April 2024 - Linux-stable-mirror

[PATCH] misc/pvpanic-pci: register attributes via pci_driver

by Thomas Weißschuh

In __pci_register_driver(), the pci core overwrites the dev_groups field of the embedded struct device_driver with the dev_groups from the outer struct pci_driver unconditionally. Set dev_groups in the pci_driver to make sure it is used. This was broken since the introduction of pvpanic-pci. Fixes: db3a4f0abefd ("misc/pvpanic: add PCI driver") Cc: stable(a)vger.kernel.org Signed-off-by: Thomas Weißschuh <linux(a)weissschuh.net> --- Greg, does it make sense to duplicate fields between struct pci_driver and struct device_driver? The fields "name", "groups" and "dev_groups" are duplicated. pci_driver::dev_groups was introduced in commit ded13b9cfd59 ("PCI: Add support for dev_groups to struct pci_driver") because "this helps converting PCI drivers sysfs attributes to static" I don't understand the reasoning. The embedded device_driver shares the same storage lifetime and the fields have the exact same type. --- drivers/misc/pvpanic/pvpanic-pci.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/drivers/misc/pvpanic/pvpanic-pci.c b/drivers/misc/pvpanic/pvpanic-pci.c index 9ad20e82785b..b21598a18f6d 100644 --- a/drivers/misc/pvpanic/pvpanic-pci.c +++ b/drivers/misc/pvpanic/pvpanic-pci.c @@ -44,8 +44,6 @@ static struct pci_driver pvpanic_pci_driver = { .name = "pvpanic-pci", .id_table = pvpanic_pci_id_tbl, .probe = pvpanic_pci_probe, - .driver = { - .dev_groups = pvpanic_dev_groups, - }, + .dev_groups = pvpanic_dev_groups, }; module_pci_driver(pvpanic_pci_driver); --- base-commit: 00dcf5d862e86e57f5ce46344039f11bb1ad61f6 change-id: 20240411-pvpanic-pci-dev-groups-e3beebcbc4e4 Best regards, -- Thomas Weißschuh <linux(a)weissschuh.net>

1 year, 6 months

2
2
0 0

FAILED: patch "[PATCH] drm/i915/vma: Fix UAF on destroy against retire race" failed to apply to 6.7-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.7-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.7.y git checkout FETCH_HEAD git cherry-pick -x 0e45882ca829b26b915162e8e86dbb1095768e9e # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2024033027-expensive-footage-f3ea@gregkh' --subject-prefix 'PATCH 6.7.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 0e45882ca829b26b915162e8e86dbb1095768e9e Mon Sep 17 00:00:00 2001 From: Janusz Krzysztofik <janusz.krzysztofik(a)linux.intel.com> Date: Tue, 5 Mar 2024 15:35:06 +0100 Subject: [PATCH] drm/i915/vma: Fix UAF on destroy against retire race MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Object debugging tools were sporadically reporting illegal attempts to free a still active i915 VMA object when parking a GT believed to be idle. [161.359441] ODEBUG: free active (active state 0) object: ffff88811643b958 object type: i915_active hint: __i915_vma_active+0x0/0x50 [i915] [161.360082] WARNING: CPU: 5 PID: 276 at lib/debugobjects.c:514 debug_print_object+0x80/0xb0 ... [161.360304] CPU: 5 PID: 276 Comm: kworker/5:2 Not tainted 6.5.0-rc1-CI_DRM_13375-g003f860e5577+ #1 [161.360314] Hardware name: Intel Corporation Rocket Lake Client Platform/RocketLake S UDIMM 6L RVP, BIOS RKLSFWI1.R00.3173.A03.2204210138 04/21/2022 [161.360322] Workqueue: i915-unordered __intel_wakeref_put_work [i915] [161.360592] RIP: 0010:debug_print_object+0x80/0xb0 ... [161.361347] debug_object_free+0xeb/0x110 [161.361362] i915_active_fini+0x14/0x130 [i915] [161.361866] release_references+0xfe/0x1f0 [i915] [161.362543] i915_vma_parked+0x1db/0x380 [i915] [161.363129] __gt_park+0x121/0x230 [i915] [161.363515] ____intel_wakeref_put_last+0x1f/0x70 [i915] That has been tracked down to be happening when another thread is deactivating the VMA inside __active_retire() helper, after the VMA's active counter has been already decremented to 0, but before deactivation of the VMA's object is reported to the object debugging tool. We could prevent from that race by serializing i915_active_fini() with __active_retire() via ref->tree_lock, but that wouldn't stop the VMA from being used, e.g. from __i915_vma_retire() called at the end of __active_retire(), after that VMA has been already freed by a concurrent i915_vma_destroy() on return from the i915_active_fini(). Then, we should rather fix the issue at the VMA level, not in i915_active. Since __i915_vma_parked() is called from __gt_park() on last put of the GT's wakeref, the issue could be addressed by holding the GT wakeref long enough for __active_retire() to complete before that wakeref is released and the GT parked. I believe the issue was introduced by commit d93939730347 ("drm/i915: Remove the vma refcount") which moved a call to i915_active_fini() from a dropped i915_vma_release(), called on last put of the removed VMA kref, to i915_vma_parked() processing path called on last put of a GT wakeref. However, its visibility to the object debugging tool was suppressed by a bug in i915_active that was fixed two weeks later with commit e92eb246feb9 ("drm/i915/active: Fix missing debug object activation"). A VMA associated with a request doesn't acquire a GT wakeref by itself. Instead, it depends on a wakeref held directly by the request's active intel_context for a GT associated with its VM, and indirectly on that intel_context's engine wakeref if the engine belongs to the same GT as the VMA's VM. Those wakerefs are released asynchronously to VMA deactivation. Fix the issue by getting a wakeref for the VMA's GT when activating it, and putting that wakeref only after the VMA is deactivated. However, exclude global GTT from that processing path, otherwise the GPU never goes idle. Since __i915_vma_retire() may be called from atomic contexts, use async variant of wakeref put. Also, to avoid circular locking dependency, take care of acquiring the wakeref before VM mutex when both are needed. v7: Add inline comments with justifications for: - using untracked variants of intel_gt_pm_get/put() (Nirmoy), - using async variant of _put(), - not getting the wakeref in case of a global GTT, - always getting the first wakeref outside vm->mutex. v6: Since __i915_vma_active/retire() callbacks are not serialized, storing a wakeref tracking handle inside struct i915_vma is not safe, and there is no other good place for that. Use untracked variants of intel_gt_pm_get/put_async(). v5: Replace "tile" with "GT" across commit description (Rodrigo), - avoid mentioning multi-GT case in commit description (Rodrigo), - explain why we need to take a temporary wakeref unconditionally inside i915_vma_pin_ww() (Rodrigo). v4: Refresh on top of commit 5e4e06e4087e ("drm/i915: Track gt pm wakerefs") (Andi), - for more easy backporting, split out removal of former insufficient workarounds and move them to separate patches (Nirmoy). - clean up commit message and description a bit. v3: Identify root cause more precisely, and a commit to blame, - identify and drop former workarounds, - update commit message and description. v2: Get the wakeref before VM mutex to avoid circular locking dependency, - drop questionable Fixes: tag. Fixes: d93939730347 ("drm/i915: Remove the vma refcount") Closes: https://gitlab.freedesktop.org/drm/intel/issues/8875 Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik(a)linux.intel.com> Cc: Thomas Hellström <thomas.hellstrom(a)linux.intel.com> Cc: Nirmoy Das <nirmoy.das(a)intel.com> Cc: Andi Shyti <andi.shyti(a)linux.intel.com> Cc: Rodrigo Vivi <rodrigo.vivi(a)intel.com> Cc: stable(a)vger.kernel.org # v5.19+ Reviewed-by: Nirmoy Das <nirmoy.das(a)intel.com> Signed-off-by: Andi Shyti <andi.shyti(a)linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240305143747.335367-6-janus… (cherry picked from commit f3c71b2ded5c4367144a810ef25f998fd1d6c381) Signed-off-by: Rodrigo Vivi <rodrigo.vivi(a)intel.com> diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index d09aad34ba37..b70715b1411d 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -34,6 +34,7 @@ #include "gt/intel_engine.h" #include "gt/intel_engine_heartbeat.h" #include "gt/intel_gt.h" +#include "gt/intel_gt_pm.h" #include "gt/intel_gt_requests.h" #include "gt/intel_tlb.h" @@ -103,12 +104,42 @@ static inline struct i915_vma *active_to_vma(struct i915_active *ref) static int __i915_vma_active(struct i915_active *ref) { - return i915_vma_tryget(active_to_vma(ref)) ? 0 : -ENOENT; + struct i915_vma *vma = active_to_vma(ref); + + if (!i915_vma_tryget(vma)) + return -ENOENT; + + /* + * Exclude global GTT VMA from holding a GT wakeref + * while active, otherwise GPU never goes idle. + */ + if (!i915_vma_is_ggtt(vma)) { + /* + * Since we and our _retire() counterpart can be + * called asynchronously, storing a wakeref tracking + * handle inside struct i915_vma is not safe, and + * there is no other good place for that. Hence, + * use untracked variants of intel_gt_pm_get/put(). + */ + intel_gt_pm_get_untracked(vma->vm->gt); + } + + return 0; } static void __i915_vma_retire(struct i915_active *ref) { - i915_vma_put(active_to_vma(ref)); + struct i915_vma *vma = active_to_vma(ref); + + if (!i915_vma_is_ggtt(vma)) { + /* + * Since we can be called from atomic contexts, + * use an async variant of intel_gt_pm_put(). + */ + intel_gt_pm_put_async_untracked(vma->vm->gt); + } + + i915_vma_put(vma); } static struct i915_vma * @@ -1404,7 +1435,7 @@ int i915_vma_pin_ww(struct i915_vma *vma, struct i915_gem_ww_ctx *ww, struct i915_vma_work *work = NULL; struct dma_fence *moving = NULL; struct i915_vma_resource *vma_res = NULL; - intel_wakeref_t wakeref = 0; + intel_wakeref_t wakeref; unsigned int bound; int err; @@ -1424,8 +1455,14 @@ int i915_vma_pin_ww(struct i915_vma *vma, struct i915_gem_ww_ctx *ww, if (err) return err; - if (flags & PIN_GLOBAL) - wakeref = intel_runtime_pm_get(&vma->vm->i915->runtime_pm); + /* + * In case of a global GTT, we must hold a runtime-pm wakeref + * while global PTEs are updated. In other cases, we hold + * the rpm reference while the VMA is active. Since runtime + * resume may require allocations, which are forbidden inside + * vm->mutex, get the first rpm wakeref outside of the mutex. + */ + wakeref = intel_runtime_pm_get(&vma->vm->i915->runtime_pm); if (flags & vma->vm->bind_async_flags) { /* lock VM */ @@ -1561,8 +1598,7 @@ int i915_vma_pin_ww(struct i915_vma *vma, struct i915_gem_ww_ctx *ww, if (work) dma_fence_work_commit_imm(&work->base); err_rpm: - if (wakeref) - intel_runtime_pm_put(&vma->vm->i915->runtime_pm, wakeref); + intel_runtime_pm_put(&vma->vm->i915->runtime_pm, wakeref); if (moving) dma_fence_put(moving);

1 year, 6 months

3
2
0 0

[PATCH v4] mtd: limit OTP NVMEM Cell parse to non Nand devices

by Christian Marangi

MTD OTP logic is very fragile on parsing NVMEM Cell and can be problematic with some specific kind of devices. The problem was discovered by e87161321a40 ("mtd: rawnand: macronix: OTP access for MX30LFxG18AC") where OTP support was added to a NAND device. With the case of NAND devices, it does require a node where ECC info are declared and all the fixed partitions, and this cause the OTP codepath to parse this node as OTP NVMEM Cells, making probe fail and the NAND device registration fail. MTD OTP parsing should have been limited to always using compatible to prevent this error by using node with compatible "otp-user" or "otp-factory". NVMEM across the years had various iteration on how Cells could be declared in DT, in some old implementation, no_of_node should have been enabled but now add_legacy_fixed_of_cells should be used to disable NVMEM to parse child node as NVMEM Cell. To fix this and limit any regression with other MTD that makes use of declaring OTP as direct child of the dev node, disable add_legacy_fixed_of_cells if we detect the MTD type is Nand. With the following logic, the OTP NVMEM entry is correctly created with no Cells and the MTD Nand is correctly probed and partitions are correctly exposed. Fixes: 4b361cfa8624 ("mtd: core: add OTP nvmem provider support") Cc: <stable(a)vger.kernel.org> # v6.7+ Signed-off-by: Christian Marangi <ansuelsmth(a)gmail.com> --- To backport this to v6.6 and previous, config.no_of_node = mtd_type_is_nand(mtd); should be used as it does pose the same usage of add_legacy_fixed_of_cells. Changes v4: - Add info on how to backport this to previous kernel - Fix Fixes tag - Reformat commit description as it was unprecise and had false statement Changes v3: - Fix commit description Changes v2: - Use mtd_type_is_nand instead of node name check drivers/mtd/mtdcore.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/mtd/mtdcore.c b/drivers/mtd/mtdcore.c index 5887feb347a4..0de87bc63840 100644 --- a/drivers/mtd/mtdcore.c +++ b/drivers/mtd/mtdcore.c @@ -900,7 +900,7 @@ static struct nvmem_device *mtd_otp_nvmem_register(struct mtd_info *mtd, config.name = compatible; config.id = NVMEM_DEVID_AUTO; config.owner = THIS_MODULE; - config.add_legacy_fixed_of_cells = true; + config.add_legacy_fixed_of_cells = !mtd_type_is_nand(mtd); config.type = NVMEM_TYPE_OTP; config.root_only = true; config.ignore_wp = true; -- 2.43.0

1 year, 6 months

3
3
0 0

[PATCH v2 5.10.y 1/1] mailbox: imx: fix suspend failure

by Daisuke Mizobuchi

imx_mu_isr() always calls pm_system_wakeup() even when it should not, making the system unable to enter sleep. Suspend fails as follows: armadillo:~# echo mem > /sys/power/state [ 2614.602432] PM: suspend entry (deep) [ 2614.610640] Filesystems sync: 0.004 seconds [ 2614.618016] Freezing user space processes ... (elapsed 0.001 seconds) done. [ 2614.626555] OOM killer disabled. [ 2614.629792] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done. [ 2614.638456] printk: Suspending console(s) (use no_console_suspend to debug) [ 2614.649504] PM: Some devices failed to suspend, or early wake event detected [ 2614.730103] PM: resume devices took 0.080 seconds [ 2614.741924] OOM killer enabled. [ 2614.745073] Restarting tasks ... done. [ 2614.754532] PM: suspend exit ash: write error: Resource busy armadillo:~# Upstream commit 892cb524ae8a is correct, so this seems to be a mistake during cherry-pick. Cc: <stable(a)vger.kernel.org> Fixes: a16f5ae8ade1 ("mailbox: imx: fix wakeup failure from freeze mode") Signed-off-by: Daisuke Mizobuchi <mizo(a)atmark-techno.com> Reviewed-by: Dominique Martinet <dominique.martinet(a)atmark-techno.com> --- drivers/mailbox/imx-mailbox.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/mailbox/imx-mailbox.c b/drivers/mailbox/imx-mailbox.c index c5663398c6b7..28f5450e4130 100644 --- a/drivers/mailbox/imx-mailbox.c +++ b/drivers/mailbox/imx-mailbox.c @@ -331,8 +331,6 @@ static int imx_mu_startup(struct mbox_chan *chan) break; } - priv->suspend = true; - return 0; } @@ -550,8 +548,6 @@ static int imx_mu_probe(struct platform_device *pdev) clk_disable_unprepare(priv->clk); - priv->suspend = false; - return 0; disable_runtime_pm: @@ -614,6 +610,8 @@ static int __maybe_unused imx_mu_suspend_noirq(struct device *dev) if (!priv->clk) priv->xcr = imx_mu_read(priv, priv->dcfg->xCR); + priv->suspend = true; + return 0; } @@ -632,6 +630,8 @@ static int __maybe_unused imx_mu_resume_noirq(struct device *dev) if (!imx_mu_read(priv, priv->dcfg->xCR) && !priv->clk) imx_mu_write(priv, priv->xcr, priv->dcfg->xCR); + priv->suspend = false; + return 0; } -- 2.30.2

1 year, 6 months

3
2
0 0

[PATCH v2 5/5] drm/vmwgfx: Sort primary plane formats by order of preference

by Zack Rusin

The table of primary plane formats wasn't sorted at all, leading to applications picking our least desirable formats by defaults. Sort the primary plane formats according to our order of preference. Nice side-effect of this change is that it makes IGT's kms_atomic plane-invalid-params pass because the test picks the first format which for vmwgfx was DRM_FORMAT_XRGB1555 and uses fb's with odd sizes which make Pixman, which IGT depends on assert due to the fact that our 16bpp formats aren't 32 bit aligned like Pixman requires all formats to be. Signed-off-by: Zack Rusin <zack.rusin(a)broadcom.com> Fixes: 36cc79bc9077 ("drm/vmwgfx: Add universal plane support") Cc: Broadcom internal kernel review list <bcm-kernel-feedback-list(a)broadcom.com> Cc: dri-devel(a)lists.freedesktop.org Cc: <stable(a)vger.kernel.org> # v4.12+ Acked-by: Pekka Paalanen <pekka.paalanen(a)collabora.com> --- drivers/gpu/drm/vmwgfx/vmwgfx_kms.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h index bf9931e3a728..bf24f2f0dcfc 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h @@ -233,10 +233,10 @@ struct vmw_framebuffer_bo { static const uint32_t __maybe_unused vmw_primary_plane_formats[] = { - DRM_FORMAT_XRGB1555, - DRM_FORMAT_RGB565, DRM_FORMAT_XRGB8888, DRM_FORMAT_ARGB8888, + DRM_FORMAT_RGB565, + DRM_FORMAT_XRGB1555, }; static const uint32_t __maybe_unused vmw_cursor_plane_formats[] = { -- 2.40.1

1 year, 6 months

1
0
0 0

[PATCH v2 4/5] drm/vmwgfx: Fix crtc's atomic check conditional

by Zack Rusin

The conditional was supposed to prevent enabling of a crtc state without a set primary plane. Accidently it also prevented disabling crtc state with a set primary plane. Neither is correct. Fix the conditional and just driver-warn when a crtc state has been enabled without a primary plane which will help debug broken userspace. Fixes IGT's kms_atomic_interruptible and kms_atomic_transition tests. Signed-off-by: Zack Rusin <zack.rusin(a)broadcom.com> Fixes: 06ec41909e31 ("drm/vmwgfx: Add and connect CRTC helper functions") Cc: Broadcom internal kernel review list <bcm-kernel-feedback-list(a)broadcom.com> Cc: dri-devel(a)lists.freedesktop.org Cc: <stable(a)vger.kernel.org> # v4.12+ Reviewed-by: Ian Forbes <ian.forbes(a)broadcom.com> Reviewed-by: Martin Krastev <martin.krastev(a)broadcom.com> --- drivers/gpu/drm/vmwgfx/vmwgfx_kms.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c index e33e5993d8fc..13b2820cae51 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c @@ -931,6 +931,7 @@ int vmw_du_cursor_plane_atomic_check(struct drm_plane *plane, int vmw_du_crtc_atomic_check(struct drm_crtc *crtc, struct drm_atomic_state *state) { + struct vmw_private *vmw = vmw_priv(crtc->dev); struct drm_crtc_state *new_state = drm_atomic_get_new_crtc_state(state, crtc); struct vmw_display_unit *du = vmw_crtc_to_du(new_state->crtc); @@ -938,9 +939,13 @@ int vmw_du_crtc_atomic_check(struct drm_crtc *crtc, bool has_primary = new_state->plane_mask & drm_plane_mask(crtc->primary); - /* We always want to have an active plane with an active CRTC */ - if (has_primary != new_state->enable) - return -EINVAL; + /* + * This is fine in general, but broken userspace might expect + * some actual rendering so give a clue as why it's blank. + */ + if (new_state->enable && !has_primary) + drm_dbg_driver(&vmw->drm, + "CRTC without a primary plane will be blank.\n"); if (new_state->connector_mask != connector_mask && -- 2.40.1

1 year, 6 months

1
0
0 0

[PATCH v2 3/5] drm/vmwgfx: Fix prime import/export

by Zack Rusin

vmwgfx never supported prime import of external buffers. Furthermore the driver exposes two different objects to userspace: vmw_surface's and gem buffers but prime import/export only worked with vmw_surfaces. Because gem buffers are used through the dumb_buffer interface this meant that the driver created buffers couldn't have been prime exported or imported. Fix prime import/export. Makes IGT's kms_prime pass. Signed-off-by: Zack Rusin <zack.rusin(a)broadcom.com> Fixes: 8afa13a0583f ("drm/vmwgfx: Implement DRIVER_GEM") Cc: <stable(a)vger.kernel.org> # v6.6+ Reviewed-by: Martin Krastev <martin.krastev(a)broadcom.com> --- drivers/gpu/drm/vmwgfx/vmwgfx_blit.c | 35 +++++++++++++++-- drivers/gpu/drm/vmwgfx/vmwgfx_bo.c | 7 ++-- drivers/gpu/drm/vmwgfx/vmwgfx_bo.h | 2 + drivers/gpu/drm/vmwgfx/vmwgfx_drv.c | 1 + drivers/gpu/drm/vmwgfx/vmwgfx_drv.h | 3 ++ drivers/gpu/drm/vmwgfx/vmwgfx_gem.c | 32 ++++++++++++++++ drivers/gpu/drm/vmwgfx/vmwgfx_prime.c | 15 +++++++- drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c | 44 +++++++++++++++------- 8 files changed, 117 insertions(+), 22 deletions(-) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_blit.c b/drivers/gpu/drm/vmwgfx/vmwgfx_blit.c index c52c7bf1485b..717d624e9a05 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_blit.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_blit.c @@ -456,8 +456,10 @@ int vmw_bo_cpu_blit(struct ttm_buffer_object *dst, .no_wait_gpu = false }; u32 j, initial_line = dst_offset / dst_stride; - struct vmw_bo_blit_line_data d; + struct vmw_bo_blit_line_data d = {0}; int ret = 0; + struct page **dst_pages = NULL; + struct page **src_pages = NULL; /* Buffer objects need to be either pinned or reserved: */ if (!(dst->pin_count)) @@ -477,12 +479,35 @@ int vmw_bo_cpu_blit(struct ttm_buffer_object *dst, return ret; } + if (!src->ttm->pages && src->ttm->sg) { + src_pages = kvmalloc_array(src->ttm->num_pages, + sizeof(struct page *), GFP_KERNEL); + if (!src_pages) + return -ENOMEM; + ret = drm_prime_sg_to_page_array(src->ttm->sg, src_pages, + src->ttm->num_pages); + if (ret) + goto out; + } + if (!dst->ttm->pages && dst->ttm->sg) { + dst_pages = kvmalloc_array(dst->ttm->num_pages, + sizeof(struct page *), GFP_KERNEL); + if (!dst_pages) { + ret = -ENOMEM; + goto out; + } + ret = drm_prime_sg_to_page_array(dst->ttm->sg, dst_pages, + dst->ttm->num_pages); + if (ret) + goto out; + } + d.mapped_dst = 0; d.mapped_src = 0; d.dst_addr = NULL; d.src_addr = NULL; - d.dst_pages = dst->ttm->pages; - d.src_pages = src->ttm->pages; + d.dst_pages = dst->ttm->pages ? dst->ttm->pages : dst_pages; + d.src_pages = src->ttm->pages ? src->ttm->pages : src_pages; d.dst_num_pages = PFN_UP(dst->resource->size); d.src_num_pages = PFN_UP(src->resource->size); d.dst_prot = ttm_io_prot(dst, dst->resource, PAGE_KERNEL); @@ -504,6 +529,10 @@ int vmw_bo_cpu_blit(struct ttm_buffer_object *dst, kunmap_atomic(d.src_addr); if (d.dst_addr) kunmap_atomic(d.dst_addr); + if (src_pages) + kvfree(src_pages); + if (dst_pages) + kvfree(dst_pages); return ret; } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c index bfd41ce3c8f4..e5eb21a471a6 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c @@ -377,7 +377,8 @@ static int vmw_bo_init(struct vmw_private *dev_priv, { struct ttm_operation_ctx ctx = { .interruptible = params->bo_type != ttm_bo_type_kernel, - .no_wait_gpu = false + .no_wait_gpu = false, + .resv = params->resv, }; struct ttm_device *bdev = &dev_priv->bdev; struct drm_device *vdev = &dev_priv->drm; @@ -394,8 +395,8 @@ static int vmw_bo_init(struct vmw_private *dev_priv, vmw_bo_placement_set(vmw_bo, params->domain, params->busy_domain); ret = ttm_bo_init_reserved(bdev, &vmw_bo->tbo, params->bo_type, - &vmw_bo->placement, 0, &ctx, NULL, - NULL, destroy); + &vmw_bo->placement, 0, &ctx, + params->sg, params->resv, destroy); if (unlikely(ret)) return ret; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.h b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.h index 0d496dc9c6af..f349642e6190 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.h @@ -55,6 +55,8 @@ struct vmw_bo_params { enum ttm_bo_type bo_type; size_t size; bool pin; + struct dma_resv *resv; + struct sg_table *sg; }; /** diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c index 89d3679d2608..41ad13e45554 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c @@ -1631,6 +1631,7 @@ static const struct drm_driver driver = { .prime_fd_to_handle = vmw_prime_fd_to_handle, .prime_handle_to_fd = vmw_prime_handle_to_fd, + .gem_prime_import_sg_table = vmw_prime_import_sg_table, .fops = &vmwgfx_driver_fops, .name = VMWGFX_DRIVER_NAME, diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h index ddbceaa31b59..4ecaea0026fc 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -1107,6 +1107,9 @@ extern int vmw_prime_handle_to_fd(struct drm_device *dev, struct drm_file *file_priv, uint32_t handle, uint32_t flags, int *prime_fd); +struct drm_gem_object *vmw_prime_import_sg_table(struct drm_device *dev, + struct dma_buf_attachment *attach, + struct sg_table *table); /* * MemoryOBject management - vmwgfx_mob.c diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_gem.c b/drivers/gpu/drm/vmwgfx/vmwgfx_gem.c index 186150f41fbc..2132a8ad8c0c 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_gem.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_gem.c @@ -136,6 +136,38 @@ int vmw_gem_object_create_with_handle(struct vmw_private *dev_priv, return ret; } +struct drm_gem_object *vmw_prime_import_sg_table(struct drm_device *dev, + struct dma_buf_attachment *attach, + struct sg_table *table) +{ + int ret; + struct vmw_private *dev_priv = vmw_priv(dev); + struct drm_gem_object *gem = NULL; + struct vmw_bo *vbo; + struct vmw_bo_params params = { + .domain = (dev_priv->has_mob) ? VMW_BO_DOMAIN_SYS : VMW_BO_DOMAIN_VRAM, + .busy_domain = VMW_BO_DOMAIN_SYS, + .bo_type = ttm_bo_type_sg, + .size = attach->dmabuf->size, + .pin = false, + .resv = attach->dmabuf->resv, + .sg = table, + + }; + + dma_resv_lock(params.resv, NULL); + + ret = vmw_bo_create(dev_priv, &params, &vbo); + if (ret != 0) + goto out_no_bo; + + vbo->tbo.base.funcs = &vmw_gem_object_funcs; + + gem = &vbo->tbo.base; +out_no_bo: + dma_resv_unlock(params.resv); + return gem; +} int vmw_gem_object_create_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_prime.c b/drivers/gpu/drm/vmwgfx/vmwgfx_prime.c index 2d72a5ee7c0c..c99cad444991 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_prime.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_prime.c @@ -75,8 +75,12 @@ int vmw_prime_fd_to_handle(struct drm_device *dev, int fd, u32 *handle) { struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile; + int ret = ttm_prime_fd_to_handle(tfile, fd, handle); - return ttm_prime_fd_to_handle(tfile, fd, handle); + if (ret) + ret = drm_gem_prime_fd_to_handle(dev, file_priv, fd, handle); + + return ret; } int vmw_prime_handle_to_fd(struct drm_device *dev, @@ -85,5 +89,12 @@ int vmw_prime_handle_to_fd(struct drm_device *dev, int *prime_fd) { struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile; - return ttm_prime_handle_to_fd(tfile, handle, flags, prime_fd); + int ret; + + if (handle > VMWGFX_NUM_MOB) + ret = ttm_prime_handle_to_fd(tfile, handle, flags, prime_fd); + else + ret = drm_gem_prime_handle_to_fd(dev, file_priv, handle, flags, prime_fd); + + return ret; } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c index 4d23d0a70bcb..621d98b376bb 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c @@ -188,13 +188,18 @@ static int vmw_ttm_map_dma(struct vmw_ttm_tt *vmw_tt) switch (dev_priv->map_mode) { case vmw_dma_map_bind: case vmw_dma_map_populate: - vsgt->sgt = &vmw_tt->sgt; - ret = sg_alloc_table_from_pages_segment( - &vmw_tt->sgt, vsgt->pages, vsgt->num_pages, 0, - (unsigned long)vsgt->num_pages << PAGE_SHIFT, - dma_get_max_seg_size(dev_priv->drm.dev), GFP_KERNEL); - if (ret) - goto out_sg_alloc_fail; + if (vmw_tt->dma_ttm.page_flags & TTM_TT_FLAG_EXTERNAL) { + vsgt->sgt = vmw_tt->dma_ttm.sg; + } else { + vsgt->sgt = &vmw_tt->sgt; + ret = sg_alloc_table_from_pages_segment(&vmw_tt->sgt, + vsgt->pages, vsgt->num_pages, 0, + (unsigned long)vsgt->num_pages << PAGE_SHIFT, + dma_get_max_seg_size(dev_priv->drm.dev), + GFP_KERNEL); + if (ret) + goto out_sg_alloc_fail; + } ret = vmw_ttm_map_for_dma(vmw_tt); if (unlikely(ret != 0)) @@ -209,8 +214,9 @@ static int vmw_ttm_map_dma(struct vmw_ttm_tt *vmw_tt) return 0; out_map_fail: - sg_free_table(vmw_tt->vsgt.sgt); - vmw_tt->vsgt.sgt = NULL; + drm_warn(&dev_priv->drm, "VSG table map failed!"); + sg_free_table(vsgt->sgt); + vsgt->sgt = NULL; out_sg_alloc_fail: return ret; } @@ -356,15 +362,17 @@ static void vmw_ttm_destroy(struct ttm_device *bdev, struct ttm_tt *ttm) static int vmw_ttm_populate(struct ttm_device *bdev, struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) { - int ret; + bool external = (ttm->page_flags & TTM_TT_FLAG_EXTERNAL) != 0; - /* TODO: maybe completely drop this ? */ if (ttm_tt_is_populated(ttm)) return 0; - ret = ttm_pool_alloc(&bdev->pool, ttm, ctx); + if (external && ttm->sg) + return drm_prime_sg_to_dma_addr_array(ttm->sg, + ttm->dma_address, + ttm->num_pages); - return ret; + return ttm_pool_alloc(&bdev->pool, ttm, ctx); } static void vmw_ttm_unpopulate(struct ttm_device *bdev, @@ -372,6 +380,10 @@ static void vmw_ttm_unpopulate(struct ttm_device *bdev, { struct vmw_ttm_tt *vmw_tt = container_of(ttm, struct vmw_ttm_tt, dma_ttm); + bool external = (ttm->page_flags & TTM_TT_FLAG_EXTERNAL) != 0; + + if (external) + return; vmw_ttm_unbind(bdev, ttm); @@ -390,6 +402,7 @@ static struct ttm_tt *vmw_ttm_tt_create(struct ttm_buffer_object *bo, { struct vmw_ttm_tt *vmw_be; int ret; + bool external = bo->type == ttm_bo_type_sg; vmw_be = kzalloc(sizeof(*vmw_be), GFP_KERNEL); if (!vmw_be) @@ -398,7 +411,10 @@ static struct ttm_tt *vmw_ttm_tt_create(struct ttm_buffer_object *bo, vmw_be->dev_priv = vmw_priv_from_ttm(bo->bdev); vmw_be->mob = NULL; - if (vmw_be->dev_priv->map_mode == vmw_dma_alloc_coherent) + if (external) + page_flags |= TTM_TT_FLAG_EXTERNAL | TTM_TT_FLAG_EXTERNAL_MAPPABLE; + + if (vmw_be->dev_priv->map_mode == vmw_dma_alloc_coherent || external) ret = ttm_sg_tt_init(&vmw_be->dma_ttm, bo, page_flags, ttm_cached); else -- 2.40.1

1 year, 6 months

1
0
0 0

[PATCH v2] firmware: qcom: uefisecapp: Fix memory related IO errors and crashes

by Maximilian Luz

It turns out that while the QSEECOM APP_SEND command has specific fields for request and response buffers, uefisecapp expects them both to be in a single memory region. Failure to adhere to this has (so far) resulted in either no response being written to the response buffer (causing an EIO to be emitted down the line), the SCM call to fail with EINVAL (i.e., directly from TZ/firmware), or the device to be hard-reset. While this issue can be triggered deterministically, in the current form it seems to happen rather sporadically (which is why it has gone unnoticed during earlier testing). This is likely due to the two kzalloc() calls (for request and response) being directly after each other. Which means that those likely return consecutive regions most of the time, especially when not much else is going on in the system. Fix this by allocating a single memory region for both request and response buffers, properly aligning both structs inside it. This unfortunately also means that the qcom_scm_qseecom_app_send() interface needs to be restructured, as it should no longer map the DMA regions separately. Therefore, move the responsibility of DMA allocation (or mapping) to the caller. Fixes: 759e7a2b62eb ("firmware: Add support for Qualcomm UEFI Secure Application") Cc: stable(a)vger.kernel.org # 6.7 Tested-by: Johan Hovold <johan+linaro(a)kernel.org> Reviewed-by: Johan Hovold <johan+linaro(a)kernel.org> Signed-off-by: Maximilian Luz <luzmaximilian(a)gmail.com> --- Changes in v2: - rename DMA related variables - replace _phys suffix with _dma - drop _virt suffix - use DMA-based naming in comments instead of referring to physical addresses/memory --- .../firmware/qcom/qcom_qseecom_uefisecapp.c | 137 ++++++++++++------ drivers/firmware/qcom/qcom_scm.c | 37 +---- include/linux/firmware/qcom/qcom_qseecom.h | 55 ++++++- include/linux/firmware/qcom/qcom_scm.h | 10 +- 4 files changed, 153 insertions(+), 86 deletions(-) diff --git a/drivers/firmware/qcom/qcom_qseecom_uefisecapp.c b/drivers/firmware/qcom/qcom_qseecom_uefisecapp.c index 32188f098ef34..bc550ad0dbe0c 100644 --- a/drivers/firmware/qcom/qcom_qseecom_uefisecapp.c +++ b/drivers/firmware/qcom/qcom_qseecom_uefisecapp.c @@ -221,6 +221,19 @@ struct qsee_rsp_uefi_query_variable_info { * alignment of 8 bytes (64 bits) for GUIDs. Our definition of efi_guid_t, * however, has an alignment of 4 byte (32 bits). So far, this seems to work * fine here. See also the comment on the typedef of efi_guid_t. + * + * Note: It looks like uefisecapp is quite picky about how the memory passed to + * it is structured and aligned. In particular the request/response setup used + * for QSEE_CMD_UEFI_GET_VARIABLE. While qcom_qseecom_app_send(), in theory, + * accepts separate buffers/addresses for the request and response parts, in + * practice, however, it seems to expect them to be both part of a larger + * contiguous block. We initially allocated separate buffers for the request + * and response but this caused the QSEE_CMD_UEFI_GET_VARIABLE command to + * either not write any response to the response buffer or outright crash the + * device. Therefore, we now allocate a single contiguous block of DMA memory + * for both and properly align the data using the macros below. In particular, + * request and response structs are aligned at 8 byte (via __reqdata_offs()), + * following the driver that this has been reverse-engineered from. */ #define qcuefi_buf_align_fields(fields...) \ ({ \ @@ -244,6 +257,12 @@ struct qsee_rsp_uefi_query_variable_info { #define __array_offs(type, count, offset) \ __field_impl(sizeof(type) * (count), __alignof__(type), offset) +#define __array_offs_aligned(type, count, align, offset) \ + __field_impl(sizeof(type) * (count), align, offset) + +#define __reqdata_offs(size, offset) \ + __array_offs_aligned(u8, size, 8, offset) + #define __array(type, count) __array_offs(type, count, NULL) #define __field_offs(type, offset) __array_offs(type, 1, offset) #define __field(type) __array_offs(type, 1, NULL) @@ -277,10 +296,15 @@ static efi_status_t qsee_uefi_get_variable(struct qcuefi_client *qcuefi, const e unsigned long buffer_size = *data_size; efi_status_t efi_status = EFI_SUCCESS; unsigned long name_length; + dma_addr_t cmd_buf_dma; + size_t cmd_buf_size; + void *cmd_buf; size_t guid_offs; size_t name_offs; size_t req_size; size_t rsp_size; + size_t req_offs; + size_t rsp_offs; ssize_t status; if (!name || !guid) @@ -304,17 +328,19 @@ static efi_status_t qsee_uefi_get_variable(struct qcuefi_client *qcuefi, const e __array(u8, buffer_size) ); - req_data = kzalloc(req_size, GFP_KERNEL); - if (!req_data) { + cmd_buf_size = qcuefi_buf_align_fields( + __reqdata_offs(req_size, &req_offs) + __reqdata_offs(rsp_size, &rsp_offs) + ); + + cmd_buf = qseecom_dma_alloc(qcuefi->client, cmd_buf_size, &cmd_buf_dma, GFP_KERNEL); + if (!cmd_buf) { efi_status = EFI_OUT_OF_RESOURCES; goto out; } - rsp_data = kzalloc(rsp_size, GFP_KERNEL); - if (!rsp_data) { - efi_status = EFI_OUT_OF_RESOURCES; - goto out_free_req; - } + req_data = cmd_buf + req_offs; + rsp_data = cmd_buf + rsp_offs; req_data->command_id = QSEE_CMD_UEFI_GET_VARIABLE; req_data->data_size = buffer_size; @@ -332,7 +358,9 @@ static efi_status_t qsee_uefi_get_variable(struct qcuefi_client *qcuefi, const e memcpy(((void *)req_data) + req_data->guid_offset, guid, req_data->guid_size); - status = qcom_qseecom_app_send(qcuefi->client, req_data, req_size, rsp_data, rsp_size); + status = qcom_qseecom_app_send(qcuefi->client, + cmd_buf_dma + req_offs, req_size, + cmd_buf_dma + rsp_offs, rsp_size); if (status) { efi_status = EFI_DEVICE_ERROR; goto out_free; @@ -407,9 +435,7 @@ static efi_status_t qsee_uefi_get_variable(struct qcuefi_client *qcuefi, const e memcpy(data, ((void *)rsp_data) + rsp_data->data_offset, rsp_data->data_size); out_free: - kfree(rsp_data); -out_free_req: - kfree(req_data); + qseecom_dma_free(qcuefi->client, cmd_buf_size, cmd_buf, cmd_buf_dma); out: return efi_status; } @@ -422,10 +448,15 @@ static efi_status_t qsee_uefi_set_variable(struct qcuefi_client *qcuefi, const e struct qsee_rsp_uefi_set_variable *rsp_data; efi_status_t efi_status = EFI_SUCCESS; unsigned long name_length; + dma_addr_t cmd_buf_dma; + size_t cmd_buf_size; + void *cmd_buf; size_t name_offs; size_t guid_offs; size_t data_offs; size_t req_size; + size_t req_offs; + size_t rsp_offs; ssize_t status; if (!name || !guid) @@ -450,17 +481,19 @@ static efi_status_t qsee_uefi_set_variable(struct qcuefi_client *qcuefi, const e __array_offs(u8, data_size, &data_offs) ); - req_data = kzalloc(req_size, GFP_KERNEL); - if (!req_data) { + cmd_buf_size = qcuefi_buf_align_fields( + __reqdata_offs(req_size, &req_offs) + __reqdata_offs(sizeof(*rsp_data), &rsp_offs) + ); + + cmd_buf = qseecom_dma_alloc(qcuefi->client, cmd_buf_size, &cmd_buf_dma, GFP_KERNEL); + if (!cmd_buf) { efi_status = EFI_OUT_OF_RESOURCES; goto out; } - rsp_data = kzalloc(sizeof(*rsp_data), GFP_KERNEL); - if (!rsp_data) { - efi_status = EFI_OUT_OF_RESOURCES; - goto out_free_req; - } + req_data = cmd_buf + req_offs; + rsp_data = cmd_buf + rsp_offs; req_data->command_id = QSEE_CMD_UEFI_SET_VARIABLE; req_data->attributes = attributes; @@ -483,8 +516,9 @@ static efi_status_t qsee_uefi_set_variable(struct qcuefi_client *qcuefi, const e if (data_size) memcpy(((void *)req_data) + req_data->data_offset, data, req_data->data_size); - status = qcom_qseecom_app_send(qcuefi->client, req_data, req_size, rsp_data, - sizeof(*rsp_data)); + status = qcom_qseecom_app_send(qcuefi->client, + cmd_buf_dma + req_offs, req_size, + cmd_buf_dma + rsp_offs, sizeof(*rsp_data)); if (status) { efi_status = EFI_DEVICE_ERROR; goto out_free; @@ -507,9 +541,7 @@ static efi_status_t qsee_uefi_set_variable(struct qcuefi_client *qcuefi, const e } out_free: - kfree(rsp_data); -out_free_req: - kfree(req_data); + qseecom_dma_free(qcuefi->client, cmd_buf_size, cmd_buf, cmd_buf_dma); out: return efi_status; } @@ -521,10 +553,15 @@ static efi_status_t qsee_uefi_get_next_variable(struct qcuefi_client *qcuefi, struct qsee_req_uefi_get_next_variable *req_data; struct qsee_rsp_uefi_get_next_variable *rsp_data; efi_status_t efi_status = EFI_SUCCESS; + dma_addr_t cmd_buf_dma; + size_t cmd_buf_size; + void *cmd_buf; size_t guid_offs; size_t name_offs; size_t req_size; size_t rsp_size; + size_t req_offs; + size_t rsp_offs; ssize_t status; if (!name_size || !name || !guid) @@ -545,17 +582,19 @@ static efi_status_t qsee_uefi_get_next_variable(struct qcuefi_client *qcuefi, __array(*name, *name_size / sizeof(*name)) ); - req_data = kzalloc(req_size, GFP_KERNEL); - if (!req_data) { + cmd_buf_size = qcuefi_buf_align_fields( + __reqdata_offs(req_size, &req_offs) + __reqdata_offs(rsp_size, &rsp_offs) + ); + + cmd_buf = qseecom_dma_alloc(qcuefi->client, cmd_buf_size, &cmd_buf_dma, GFP_KERNEL); + if (!cmd_buf) { efi_status = EFI_OUT_OF_RESOURCES; goto out; } - rsp_data = kzalloc(rsp_size, GFP_KERNEL); - if (!rsp_data) { - efi_status = EFI_OUT_OF_RESOURCES; - goto out_free_req; - } + req_data = cmd_buf + req_offs; + rsp_data = cmd_buf + rsp_offs; req_data->command_id = QSEE_CMD_UEFI_GET_NEXT_VARIABLE; req_data->guid_offset = guid_offs; @@ -572,7 +611,9 @@ static efi_status_t qsee_uefi_get_next_variable(struct qcuefi_client *qcuefi, goto out_free; } - status = qcom_qseecom_app_send(qcuefi->client, req_data, req_size, rsp_data, rsp_size); + status = qcom_qseecom_app_send(qcuefi->client, + cmd_buf_dma + req_offs, req_size, + cmd_buf_dma + rsp_offs, rsp_size); if (status) { efi_status = EFI_DEVICE_ERROR; goto out_free; @@ -645,9 +686,7 @@ static efi_status_t qsee_uefi_get_next_variable(struct qcuefi_client *qcuefi, } out_free: - kfree(rsp_data); -out_free_req: - kfree(req_data); + qseecom_dma_free(qcuefi->client, cmd_buf_size, cmd_buf, cmd_buf_dma); out: return efi_status; } @@ -659,26 +698,34 @@ static efi_status_t qsee_uefi_query_variable_info(struct qcuefi_client *qcuefi, struct qsee_req_uefi_query_variable_info *req_data; struct qsee_rsp_uefi_query_variable_info *rsp_data; efi_status_t efi_status = EFI_SUCCESS; + dma_addr_t cmd_buf_dma; + size_t cmd_buf_size; + void *cmd_buf; + size_t req_offs; + size_t rsp_offs; int status; - req_data = kzalloc(sizeof(*req_data), GFP_KERNEL); - if (!req_data) { + cmd_buf_size = qcuefi_buf_align_fields( + __reqdata_offs(sizeof(*req_data), &req_offs) + __reqdata_offs(sizeof(*rsp_data), &rsp_offs) + ); + + cmd_buf = qseecom_dma_alloc(qcuefi->client, cmd_buf_size, &cmd_buf_dma, GFP_KERNEL); + if (!cmd_buf) { efi_status = EFI_OUT_OF_RESOURCES; goto out; } - rsp_data = kzalloc(sizeof(*rsp_data), GFP_KERNEL); - if (!rsp_data) { - efi_status = EFI_OUT_OF_RESOURCES; - goto out_free_req; - } + req_data = cmd_buf + req_offs; + rsp_data = cmd_buf + rsp_offs; req_data->command_id = QSEE_CMD_UEFI_QUERY_VARIABLE_INFO; req_data->attributes = attr; req_data->length = sizeof(*req_data); - status = qcom_qseecom_app_send(qcuefi->client, req_data, sizeof(*req_data), rsp_data, - sizeof(*rsp_data)); + status = qcom_qseecom_app_send(qcuefi->client, + cmd_buf_dma + req_offs, sizeof(*req_data), + cmd_buf_dma + rsp_offs, sizeof(*rsp_data)); if (status) { efi_status = EFI_DEVICE_ERROR; goto out_free; @@ -711,9 +758,7 @@ static efi_status_t qsee_uefi_query_variable_info(struct qcuefi_client *qcuefi, *max_variable_size = rsp_data->max_variable_size; out_free: - kfree(rsp_data); -out_free_req: - kfree(req_data); + qseecom_dma_free(qcuefi->client, cmd_buf_size, cmd_buf, cmd_buf_dma); out: return efi_status; } diff --git a/drivers/firmware/qcom/qcom_scm.c b/drivers/firmware/qcom/qcom_scm.c index 81c15aeff934a..f9c9fcfeaadaf 100644 --- a/drivers/firmware/qcom/qcom_scm.c +++ b/drivers/firmware/qcom/qcom_scm.c @@ -1579,9 +1579,9 @@ EXPORT_SYMBOL_GPL(qcom_scm_qseecom_app_get_id); /** * qcom_scm_qseecom_app_send() - Send to and receive data from a given QSEE app. * @app_id: The ID of the target app. - * @req: Request buffer sent to the app (must be DMA-mappable). + * @req: DMA address of the request buffer sent to the app. * @req_size: Size of the request buffer. - * @rsp: Response buffer, written to by the app (must be DMA-mappable). + * @rsp: DMA address of the response buffer, written to by the app. * @rsp_size: Size of the response buffer. * * Sends a request to the QSEE app associated with the given ID and read back @@ -1592,33 +1592,13 @@ EXPORT_SYMBOL_GPL(qcom_scm_qseecom_app_get_id); * * Return: Zero on success, nonzero on failure. */ -int qcom_scm_qseecom_app_send(u32 app_id, void *req, size_t req_size, void *rsp, - size_t rsp_size) +int qcom_scm_qseecom_app_send(u32 app_id, dma_addr_t req, size_t req_size, + dma_addr_t rsp, size_t rsp_size) { struct qcom_scm_qseecom_resp res = {}; struct qcom_scm_desc desc = {}; - dma_addr_t req_phys; - dma_addr_t rsp_phys; int status; - /* Map request buffer */ - req_phys = dma_map_single(__scm->dev, req, req_size, DMA_TO_DEVICE); - status = dma_mapping_error(__scm->dev, req_phys); - if (status) { - dev_err(__scm->dev, "qseecom: failed to map request buffer\n"); - return status; - } - - /* Map response buffer */ - rsp_phys = dma_map_single(__scm->dev, rsp, rsp_size, DMA_FROM_DEVICE); - status = dma_mapping_error(__scm->dev, rsp_phys); - if (status) { - dma_unmap_single(__scm->dev, req_phys, req_size, DMA_TO_DEVICE); - dev_err(__scm->dev, "qseecom: failed to map response buffer\n"); - return status; - } - - /* Set up SCM call data */ desc.owner = QSEECOM_TZ_OWNER_TZ_APPS; desc.svc = QSEECOM_TZ_SVC_APP_ID_PLACEHOLDER; desc.cmd = QSEECOM_TZ_CMD_APP_SEND; @@ -1626,18 +1606,13 @@ int qcom_scm_qseecom_app_send(u32 app_id, void *req, size_t req_size, void *rsp, QCOM_SCM_RW, QCOM_SCM_VAL, QCOM_SCM_RW, QCOM_SCM_VAL); desc.args[0] = app_id; - desc.args[1] = req_phys; + desc.args[1] = req; desc.args[2] = req_size; - desc.args[3] = rsp_phys; + desc.args[3] = rsp; desc.args[4] = rsp_size; - /* Perform call */ status = qcom_scm_qseecom_call(&desc, &res); - /* Unmap buffers */ - dma_unmap_single(__scm->dev, rsp_phys, rsp_size, DMA_FROM_DEVICE); - dma_unmap_single(__scm->dev, req_phys, req_size, DMA_TO_DEVICE); - if (status) return status; diff --git a/include/linux/firmware/qcom/qcom_qseecom.h b/include/linux/firmware/qcom/qcom_qseecom.h index 5c28298a98bec..366243ee96096 100644 --- a/include/linux/firmware/qcom/qcom_qseecom.h +++ b/include/linux/firmware/qcom/qcom_qseecom.h @@ -10,6 +10,7 @@ #define __QCOM_QSEECOM_H #include <linux/auxiliary_bus.h> +#include <linux/dma-mapping.h> #include <linux/types.h> #include <linux/firmware/qcom/qcom_scm.h> @@ -24,12 +25,57 @@ struct qseecom_client { u32 app_id; }; +/** + * qseecom_scm_dev() - Get the SCM device associated with the QSEECOM client. + * @client: The QSEECOM client device. + * + * Returns the SCM device under which the provided QSEECOM client device + * operates. This function is intended to be used for DMA allocations. + */ +static inline struct device *qseecom_scm_dev(struct qseecom_client *client) +{ + return client->aux_dev.dev.parent->parent; +} + +/** + * qseecom_dma_alloc() - Allocate DMA memory for a QSEECOM client. + * @client: The QSEECOM client to allocate the memory for. + * @size: The number of bytes to allocate. + * @dma_handle: Pointer to where the DMA address should be stored. + * @gfp: Allocation flags. + * + * Wrapper function for dma_alloc_coherent(), allocating DMA memory usable for + * TZ/QSEECOM communication. Refer to dma_alloc_coherent() for details. + */ +static inline void *qseecom_dma_alloc(struct qseecom_client *client, size_t size, + dma_addr_t *dma_handle, gfp_t gfp) +{ + return dma_alloc_coherent(qseecom_scm_dev(client), size, dma_handle, gfp); +} + +/** + * dma_free_coherent() - Free QSEECOM DMA memory. + * @client: The QSEECOM client for which the memory has been allocated. + * @size: The number of bytes allocated. + * @cpu_addr: Virtual memory address to free. + * @dma_handle: DMA memory address to free. + * + * Wrapper function for dma_free_coherent(), freeing memory previously + * allocated with qseecom_dma_alloc(). Refer to dma_free_coherent() for + * details. + */ +static inline void qseecom_dma_free(struct qseecom_client *client, size_t size, + void *cpu_addr, dma_addr_t dma_handle) +{ + return dma_free_coherent(qseecom_scm_dev(client), size, cpu_addr, dma_handle); +} + /** * qcom_qseecom_app_send() - Send to and receive data from a given QSEE app. * @client: The QSEECOM client associated with the target app. - * @req: Request buffer sent to the app (must be DMA-mappable). + * @req: DMA address of the request buffer sent to the app. * @req_size: Size of the request buffer. - * @rsp: Response buffer, written to by the app (must be DMA-mappable). + * @rsp: DMA address of the response buffer, written to by the app. * @rsp_size: Size of the response buffer. * * Sends a request to the QSEE app associated with the given client and read @@ -43,8 +89,9 @@ struct qseecom_client { * * Return: Zero on success, nonzero on failure. */ -static inline int qcom_qseecom_app_send(struct qseecom_client *client, void *req, size_t req_size, - void *rsp, size_t rsp_size) +static inline int qcom_qseecom_app_send(struct qseecom_client *client, + dma_addr_t req, size_t req_size, + dma_addr_t rsp, size_t rsp_size) { return qcom_scm_qseecom_app_send(client->app_id, req, req_size, rsp, rsp_size); } diff --git a/include/linux/firmware/qcom/qcom_scm.h b/include/linux/firmware/qcom/qcom_scm.h index ccaf288460546..aaa19f93ac430 100644 --- a/include/linux/firmware/qcom/qcom_scm.h +++ b/include/linux/firmware/qcom/qcom_scm.h @@ -118,8 +118,8 @@ bool qcom_scm_lmh_dcvsh_available(void); #ifdef CONFIG_QCOM_QSEECOM int qcom_scm_qseecom_app_get_id(const char *app_name, u32 *app_id); -int qcom_scm_qseecom_app_send(u32 app_id, void *req, size_t req_size, void *rsp, - size_t rsp_size); +int qcom_scm_qseecom_app_send(u32 app_id, dma_addr_t req, size_t req_size, + dma_addr_t rsp, size_t rsp_size); #else /* CONFIG_QCOM_QSEECOM */ @@ -128,9 +128,9 @@ static inline int qcom_scm_qseecom_app_get_id(const char *app_name, u32 *app_id) return -EINVAL; } -static inline int qcom_scm_qseecom_app_send(u32 app_id, void *req, - size_t req_size, void *rsp, - size_t rsp_size) +static inline int qcom_scm_qseecom_app_send(u32 app_id, + dma_addr_t req, size_t req_size, + dma_addr_t rsp, size_t rsp_size) { return -EINVAL; } -- 2.44.0

1 year, 6 months

3
2
0 0

[PATCH v2 5.10/5.15] ata: libata-scsi: check cdb length for VARIABLE_LENGTH_CMD commands

by Mikhail Ukhin

No upstream commit exists for this patch. Fuzzing of 5.10 stable branch reports a slab-out-of-bounds error in ata_scsi_pass_thru. The error is fixed in 5.18 by commit ce70fd9a551a ("scsi: core: Remove the cmd field from struct scsi_request") upstream. Backporting this commit would require significant changes to the code so it is bettter to use a simple fix for that particular error. The problem is that the length of the received SCSI command is not validated if scsi_op == VARIABLE_LENGTH_CMD. It can lead to out-of-bounds reading if the user sends a request with SCSI command of length less than 32. Found by Linux Verification Center (linuxtesting.org) with Syzkaller. Signed-off-by: Artem Sadovnikov <ancowi69(a)gmail.com> Signed-off-by: Mikhail Ivanov <iwanov-23(a)bk.ru> Signed-off-by: Mikhail Ukhin <mish.uxin2012(a)yandex.ru> --- v2: The new addresses were added and the text was updated. drivers/ata/libata-scsi.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c index dfa090ccd21c..77589e911d3d 100644 --- a/drivers/ata/libata-scsi.c +++ b/drivers/ata/libata-scsi.c @@ -4065,6 +4065,9 @@ int __ata_scsi_queuecmd(struct scsi_cmnd *scmd, struct ata_device *dev) if (unlikely(!scmd->cmd_len)) goto bad_cdb_len; + + if (scsi_op == VARIABLE_LENGTH_CMD && scmd->cmd_len < 32) + goto bad_cdb_len; if (dev->class == ATA_DEV_ATA || dev->class == ATA_DEV_ZAC) { if (unlikely(scmd->cmd_len > dev->cdb_len)) -- 2.25.1

1 year, 6 months

2
1
0 0

[PATCH v2] ACPI: CPPC: Fix access width used for PCC registers

by Vanshidhar Konda

commit 2f4a4d63a193be6fd530d180bb13c3592052904c modified cpc_read/cpc_write to use access_width to read CPC registers. For PCC registers the access width field in the ACPI register macro specifies the PCC subspace id. For non-zero PCC subspace id the access width is incorrectly treated as access width. This causes errors when reading from PCC registers in the CPPC driver. For PCC registers base the size of read/write on the bit width field. The debug message in cpc_read/cpc_write is updated to print relevant information for the address space type used to read the register. Signed-off-by: Vanshidhar Konda <vanshikonda(a)os.amperecomputing.com> Tested-by: Jarred White <jarredwhite(a)linux.microsoft.com> Reviewed-by: Jarred White <jarredwhite(a)linux.microsoft.com> Cc: 5.15+ <stable(a)vger.kernel.org> # 5.15+ --- When testing v6.9-rc1 kernel on AmpereOne system dmesg showed that cpufreq policy had failed to initialize on some cores during boot because cpufreq->get() always returned 0. On this system CPPC registers are in PCC subspace index 2 that are 32 bits wide. With this patch the CPPC driver interpreted the access width field as 16 bits, causing the register read to roll over too quickly to provide valid values during frequency computation. v2: - Use size variable in debug print message - Use size instead of reg->bit_width for acpi_os_read_memory and acpi_os_write_memory drivers/acpi/cppc_acpi.c | 53 ++++++++++++++++++++++++++++------------ 1 file changed, 37 insertions(+), 16 deletions(-) diff --git a/drivers/acpi/cppc_acpi.c b/drivers/acpi/cppc_acpi.c index 4bfbe55553f4..a037e9d15f48 100644 --- a/drivers/acpi/cppc_acpi.c +++ b/drivers/acpi/cppc_acpi.c @@ -1002,14 +1002,14 @@ static int cpc_read(int cpu, struct cpc_register_resource *reg_res, u64 *val) } *val = 0; + size = GET_BIT_WIDTH(reg); if (reg->space_id == ACPI_ADR_SPACE_SYSTEM_IO) { - u32 width = GET_BIT_WIDTH(reg); u32 val_u32; acpi_status status; status = acpi_os_read_port((acpi_io_address)reg->address, - &val_u32, width); + &val_u32, size); if (ACPI_FAILURE(status)) { pr_debug("Error: Failed to read SystemIO port %llx\n", reg->address); @@ -1018,17 +1018,22 @@ static int cpc_read(int cpu, struct cpc_register_resource *reg_res, u64 *val) *val = val_u32; return 0; - } else if (reg->space_id == ACPI_ADR_SPACE_PLATFORM_COMM && pcc_ss_id >= 0) + } else if (reg->space_id == ACPI_ADR_SPACE_PLATFORM_COMM && pcc_ss_id >= 0) { + /* + * For registers in PCC space, the register size is determined + * by the bit width field; the access size is used to indicate + * the PCC subspace id. + */ + size = reg->bit_width; vaddr = GET_PCC_VADDR(reg->address, pcc_ss_id); + } else if (reg->space_id == ACPI_ADR_SPACE_SYSTEM_MEMORY) vaddr = reg_res->sys_mem_vaddr; else if (reg->space_id == ACPI_ADR_SPACE_FIXED_HARDWARE) return cpc_read_ffh(cpu, reg, val); else return acpi_os_read_memory((acpi_physical_address)reg->address, - val, reg->bit_width); - - size = GET_BIT_WIDTH(reg); + val, size); switch (size) { case 8: @@ -1044,8 +1049,13 @@ static int cpc_read(int cpu, struct cpc_register_resource *reg_res, u64 *val) *val = readq_relaxed(vaddr); break; default: - pr_debug("Error: Cannot read %u bit width from PCC for ss: %d\n", - reg->bit_width, pcc_ss_id); + if (reg->space_id == ACPI_ADR_SPACE_SYSTEM_MEMORY) { + pr_debug("Error: Cannot read %u width from for system memory: 0x%llx\n", + size, reg->address); + } else if (reg->space_id == ACPI_ADR_SPACE_PLATFORM_COMM) { + pr_debug("Error: Cannot read %u bit width to PCC for ss: %d\n", + size, pcc_ss_id); + } return -EFAULT; } @@ -1063,12 +1073,13 @@ static int cpc_write(int cpu, struct cpc_register_resource *reg_res, u64 val) int pcc_ss_id = per_cpu(cpu_pcc_subspace_idx, cpu); struct cpc_reg *reg = &reg_res->cpc_entry.reg; + size = GET_BIT_WIDTH(reg); + if (reg->space_id == ACPI_ADR_SPACE_SYSTEM_IO) { - u32 width = GET_BIT_WIDTH(reg); acpi_status status; status = acpi_os_write_port((acpi_io_address)reg->address, - (u32)val, width); + (u32)val, size); if (ACPI_FAILURE(status)) { pr_debug("Error: Failed to write SystemIO port %llx\n", reg->address); @@ -1076,17 +1087,22 @@ static int cpc_write(int cpu, struct cpc_register_resource *reg_res, u64 val) } return 0; - } else if (reg->space_id == ACPI_ADR_SPACE_PLATFORM_COMM && pcc_ss_id >= 0) + } else if (reg->space_id == ACPI_ADR_SPACE_PLATFORM_COMM && pcc_ss_id >= 0) { + /* + * For registers in PCC space, the register size is determined + * by the bit width field; the access size is used to indicate + * the PCC subspace id. + */ + size = reg->bit_width; vaddr = GET_PCC_VADDR(reg->address, pcc_ss_id); + } else if (reg->space_id == ACPI_ADR_SPACE_SYSTEM_MEMORY) vaddr = reg_res->sys_mem_vaddr; else if (reg->space_id == ACPI_ADR_SPACE_FIXED_HARDWARE) return cpc_write_ffh(cpu, reg, val); else return acpi_os_write_memory((acpi_physical_address)reg->address, - val, reg->bit_width); - - size = GET_BIT_WIDTH(reg); + val, size); if (reg->space_id == ACPI_ADR_SPACE_SYSTEM_MEMORY) val = MASK_VAL(reg, val); @@ -1105,8 +1121,13 @@ static int cpc_write(int cpu, struct cpc_register_resource *reg_res, u64 val) writeq_relaxed(val, vaddr); break; default: - pr_debug("Error: Cannot write %u bit width to PCC for ss: %d\n", - reg->bit_width, pcc_ss_id); + if (reg->space_id == ACPI_ADR_SPACE_SYSTEM_MEMORY) { + pr_debug("Error: Cannot write %u width from for system memory: 0x%llx\n", + size, reg->address); + } else if (reg->space_id == ACPI_ADR_SPACE_PLATFORM_COMM) { + pr_debug("Error: Cannot write %u bit width to PCC for ss: %d\n", + size, pcc_ss_id); + } ret_val = -EFAULT; break; } -- 2.43.1

1 year, 6 months

2
1
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror April 2024