FIONREAD) we return zero because this is calculated from
the copied_seq value.
To fix this we move tcp->copied_seq accounting into the recv handler so
that we update these when the recvmsg() hook is called and data is in
fact copied into user buffers. This gives an accurate FIONREAD value
as expected and improves ACK handling. Before we were calling the
tcp_rcv_space_adjust() which would update 'number of bytes copied to
user in last RTT' which is wrong for programs returning SK_PASS. The
bytes are only copied to the user when recvmsg is handled.
Doing the fix for recvmsg is straightforward, but fixing redirect and
SK_DROP pkts is a bit tricker. Build a tcp_psock_eat() helper and then
call this from skmsg handlers. This fixes another issue where a broken
socket with a BPF program doing a resubmit could hang the receiver. This
happened because although read_skb() consumed the skb through sock_drop()
it did not update the copied_seq. Now if a single reccv socket is
redirecting to many sockets (for example for lb) the receiver sk will be
hung even though we might expect it to continue. The hang comes from
not updating the copied_seq numbers and memory pressure resulting from
that.
We have a slight layer problem of calling tcp_eat_skb even if its not
a TCP socket. To fix we could refactor and create per type receiver
handlers. I decided this is more work than we want in the fix and we
already have some small tweaks depending on caller that use the
helper skb_bpf_strparser(). So we extend that a bit and always set
the strparser bit when it is in use and then we can gate the
seq_copied updates on this.
Fixes: 04919bed948dc ("tcp: Introduce tcp_read_skb()") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/bpf/20230523025618.113937-9-john.fastabend@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
When TCP stack has data ready to read sk_data_ready() is called. Sockmap
overwrites this with its own handler to call into BPF verdict program.
But, the original TCP socket had sock_def_readable that would additionally
wake up any user space waiters with sk_wake_async().
Sockmap saved the callback when the socket was created so call the saved
data ready callback and then we can wake up any epoll() logic waiting
on the read.
Note we call on 'copied >= 0' to account for returning 0 when a FIN is
received because we need to wake up user for this as well so they
can do the recvmsg() -> 0 and detect the shutdown.
Fixes: 04919bed948dc ("tcp: Introduce tcp_read_skb()") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/bpf/20230523025618.113937-8-john.fastabend@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
A common mechanism to put a TCP socket into the sockmap is to hook the
BPF_SOCK_OPS_{ACTIVE_PASSIVE}_ESTABLISHED_CB event with a BPF program
that can map the socket info to the correct BPF verdict parser. When
the user adds the socket to the map the psock is created and the new
ops are assigned to ensure the verdict program will 'see' the sk_buffs
as they arrive.
Part of this process hooks the sk_data_ready op with a BPF specific
handler to wake up the BPF verdict program when data is ready to read.
The logic is simple enough (posted here for easy reading)
if (unlikely(!sock || !sock->ops || !sock->ops->read_skb))
return;
sock->ops->read_skb(sk, sk_psock_verdict_recv);
}
The oversight here is sk->sk_socket is not assigned until the application
accepts() the new socket. However, its entirely ok for the peer application
to do a connect() followed immediately by sends. The socket on the receiver
is sitting on the backlog queue of the listening socket until its accepted
and the data is queued up. If the peer never accepts the socket or is slow
it will eventually hit data limits and rate limit the session. But,
important for BPF sockmap hooks when this data is received TCP stack does
the sk_data_ready() call but the read_skb() for this data is never called
because sk_socket is missing. The data sits on the sk_receive_queue.
Then once the socket is accepted if we never receive more data from the
peer there will be no further sk_data_ready calls and all the data
is still on the sk_receive_queue(). Then user calls recvmsg after accept()
and for TCP sockets in sockmap we use the tcp_bpf_recvmsg_parser() handler.
The handler checks for data in the sk_msg ingress queue expecting that
the BPF program has already run from the sk_data_ready hook and enqueued
the data as needed. So we are stuck.
To fix do an unlikely check in recvmsg handler for data on the
sk_receive_queue and if it exists wake up data_ready. We have the sock
locked in both read_skb and recvmsg so should avoid having multiple
runners.
Fixes: 04919bed948dc ("tcp: Introduce tcp_read_skb()") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/bpf/20230523025618.113937-7-john.fastabend@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
The sockmap code is returning EAGAIN after a FIN packet is received and no
more data is on the receive queue. Correct behavior is to return 0 to the
user and the user can then close the socket. The EAGAIN causes many apps
to retry which masks the problem. Eventually the socket is evicted from
the sockmap because its released from sockmap sock free handling. The
issue creates a delay and can cause some errors on application side.
To fix this check on sk_msg_recvmsg side if length is zero and FIN flag
is set then set return to zero. A selftest will be added to check this
condition.
Fixes: 04919bed948dc ("tcp: Introduce tcp_read_skb()") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Tested-by: William Findlay <will@isovalent.com> Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/bpf/20230523025618.113937-6-john.fastabend@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
We noticed some rare sk_buffs were stepping past the queue when system was
under memory pressure. The general theory is to skip enqueueing
sk_buffs when its not necessary which is the normal case with a system
that is properly provisioned for the task, no memory pressure and enough
cpu assigned.
But, if we can't allocate memory due to an ENOMEM error when enqueueing
the sk_buff into the sockmap receive queue we push it onto a delayed
workqueue to retry later. When a new sk_buff is received we then check
if that queue is empty. However, there is a problem with simply checking
the queue length. When a sk_buff is being processed from the ingress queue
but not yet on the sockmap msg receive queue its possible to also recv
a sk_buff through normal path. It will check the ingress queue which is
zero and then skip ahead of the pkt being processed.
Previously we used sock lock from both contexts which made the problem
harder to hit, but not impossible.
To fix instead of popping the skb from the queue entirely we peek the
skb from the queue and do the copy there. This ensures checks to the
queue length are non-zero while skb is being processed. Then finally
when the entire skb has been copied to user space queue or another
socket we pop it off the queue. This way the queue length check allows
bypassing the queue only after the list has been completely processed.
To reproduce issue we run NGINX compliance test with sockmap running and
observe some flakes in our testing that we attributed to this issue.
Fixes: 04919bed948dc ("tcp: Introduce tcp_read_skb()") Suggested-by: Jakub Sitnicki <jakub@cloudflare.com> Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Tested-by: William Findlay <will@isovalent.com> Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/bpf/20230523025618.113937-5-john.fastabend@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Now that the backlog manages the reschedule() logic correctly we can drop
the partial fix to reschedule from recvmsg hook.
Rescheduling on recvmsg hook was added to address a corner case where we
still had data in the backlog state but had nothing to kick it and
reschedule the backlog worker to run and finish copying data out of the
state. This had a couple limitations, first it required user space to
kick it introducing an unnecessary EBUSY and retry. Second it only
handled the ingress case and egress redirects would still be hung.
With the correct fix, pushing the reschedule logic down to where the
enomem error occurs we can drop this fix.
Fixes: bec217197b412 ("skmsg: Schedule psock work if the cached skb exists on the psock") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/bpf/20230523025618.113937-4-john.fastabend@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Sk_buffs are fed into sockmap verdict programs either from a strparser
(when the user might want to decide how framing of skb is done by attaching
another parser program) or directly through tcp_read_sock. The
tcp_read_sock is the preferred method for performance when the BPF logic is
a stream parser.
The flow for Cilium's common use case with a stream parser is,
tcp_read_sock()
sk_psock_verdict_recv
ret = bpf_prog_run_pin_on_cpu()
sk_psock_verdict_apply(sock, skb, ret)
// if system is under memory pressure or app is slow we may
// need to queue skb. Do this queuing through ingress_skb and
// then kick timer to wake up handler
skb_queue_tail(ingress_skb, skb)
schedule_work(work);
The work queue is wired up to sk_psock_backlog(). This will then walk the
ingress_skb skb list that holds our sk_buffs that could not be handled,
but should be OK to run at some later point. However, its possible that
the workqueue doing this work still hits an error when sending the skb.
When this happens the skbuff is requeued on a temporary 'state' struct
kept with the workqueue. This is necessary because its possible to
partially send an skbuff before hitting an error and we need to know how
and where to restart when the workqueue runs next.
Now for the trouble, we don't rekick the workqueue. This can cause a
stall where the skbuff we just cached on the state variable might never
be sent. This happens when its the last packet in a flow and no further
packets come along that would cause the system to kick the workqueue from
that side.
To fix we could do simple schedule_work(), but while under memory pressure
it makes sense to back off some instead of continue to retry repeatedly. So
instead to fix convert schedule_work to schedule_delayed_work and add
backoff logic to reschedule from backlog queue on errors. Its not obvious
though what a good backoff is so use '1'.
To test we observed some flakes whil running NGINX compliance test with
sockmap we attributed these failed test to this bug and subsequent issue.
>From on list discussion. This commit
bec217197b41("skmsg: Schedule psock work if the cached skb exists on the psock")
was intended to address similar race, but had a couple cases it missed.
Most obvious it only accounted for receiving traffic on the local socket
so if redirecting into another socket we could still get an sk_buff stuck
here. Next it missed the case where copied=0 in the recv() handler and
then we wouldn't kick the scheduler. Also its sub-optimal to require
userspace to kick the internal mechanisms of sockmap to wake it up and
copy data to user. It results in an extra syscall and requires the app
to actual handle the EAGAIN correctly.
Fixes: 04919bed948dc ("tcp: Introduce tcp_read_skb()") Signed-off-by: John Fastabend <john.fastabend@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Tested-by: William Findlay <will@isovalent.com> Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com> Link: https://lore.kernel.org/bpf/20230523025618.113937-3-john.fastabend@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
The read_skb hook calls consume_skb() now, but this means that if the
recv_actor program wants to use the skb it needs to inc the ref cnt
so that the consume_skb() doesn't kfree the sk_buff.
This is problematic because in some error cases under memory pressure
we may need to linearize the sk_buff from sk_psock_skb_ingress_enqueue().
Then we get this,
Because we incremented users refcnt from sk_psock_verdict_recv() we
hit the bug on with refcnt > 1 and trip it.
To fix lets simply pass ownership of the sk_buff through the skb_read
call. Then we can drop the consume from read_skb handlers and assume
the verdict recv does any required kfree.
Bug found while testing in our CI which runs in VMs that hit memory
constraints rather regularly. William tested TCP read_skb handlers.
In fact the device with chip id 0xD283 is called NCT6126D, and that is
the chip id the Nuvoton code was written for. Correct that name to avoid
confusion, because a NCT6116D in fact exists as well but has another
chip id, and is currently not supported.
The look at the spec also revealed that GPIO group7 in fact has 8 pins,
so correct the pin count in that group as well.
Fixes: d0918a84aff0 ("gpio-f7188x: Add GPIO support for Nuvoton NCT6116") Reported-by: Xing Tong Wu <xingtong.wu@siemens.com> Signed-off-by: Henning Schild <henning.schild@siemens.com> Acked-by: Simon Guinot <simon.guinot@sequanux.org> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
devcom events are sent to all registered component. Following the
cited patch, it is possible for two components, e.g.: two eswitches,
to send devcom events, while both components are registered. This
means eswitch layer will do double un/pairing, which is double
allocation and free of resources, even though only one un/pairing is
needed. flow example:
When receive buffer is small we try to copy out the data from
TCP into a skb maintained by TLS to prevent connection from
stalling. Unfortunately if a single record is made up of a mix
of decrypted and non-decrypted skbs combining them into a single
skb leads to loss of decryption status, resulting in decryption
errors or data corruption.
Similarly when trying to use TCP receive queue directly we need
to make sure that all the skbs within the record have the same
status. If we don't the mixed status will be detected correctly
but we'll CoW the anchor, again collapsing it into a single paged
skb without decrypted status preserved. So the "fixup" code will
not know which parts of skb to re-encrypt.
Fixes: 84c61fe1a75b ("tls: rx: do not use the standard strparser") Tested-by: Shai Amiram <samiram@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org> Reviewed-by: Simon Horman <simon.horman@corigine.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
We'll need to copy input skbs individually in the next patch.
Factor that code out (without assuming we're copying a full record).
Tested-by: Shai Amiram <samiram@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org> Reviewed-by: Simon Horman <simon.horman@corigine.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Stable-dep-of: eca9bfafee3a ("tls: rx: strp: preserve decryption status of skbs when needed") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
If a record is partially decrypted we'll have to CoW it, anyway,
so go into copy mode and allocate a writable skb right away.
This will make subsequent fix simpler because we won't have to
teach tls_strp_msg_make_copy() how to copy skbs while preserving
decrypt status.
Tested-by: Shai Amiram <samiram@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org> Reviewed-by: Simon Horman <simon.horman@corigine.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Stable-dep-of: eca9bfafee3a ("tls: rx: strp: preserve decryption status of skbs when needed") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
We call tls_rx_msg_size(skb) before doing skb->len += chunk.
So the tls_rx_msg_size() code will see old skb->len, most
likely leading to an over-read.
Worst case we will over read an entire record, next iteration
will try to trim the skb but may end up turning frag len negative
or discarding the subsequent record (since we already told TCP
we've read it during previous read but now we'll trim it out of
the skb).
Fixes: 84c61fe1a75b ("tls: rx: do not use the standard strparser") Tested-by: Shai Amiram <samiram@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org> Reviewed-by: Simon Horman <simon.horman@corigine.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
alloc_skb_with_frags() fills in page frag sizes but does not
set skb->len and skb->data_len. Set those correctly otherwise
device offload will most likely generate an empty skb and
hit the BUG() at the end of __skb_nsg().
Fixes: 84c61fe1a75b ("tls: rx: do not use the standard strparser") Tested-by: Shai Amiram <samiram@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org> Reviewed-by: Simon Horman <simon.horman@corigine.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
skb->len covers the entire skb, including the frag_list.
In fact we're guaranteed that rxm->full_len <= skb->len,
so since the change under Fixes we were not checking decrypt
status of any skb but the first.
Note that the skb_pagelen() added here may feel a bit costly,
but it's removed by subsequent fixes, anyway.
Reported-by: Tariq Toukan <tariqt@nvidia.com> Fixes: 86b259f6f888 ("tls: rx: device: bound the frag walk") Tested-by: Shai Amiram <samiram@nvidia.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org> Reviewed-by: Simon Horman <simon.horman@corigine.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
This is because the DRAM address used for accessing metrics table
needs to be refreshed after a suspend resume cycle. Add a resume
callback to reset this again.
Fixes: 1a409b35c995 ("platform/x86/amd/pmf: Get performance metrics from PMFW") Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Link: https://lore.kernel.org/r/20230513011408.958-1-mario.limonciello@amd.com Reviewed-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Commit bb1be7498500 ("firmware: arm_ffa: Add v1.1 get_partition_info support")
adds support to discovery the UUIDs of the partitions or just fetch the
partition count using the PARTITION_INFO_GET_RETURN_COUNT_ONLY flag.
However the commit doesn't handle the fact that the older version doesn't
understand the flag and must be MBZ which results in firmware returning
invalid parameter error. That results in the failure of the driver probe
which is in correct.
Limit the usage of the PARTITION_INFO_GET_RETURN_COUNT_ONLY flag for the
versions above v1.0(i.e v1.1 and onwards) which fixes the issue.
With a raw socket bound to IPPROTO_RAW (ie with hdrincl enabled), the
protocol field of the flow structure, build by raw_sendmsg() /
rawv6_sendmsg()), is set to IPPROTO_RAW. This breaks the ipsec policy
lookup when some policies are defined with a protocol in the selector.
For ipv6, the sin6_port field from 'struct sockaddr_in6' could be used to
specify the protocol. Just accept all values for IPPROTO_RAW socket.
For ipv4, the sin_port field of 'struct sockaddr_in' could not be used
without breaking backward compatibility (the value of this field was never
checked). Let's add a new kind of control message, so that the userland
could specify which protocol is used.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") CC: stable@vger.kernel.org Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com> Link: https://lore.kernel.org/r/20230522120820.1319391-1-nicolas.dichtel@6wind.com Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Users who want to share a single public IP address for outgoing connections
between several hosts traditionally reach for SNAT. However, SNAT requires
state keeping on the node(s) performing the NAT.
A stateless alternative exists, where a single IP address used for egress
can be shared between several hosts by partitioning the available ephemeral
port range. In such a setup:
1. Each host gets assigned a disjoint range of ephemeral ports.
2. Applications open connections from the host-assigned port range.
3. Return traffic gets routed to the host based on both, the destination IP
and the destination port.
An application which wants to open an outgoing connection (connect) from a
given port range today can choose between two solutions:
1. Manually pick the source port by bind()'ing to it before connect()'ing
the socket.
This approach has a couple of downsides:
a) Search for a free port has to be implemented in the user-space. If
the chosen 4-tuple happens to be busy, the application needs to retry
from a different local port number.
Detecting if 4-tuple is busy can be either easy (TCP) or hard
(UDP). In TCP case, the application simply has to check if connect()
returned an error (EADDRNOTAVAIL). That is assuming that the local
port sharing was enabled (REUSEADDR) by all the sockets.
# Assume desired local port range is 60_000-60_511
s = socket(AF_INET, SOCK_STREAM)
s.setsockopt(SOL_SOCKET, SO_REUSEADDR, 1)
s.bind(("192.0.2.1", 60_000))
s.connect(("1.1.1.1", 53))
# Fails only if 192.0.2.1:60000 -> 1.1.1.1:53 is busy
# Application must retry with another local port
In case of UDP, the network stack allows binding more than one socket
to the same 4-tuple, when local port sharing is enabled
(REUSEADDR). Hence detecting the conflict is much harder and involves
querying sock_diag and toggling the REUSEADDR flag [1].
b) For TCP, bind()-ing to a port within the ephemeral port range means
that no connecting sockets, that is those which leave it to the
network stack to find a free local port at connect() time, can use
the this port.
IOW, the bind hash bucket tb->fastreuse will be 0 or 1, and the port
will be skipped during the free port search at connect() time.
2. Isolate the app in a dedicated netns and use the use the per-netns
ip_local_port_range sysctl to adjust the ephemeral port range bounds.
The per-netns setting affects all sockets, so this approach can be used
only if:
- there is just one egress IP address, or
- the desired egress port range is the same for all egress IP addresses
used by the application.
For TCP, this approach avoids the downsides of (1). Free port search and
4-tuple conflict detection is done by the network stack:
s = socket(AF_INET, SOCK_STREAM)
s.setsockopt(SOL_IP, IP_BIND_ADDRESS_NO_PORT, 1)
s.bind(("192.0.2.1", 0))
s.connect(("1.1.1.1", 53))
# Fails if all 4-tuples 192.0.2.1:60000-60511 -> 1.1.1.1:53 are busy
For UDP this approach has limited applicability. Setting the
IP_BIND_ADDRESS_NO_PORT socket option does not result in local source
port being shared with other connected UDP sockets.
Hence relying on the network stack to find a free source port, limits the
number of outgoing UDP flows from a single IP address down to the number
of available ephemeral ports.
To put it another way, partitioning the ephemeral port range between hosts
using the existing Linux networking API is cumbersome.
To address this use case, add a new socket option at the SOL_IP level,
named IP_LOCAL_PORT_RANGE. The new option can be used to clamp down the
ephemeral port range for each socket individually.
The option can be used only to narrow down the per-netns local port
range. If the per-socket range lies outside of the per-netns range, the
latter takes precedence.
UAPI-wise, the low and high range bounds are passed to the kernel as a pair
of u16 values in host byte order packed into a u32. This avoids pointer
passing.
PORT_LO = 40_000
PORT_HI = 40_511
s = socket(AF_INET, SOCK_STREAM)
v = struct.pack("I", PORT_HI << 16 | PORT_LO)
s.setsockopt(SOL_IP, IP_LOCAL_PORT_RANGE, v)
s.bind(("127.0.0.1", 0))
s.getsockname()
# Local address between ("127.0.0.1", 40_000) and ("127.0.0.1", 40_511),
# if there is a free port. EADDRINUSE otherwise.
Kamal Mostafa [Fri, 28 Jul 2023 19:08:58 +0000 (12:08 -0700)]
UBUNTU: Upstream stable to v6.1.31, v6.3.5
BugLink: https://bugs.launchpad.net/bugs/2028979 Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
The commit c6d96df9fa2c ("net: ethernet: mtk_eth_soc: drop generic vlan rx
offload, only use DSA untagging") makes VLAN RX offloading to be only used
on the SoCs without the MTK_NETSYS_V2 ability (which are not just MT7621
and MT7622). The commit disables the proper handling of special tagged
(DSA) frames, added with commit 87e3df4961f4 ("net-next: ethernet:
mediatek: add CDM able to recognize the tag for DSA"), for non
MTK_NETSYS_V2 SoCs when it finds a MAC that does not use DSA. So if the
other MAC uses DSA, the CDMQ component transmits DSA tagged frames to the
CPU improperly. This issue can be observed on frames with TCP, for example,
a TCP speed test using iperf3 won't work.
The commit disables the proper handling of special tagged (DSA) frames
because it assumes that these SoCs don't use more than one MAC, which is
wrong. Although I made Frank address this false assumption on the patch log
when they sent the patch on behalf of Felix, the code still made changes
with this assumption.
Therefore, the proper handling of special tagged (DSA) frames must be kept
enabled in all circumstances as it doesn't affect non DSA tagged frames.
Hardware DSA untagging, introduced with the commit 2d7605a72906 ("net:
ethernet: mtk_eth_soc: enable hardware DSA untagging"), and VLAN RX
offloading are operations on the two CDM components of the frame engine,
CDMP and CDMQ, which connect to Packet DMA (PDMA) and QoS DMA (QDMA) and
are between the MACs and the CPU. These operations apply to all MACs of the
SoC so if one MAC uses DSA and the other doesn't, the hardware DSA
untagging operation will cause the CDMP component to transmit non DSA
tagged frames to the CPU improperly.
Since the VLAN RX offloading feature configuration was dropped, VLAN RX
offloading can only be used along with hardware DSA untagging. So, for the
case above, we need to disable both features and leave it to the CPU,
therefore software, to untag the DSA and VLAN tags.
So the correct way to handle this is:
For all SoCs:
Enable the proper handling of special tagged (DSA) frames
(MTK_CDMQ_IG_CTRL).
Fixes: c6d96df9fa2c ("net: ethernet: mtk_eth_soc: drop generic vlan rx offload, only use DSA untagging") Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Move cxl_await_media_ready() to cxl_pci probe before driver starts issuing
IDENTIFY and retrieving memory device information to ensure that the
device is ready to provide the information. Allow cxl_pci_probe() to succeed
even if media is not ready. Cache the media failure in cxlds and don't ask
the device for any media information.
The rationale for proceeding in the !media_ready case is to allow for
mailbox operations to interrogate and/or remediate the device. After
media is repaired then rebinding the cxl_pci driver is expected to
restart the capacity scan.
Suggested-by: Dan Williams <dan.j.williams@intel.com> Fixes: b39cb1052a5c ("cxl/mem: Register CXL memX devices") Reviewed-by: Ira Weiny <ira.weiny@intel.com> Signed-off-by: Dave Jiang <dave.jiang@intel.com> Link: https://lore.kernel.org/r/168445310026.3251520.8124296540679268206.stgit@djiang5-mobl3
[djbw: fixup cxl_test] Signed-off-by: Dan Williams <dan.j.williams@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
cifs.ko maps NT_STATUS_NOT_FOUND to -EIO when SMB1 servers couldn't
resolve referral paths. Proceed to tree connect when we get -EIO from
dfs_get_referral() as well.
Reported-by: Kris Karas (Bug Reporting) <bugs-a21@moonlit-rail.com> Tested-by: Woody Suwalski <terraluna977@gmail.com> Fixes: 8e3554150d6c ("cifs: fix sharing of DFS connections") Cc: stable@vger.kernel.org # v6.2+ Signed-off-by: Paulo Alcantara (SUSE) <pc@manguebit.com> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
commit 4917f55b4ef9 ("mm/sparse-vmemmap: improve memory savings for
compound devmaps") added support for using optimized vmmemap for devdax
devices. But how vmemmap mappings are created are architecture specific.
For example, powerpc with hash translation doesn't have vmemmap mappings
in init_mm page table instead they are bolted table entries in the
hardware page table
vmemmap_populate_compound_pages() used by vmemmap optimization code is not
aware of these architecture-specific mapping. Hence allow architecture to
opt for this feature. I selected architectures supporting
HUGETLB_PAGE_OPTIMIZE_VMEMMAP option as also supporting this feature.
After some investigation, I have found the following issue: unlike other
zswap backends, zsmalloc performs the LRU list update at the object
mapping time, rather than when the slot for the object is allocated.
This deviation was discussed and agreed upon during the review process
of the zsmalloc writeback patch series:
With the current upstream code, this issue rarely happens. zswap only
triggers writeback when the pool is already full, at which point all
further store attempts are short-circuited. This creates an implicit
pseudo-serialization between reclaim and store. I am working on a new
zswap shrinking mechanism, which makes interleaving reclaim and store
more likely, exposing this bug.
zbud and z3fold do not have this problem, because they perform the LRU
list update in the alloc function, while still holding the pool's lock.
This patch fixes the aforementioned bug by moving the LRU update back to
zs_malloc(), analogous to zbud and z3fold.
Link: https://lkml.kernel.org/r/20230505185054.2417128-1-nphamcs@gmail.com Fixes: 64f768c6b32e ("zsmalloc: add a LRU to zs_pool to keep track of zspages in LRU order") Signed-off-by: Nhat Pham <nphamcs@gmail.com> Suggested-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Sergey Senozhatsky <senozhatsky@chromium.org> Acked-by: Minchan Kim <minchan@kernel.org> Cc: Dan Streetman <ddstreet@ieee.org> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Seth Jennings <sjenning@redhat.com> Cc: Vitaly Wool <vitaly.wool@konsulko.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
The mscc driver implements support for VSC8502, so its ID should be in
the MODULE_DEVICE_TABLE for automatic loading.
Signed-off-by: David Epping <david.epping@missinglinkelectronics.com> Fixes: d3169863310d ("net: phy: mscc: add support for VSC8502") Reviewed-by: Vladimir Oltean <olteanv@gmail.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
During the CLC handshake, server sequentially tries available SMCRv2
and SMCRv1 devices in smc_listen_work().
If an SMCRv2 device is found. SMCv2 based link group and link will be
assigned to the connection. Then assumed that some buffer assignment
errors happen later in the CLC handshake, such as RMB registration
failure, server will give up SMCRv2 and try SMCRv1 device instead. But
the resources assigned to the connection won't be reset.
When server tries SMCRv1 device, the connection creation process will
be executed again. Since conn->lnk has been assigned when trying SMCRv2,
it will not be set to the correct SMCRv1 link in
smcr_lgr_conn_assign_link(). So in such situation, conn->lgr points to
correct SMCRv1 link group but conn->lnk points to the SMCRv2 link
mistakenly.
Then in smc_clc_send_confirm_accept(), conn->rmb_desc->mr[link->link_idx]
will be accessed. Since the link->link_idx is not correct, the related
MR may not have been initialized, so crash happens.
| Try SMCRv2 device first
| |-> conn->lgr: assign existed SMCRv2 link group;
| |-> conn->link: assign existed SMCRv2 link (link_idx may be 1 in SMC_LGR_SYMMETRIC);
| |-> sndbuf & RMB creation fails, quit;
|
| Try SMCRv1 device then
| |-> conn->lgr: create SMCRv1 link group and assign;
| |-> conn->link: keep SMCRv2 link mistakenly;
| |-> sndbuf & RMB creation succeed, only RMB->mr[link_idx = 0]
| initialized.
|
| Then smc_clc_send_confirm_accept() accesses
| conn->rmb_desc->mr[conn->link->link_idx, which is 1], then crash.
v
This patch tries to fix this by cleaning conn->lnk before assigning
link. In addition, it is better to reset the connection and clean the
resources assigned if trying SMCRv2 failed in buffer creation or
registration.
Fixes: e49300a6bf62 ("net/smc: add listen processing for SMC-Rv2") Link: https://lore.kernel.org/r/20220523055056.2078994-1-liuyacan@corp.netease.com/ Signed-off-by: Wen Gu <guwen@linux.alibaba.com> Reviewed-by: Tony Lu <tonylu@linux.alibaba.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Signed-off-by: Sen Chu <sen.chu@mediatek.com Fixes: 4cfc96547512 ("regulator: mt6359: Add support for MT6359P regulator") Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com Link: https://lore.kernel.org/r/20230518040646.8730-1-sen.chu@mediatek.com Signed-off-by: Mark Brown <broonie@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
The transmit buffers allocated by the driver can be used to transmit data
by any messages/commands needing the buffer. However, it is not guaranteed
to have been zero-ed before every new transmission and hence it will just
contain residual value from the previous transmission. There are several
reserved fields in the memory descriptors that must be zero(MBZ). The
receiver can reject the transmission if any such MBZ fields are non-zero.
While we can set the whole page to zero, it is not optimal as most of the
fields get initialised to the value required for the current transmission.
So, just set the reserved/MBZ fields to zero in the memory descriptors
explicitly to honour the requirement and keep the receiver happy.
Fixes: cc2195fe536c ("firmware: arm_ffa: Add support for MEM_* interfaces") Reported-by: Marc Bonnici <marc.bonnici@arm.com> Link: https://lore.kernel.org/r/20230503131252.12585-1-sudeep.holla@arm.com Signed-off-by: Sudeep Holla <sudeep.holla@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
While testing the ethernet interface on a Variscite symphony carrier
board using an imx8mn SOM with an onboard ADIN1300 PHY (EC hardware
configuration), the ethernet PHY is not detected.
The ADIN1300 datasheet indicate that the "Management interface
active (t4)" state is reached at most 5ms after the reset signal is
deasserted.
The device tree in Variscite custom git repository uses the following
property:
phy-reset-post-delay = <20>;
Add a new MDIO property 'reset-deassert-us' of 20ms to have the same
delay inside the ethphy node. Adding this property fixes the problem
with the PHY detection.
Note that this SOM can also have an Atheros AR8033 PHY. In this case,
a 1ms deassert delay is sufficient. Add a comment to that effect.
Fixes: ade0176dd8a0 ("arm64: dts: imx8mn-var-som: Add Variscite VAR-SOM-MX8MN System on Module") Signed-off-by: Hugo Villeneuve <hvilleneuve@dimonoff.com> Signed-off-by: Shawn Guo <shawnguo@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
From one hand, mlx5 driver is allowing to probe PFs in parallel.
From the other hand, devcom, which is a share resource between PFs, is
registered without any lock. This might resulted in memory problems.
Hence, use the global mlx5_dev_list_lock in order to serialize devcom
registration.
Fixes: fadd59fc50d0 ("net/mlx5: Introduce inter-device communication mechanism") Signed-off-by: Shay Drory <shayd@nvidia.com> Reviewed-by: Mark Bloch <mbloch@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
In case devcom allocation is failed, mlx5 is always freeing the priv.
However, this priv might have been allocated by a different thread,
and freeing it might lead to use-after-free bugs.
Fix it by freeing the priv only in case it was allocated by the
running thread.
Fixes: fadd59fc50d0 ("net/mlx5: Introduce inter-device communication mechanism") Signed-off-by: Shay Drory <shayd@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
DEVX can issue a general command, which is not used by mlx5 driver.
In case such command is failed, mlx5 is trying to collect the failure
data, However, mlx5 doesn't create a storage for this command, since
mlx5 doesn't use it. This lead to array-index-out-of-bounds error.
Fix it by checking whether the command is known before collecting the
failure data.
Fixes: 34f46ae0d4b3 ("net/mlx5: Add command failures data to debugfs") Signed-off-by: Shay Drory <shayd@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
SW Steering uses RC QP for writing STEs to ICM. This writingis done in LB
(loopback), and FL (force-loopback) QP is preferred for performance. FL is
available when RoCE is enabled or disabled based on RoCE caps.
This patch adds reading of FL capability from HCA caps in addition to the
existing reading from RoCE caps, thus fixing the case where we didn't
have loopback enabled when RoCE was disabled.
Fixes: 7304d603a57a ("net/mlx5: DR, Add support for force-loopback QP") Signed-off-by: Itamar Gozlan <igozlan@nvidia.com> Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
In case user switch a device from switchdev mode to legacy mode, mlx5
first unpair the E-switch and afterwards unload the uplink vport.
From the other hand, in case user remove or reload a device, mlx5
first unload the uplink vport and afterwards unpair the E-switch.
The latter is causing a bug[1], hence, handle pairing of E-switch as
part of uplink un/load APIs.
[1]
In case VF_LAG is used, every tc fdb flow is duplicated to the peer
esw. However, the original esw keeps a pointer to this duplicated
flow, not the peer esw.
e.g.: if user create tc fdb flow over esw0, the flow is duplicated
over esw1, in FW/HW, but in SW, esw0 keeps a pointer to the duplicated
flow.
During module unload while a peer tc fdb flow is still offloaded, in
case the first device to be removed is the peer device (esw1 in the
example above), the peer net-dev is destroyed, and so the mlx5e_priv
is memset to 0.
Afterwards, the peer device is trying to unpair himself from the
original device (esw0 in the example above). Unpair API invoke the
original device to clear peer flow from its eswitch (esw0), but the
peer flow, which is stored over the original eswitch (esw0), is
trying to use the peer mlx5e_priv, which is memset to 0 and result in
bellow kernel-oops.
When calculating crc for hash index we use the function crc32 that
calculates for little-endian (LE) arch.
Then we convert it to network endianness using htonl(), but it's wrong
to do the conversion in BE archs since the crc32 value is already LE.
The solution is to switch the bytes from the crc result for all types
of arc.
Fixes: 40416d8ede65 ("net/mlx5: DR, Replace CRC32 implementation to use kernel lib") Signed-off-by: Erez Shitrit <erezsh@nvidia.com> Reviewed-by: Alex Vesker <valex@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
NAPI gets called with budget of 0 from netpoll, which has interrupts
disabled. We should try to free some space on Tx rings and nothing
else.
Specifically do not try to handle XDP TX or try to refill Rx buffers -
we can't use the page pool from IRQ context. Don't check if IRQs moved,
either, that makes no sense in netpoll. Netpoll calls _all_ the rings
from whatever CPU it happens to be invoked on.
In general do as little as possible, the work quickly adds up when
there's tens of rings to poll.
AFAIU page pool takes a BH lock, releases it and since BH is now
enabled tries to run softirqs.
Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Fixes: 60bbf7eeef10 ("mlx5: use page_pool for xdp_return_frame call") Signed-off-by: Jakub Kicinski <kuba@kernel.org> Reviewed-by: Simon Horman <simon.horman@corigine.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
With introduction of post action infrastructure most of the users of encap
attribute had been modified in order to obtain the correct attribute by
calling mlx5e_tc_get_encap_attr() helper instead of assuming encap action
is always on default attribute. However, the cited commit didn't modify
mlx5e_invalidate_encap() which prevents it from destroying correct modify
header action which leads to a warning [0]. Fix the issue by using correct
attribute.
[0]:
Feb 21 09:47:35 c-237-177-40-045 kernel: WARNING: CPU: 17 PID: 654 at drivers/net/ethernet/mellanox/mlx5/core/en_tc.c:684 mlx5e_tc_attach_mod_hdr+0x1cc/0x230 [mlx5_core]
Feb 21 09:47:35 c-237-177-40-045 kernel: RIP: 0010:mlx5e_tc_attach_mod_hdr+0x1cc/0x230 [mlx5_core]
Feb 21 09:47:35 c-237-177-40-045 kernel: Call Trace:
Feb 21 09:47:35 c-237-177-40-045 kernel: <TASK>
Feb 21 09:47:35 c-237-177-40-045 kernel: mlx5e_tc_fib_event_work+0x8e3/0x1f60 [mlx5_core]
Feb 21 09:47:35 c-237-177-40-045 kernel: ? mlx5e_take_all_encap_flows+0xe0/0xe0 [mlx5_core]
Feb 21 09:47:35 c-237-177-40-045 kernel: ? lock_downgrade+0x6d0/0x6d0
Feb 21 09:47:35 c-237-177-40-045 kernel: ? lockdep_hardirqs_on_prepare+0x273/0x3f0
Feb 21 09:47:35 c-237-177-40-045 kernel: ? lockdep_hardirqs_on_prepare+0x273/0x3f0
Feb 21 09:47:35 c-237-177-40-045 kernel: process_one_work+0x7c2/0x1310
Feb 21 09:47:35 c-237-177-40-045 kernel: ? lockdep_hardirqs_on_prepare+0x3f0/0x3f0
Feb 21 09:47:35 c-237-177-40-045 kernel: ? pwq_dec_nr_in_flight+0x230/0x230
Feb 21 09:47:35 c-237-177-40-045 kernel: ? rwlock_bug.part.0+0x90/0x90
Feb 21 09:47:35 c-237-177-40-045 kernel: worker_thread+0x59d/0xec0
Feb 21 09:47:35 c-237-177-40-045 kernel: ? __kthread_parkme+0xd9/0x1d0
Fixes: 8300f225268b ("net/mlx5e: Create new flow attr for multi table actions") Signed-off-by: Vlad Buslov <vladbu@nvidia.com> Reviewed-by: Roi Dayan <roid@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Cited commit causes ABBA deadlock[0] when peer flows are created while
holding the devcom rw semaphore. Due to peer flows offload implementation
the lock is taken much higher up the call chain and there is no obvious way
to easily fix the deadlock. Instead, since tc route query code needs the
peer eswitch structure only to perform a lookup in xarray and doesn't
perform any sleeping operations with it, refactor the code for lockless
execution in following ways:
- RCUify the devcom 'data' pointer. When resetting the pointer
synchronously wait for RCU grace period before returning. This is fine
since devcom is currently only used for synchronization of
pairing/unpairing of eswitches which is rare and already expensive as-is.
- Wrap all usages of 'paired' boolean in {READ|WRITE}_ONCE(). The flag has
already been used in some unlocked contexts without proper
annotations (e.g. users of mlx5_devcom_is_paired() function), but it wasn't
an issue since all relevant code paths checked it again after obtaining the
devcom semaphore. Now it is also used by mlx5_devcom_get_peer_data_rcu() as
"best effort" check to return NULL when devcom is being unpaired. Note that
while RCU read lock doesn't prevent the unpaired flag from being changed
concurrently it still guarantees that reader can continue to use 'data'.
- Refactor mlx5e_tc_query_route_vport() function to use new
mlx5_devcom_get_peer_data_rcu() API which fixes the deadlock.
[0]:
[ 164.599612] ======================================================
[ 164.600142] WARNING: possible circular locking dependency detected
[ 164.600667] 6.3.0-rc3+ #1 Not tainted
[ 164.601021] ------------------------------------------------------
[ 164.601557] handler1/3456 is trying to acquire lock:
[ 164.601998] ffff88811f1714b0 (&esw->offloads.encap_tbl_lock){+.+.}-{3:3}, at: mlx5e_attach_encap+0xd8/0x8b0 [mlx5_core]
[ 164.603078]
but task is already holding lock:
[ 164.603617] ffff88810137fc98 (&comp->sem){++++}-{3:3}, at: mlx5_devcom_get_peer_data+0x37/0x80 [mlx5_core]
[ 164.604459]
which lock already depends on the new lock.
Check in the mlx5e_ptp_poll_ts_cq context if the ptp tx sq should be woken
up. Before change, the ptp tx sq may never wake up if the ptp tx ts skb
fifo is full when mlx5e_poll_tx_cq checks if the queue should be woken up.