Skip to contentSkip to conversation

5.17.2-rc1 review

PATCH In review Greg Kroah-Hartman 16 participants 1,147 messages Apr 5, 2022 – Apr 6, 2022 Page 2 of 9

Subsystems (11)
Stable branch VFS MIPS Crypto API +7 more subsystems
lore
Share & tools
lore
author maintainer
#99 lore raw

PATCH 99/1126 ALSA: hda/realtek: Fix audio regression on Mi Notebook Pro 2020

In reply to Greg Kroah-Hartman

From: Kai-Heng Feng <[email protected]>

commit f30741cded62f87bb4b1cc58bc627f076abcaba8 upstream.

Commit 5aec98913095 ("ALSA: hda/realtek - ALC236 headset MIC recording issue") is to solve recording issue met on AL236, by matching codec variant ALC269_TYPE_ALC257 and ALC269_TYPE_ALC256.

This match can be too broad and Mi Notebook Pro 2020 is broken by the patch.

Instead, use codec ID to be narrow down the scope, in order to make ALC256 unaffected.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=215484

---
 sound/pci/hda/patch_realtek.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
sound/pci/hda/patch_realtek.c +2 -2
@@ -3617,8 +3617,8 @@ static void alc256_shutup(struct hda_cod
 	/* If disable 3k pulldown control for alc257, the Mic detection will not work correctly
 	 * when booting with headset plugged. So skip setting it for the codec alc257
 	 */
-	if (spec->codec_variant != ALC269_TYPE_ALC257 &&
-	    spec->codec_variant != ALC269_TYPE_ALC256)
+	if (codec->core.vendor_id != 0x10ec0236 &&
+	    codec->core.vendor_id != 0x10ec0257)
 		alc_update_coef_idx(codec, 0x46, 0, 3 << 12);
 
 	if (!spec->no_shutup_pins)
author maintainer
#100 lore raw

PATCH 100/1126 rtc: mc146818-lib: fix locking in mc146818_set_time

In reply to Greg Kroah-Hartman

From: Mateusz Jończyk <[email protected]>

commit 811f5559270f25c34c338d6eaa2ece2544c3d3bd upstream.

In mc146818_set_time(), CMOS_READ(RTC_CONTROL) was performed without the rtc_lock taken, which is required for CMOS accesses. Fix this.

Nothing in kernel modifies RTC_DM_BINARY, so a separate critical section is allowed here.

Fixes dcf257e92622 ("rtc: mc146818: Reduce spinlock section in mc146818_set_time()")
Cc Alessandro Zummo <[email protected]>
Cc Alexandre Belloni <[email protected]>
Cc Thomas Gleixner <[email protected]>
---
 drivers/rtc/rtc-mc146818-lib.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/rtc/rtc-mc146818-lib.c +4 -2
@@ -232,8 +232,10 @@ int mc146818_set_time(struct rtc_time *t
 	if (yrs >= 100)
 		yrs -= 100;
 
-	if (!(CMOS_READ(RTC_CONTROL) & RTC_DM_BINARY)
-	    || RTC_ALWAYS_BCD) {
+	spin_lock_irqsave(&rtc_lock, flags);
+	save_control = CMOS_READ(RTC_CONTROL);
+	spin_unlock_irqrestore(&rtc_lock, flags);
+	if (!(save_control & RTC_DM_BINARY) || RTC_ALWAYS_BCD) {
 		sec = bin2bcd(sec);
 		min = bin2bcd(min);
 		hrs = bin2bcd(hrs);
author maintainer
#101 lore raw

PATCH 101/1126 rtc: pl031: fix rtc features null pointer dereference

In reply to Greg Kroah-Hartman

From: Ali Pouladi <[email protected]>

commit ea6af39f3da50c86367a71eb3cc674ade3ed244c upstream.

When there is no interrupt line, rtc alarm feature is disabled.

The clearing of the alarm feature bit was being done prior to allocations of ldata->rtc device, resulting in a null pointer dereference.

Clear RTC_FEATURE_ALARM after the rtc device is allocated.

---
 drivers/rtc/rtc-pl031.c |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/rtc/rtc-pl031.c +3 -3
@@ -350,9 +350,6 @@ static int pl031_probe(struct amba_devic
 		}
 	}
 
-	if (!adev->irq[0])
-		clear_bit(RTC_FEATURE_ALARM, ldata->rtc->features);
-
 	device_init_wakeup(&adev->dev, true);
 	ldata->rtc = devm_rtc_allocate_device(&adev->dev);
 	if (IS_ERR(ldata->rtc)) {
@@ -360,6 +357,9 @@ static int pl031_probe(struct amba_devic
 		goto out;
 	}
 
+	if (!adev->irq[0])
+		clear_bit(RTC_FEATURE_ALARM, ldata->rtc->features);
+
 	ldata->rtc->ops = ops;
 	ldata->rtc->range_min = vendor->range_min;
 	ldata->rtc->range_max = vendor->range_max;
author maintainer
#102 lore raw

PATCH 102/1126 io_uring: ensure that fsnotify is always called

In reply to Greg Kroah-Hartman

From: Jens Axboe <[email protected]>

commit f63cf5192fe3418ad5ae1a4412eba5694b145f79 upstream.

Ensure that we call fsnotify_modify() if we write a file, and that we do fsnotify_access() if we read it. This enables anyone using inotify on the file to get notified.

Ditto for fallocate, ensure that fsnotify_modify() is called.

---
 fs/io_uring.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
fs/io_uring.c +7 -1
@@ -2813,8 +2813,12 @@ static bool io_rw_should_reissue(struct
 
 static bool __io_complete_rw_common(struct io_kiocb *req, long res)
 {
-	if (req->rw.kiocb.ki_flags & IOCB_WRITE)
+	if (req->rw.kiocb.ki_flags & IOCB_WRITE) {
 		kiocb_end_write(req);
+		fsnotify_modify(req->file);
+	} else {
+		fsnotify_access(req->file);
+	}
 	if (unlikely(res != req->result)) {
 		if ((res == -EAGAIN || res == -EOPNOTSUPP) &&
 		    io_rw_should_reissue(req)) {
@@ -4301,6 +4305,8 @@ static int io_fallocate(struct io_kiocb
 				req->sync.len);
 	if (ret < 0)
 		req_set_fail(req);
+	else
+		fsnotify_modify(req->file);
 	io_req_complete(req, ret);
 	return 0;
 }
author maintainer
#103 lore raw

PATCH 103/1126 ocfs2: fix crash when mount with quota enabled

In reply to Greg Kroah-Hartman

From: Joseph Qi <[email protected]>

commit de19433423c7bedabbd4f9a25f7dbc62c5e78921 upstream.

There is a reported crash when mounting ocfs2 with quota enabled.

  RIP: 0010:ocfs2_qinfo_lock_res_init+0x44/0x50 [ocfs2]
  Call Trace:
    ocfs2_local_read_info+0xb9/0x6f0 [ocfs2]
    dquot_load_quota_sb+0x216/0x470
    dquot_load_quota_inode+0x85/0x100
    ocfs2_enable_quotas+0xa0/0x1c0 [ocfs2]
    ocfs2_fill_super.cold+0xc8/0x1bf [ocfs2]
    mount_bdev+0x185/0x1b0
    legacy_get_tree+0x27/0x40
    vfs_get_tree+0x25/0xb0
    path_mount+0x465/0xac0
    __x64_sys_mount+0x103/0x140

It is caused by when initializing dqi_gqlock, the corresponding dqi_type and dqi_sb are not properly initialized.

This issue is introduced by commit 6c85c2c72819, which wants to avoid accessing uninitialized variables in error cases. So make global quota info properly initialized.

---
 fs/ocfs2/quota_global.c |   23 ++++++++++++-----------
 fs/ocfs2/quota_local.c  |    2 --
 2 files changed, 12 insertions(+), 13 deletions(-)
fs/ocfs2/quota_global.c +12 -11
@@ -337,7 +337,6 @@ void ocfs2_unlock_global_qf(struct ocfs2
 /* Read information header from global quota file */
 int ocfs2_global_read_info(struct super_block *sb, int type)
 {
-	struct inode *gqinode = NULL;
 	unsigned int ino[OCFS2_MAXQUOTAS] = { USER_QUOTA_SYSTEM_INODE,
 					      GROUP_QUOTA_SYSTEM_INODE };
 	struct ocfs2_global_disk_dqinfo dinfo;
@@ -346,29 +345,31 @@ int ocfs2_global_read_info(struct super_
 	u64 pcount;
 	int status;
 
+	oinfo->dqi_gi.dqi_sb = sb;
+	oinfo->dqi_gi.dqi_type = type;
+	ocfs2_qinfo_lock_res_init(&oinfo->dqi_gqlock, oinfo);
+	oinfo->dqi_gi.dqi_entry_size = sizeof(struct ocfs2_global_disk_dqblk);
+	oinfo->dqi_gi.dqi_ops = &ocfs2_global_ops;
+	oinfo->dqi_gqi_bh = NULL;
+	oinfo->dqi_gqi_count = 0;
+
 	/* Read global header */
-	gqinode = ocfs2_get_system_file_inode(OCFS2_SB(sb), ino[type],
+	oinfo->dqi_gqinode = ocfs2_get_system_file_inode(OCFS2_SB(sb), ino[type],
 			OCFS2_INVALID_SLOT);
-	if (!gqinode) {
+	if (!oinfo->dqi_gqinode) {
 		mlog(ML_ERROR, "failed to get global quota inode (type=%d)\n",
 			type);
 		status = -EINVAL;
 		goto out_err;
 	}
-	oinfo->dqi_gi.dqi_sb = sb;
-	oinfo->dqi_gi.dqi_type = type;
-	oinfo->dqi_gi.dqi_entry_size = sizeof(struct ocfs2_global_disk_dqblk);
-	oinfo->dqi_gi.dqi_ops = &ocfs2_global_ops;
-	oinfo->dqi_gqi_bh = NULL;
-	oinfo->dqi_gqi_count = 0;
-	oinfo->dqi_gqinode = gqinode;
+
 	status = ocfs2_lock_global_qf(oinfo, 0);
 	if (status < 0) {
 		mlog_errno(status);
 		goto out_err;
 	}
 
-	status = ocfs2_extent_map_get_blocks(gqinode, 0, &oinfo->dqi_giblk,
+	status = ocfs2_extent_map_get_blocks(oinfo->dqi_gqinode, 0, &oinfo->dqi_giblk,
 					     &pcount, NULL);
 	if (status < 0)
 		goto out_unlock;
fs/ocfs2/quota_local.c +0 -2
@@ -702,8 +702,6 @@ static int ocfs2_local_read_info(struct
 	info->dqi_priv = oinfo;
 	oinfo->dqi_type = type;
 	INIT_LIST_HEAD(&oinfo->dqi_chunk);
-	oinfo->dqi_gqinode = NULL;
-	ocfs2_qinfo_lock_res_init(&oinfo->dqi_gqlock, oinfo);
 	oinfo->dqi_rec = NULL;
 	oinfo->dqi_lqi_bh = NULL;
 	oinfo->dqi_libh = NULL;
author maintainer
#104 lore raw

PATCH 104/1126 drm/simpledrm: Add "panel orientation" property on non-upright mounted LCD panels

In reply to Greg Kroah-Hartman

From: Hans de Goede <[email protected]>

commit 94fa115f7b28a3f02611499175e134f0a823b686 upstream.

Some devices use e.g. a portrait panel in a standard laptop casing made for landscape panels. efifb calls drm_get_panel_orientation_quirk() and sets fb_info.fbcon_rotate_hint to make fbcon rotate the console so that it shows up-right instead of on its side.

When switching to simpledrm the fbcon renders on its side. Call the drm_connector_set_panel_orientation_with_quirk() helper to add a "panel orientation" property on devices listed in the quirk table, to make the fbcon (and aware userspace apps) rotate the image to display properly.

---
 drivers/gpu/drm/tiny/simpledrm.c |    3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/tiny/simpledrm.c +3 -0
@@ -798,6 +798,9 @@ static int simpledrm_device_init_modeset
 	if (ret)
 		return ret;
 	drm_connector_helper_add(connector, &simpledrm_connector_helper_funcs);
+	drm_connector_set_panel_orientation_with_quirk(connector,
+						       DRM_MODE_PANEL_ORIENTATION_UNKNOWN,
+						       mode->hdisplay, mode->vdisplay);
 
 	formats = simpledrm_device_formats(sdev, &nformats);
 
author maintainer
#105 lore raw

PATCH 105/1126 mm: madvise: skip unmapped vma holes passed to process_madvise

In reply to Greg Kroah-Hartman

From: Charan Teja Kalla <[email protected]>

commit 08095d6310a7ce43256b4251577bc66a25c6e1a6 upstream.

The process_madvise() system call is expected to skip holes in vma passed through 'struct iovec' vector list. But do_madvise, which process_madvise() calls for each vma, returns ENOMEM in case of unmapped holes, despite the VMA is processed.

Thus process_madvise() should treat ENOMEM as expected and consider the VMA passed to as processed and continue processing other vma's in the vector list. Returning -ENOMEM to user, despite the VMA is processed, will be unable to figure out where to start the next madvise.

Fixes ecb8ac8b1f14("mm/madvise: introduce process_madvise() syscall: an external memory hinting API")
Cc David Rientjes <[email protected]>
Cc Michal Hocko <[email protected]>
Cc Minchan Kim <[email protected]>
Cc Nadav Amit <[email protected]>
Cc Stephen Rothwell <[email protected]>
Cc Suren Baghdasaryan <[email protected]>
Cc Vlastimil Babka <[email protected]>
---
 mm/madvise.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
mm/madvise.c +8 -1
@@ -1426,9 +1426,16 @@ SYSCALL_DEFINE5(process_madvise, int, pi
 
 	while (iov_iter_count(&iter)) {
 		iovec = iov_iter_iovec(&iter);
+		/*
+		 * do_madvise returns ENOMEM if unmapped holes are present
+		 * in the passed VMA. process_madvise() is expected to skip
+		 * unmapped holes passed to it in the 'struct iovec' list
+		 * and not fail because of them. Thus treat -ENOMEM return
+		 * from do_madvise as valid and continue processing.
+		 */
 		ret = do_madvise(mm, (unsigned long)iovec.iov_base,
 					iovec.iov_len, behavior);
-		if (ret < 0)
+		if (ret < 0 && ret != -ENOMEM)
 			break;
 		iov_iter_advance(&iter, iovec.iov_len);
 	}
author maintainer
#106 lore raw

PATCH 106/1126 mm: madvise: return correct bytes advised with process_madvise

In reply to Greg Kroah-Hartman

From: Charan Teja Kalla <[email protected]>

commit 5bd009c7c9a9e888077c07535dc0c70aeab242c3 upstream.

Patch series "mm: madvise: return correct bytes processed with process_madvise", v2. With the process_madvise(), always choose to return non zero processed bytes over an error. This can help the user to know on which VMA, passed in the 'struct iovec' vector list, is failed to advise thus can take the decission of retrying/skipping on that VMA.

This patch (of 2):

The process_madvise() system call returns error even after processing some VMA's passed in the 'struct iovec' vector list which leaves the user confused to know where to restart the advise next. It is also against this syscall man page[1] documentation where it mentions that "return value may be less than the total number of requested bytes, if an error occurred after some iovec elements were already processed.".

Consider a user passed 10 VMA's in the 'struct iovec' vector list of which 9 are processed but one. Then it just returns the error caused on that failed VMA despite the first 9 VMA's processed, leaving the user confused about on which VMA it is failed. Returning the number of bytes processed here can help the user to know which VMA it is failed on and thus can retry/skip the advise on that VMA.

[1]https://man7.org/linux/man-pages/man2/process_madvise.2.html.

Fixes ecb8ac8b1f14("mm/madvise: introduce process_madvise() syscall: an external memory hinting API")
Cc Suren Baghdasaryan <[email protected]>
Cc Vlastimil Babka <[email protected]>
Cc David Rientjes <[email protected]>
Cc Stephen Rothwell <[email protected]>
Cc Minchan Kim <[email protected]>
Cc Nadav Amit <[email protected]>
Cc Michal Hocko <[email protected]>
---
 mm/madvise.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
mm/madvise.c +1 -2
@@ -1440,8 +1440,7 @@ SYSCALL_DEFINE5(process_madvise, int, pi
 		iov_iter_advance(&iter, iovec.iov_len);
 	}
 
-	if (ret == 0)
-		ret = total_len - iov_iter_count(&iter);
+	ret = (total_len - iov_iter_count(&iter)) ? : ret;
 
 release_mm:
 	mmput(mm);
author maintainer
#107 lore raw

PATCH 107/1126 Revert "mm: madvise: skip unmapped vma holes passed to process_madvise"

In reply to Greg Kroah-Hartman

From: Charan Teja Kalla <[email protected]>

commit e6b0a7b357659c332231621e4315658d062c23ee upstream.

This reverts commit 08095d6310a7 ("mm: madvise: skip unmapped vma holes passed to process_madvise") as process_madvise() fails to return the exact processed bytes in other cases too.

As an example: if process_madvise() hits mlocked pages after processing some initial bytes passed in [start, end), it just returns EINVAL although some bytes are processed. Thus making an exception only for ENOMEM is partially fixing the problem of returning the proper advised bytes.

Thus revert this patch and return proper bytes advised.

---
 mm/madvise.c |    9 +--------
 1 file changed, 1 insertion(+), 8 deletions(-)
mm/madvise.c +1 -8
@@ -1426,16 +1426,9 @@ SYSCALL_DEFINE5(process_madvise, int, pi
 
 	while (iov_iter_count(&iter)) {
 		iovec = iov_iter_iovec(&iter);
-		/*
-		 * do_madvise returns ENOMEM if unmapped holes are present
-		 * in the passed VMA. process_madvise() is expected to skip
-		 * unmapped holes passed to it in the 'struct iovec' list
-		 * and not fail because of them. Thus treat -ENOMEM return
-		 * from do_madvise as valid and continue processing.
-		 */
 		ret = do_madvise(mm, (unsigned long)iovec.iov_base,
 					iovec.iov_len, behavior);
-		if (ret < 0 && ret != -ENOMEM)
+		if (ret < 0)
 			break;
 		iov_iter_advance(&iter, iovec.iov_len);
 	}
author maintainer
#108 lore raw

PATCH 108/1126 mm,hwpoison: unmap poisoned page before invalidation

In reply to Greg Kroah-Hartman

From: Rik van Riel <[email protected]>

commit 3149c79f3cb0e2e3bafb7cfadacec090cbd250d3 upstream.

In some cases it appears the invalidation of a hwpoisoned page fails because the page is still mapped in another process. This can cause a program to be continuously restarted and die when it page faults on the page that was not invalidated. Avoid that problem by unmapping the hwpoisoned page when we find it.

Another issue is that sometimes we end up oopsing in finish_fault, if the code tries to do something with the now-NULL vmf->page. I did not hit this error when submitting the previous patch because there are several opportunities for alloc_set_pte to bail out before accessing vmf->page, and that apparently happened on those systems, and most of the time on other systems, too.

However, across several million systems that error does occur a handful of times a day. It can be avoided by returning VM_FAULT_NOPAGE which will cause do_read_fault to return before calling finish_fault.

---
 mm/memory.c |   12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)
mm/memory.c +8 -4
@@ -3893,14 +3893,18 @@ static vm_fault_t __do_fault(struct vm_f
 		return ret;
 
 	if (unlikely(PageHWPoison(vmf->page))) {
+		struct page *page = vmf->page;
 		vm_fault_t poisonret = VM_FAULT_HWPOISON;
 		if (ret & VM_FAULT_LOCKED) {
+			if (page_mapped(page))
+				unmap_mapping_pages(page_mapping(page),
+						    page->index, 1, false);
 			/* Retry if a clean page was removed from the cache. */
-			if (invalidate_inode_page(vmf->page))
-				poisonret = 0;
-			unlock_page(vmf->page);
+			if (invalidate_inode_page(page))
+				poisonret = VM_FAULT_NOPAGE;
+			unlock_page(page);
 		}
-		put_page(vmf->page);
+		put_page(page);
 		vmf->page = NULL;
 		return poisonret;
 	}
author maintainer
#109 lore raw

PATCH 109/1126 mm: only re-generate demotion targets when a numa node changes its N_CPU state

In reply to Greg Kroah-Hartman

From: Oscar Salvador <[email protected]>

commit 734c15700cdf9062ae98d8b131c6fe873dfad26d upstream.

Abhishek reported that after patch [1], hotplug operations are taking roughly double the expected time. [2]

The reason behind is that the CPU callbacks that migrate_on_reclaim_init() sets always call set_migration_target_nodes() whenever a CPU is brought up/down.

But we only care about numa nodes going from having cpus to become cpuless, and vice versa, as that influences the demotion_target order.

We do already have two CPU callbacks (vmstat_cpu_online() and vmstat_cpu_dead()) that check exactly that, so get rid of the CPU callbacks in migrate_on_reclaim_init() and only call set_migration_target_nodes() from vmstat_cpu_{dead,online}() whenever a numa node change its N_CPU state.

[1] https://lore.kernel.org/linux-mm/[email protected]/
[2] https://lore.kernel.org/linux-mm/[email protected]/

---
 include/linux/migrate.h |    8 ++++++++
 mm/migrate.c            |   47 ++++++++++-------------------------------------
 mm/vmstat.c             |   13 ++++++++++++-
 3 files changed, 30 insertions(+), 38 deletions(-)
include/linux/migrate.h +8 -0
@@ -48,7 +48,15 @@ int folio_migrate_mapping(struct address
 		struct folio *newfolio, struct folio *folio, int extra_count);
 
 extern bool numa_demotion_enabled;
+extern void migrate_on_reclaim_init(void);
+#ifdef CONFIG_HOTPLUG_CPU
+extern void set_migration_target_nodes(void);
 #else
+static inline void set_migration_target_nodes(void) {}
+#endif
+#else
+
+static inline void set_migration_target_nodes(void) {}
 
 static inline void putback_movable_pages(struct list_head *l) {}
 static inline int migrate_pages(struct list_head *l, new_page_t new,
mm/migrate.c +10 -37

68 lines not on this page · raw message

mm/vmstat.c +12 -1
@@ -28,6 +28,7 @@
 #include <linux/mm_inline.h>
 #include <linux/page_ext.h>
 #include <linux/page_owner.h>
+#include <linux/migrate.h>
 
 #include "internal.h"
 
@@ -2043,7 +2044,12 @@ static void __init init_cpu_node_state(v
 static int vmstat_cpu_online(unsigned int cpu)
 {
 	refresh_zone_stat_thresholds();
-	node_set_state(cpu_to_node(cpu), N_CPU);
+
+	if (!node_state(cpu_to_node(cpu), N_CPU)) {
+		node_set_state(cpu_to_node(cpu), N_CPU);
+		set_migration_target_nodes();
+	}
+
 	return 0;
 }
 
@@ -2066,6 +2072,8 @@ static int vmstat_cpu_dead(unsigned int
 		return 0;
 
 	node_clear_state(node, N_CPU);
+	set_migration_target_nodes();
+
 	return 0;
 }
 
@@ -2097,6 +2105,9 @@ void __init init_mm_internals(void)
 
 	start_shepherd_timer();
 #endif
+#if defined(CONFIG_MIGRATION) && defined(CONFIG_HOTPLUG_CPU)
+	migrate_on_reclaim_init();
+#endif
 #ifdef CONFIG_PROC_FS
 	proc_create_seq("buddyinfo", 0444, NULL, &fragmentation_op);
 	proc_create_seq("pagetypeinfo", 0400, NULL, &pagetypeinfo_op);
author maintainer
#110 lore raw

PATCH 110/1126 mm/kmemleak: reset tag when compare object pointer

In reply to Greg Kroah-Hartman

From: Kuan-Ying Lee <[email protected]>

commit bfc8089f00fa526dea983844c880fa8106c33ac4 upstream.

When we use HW-tag based kasan and enable vmalloc support, we hit the following bug. It is due to comparison between tagged object and non-tagged pointer.

We need to reset the kasan tag when we need to compare tagged object and non-tagged pointer.

kmemleak: [name:kmemleak&]Scan area larger than object 0xffffffe77076f440

  CPU: 4 PID: 1 Comm: init Tainted: G S      W         5.15.25-android13-0-g5cacf919c2bc #1
  Hardware name: MT6983(ENG) (DT)
  Call trace:
   add_scan_area+0xc4/0x244
   kmemleak_scan_area+0x40/0x9c
   layout_and_allocate+0x1e8/0x288
   load_module+0x2c8/0xf00
   __se_sys_finit_module+0x190/0x1d0
   __arm64_sys_finit_module+0x20/0x30
   invoke_syscall+0x60/0x170
   el0_svc_common+0xc8/0x114
   do_el0_svc+0x28/0xa0
   el0_svc+0x60/0xf8
   el0t_64_sync_handler+0x88/0xec
   el0t_64_sync+0x1b4/0x1b8
  kmemleak: [name:kmemleak&]Object 0xf5ffffe77076b000 (size 32768):
  kmemleak: [name:kmemleak&]  comm "init", pid 1, jiffies 4294894197
  kmemleak: [name:kmemleak&]  min_count = 0
  kmemleak: [name:kmemleak&]  count = 0
  kmemleak: [name:kmemleak&]  flags = 0x1
  kmemleak: [name:kmemleak&]  checksum = 0
  kmemleak: [name:kmemleak&]  backtrace:
       module_alloc+0x9c/0x120
       move_module+0x34/0x19c
       layout_and_allocate+0x1c4/0x288
       load_module+0x2c8/0xf00
       __se_sys_finit_module+0x190/0x1d0
       __arm64_sys_finit_module+0x20/0x30
       invoke_syscall+0x60/0x170
       el0_svc_common+0xc8/0x114
       do_el0_svc+0x28/0xa0
       el0_svc+0x60/0xf8
       el0t_64_sync_handler+0x88/0xec
       el0t_64_sync+0x1b4/0x1b8
---
 mm/kmemleak.c |    9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
mm/kmemleak.c +7 -2
@@ -796,6 +796,8 @@ static void add_scan_area(unsigned long
 	unsigned long flags;
 	struct kmemleak_object *object;
 	struct kmemleak_scan_area *area = NULL;
+	unsigned long untagged_ptr;
+	unsigned long untagged_objp;
 
 	object = find_and_get_object(ptr, 1);
 	if (!object) {
@@ -804,6 +806,9 @@ static void add_scan_area(unsigned long
 		return;
 	}
 
+	untagged_ptr = (unsigned long)kasan_reset_tag((void *)ptr);
+	untagged_objp = (unsigned long)kasan_reset_tag((void *)object->pointer);
+
 	if (scan_area_cache)
 		area = kmem_cache_alloc(scan_area_cache, gfp_kmemleak_mask(gfp));
 
@@ -815,8 +820,8 @@ static void add_scan_area(unsigned long
 		goto out_unlock;
 	}
 	if (size == SIZE_MAX) {
-		size = object->pointer + object->size - ptr;
-	} else if (ptr + size > object->pointer + object->size) {
+		size = untagged_objp + object->size - untagged_ptr;
+	} else if (untagged_ptr + size > untagged_objp + object->size) {
 		kmemleak_warn("Scan area larger than object 0x%08lx\n", ptr);
 		dump_object_info(object);
 		kmem_cache_free(scan_area_cache, area);
author maintainer
#111 lore raw

PATCH 111/1126 dm stats: fix too short end duration_ns when using precise_timestamps

In reply to Greg Kroah-Hartman

From: Mike Snitzer <[email protected]>

commit 0cdb90f0f306384ecbc60dfd6dc48cdbc1f2d0d8 upstream.

dm_stats_account_io()'s STAT_PRECISE_TIMESTAMPS support doesn't handle the fact that with commit b879f915bc48 ("dm: properly fix redundant bio-based IO accounting") io->start_time _may_ be in the past (meaning the start_io_acct() was deferred until later).

Add a new dm_stats_recalc_precise_timestamps() helper that will set/clear a new 'precise_timestamps' flag in the dm_stats struct based on whether any configured stats enable STAT_PRECISE_TIMESTAMPS. And update DM core's alloc_io() to use dm_stats_record_start() to set stats_aux.duration_ns if stats->precise_timestamps is true.

Also, remove unused 'last_sector' and 'last_rw' members from the dm_stats struct.

Fixes b879f915bc48 ("dm: properly fix redundant bio-based IO accounting")
---
 drivers/md/dm-stats.c |   28 +++++++++++++++++++++++++---
 drivers/md/dm-stats.h |    9 +++++++--
 drivers/md/dm.c       |    2 ++
 3 files changed, 34 insertions(+), 5 deletions(-)
drivers/md/dm-stats.c +25 -3
@@ -195,6 +195,7 @@ void dm_stats_init(struct dm_stats *stat
 
 	mutex_init(&stats->mutex);
 	INIT_LIST_HEAD(&stats->list);
+	stats->precise_timestamps = false;
 	stats->last = alloc_percpu(struct dm_stats_last_position);
 	for_each_possible_cpu(cpu) {
 		last = per_cpu_ptr(stats->last, cpu);
@@ -231,6 +232,22 @@ void dm_stats_cleanup(struct dm_stats *s
 	mutex_destroy(&stats->mutex);
 }
 
+static void dm_stats_recalc_precise_timestamps(struct dm_stats *stats)
+{
+	struct list_head *l;
+	struct dm_stat *tmp_s;
+	bool precise_timestamps = false;
+
+	list_for_each(l, &stats->list) {
+		tmp_s = container_of(l, struct dm_stat, list_entry);
+		if (tmp_s->stat_flags & STAT_PRECISE_TIMESTAMPS) {
+			precise_timestamps = true;
+			break;
+		}
+	}
+	stats->precise_timestamps = precise_timestamps;
+}
+
 static int dm_stats_create(struct dm_stats *stats, sector_t start, sector_t end,
 			   sector_t step, unsigned stat_flags,
 			   unsigned n_histogram_entries,
@@ -376,6 +393,9 @@ static int dm_stats_create(struct dm_sta
 	}
 	ret_id = s->id;
 	list_add_tail_rcu(&s->list_entry, l);
+
+	dm_stats_recalc_precise_timestamps(stats);
+
 	mutex_unlock(&stats->mutex);
 
 	resume_callback(md);
@@ -418,6 +438,9 @@ static int dm_stats_delete(struct dm_sta
 	}
 
 	list_del_rcu(&s->list_entry);
+
+	dm_stats_recalc_precise_timestamps(stats);
+
 	mutex_unlock(&stats->mutex);
 
 	/*
@@ -654,9 +677,8 @@ void dm_stats_account_io(struct dm_stats
 	got_precise_time = false;
 	list_for_each_entry_rcu(s, &stats->list, list_entry) {
 		if (s->stat_flags & STAT_PRECISE_TIMESTAMPS && !got_precise_time) {
-			if (!end)
-				stats_aux->duration_ns = ktime_to_ns(ktime_get());
-			else
+			/* start (!end) duration_ns is set by DM core's alloc_io() */
+			if (end)
 				stats_aux->duration_ns = ktime_to_ns(ktime_get()) - stats_aux->duration_ns;
 			got_precise_time = true;
 		}
drivers/md/dm-stats.h +7 -2
@@ -13,8 +13,7 @@ struct dm_stats {
 	struct mutex mutex;
 	struct list_head list;	/* list of struct dm_stat */
 	struct dm_stats_last_position __percpu *last;
-	sector_t last_sector;
-	unsigned last_rw;
+	bool precise_timestamps;
 };
 
 struct dm_stats_aux {
@@ -40,4 +39,10 @@ static inline bool dm_stats_used(struct
 	return !list_empty(&st->list);
 }
 
+static inline void dm_stats_record_start(struct dm_stats *stats, struct dm_stats_aux *aux)
+{
+	if (unlikely(stats->precise_timestamps))
+		aux->duration_ns = ktime_to_ns(ktime_get());
+}
+
 #endif
drivers/md/dm.c +2 -0
@@ -537,6 +537,8 @@ static struct dm_io *alloc_io(struct map
 
 	io->start_time = jiffies;
 
+	dm_stats_record_start(&md->stats, &io->stats_aux);
+
 	return io;
 }
 
author maintainer
#112 lore raw

PATCH 112/1126 dm: fix use-after-free in dm_cleanup_zoned_dev()

In reply to Greg Kroah-Hartman

From: Kirill Tkhai <[email protected]>

commit 588b7f5df0cb64f281290c7672470c006abe7160 upstream.

dm_cleanup_zoned_dev() uses queue, so it must be called before blk_cleanup_disk() starts its killing:

blk_cleanup_disk->blk_cleanup_queue()->kobject_put()->blk_release_queue()-> ->...RCU...->blk_free_queue_rcu()->kmem_cache_free()

Otherwise, RCU callback may be executed first and
dm_cleanup_zoned_dev() will touch free'd memory:

 BUG: KASAN: use-after-free in dm_cleanup_zoned_dev+0x33/0xd0
 Read of size 8 at addr ffff88805ac6e430 by task dmsetup/681

 CPU: 4 PID: 681 Comm: dmsetup Not tainted 5.17.0-rc2+ #6
 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-2 04/01/2014
 Call Trace:
  <TASK>
  dump_stack_lvl+0x57/0x7d
  print_address_description.constprop.0+0x1f/0x150
  ? dm_cleanup_zoned_dev+0x33/0xd0
  kasan_report.cold+0x7f/0x11b
  ? dm_cleanup_zoned_dev+0x33/0xd0
  dm_cleanup_zoned_dev+0x33/0xd0
  __dm_destroy+0x26a/0x400
  ? dm_blk_ioctl+0x230/0x230
  ? up_write+0xd8/0x270
  dev_remove+0x156/0x1d0
  ctl_ioctl+0x269/0x530
  ? table_clear+0x140/0x140
  ? lock_release+0xb2/0x750
  ? remove_all+0x40/0x40
  ? rcu_read_lock_sched_held+0x12/0x70
  ? lock_downgrade+0x3c0/0x3c0
  ? rcu_read_lock_sched_held+0x12/0x70
  dm_ctl_ioctl+0xa/0x10
  __x64_sys_ioctl+0xb9/0xf0
  do_syscall_64+0x3b/0x90
  entry_SYSCALL_64_after_hwframe+0x44/0xae
 RIP: 0033:0x7fb6dfa95c27
Fixes bb37d77239af ("dm: introduce zone append emulation")
---
 drivers/md/dm.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/md/dm.c +1 -1
@@ -1609,6 +1609,7 @@ static void cleanup_mapped_device(struct
 		md->dax_dev = NULL;
 	}
 
+	dm_cleanup_zoned_dev(md);
 	if (md->disk) {
 		spin_lock(&_minor_lock);
 		md->disk->private_data = NULL;
@@ -1629,7 +1630,6 @@ static void cleanup_mapped_device(struct
 	mutex_destroy(&md->swap_bios_lock);
 
 	dm_mq_cleanup_mapped_device(md);
-	dm_cleanup_zoned_dev(md);
 }
 
 /*
author maintainer
#113 lore raw

PATCH 113/1126 dm: interlock pending dm_io and dm_wait_for_bios_completion

In reply to Greg Kroah-Hartman

From: Mike Snitzer <[email protected]>

commit 9f6dc633761006f974701d4c88da71ab68670749 upstream.

Commit d208b89401e0 ("dm: fix mempool NULL pointer race when completing IO") didn't go far enough.

When bio_end_io_acct ends the count of in-flight I/Os may reach zero and the DM device may be suspended. There is a possibility that the suspend races with dm_stats_account_io.

Fix this by adding percpu "pending_io" counters to track outstanding dm_io. Move kicking of suspend queue to dm_io_dec_pending(). Also, rename md_in_flight_bios() to dm_in_flight_bios() and update it to iterate all pending_io counters.

Fixes d208b89401e0 ("dm: fix mempool NULL pointer race when completing IO")
---
 drivers/md/dm-core.h |    2 ++
 drivers/md/dm.c      |   35 +++++++++++++++++++++++------------
 2 files changed, 25 insertions(+), 12 deletions(-)
drivers/md/dm-core.h +2 -0
@@ -65,6 +65,8 @@ struct mapped_device {
 	struct gendisk *disk;
 	struct dax_device *dax_dev;
 
+	unsigned long __percpu *pending_io;
+
 	/*
 	 * A list of ios that arrived while we were suspended.
 	 */
drivers/md/dm.c +23 -12
@@ -507,10 +507,6 @@ static void end_io_acct(struct mapped_de
 		dm_stats_account_io(&md->stats, bio_data_dir(bio),
 				    bio->bi_iter.bi_sector, bio_sectors(bio),
 				    true, duration, stats_aux);
-
-	/* nudge anyone waiting on suspend queue */
-	if (unlikely(wq_has_sleeper(&md->wait)))
-		wake_up(&md->wait);
 }
 
 static struct dm_io *alloc_io(struct mapped_device *md, struct bio *bio)
@@ -531,6 +527,7 @@ static struct dm_io *alloc_io(struct map
 	io->magic = DM_IO_MAGIC;
 	io->status = 0;
 	atomic_set(&io->io_count, 1);
+	this_cpu_inc(*md->pending_io);
 	io->orig_bio = bio;
 	io->md = md;
 	spin_lock_init(&io->endio_lock);
@@ -828,6 +825,12 @@ void dm_io_dec_pending(struct dm_io *io,
 		stats_aux = io->stats_aux;
 		free_io(md, io);
 		end_io_acct(md, bio, start_time, &stats_aux);
+		smp_wmb();
+		this_cpu_dec(*md->pending_io);
+
+		/* nudge anyone waiting on suspend queue */
+		if (unlikely(wq_has_sleeper(&md->wait)))
+			wake_up(&md->wait);
 
 		if (io_error == BLK_STS_DM_REQUEUE)
 			return;
@@ -1622,6 +1625,11 @@ static void cleanup_mapped_device(struct
 		blk_cleanup_disk(md->disk);
 	}
 
+	if (md->pending_io) {
+		free_percpu(md->pending_io);
+		md->pending_io = NULL;
+	}
+
 	cleanup_srcu_struct(&md->io_barrier);
 
 	mutex_destroy(&md->suspend_lock);
@@ -1723,6 +1731,10 @@ static struct mapped_device *alloc_dev(i
 	if (!md->wq)
 		goto bad;
 
+	md->pending_io = alloc_percpu(unsigned long);
+	if (!md->pending_io)
+		goto bad;
+
 	dm_stats_init(&md->stats);
 
 	/* Populate the mapping, nobody knows we exist yet */
@@ -2130,16 +2142,13 @@ void dm_put(struct mapped_device *md)
 }
 EXPORT_SYMBOL_GPL(dm_put);
 
-static bool md_in_flight_bios(struct mapped_device *md)
+static bool dm_in_flight_bios(struct mapped_device *md)
 {
 	int cpu;
-	struct block_device *part = dm_disk(md)->part0;
-	long sum = 0;
+	unsigned long sum = 0;
 
-	for_each_possible_cpu(cpu) {
-		sum += part_stat_local_read_cpu(part, in_flight[0], cpu);
-		sum += part_stat_local_read_cpu(part, in_flight[1], cpu);
-	}
+	for_each_possible_cpu(cpu)
+		sum += *per_cpu_ptr(md->pending_io, cpu);
 
 	return sum != 0;
 }
@@ -2152,7 +2161,7 @@ static int dm_wait_for_bios_completion(s
 	while (true) {
 		prepare_to_wait(&md->wait, &wait, task_state);
 
-		if (!md_in_flight_bios(md))
+		if (!dm_in_flight_bios(md))
 			break;
 
 		if (signal_pending_state(task_state, current)) {
@@ -2164,6 +2173,8 @@ static int dm_wait_for_bios_completion(s
 	}
 	finish_wait(&md->wait, &wait);
 
+	smp_rmb();
+
 	return r;
 }
 
author maintainer
#114 lore raw

PATCH 114/1126 dm: fix double accounting of flush with data

In reply to Greg Kroah-Hartman

From: Mike Snitzer <[email protected]>

commit 8d394bc4adf588ca4a0650745167cb83f86c18c9 upstream.

DM handles a flush with data by first issuing an empty flush and then once it completes the REQ_PREFLUSH flag is removed and the payload is issued. The problem fixed by this commit is that both the empty flush bio and the data payload will account the full extent of the data payload.

Fix this by factoring out dm_io_acct() and having it wrap all IO accounting to set the size of bio with REQ_PREFLUSH to 0, account the IO, and then restore the original size.

---
 drivers/md/dm-stats.c |    6 ++++--
 drivers/md/dm-stats.h |    2 +-
 drivers/md/dm.c       |   47 +++++++++++++++++++++++++++++++++--------------
 3 files changed, 38 insertions(+), 17 deletions(-)
drivers/md/dm-stats.c +4 -2
@@ -644,13 +644,14 @@ static void __dm_stat_bio(struct dm_stat
 
 void dm_stats_account_io(struct dm_stats *stats, unsigned long bi_rw,
 			 sector_t bi_sector, unsigned bi_sectors, bool end,
-			 unsigned long duration_jiffies,
+			 unsigned long start_time,
 			 struct dm_stats_aux *stats_aux)
 {
 	struct dm_stat *s;
 	sector_t end_sector;
 	struct dm_stats_last_position *last;
 	bool got_precise_time;
+	unsigned long duration_jiffies = 0;
 
 	if (unlikely(!bi_sectors))
 		return;
@@ -670,7 +671,8 @@ void dm_stats_account_io(struct dm_stats
 				       ));
 		WRITE_ONCE(last->last_sector, end_sector);
 		WRITE_ONCE(last->last_rw, bi_rw);
-	}
+	} else
+		duration_jiffies = jiffies - start_time;
 
 	rcu_read_lock();
 
drivers/md/dm-stats.h +1 -1
@@ -31,7 +31,7 @@ int dm_stats_message(struct mapped_devic
 
 void dm_stats_account_io(struct dm_stats *stats, unsigned long bi_rw,
 			 sector_t bi_sector, unsigned bi_sectors, bool end,
-			 unsigned long duration_jiffies,
+			 unsigned long start_time,
 			 struct dm_stats_aux *aux);
 
 static inline bool dm_stats_used(struct dm_stats *st)
drivers/md/dm.c +33 -14

69 lines not on this page · raw message

author maintainer
#115 lore raw

PATCH 115/1126 dm integrity: set journal entry unused when shrinking device

In reply to Greg Kroah-Hartman

From: Mikulas Patocka <[email protected]>

commit cc09e8a9dec4f0e8299e80a7a2a8e6f54164a10b upstream.

Commit f6f72f32c22c ("dm integrity: don't replay journal data past the end of the device") skips journal replay if the target sector points beyond the end of the device. Unfortunatelly, it doesn't set the journal entry unused, which resulted in this BUG being triggered:
BUG_ON(!journal_entry_is_unused(je))

Fix this by calling journal_entry_set_unused() for this case.

Fixes f6f72f32c22c ("dm integrity: don't replay journal data past the end of the device")
[snitzer: revised header]
---
 drivers/md/dm-integrity.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/md/dm-integrity.c +4 -2
@@ -2473,9 +2473,11 @@ static void do_journal_write(struct dm_i
 					dm_integrity_io_error(ic, "invalid sector in journal", -EIO);
 					sec &= ~(sector_t)(ic->sectors_per_block - 1);
 				}
+				if (unlikely(sec >= ic->provided_data_sectors)) {
+					journal_entry_set_unused(je);
+					continue;
+				}
 			}
-			if (unlikely(sec >= ic->provided_data_sectors))
-				continue;
 			get_area_and_offset(ic, sec, &area, &offset);
 			restore_last_bytes(ic, access_journal_data(ic, i, j), je);
 			for (k = j + 1; k < ic->journal_section_entries; k++) {
author maintainer
#116 lore raw

PATCH 116/1126 tracing: Have trace event string test handle zero length strings

In reply to Greg Kroah-Hartman

From: Steven Rostedt (Google) <[email protected]>

commit eca344a7362e0f34f179298fd8366bcd556eede1 upstream.

If a trace event has in its TP_printk():

"%*.s", len, len ? __get_str(string) : NULL

It is perfectly valid if len is zero and passing in the NULL. Unfortunately, the runtime string check at time of reading the trace sees the NULL and flags it as a bad string and produces a WARN_ON().

Handle this case by passing into the test function if the format has an asterisk (star) and if so, if the length is zero, then mark it as safe.

Fixes 9a6944fee68e2 ("tracing: Add a verifier to check string pointers for trace events")
---
 kernel/trace/trace.c |    9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
kernel/trace/trace.c +7 -2
@@ -3663,12 +3663,17 @@ static char *trace_iter_expand_format(st
 }
 
 /* Returns true if the string is safe to dereference from an event */
-static bool trace_safe_str(struct trace_iterator *iter, const char *str)
+static bool trace_safe_str(struct trace_iterator *iter, const char *str,
+			   bool star, int len)
 {
 	unsigned long addr = (unsigned long)str;
 	struct trace_event *trace_event;
 	struct trace_event_call *event;
 
+	/* Ignore strings with no length */
+	if (star && !len)
+		return true;
+
 	/* OK if part of the event data */
 	if ((addr >= (unsigned long)iter->ent) &&
 	    (addr < (unsigned long)iter->ent + iter->ent_size))
@@ -3854,7 +3859,7 @@ void trace_check_vprintf(struct trace_it
 		 * instead. See samples/trace_events/trace-events-sample.h
 		 * for reference.
 		 */
-		if (WARN_ONCE(!trace_safe_str(iter, str),
+		if (WARN_ONCE(!trace_safe_str(iter, str, star, len),
 			      "fmt: '%s' current_buffer: '%s'",
 			      fmt, show_buffer(&iter->seq))) {
 			int ret;
author maintainer
#117 lore raw

PATCH 117/1126 drbd: fix potential silent data corruption

In reply to Greg Kroah-Hartman

From: Lars Ellenberg <[email protected]>

commit f4329d1f848ac35757d9cc5487669d19dfc5979c upstream.

Scenario:
---------

bio chain generated by blk_queue_split().
Some split bio fails and propagates its error status to the "parent" bio. But then the (last part of the) parent bio itself completes without error.

We would clobber the already recorded error status with BLK_STS_OK, causing silent data corruption.

Reproducer:
-----------

How to trigger this in the real world within seconds:

DRBD on top of degraded parity raid,
small stripe_cache_size, large read_ahead setting. Drop page cache (sysctl vm.drop_caches=1, fadvise "DONTNEED", umount and mount again, "reboot").

Cause significant read ahead.

Large read ahead request is split by blk_queue_split(). Parts of the read ahead that are already in the stripe cache, or find an available stripe cache to use, can be serviced. Parts of the read ahead that would need "too much work", would need to wait for a "stripe_head" to become available, are rejected immediately.

For larger read ahead requests that are split in many pieces, it is very likely that some "splits" will be serviced, but then the stripe cache is exhausted/busy, and the remaining ones will be rejected.

---
 drivers/block/drbd/drbd_req.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/block/drbd/drbd_req.c +2 -1
@@ -180,7 +180,8 @@ void start_new_tl_epoch(struct drbd_conn
 void complete_master_bio(struct drbd_device *device,
 		struct bio_and_error *m)
 {
-	m->bio->bi_status = errno_to_blk_status(m->error);
+	if (unlikely(m->error))
+		m->bio->bi_status = errno_to_blk_status(m->error);
 	bio_endio(m->bio);
 	dec_ap_bio(device);
 }
author maintainer
#118 lore raw

PATCH 118/1126 can: isotp: sanitize CAN ID checks in isotp_bind()

In reply to Greg Kroah-Hartman

From: Oliver Hartkopp <[email protected]>

commit 3ea566422cbde9610c2734980d1286ab681bb40e upstream.

Syzbot created an environment that lead to a state machine status that can not be reached with a compliant CAN ID address configuration. The provided address information consisted of CAN ID 0x6000001 and 0xC28001 which both boil down to 11 bit CAN IDs 0x001 in sending and receiving.

Sanitize the SFF/EFF CAN ID values before performing the address checks.

Fixes e057dd3fc20f ("can: add ISO 15765-2:2016 transport protocol")
Reported-by [email protected]
---
 net/can/isotp.c |   38 ++++++++++++++++++++------------------
 1 file changed, 20 insertions(+), 18 deletions(-)
net/can/isotp.c +20 -18
@@ -1104,6 +1104,7 @@ static int isotp_bind(struct socket *soc
 	struct net *net = sock_net(sk);
 	int ifindex;
 	struct net_device *dev;
+	canid_t tx_id, rx_id;
 	int err = 0;
 	int notify_enetdown = 0;
 	int do_rx_reg = 1;
@@ -1111,8 +1112,18 @@ static int isotp_bind(struct socket *soc
 	if (len < ISOTP_MIN_NAMELEN)
 		return -EINVAL;
 
-	if (addr->can_addr.tp.tx_id & (CAN_ERR_FLAG | CAN_RTR_FLAG))
-		return -EADDRNOTAVAIL;
+	/* sanitize tx/rx CAN identifiers */
+	tx_id = addr->can_addr.tp.tx_id;
+	if (tx_id & CAN_EFF_FLAG)
+		tx_id &= (CAN_EFF_FLAG | CAN_EFF_MASK);
+	else
+		tx_id &= CAN_SFF_MASK;
+
+	rx_id = addr->can_addr.tp.rx_id;
+	if (rx_id & CAN_EFF_FLAG)
+		rx_id &= (CAN_EFF_FLAG | CAN_EFF_MASK);
+	else
+		rx_id &= CAN_SFF_MASK;
 
 	if (!addr->can_ifindex)
 		return -ENODEV;
@@ -1124,21 +1135,13 @@ static int isotp_bind(struct socket *soc
 		do_rx_reg = 0;
 
 	/* do not validate rx address for functional addressing */
-	if (do_rx_reg) {
-		if (addr->can_addr.tp.rx_id == addr->can_addr.tp.tx_id) {
-			err = -EADDRNOTAVAIL;
-			goto out;
-		}
-
-		if (addr->can_addr.tp.rx_id & (CAN_ERR_FLAG | CAN_RTR_FLAG)) {
-			err = -EADDRNOTAVAIL;
-			goto out;
-		}
+	if (do_rx_reg && rx_id == tx_id) {
+		err = -EADDRNOTAVAIL;
+		goto out;
 	}
 
 	if (so->bound && addr->can_ifindex == so->ifindex &&
-	    addr->can_addr.tp.rx_id == so->rxid &&
-	    addr->can_addr.tp.tx_id == so->txid)
+	    rx_id == so->rxid && tx_id == so->txid)
 		goto out;
 
 	dev = dev_get_by_index(net, addr->can_ifindex);
@@ -1162,8 +1165,7 @@ static int isotp_bind(struct socket *soc
 	ifindex = dev->ifindex;
 
 	if (do_rx_reg)
-		can_rx_register(net, dev, addr->can_addr.tp.rx_id,
-				SINGLE_MASK(addr->can_addr.tp.rx_id),
+		can_rx_register(net, dev, rx_id, SINGLE_MASK(rx_id),
 				isotp_rcv, sk, "isotp", sk);
 
 	dev_put(dev);
@@ -1183,8 +1185,8 @@ static int isotp_bind(struct socket *soc
 
 	/* switch to new settings */
 	so->ifindex = ifindex;
-	so->rxid = addr->can_addr.tp.rx_id;
-	so->txid = addr->can_addr.tp.tx_id;
+	so->rxid = rx_id;
+	so->txid = tx_id;
 	so->bound = 1;
 
 out:
author maintainer
#119 lore raw

PATCH 119/1126 PCI: fu740: Force 2.5GT/s for initial device probe

In reply to Greg Kroah-Hartman

From: Ben Dooks <[email protected]>

commit a382c757ec5ef83137a86125f43a4c43dc2ab50b upstream.

The fu740 PCIe core does not probe any devices on the SiFive Unmatched board without this fix (or having U-Boot explicitly start the PCIe via either boot-script or user command). The fix is to start the link at 2.5GT/s speeds and once the link is up then change the maximum speed back to the default.

The U-Boot driver claims to set the link-speed to 2.5GT/s to get the probe to work (and U-Boot does print link up at 2.5GT/s) in the following code:
https://source.denx.de/u-boot/u-boot/-/blob/master/drivers/pci/pcie_dw_sifive.c?id=v2022.01#L271

---
 drivers/pci/controller/dwc/pcie-fu740.c |   51 +++++++++++++++++++++++++++++++-
 1 file changed, 50 insertions(+), 1 deletion(-)
drivers/pci/controller/dwc/pcie-fu740.c +50 -1
@@ -181,10 +181,59 @@ static int fu740_pcie_start_link(struct
 {
 	struct device *dev = pci->dev;
 	struct fu740_pcie *afp = dev_get_drvdata(dev);
+	u8 cap_exp = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
+	int ret;
+	u32 orig, tmp;
+
+	/*
+	 * Force 2.5GT/s when starting the link, due to some devices not
+	 * probing at higher speeds. This happens with the PCIe switch
+	 * on the Unmatched board when U-Boot has not initialised the PCIe.
+	 * The fix in U-Boot is to force 2.5GT/s, which then gets cleared
+	 * by the soft reset done by this driver.
+	 */
+	dev_dbg(dev, "cap_exp at %x\n", cap_exp);
+	dw_pcie_dbi_ro_wr_en(pci);
+
+	tmp = dw_pcie_readl_dbi(pci, cap_exp + PCI_EXP_LNKCAP);
+	orig = tmp & PCI_EXP_LNKCAP_SLS;
+	tmp &= ~PCI_EXP_LNKCAP_SLS;
+	tmp |= PCI_EXP_LNKCAP_SLS_2_5GB;
+	dw_pcie_writel_dbi(pci, cap_exp + PCI_EXP_LNKCAP, tmp);
 
 	/* Enable LTSSM */
 	writel_relaxed(0x1, afp->mgmt_base + PCIEX8MGMT_APP_LTSSM_ENABLE);
-	return 0;
+
+	ret = dw_pcie_wait_for_link(pci);
+	if (ret) {
+		dev_err(dev, "error: link did not start\n");
+		goto err;
+	}
+
+	tmp = dw_pcie_readl_dbi(pci, cap_exp + PCI_EXP_LNKCAP);
+	if ((tmp & PCI_EXP_LNKCAP_SLS) != orig) {
+		dev_dbg(dev, "changing speed back to original\n");
+
+		tmp &= ~PCI_EXP_LNKCAP_SLS;
+		tmp |= orig;
+		dw_pcie_writel_dbi(pci, cap_exp + PCI_EXP_LNKCAP, tmp);
+
+		tmp = dw_pcie_readl_dbi(pci, PCIE_LINK_WIDTH_SPEED_CONTROL);
+		tmp |= PORT_LOGIC_SPEED_CHANGE;
+		dw_pcie_writel_dbi(pci, PCIE_LINK_WIDTH_SPEED_CONTROL, tmp);
+
+		ret = dw_pcie_wait_for_link(pci);
+		if (ret) {
+			dev_err(dev, "error: link did not start at new speed\n");
+			goto err;
+		}
+	}
+
+	ret = 0;
+err:
+	WARN_ON(ret);	/* we assume that errors will be very rare */
+	dw_pcie_dbi_ro_wr_dis(pci);
+	return ret;
 }
 
 static int fu740_pcie_host_init(struct pcie_port *pp)
author maintainer
#120 lore raw

PATCH 120/1126 arm64: signal: nofpsimd: Do not allocate fp/simd context when not available

In reply to Greg Kroah-Hartman

From: David Engraf <[email protected]>

commit 0a32c88ddb9af30e8a16d41d7b9b824c27d29459 upstream.

Commit 6d502b6ba1b2 ("arm64: signal: nofpsimd: Handle fp/simd context for signal frames") introduced saving the fp/simd context for signal handling only when support is available. But setup_sigframe_layout() always reserves memory for fp/simd context. The additional memory is not touched because preserve_fpsimd_context() is not called and thus the magic is invalid.

This may lead to an error when parse_user_sigframe() checks the fp/simd area and does not find a valid magic number.

Fixes 6d502b6ba1b267b3 ("arm64: signal: nofpsimd: Handle fp/simd context for signal frames")
Cc <[email protected]> # 5.6.x
---
 arch/arm64/kernel/signal.c |   10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)
arch/arm64/kernel/signal.c +6 -4
@@ -577,10 +577,12 @@ static int setup_sigframe_layout(struct
 {
 	int err;
 
-	err = sigframe_alloc(user, &user->fpsimd_offset,
-			     sizeof(struct fpsimd_context));
-	if (err)
-		return err;
+	if (system_supports_fpsimd()) {
+		err = sigframe_alloc(user, &user->fpsimd_offset,
+				     sizeof(struct fpsimd_context));
+		if (err)
+			return err;
+	}
 
 	/* fault information, if valid */
 	if (add_all || current->thread.fault_code) {
author maintainer
#121 lore raw

PATCH 121/1126 arm64: Do not defer reserve_crashkernel() for platforms with no DMA memory zones

In reply to Greg Kroah-Hartman

From: Vijay Balakrishna <[email protected]>

commit 031495635b4668f94e964e037ca93d0d38bfde58 upstream.

The following patches resulted in deferring crash kernel reservation to mem_init(), mainly aimed at platforms with DMA memory zones (no IOMMU), in particular Raspberry Pi 4.

commit 1a8e1cef7603 ("arm64: use both ZONE_DMA and ZONE_DMA32")
commit 8424ecdde7df ("arm64: mm: Set ZONE_DMA size based on devicetree's dma-ranges") commit 0a30c53573b0 ("arm64: mm: Move reserve_crashkernel() into mem_init()") commit 2687275a5843 ("arm64: Force NO_BLOCK_MAPPINGS if crashkernel reservation is required")

Above changes introduced boot slowdown due to linear map creation for all the memory banks with NO_BLOCK_MAPPINGS, see discussion[1]. The proposed changes restore crash kernel reservation to earlier behavior thus avoids slow boot, particularly for platforms with IOMMU (no DMA memory zones).

Tested changes to confirm no ~150ms boot slowdown on our SoC with IOMMU and 8GB memory. Also tested with ZONE_DMA and/or ZONE_DMA32 configs to confirm no regression to deferring scheme of crash kernel memory reservation. In both cases successfully collected kernel crash dump.

[1] https://lore.kernel.org/all/[email protected]/

[will: Add #ifdef CONFIG_KEXEC_CORE guards to fix 'crashk_res' references in allnoconfig build]
---
 arch/arm64/mm/init.c |   36 ++++++++++++++++++++++++++++++++----
 arch/arm64/mm/mmu.c  |   32 +++++++++++++++++++++++++++++++-
 2 files changed, 63 insertions(+), 5 deletions(-)
arch/arm64/mm/init.c +32 -4
@@ -61,8 +61,34 @@ EXPORT_SYMBOL(memstart_addr);
  * unless restricted on specific platforms (e.g. 30-bit on Raspberry Pi 4).
  * In such case, ZONE_DMA32 covers the rest of the 32-bit addressable memory,
  * otherwise it is empty.
+ *
+ * Memory reservation for crash kernel either done early or deferred
+ * depending on DMA memory zones configs (ZONE_DMA) --
+ *
+ * In absence of ZONE_DMA configs arm64_dma_phys_limit initialized
+ * here instead of max_zone_phys().  This lets early reservation of
+ * crash kernel memory which has a dependency on arm64_dma_phys_limit.
+ * Reserving memory early for crash kernel allows linear creation of block
+ * mappings (greater than page-granularity) for all the memory bank rangs.
+ * In this scheme a comparatively quicker boot is observed.
+ *
+ * If ZONE_DMA configs are defined, crash kernel memory reservation
+ * is delayed until DMA zone memory range size initilazation performed in
+ * zone_sizes_init().  The defer is necessary to steer clear of DMA zone
+ * memory range to avoid overlap allocation.  So crash kernel memory boundaries
+ * are not known when mapping all bank memory ranges, which otherwise means
+ * not possible to exclude crash kernel range from creating block mappings
+ * so page-granularity mappings are created for the entire memory range.
+ * Hence a slightly slower boot is observed.
+ *
+ * Note: Page-granularity mapppings are necessary for crash kernel memory
+ * range for shrinking its size via /sys/kernel/kexec_crash_size interface.
  */
-phys_addr_t arm64_dma_phys_limit __ro_after_init;
+#if IS_ENABLED(CONFIG_ZONE_DMA) || IS_ENABLED(CONFIG_ZONE_DMA32)
+phys_addr_t __ro_after_init arm64_dma_phys_limit;
+#else
+const phys_addr_t arm64_dma_phys_limit = PHYS_MASK + 1;
+#endif
 
 #ifdef CONFIG_KEXEC_CORE
 /*
@@ -153,8 +179,6 @@ static void __init zone_sizes_init(unsig
 	if (!arm64_dma_phys_limit)
 		arm64_dma_phys_limit = dma32_phys_limit;
 #endif
-	if (!arm64_dma_phys_limit)
-		arm64_dma_phys_limit = PHYS_MASK + 1;
 	max_zone_pfns[ZONE_NORMAL] = max;
 
 	free_area_init(max_zone_pfns);
@@ -315,6 +339,9 @@ void __init arm64_memblock_init(void)
 
 	early_init_fdt_scan_reserved_mem();
 
+	if (!IS_ENABLED(CONFIG_ZONE_DMA) && !IS_ENABLED(CONFIG_ZONE_DMA32))
+		reserve_crashkernel();
+
 	high_memory = __va(memblock_end_of_DRAM() - 1) + 1;
 }
 
@@ -361,7 +388,8 @@ void __init bootmem_init(void)
 	 * request_standard_resources() depends on crashkernel's memory being
 	 * reserved, so do it here.
 	 */
-	reserve_crashkernel();
+	if (IS_ENABLED(CONFIG_ZONE_DMA) || IS_ENABLED(CONFIG_ZONE_DMA32))
+		reserve_crashkernel();
 
 	memblock_dump_all();
 }
arch/arm64/mm/mmu.c +31 -1
@@ -517,7 +517,7 @@ static void __init map_mem(pgd_t *pgdp)
 	 */
 	BUILD_BUG_ON(pgd_index(direct_map_end - 1) == pgd_index(direct_map_end));
 
-	if (can_set_direct_map() || crash_mem_map || IS_ENABLED(CONFIG_KFENCE))
+	if (can_set_direct_map() || IS_ENABLED(CONFIG_KFENCE))
 		flags |= NO_BLOCK_MAPPINGS | NO_CONT_MAPPINGS;
 
 	/*
@@ -528,6 +528,17 @@ static void __init map_mem(pgd_t *pgdp)
 	 */
 	memblock_mark_nomap(kernel_start, kernel_end - kernel_start);
 
+#ifdef CONFIG_KEXEC_CORE
+	if (crash_mem_map) {
+		if (IS_ENABLED(CONFIG_ZONE_DMA) ||
+		    IS_ENABLED(CONFIG_ZONE_DMA32))
+			flags |= NO_BLOCK_MAPPINGS | NO_CONT_MAPPINGS;
+		else if (crashk_res.end)
+			memblock_mark_nomap(crashk_res.start,
+			    resource_size(&crashk_res));
+	}
+#endif
+
 	/* map all the memory banks */
 	for_each_mem_range(i, &start, &end) {
 		if (start >= end)
@@ -554,6 +565,25 @@ static void __init map_mem(pgd_t *pgdp)
 	__map_memblock(pgdp, kernel_start, kernel_end,
 		       PAGE_KERNEL, NO_CONT_MAPPINGS);
 	memblock_clear_nomap(kernel_start, kernel_end - kernel_start);
+
+	/*
+	 * Use page-level mappings here so that we can shrink the region
+	 * in page granularity and put back unused memory to buddy system
+	 * through /sys/kernel/kexec_crash_size interface.
+	 */
+#ifdef CONFIG_KEXEC_CORE
+	if (crash_mem_map &&
+	    !IS_ENABLED(CONFIG_ZONE_DMA) && !IS_ENABLED(CONFIG_ZONE_DMA32)) {
+		if (crashk_res.end) {
+			__map_memblock(pgdp, crashk_res.start,
+				       crashk_res.end + 1,
+				       PAGE_KERNEL,
+				       NO_BLOCK_MAPPINGS | NO_CONT_MAPPINGS);
+			memblock_clear_nomap(crashk_res.start,
+					     resource_size(&crashk_res));
+		}
+	}
+#endif
 }
 
 void mark_rodata_ro(void)
author maintainer
#122 lore raw

PATCH 122/1126 arm64: dts: qcom: sm8250: Fix MSI IRQ for PCIe1 and PCIe2

In reply to Greg Kroah-Hartman

From: Manivannan Sadhasivam <[email protected]>

commit 1b7101e8124b450f2d6a35591e9cbb478c143ace upstream.

Fix the MSI IRQ used for PCIe instances 1 and 2.

---
 arch/arm64/boot/dts/qcom/sm8250.dtsi |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
arch/arm64/boot/dts/qcom/sm8250.dtsi +2 -2
@@ -1801,7 +1801,7 @@
 			ranges = <0x01000000 0x0 0x40200000 0x0 0x40200000 0x0 0x100000>,
 				 <0x02000000 0x0 0x40300000 0x0 0x40300000 0x0 0x1fd00000>;
 
-			interrupts = <GIC_SPI 306 IRQ_TYPE_EDGE_RISING>;
+			interrupts = <GIC_SPI 307 IRQ_TYPE_LEVEL_HIGH>;
 			interrupt-names = "msi";
 			#interrupt-cells = <1>;
 			interrupt-map-mask = <0 0 0 0x7>;
@@ -1907,7 +1907,7 @@
 			ranges = <0x01000000 0x0 0x64200000 0x0 0x64200000 0x0 0x100000>,
 				 <0x02000000 0x0 0x64300000 0x0 0x64300000 0x0 0x3d00000>;
 
-			interrupts = <GIC_SPI 236 IRQ_TYPE_EDGE_RISING>;
+			interrupts = <GIC_SPI 243 IRQ_TYPE_LEVEL_HIGH>;
 			interrupt-names = "msi";
 			#interrupt-cells = <1>;
 			interrupt-map-mask = <0 0 0 0x7>;
author maintainer
#123 lore raw

PATCH 123/1126 arm64: dts: ti: k3-am65: Fix gic-v3 compatible regs

In reply to Greg Kroah-Hartman

From: Nishanth Menon <[email protected]>

commit 8cae268b70f387ff9e697ccd62fb2384079124e7 upstream.

Though GIC ARE option is disabled for no GIC-v2 compatibility, Cortex-A53 is free to implement the CPU interface as long as it communicates with the GIC using the stream protocol. This requires that the SoC integration mark out the PERIPHBASE[1] as reserved area within the SoC. See longer discussion in [2] for further information.

Update the GIC register map to indicate offsets from PERIPHBASE based on [3]. Without doing this, systems like kvm will not function with gic-v2 emulation.

[1] https://developer.arm.com/documentation/ddi0500/e/system-control/aarch64-register-descriptions/configuration-base-address-register--el1
[2] https://lore.kernel.org/all/[email protected]/
[3] https://developer.arm.com/documentation/ddi0500/e/generic-interrupt-controller-cpu-interface/gic-programmers-model/memory-map

---
 arch/arm64/boot/dts/ti/k3-am65-main.dtsi |    5 ++++-
 arch/arm64/boot/dts/ti/k3-am65.dtsi      |    1 +
 2 files changed, 5 insertions(+), 1 deletion(-)
arch/arm64/boot/dts/ti/k3-am65-main.dtsi +4 -1
@@ -35,7 +35,10 @@
 		#interrupt-cells = <3>;
 		interrupt-controller;
 		reg = <0x00 0x01800000 0x00 0x10000>,	/* GICD */
-		      <0x00 0x01880000 0x00 0x90000>;	/* GICR */
+		      <0x00 0x01880000 0x00 0x90000>,	/* GICR */
+		      <0x00 0x6f000000 0x00 0x2000>,	/* GICC */
+		      <0x00 0x6f010000 0x00 0x1000>,	/* GICH */
+		      <0x00 0x6f020000 0x00 0x2000>;	/* GICV */
 		/*
 		 * vcpumntirq:
 		 * virtual CPU interface maintenance interrupt
arch/arm64/boot/dts/ti/k3-am65.dtsi +1 -0
@@ -86,6 +86,7 @@
 			 <0x00 0x46000000 0x00 0x46000000 0x00 0x00200000>,
 			 <0x00 0x47000000 0x00 0x47000000 0x00 0x00068400>,
 			 <0x00 0x50000000 0x00 0x50000000 0x00 0x8000000>,
+			 <0x00 0x6f000000 0x00 0x6f000000 0x00 0x00310000>, /* A53 PERIPHBASE */
 			 <0x00 0x70000000 0x00 0x70000000 0x00 0x200000>,
 			 <0x05 0x00000000 0x05 0x00000000 0x01 0x0000000>,
 			 <0x07 0x00000000 0x07 0x00000000 0x01 0x0000000>;
author maintainer
#124 lore raw

PATCH 124/1126 arm64: dts: ti: k3-j721e: Fix gic-v3 compatible regs

In reply to Greg Kroah-Hartman

From: Nishanth Menon <[email protected]>

commit a06ed27f3bc63ab9e10007dc0118d910908eb045 upstream.

Though GIC ARE option is disabled for no GIC-v2 compatibility, Cortex-A72 is free to implement the CPU interface as long as it communicates with the GIC using the stream protocol. This requires that the SoC integration mark out the PERIPHBASE[1] as reserved area within the SoC. See longer discussion in [2] for further information.

Update the GIC register map to indicate offsets from PERIPHBASE based on [3]. Without doing this, systems like kvm will not function with gic-v2 emulation.

[1] https://developer.arm.com/documentation/100095/0002/system-control/aarch64-register-descriptions/configuration-base-address-register--el1
[2] https://lore.kernel.org/all/[email protected]/
[3] https://developer.arm.com/documentation/100095/0002/way1382452674438

---
 arch/arm64/boot/dts/ti/k3-j721e-main.dtsi |    5 ++++-
 arch/arm64/boot/dts/ti/k3-j721e.dtsi      |    1 +
 2 files changed, 5 insertions(+), 1 deletion(-)
arch/arm64/boot/dts/ti/k3-j721e-main.dtsi +4 -1
@@ -76,7 +76,10 @@
 		#interrupt-cells = <3>;
 		interrupt-controller;
 		reg = <0x00 0x01800000 0x00 0x10000>,	/* GICD */
-		      <0x00 0x01900000 0x00 0x100000>;	/* GICR */
+		      <0x00 0x01900000 0x00 0x100000>,	/* GICR */
+		      <0x00 0x6f000000 0x00 0x2000>,	/* GICC */
+		      <0x00 0x6f010000 0x00 0x1000>,	/* GICH */
+		      <0x00 0x6f020000 0x00 0x2000>;	/* GICV */
 
 		/* vcpumntirq: virtual CPU interface maintenance interrupt */
 		interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_HIGH>;
arch/arm64/boot/dts/ti/k3-j721e.dtsi +1 -0
@@ -139,6 +139,7 @@
 			 <0x00 0x0e000000 0x00 0x0e000000 0x00 0x01800000>, /* PCIe Core*/
 			 <0x00 0x10000000 0x00 0x10000000 0x00 0x10000000>, /* PCIe DAT */
 			 <0x00 0x64800000 0x00 0x64800000 0x00 0x00800000>, /* C71 */
+			 <0x00 0x6f000000 0x00 0x6f000000 0x00 0x00310000>, /* A72 PERIPHBASE */
 			 <0x44 0x00000000 0x44 0x00000000 0x00 0x08000000>, /* PCIe2 DAT */
 			 <0x44 0x10000000 0x44 0x10000000 0x00 0x08000000>, /* PCIe3 DAT */
 			 <0x4d 0x80800000 0x4d 0x80800000 0x00 0x00800000>, /* C66_0 */
author maintainer
#125 lore raw

PATCH 125/1126 arm64: dts: ti: k3-j7200: Fix gic-v3 compatible regs

In reply to Greg Kroah-Hartman

From: Nishanth Menon <[email protected]>

commit 1a307cc299430dd7139d351a3b8941f493dfa885 upstream.

Though GIC ARE option is disabled for no GIC-v2 compatibility, Cortex-A72 is free to implement the CPU interface as long as it communicates with the GIC using the stream protocol. This requires that the SoC integration mark out the PERIPHBASE[1] as reserved area within the SoC. See longer discussion in [2] for further information.

Update the GIC register map to indicate offsets from PERIPHBASE based on [3]. Without doing this, systems like kvm will not function with gic-v2 emulation.

[1] https://developer.arm.com/documentation/100095/0002/system-control/aarch64-register-descriptions/configuration-base-address-register--el1
[2] https://lore.kernel.org/all/[email protected]/
[3] https://developer.arm.com/documentation/100095/0002/way1382452674438

---
 arch/arm64/boot/dts/ti/k3-j7200-main.dtsi |    5 ++++-
 arch/arm64/boot/dts/ti/k3-j7200.dtsi      |    1 +
 2 files changed, 5 insertions(+), 1 deletion(-)
arch/arm64/boot/dts/ti/k3-j7200-main.dtsi +4 -1
@@ -54,7 +54,10 @@
 		#interrupt-cells = <3>;
 		interrupt-controller;
 		reg = <0x00 0x01800000 0x00 0x10000>,	/* GICD */
-		      <0x00 0x01900000 0x00 0x100000>;	/* GICR */
+		      <0x00 0x01900000 0x00 0x100000>,	/* GICR */
+		      <0x00 0x6f000000 0x00 0x2000>,	/* GICC */
+		      <0x00 0x6f010000 0x00 0x1000>,	/* GICH */
+		      <0x00 0x6f020000 0x00 0x2000>;	/* GICV */
 
 		/* vcpumntirq: virtual CPU interface maintenance interrupt */
 		interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_HIGH>;
arch/arm64/boot/dts/ti/k3-j7200.dtsi +1 -0
@@ -129,6 +129,7 @@
 			 <0x00 0x00a40000 0x00 0x00a40000 0x00 0x00000800>, /* timesync router */
 			 <0x00 0x01000000 0x00 0x01000000 0x00 0x0d000000>, /* Most peripherals */
 			 <0x00 0x30000000 0x00 0x30000000 0x00 0x0c400000>, /* MAIN NAVSS */
+			 <0x00 0x6f000000 0x00 0x6f000000 0x00 0x00310000>, /* A72 PERIPHBASE */
 			 <0x00 0x70000000 0x00 0x70000000 0x00 0x00800000>, /* MSMC RAM */
 			 <0x00 0x18000000 0x00 0x18000000 0x00 0x08000000>, /* PCIe1 DAT0 */
 			 <0x41 0x00000000 0x41 0x00000000 0x01 0x00000000>, /* PCIe1 DAT1 */
author maintainer
#126 lore raw

PATCH 126/1126 arm64: dts: ti: k3-am64: Fix gic-v3 compatible regs

In reply to Greg Kroah-Hartman

From: Nishanth Menon <[email protected]>

commit de60edf1be3d42d4a1b303b41c7c53b2f865726e upstream.

Though GIC ARE option is disabled for no GIC-v2 compatibility, Cortex-A53 is free to implement the CPU interface as long as it communicates with the GIC using the stream protocol. This requires that the SoC integration mark out the PERIPHBASE[1] as reserved area within the SoC. See longer discussion in [2] for further information.

Update the GIC register map to indicate offsets from PERIPHBASE based on [3]. Without doing this, systems like kvm will not function with gic-v2 emulation.

[1] https://developer.arm.com/documentation/ddi0500/e/system-control/aarch64-register-descriptions/configuration-base-address-register--el1
[2] https://lore.kernel.org/all/[email protected]/
[3] https://developer.arm.com/documentation/ddi0500/e/generic-interrupt-controller-cpu-interface/gic-programmers-model/memory-map

---
 arch/arm64/boot/dts/ti/k3-am64-main.dtsi |    5 ++++-
 arch/arm64/boot/dts/ti/k3-am64.dtsi      |    1 +
 2 files changed, 5 insertions(+), 1 deletion(-)
arch/arm64/boot/dts/ti/k3-am64-main.dtsi +4 -1
@@ -59,7 +59,10 @@
 		#interrupt-cells = <3>;
 		interrupt-controller;
 		reg = <0x00 0x01800000 0x00 0x10000>,	/* GICD */
-		      <0x00 0x01840000 0x00 0xC0000>;	/* GICR */
+		      <0x00 0x01840000 0x00 0xC0000>,	/* GICR */
+		      <0x01 0x00000000 0x00 0x2000>,	/* GICC */
+		      <0x01 0x00010000 0x00 0x1000>,	/* GICH */
+		      <0x01 0x00020000 0x00 0x2000>;	/* GICV */
 		/*
 		 * vcpumntirq:
 		 * virtual CPU interface maintenance interrupt
arch/arm64/boot/dts/ti/k3-am64.dtsi +1 -0
@@ -87,6 +87,7 @@
 			 <0x00 0x68000000 0x00 0x68000000 0x00 0x08000000>, /* PCIe DAT0 */
 			 <0x00 0x70000000 0x00 0x70000000 0x00 0x00200000>, /* OC SRAM */
 			 <0x00 0x78000000 0x00 0x78000000 0x00 0x00800000>, /* Main R5FSS */
+			 <0x01 0x00000000 0x01 0x00000000 0x00 0x00310000>, /* A53 PERIPHBASE */
 			 <0x06 0x00000000 0x06 0x00000000 0x01 0x00000000>, /* PCIe DAT1 */
 			 <0x05 0x00000000 0x05 0x00000000 0x01 0x00000000>, /* FSS0 DAT3 */
 
author maintainer
#127 lore raw

PATCH 127/1126 arm64: dts: ti: k3-j721s2: Fix gic-v3 compatible regs

In reply to Greg Kroah-Hartman

From: Nishanth Menon <[email protected]>

commit a966803781fc5e1875511db9392b0d16174c5dd2 upstream.

Though GIC ARE option is disabled for no GIC-v2 compatibility, Cortex-A72 is free to implement the CPU interface as long as it communicates with the GIC using the stream protocol. This requires that the SoC integration mark out the PERIPHBASE[1] as reserved area within the SoC. See longer discussion in [2] for further information.

Update the GIC register map to indicate offsets from PERIPHBASE based on [3]. Without doing this, systems like kvm will not function with gic-v2 emulation.

[1] https://developer.arm.com/documentation/100095/0002/system-control/aarch64-register-descriptions/configuration-base-address-register--el1
[2] https://lore.kernel.org/all/[email protected]/
[3] https://developer.arm.com/documentation/100095/0002/way1382452674438

---
 arch/arm64/boot/dts/ti/k3-j721s2-main.dtsi |    5 ++++-
 arch/arm64/boot/dts/ti/k3-j721s2.dtsi      |    1 +
 2 files changed, 5 insertions(+), 1 deletion(-)
arch/arm64/boot/dts/ti/k3-j721s2-main.dtsi +4 -1
@@ -34,7 +34,10 @@
 		#interrupt-cells = <3>;
 		interrupt-controller;
 		reg = <0x00 0x01800000 0x00 0x200000>, /* GICD */
-		      <0x00 0x01900000 0x00 0x100000>; /* GICR */
+		      <0x00 0x01900000 0x00 0x100000>, /* GICR */
+		      <0x00 0x6f000000 0x00 0x2000>,   /* GICC */
+		      <0x00 0x6f010000 0x00 0x1000>,   /* GICH */
+		      <0x00 0x6f020000 0x00 0x2000>;   /* GICV */
 
 		/* vcpumntirq: virtual CPU interface maintenance interrupt */
 		interrupts = <GIC_PPI 9 IRQ_TYPE_LEVEL_HIGH>;
arch/arm64/boot/dts/ti/k3-j721s2.dtsi +1 -0
@@ -119,6 +119,7 @@
 			 <0x00 0x18000000 0x00 0x18000000 0x00 0x08000000>, /* PCIe1 DAT0 */
 			 <0x00 0x64800000 0x00 0x64800000 0x00 0x0070c000>, /* C71_1 */
 			 <0x00 0x65800000 0x00 0x65800000 0x00 0x0070c000>, /* C71_2 */
+			 <0x00 0x6f000000 0x00 0x6f000000 0x00 0x00310000>, /* A72 PERIPHBASE */
 			 <0x00 0x70000000 0x00 0x70000000 0x00 0x00400000>, /* MSMC RAM */
 			 <0x00 0x30000000 0x00 0x30000000 0x00 0x0c400000>, /* MAIN NAVSS */
 			 <0x41 0x00000000 0x41 0x00000000 0x01 0x00000000>, /* PCIe1 DAT1 */
author maintainer
#128 lore raw

PATCH 128/1126 ASoC: SOF: Intel: Fix NULL ptr dereference when ENOMEM

In reply to Greg Kroah-Hartman

From: Ammar Faizi <[email protected]>

commit b7fb0ae09009d076964afe4c1a2bde1ee2bd88a9 upstream.

Do not call snd_dma_free_pages() when snd_dma_alloc_pages() returns -ENOMEM because it leads to a NULL pointer dereference bug.

The dmesg says:

[ T1387] sof-audio-pci-intel-tgl 0000:00:1f.3: error: memory alloc failed: -12
[ T1387] BUG: kernel NULL pointer dereference, address: 0000000000000000
[ T1387] #PF: supervisor read access in kernel mode
[ T1387] #PF: error_code(0x0000) - not-present page
[ T1387] PGD 0 P4D 0
[ T1387] Oops: 0000 [#1] PREEMPT SMP NOPTI
[ T1387] CPU: 6 PID: 1387 Comm: alsa-sink-HDA A Tainted: G W 5.17.0-rc4-superb-owl-00055-g80d47f5de5e3
[ T1387] Hardware name: HP HP Laptop 14s-dq2xxx/87FD, BIOS F.15 09/15/2021
[ T1387] RIP: 0010:dma_free_noncontiguous+0x37/0x80
[ T1387] Code: [... snip ...]
[ T1387] RSP: 0000:ffffc90002b87770 EFLAGS: 00010246
[ T1387] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
[ T1387] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffff888101db30d0
[ T1387] RBP: 00000000fffffff4 R08: 0000000000000000 R09: 0000000000000000
[ T1387] R10: 0000000000000000 R11: ffffc90002b874d0 R12: 0000000000000001
[ T1387] R13: 0000000000058000 R14: ffff888105260c68 R15: ffff888105260828
[ T1387] FS: 00007f42e2ffd640(0000) GS:ffff888466b80000(0000) knlGS:0000000000000000
[ T1387] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ T1387] CR2: 0000000000000000 CR3: 000000014acf0003 CR4: 0000000000770ee0
[ T1387] PKRU: 55555554
[ T1387] Call Trace:
[ T1387] <TASK>
[ T1387] cl_stream_prepare+0x10a/0x120 [snd_sof_intel_hda_common 146addf995b9279ae7f509621078cccbe4f875e1] [... snip ...]
[ T1387] </TASK>

---
 sound/soc/sof/intel/hda-loader.c |   11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)
sound/soc/sof/intel/hda-loader.c +6 -5
@@ -47,7 +47,7 @@ static struct hdac_ext_stream *cl_stream
 	ret = snd_dma_alloc_pages(SNDRV_DMA_TYPE_DEV_SG, &pci->dev, size, dmab);
 	if (ret < 0) {
 		dev_err(sdev->dev, "error: memory alloc failed: %d\n", ret);
-		goto error;
+		goto out_put;
 	}
 
 	hstream->period_bytes = 0;/* initialize period_bytes */
@@ -58,22 +58,23 @@ static struct hdac_ext_stream *cl_stream
 		ret = hda_dsp_iccmax_stream_hw_params(sdev, dsp_stream, dmab, NULL);
 		if (ret < 0) {
 			dev_err(sdev->dev, "error: iccmax stream prepare failed: %d\n", ret);
-			goto error;
+			goto out_free;
 		}
 	} else {
 		ret = hda_dsp_stream_hw_params(sdev, dsp_stream, dmab, NULL);
 		if (ret < 0) {
 			dev_err(sdev->dev, "error: hdac prepare failed: %d\n", ret);
-			goto error;
+			goto out_free;
 		}
 		hda_dsp_stream_spib_config(sdev, dsp_stream, HDA_DSP_SPIB_ENABLE, size);
 	}
 
 	return dsp_stream;
 
-error:
-	hda_dsp_stream_put(sdev, direction, hstream->stream_tag);
+out_free:
 	snd_dma_free_pages(dmab);
+out_put:
+	hda_dsp_stream_put(sdev, direction, hstream->stream_tag);
 	return ERR_PTR(ret);
 }
 
author maintainer
#129 lore raw

PATCH 129/1126 mmc: core: use sysfs_emit() instead of sprintf()

In reply to Greg Kroah-Hartman

From: Sergey Shtylyov <[email protected]>

commit f5d8a5fe77ce933f53eb8f2e22bb7a1a2019ea11 upstream.

sprintf() (still used in the MMC core for the sysfs output) is vulnerable to the buffer overflow. Use the new-fangled sysfs_emit() instead.

Found by Linux Verification Center (linuxtesting.org) with the SVACE static analysis tool.

---
 drivers/mmc/core/bus.c      |    9 +++++----
 drivers/mmc/core/bus.h      |    3 ++-
 drivers/mmc/core/mmc.c      |   16 ++++++++--------
 drivers/mmc/core/sd.c       |   27 +++++++++++++--------------
 drivers/mmc/core/sdio.c     |    5 +++--
 drivers/mmc/core/sdio_bus.c |    7 ++++---
 6 files changed, 35 insertions(+), 32 deletions(-)
drivers/mmc/core/bus.c +5 -4
@@ -15,6 +15,7 @@
 #include <linux/stat.h>
 #include <linux/of.h>
 #include <linux/pm_runtime.h>
+#include <linux/sysfs.h>
 
 #include <linux/mmc/card.h>
 #include <linux/mmc/host.h>
@@ -34,13 +35,13 @@ static ssize_t type_show(struct device *
 
 	switch (card->type) {
 	case MMC_TYPE_MMC:
-		return sprintf(buf, "MMC\n");
+		return sysfs_emit(buf, "MMC\n");
 	case MMC_TYPE_SD:
-		return sprintf(buf, "SD\n");
+		return sysfs_emit(buf, "SD\n");
 	case MMC_TYPE_SDIO:
-		return sprintf(buf, "SDIO\n");
+		return sysfs_emit(buf, "SDIO\n");
 	case MMC_TYPE_SD_COMBO:
-		return sprintf(buf, "SDcombo\n");
+		return sysfs_emit(buf, "SDcombo\n");
 	default:
 		return -EFAULT;
 	}
drivers/mmc/core/bus.h +2 -1
@@ -9,6 +9,7 @@
 #define _MMC_CORE_BUS_H
 
 #include <linux/device.h>
+#include <linux/sysfs.h>
 
 struct mmc_host;
 struct mmc_card;
@@ -17,7 +18,7 @@ struct mmc_card;
 static ssize_t mmc_##name##_show (struct device *dev, struct device_attribute *attr, char *buf)	\
 {										\
 	struct mmc_card *card = mmc_dev_to_card(dev);				\
-	return sprintf(buf, fmt, args);						\
+	return sysfs_emit(buf, fmt, args);					\
 }										\
 static DEVICE_ATTR(name, S_IRUGO, mmc_##name##_show, NULL)
 
drivers/mmc/core/mmc.c +8 -8
@@ -12,6 +12,7 @@
 #include <linux/slab.h>
 #include <linux/stat.h>
 #include <linux/pm_runtime.h>
+#include <linux/sysfs.h>
 
 #include <linux/mmc/host.h>
 #include <linux/mmc/card.h>
@@ -812,12 +813,11 @@ static ssize_t mmc_fwrev_show(struct dev
 {
 	struct mmc_card *card = mmc_dev_to_card(dev);
 
-	if (card->ext_csd.rev < 7) {
-		return sprintf(buf, "0x%x\n", card->cid.fwrev);
-	} else {
-		return sprintf(buf, "0x%*phN\n", MMC_FIRMWARE_LEN,
-			       card->ext_csd.fwrev);
-	}
+	if (card->ext_csd.rev < 7)
+		return sysfs_emit(buf, "0x%x\n", card->cid.fwrev);
+	else
+		return sysfs_emit(buf, "0x%*phN\n", MMC_FIRMWARE_LEN,
+				  card->ext_csd.fwrev);
 }
 
 static DEVICE_ATTR(fwrev, S_IRUGO, mmc_fwrev_show, NULL);
@@ -830,10 +830,10 @@ static ssize_t mmc_dsr_show(struct devic
 	struct mmc_host *host = card->host;
 
 	if (card->csd.dsr_imp && host->dsr_req)
-		return sprintf(buf, "0x%x\n", host->dsr);
+		return sysfs_emit(buf, "0x%x\n", host->dsr);
 	else
 		/* return default DSR value */
-		return sprintf(buf, "0x%x\n", 0x404);
+		return sysfs_emit(buf, "0x%x\n", 0x404);
 }
 
 static DEVICE_ATTR(dsr, S_IRUGO, mmc_dsr_show, NULL);
drivers/mmc/core/sd.c +13 -14
@@ -13,6 +13,7 @@
 #include <linux/stat.h>
 #include <linux/pm_runtime.h>
 #include <linux/scatterlist.h>
+#include <linux/sysfs.h>
 
 #include <linux/mmc/host.h>
 #include <linux/mmc/card.h>
@@ -708,18 +709,16 @@ MMC_DEV_ATTR(ocr, "0x%08x\n", card->ocr)
 MMC_DEV_ATTR(rca, "0x%04x\n", card->rca);
 
 
-static ssize_t mmc_dsr_show(struct device *dev,
-                           struct device_attribute *attr,
-                           char *buf)
-{
-       struct mmc_card *card = mmc_dev_to_card(dev);
-       struct mmc_host *host = card->host;
-
-       if (card->csd.dsr_imp && host->dsr_req)
-               return sprintf(buf, "0x%x\n", host->dsr);
-       else
-               /* return default DSR value */
-               return sprintf(buf, "0x%x\n", 0x404);
+static ssize_t mmc_dsr_show(struct device *dev, struct device_attribute *attr,
+			    char *buf)
+{
+	struct mmc_card *card = mmc_dev_to_card(dev);
+	struct mmc_host *host = card->host;
+
+	if (card->csd.dsr_imp && host->dsr_req)
+		return sysfs_emit(buf, "0x%x\n", host->dsr);
+	/* return default DSR value */
+	return sysfs_emit(buf, "0x%x\n", 0x404);
 }
 
 static DEVICE_ATTR(dsr, S_IRUGO, mmc_dsr_show, NULL);
@@ -735,9 +734,9 @@ static ssize_t info##num##_show(struct d
 												\
 	if (num > card->num_info)								\
 		return -ENODATA;								\
-	if (!card->info[num-1][0])								\
+	if (!card->info[num - 1][0])								\
 		return 0;									\
-	return sprintf(buf, "%s\n", card->info[num-1]);						\
+	return sysfs_emit(buf, "%s\n", card->info[num - 1]);					\
 }												\
 static DEVICE_ATTR_RO(info##num)
 
drivers/mmc/core/sdio.c +3 -2
@@ -7,6 +7,7 @@
 
 #include <linux/err.h>
 #include <linux/pm_runtime.h>
+#include <linux/sysfs.h>
 
 #include <linux/mmc/host.h>
 #include <linux/mmc/card.h>
@@ -40,9 +41,9 @@ static ssize_t info##num##_show(struct d
 												\
 	if (num > card->num_info)								\
 		return -ENODATA;								\
-	if (!card->info[num-1][0])								\
+	if (!card->info[num - 1][0])								\
 		return 0;									\
-	return sprintf(buf, "%s\n", card->info[num-1]);						\
+	return sysfs_emit(buf, "%s\n", card->info[num - 1]);					\
 }												\
 static DEVICE_ATTR_RO(info##num)
 
drivers/mmc/core/sdio_bus.c +4 -3
@@ -14,6 +14,7 @@
 #include <linux/pm_runtime.h>
 #include <linux/pm_domain.h>
 #include <linux/acpi.h>
+#include <linux/sysfs.h>
 
 #include <linux/mmc/card.h>
 #include <linux/mmc/host.h>
@@ -35,7 +36,7 @@ field##_show(struct device *dev, struct
 	struct sdio_func *func;						\
 									\
 	func = dev_to_sdio_func (dev);					\
-	return sprintf(buf, format_string, args);			\
+	return sysfs_emit(buf, format_string, args);			\
 }									\
 static DEVICE_ATTR_RO(field)
 
@@ -52,9 +53,9 @@ static ssize_t info##num##_show(struct d
 												\
 	if (num > func->num_info)								\
 		return -ENODATA;								\
-	if (!func->info[num-1][0])								\
+	if (!func->info[num - 1][0])								\
 		return 0;									\
-	return sprintf(buf, "%s\n", func->info[num-1]);						\
+	return sysfs_emit(buf, "%s\n", func->info[num - 1]);					\
 }												\
 static DEVICE_ATTR_RO(info##num)
 
author maintainer
#130 lore raw

PATCH 130/1126 Revert "ACPI: Pass the same capabilities to the _OSC regardless of the query flag"

In reply to Greg Kroah-Hartman

From: Rafael J. Wysocki <[email protected]>

commit 2ca8e6285250c07a2e5a22ecbfd59b5a4ef73484 upstream.

Revert commit 159d8c274fd9 ("ACPI: Pass the same capabilities to the _OSC regardless of the query flag") which caused legitimate usage scenarios (when the platform firmware does not want the OS to control certain platform features controlled by the system bus scope _OSC) to break and was misguided by some misleading language in the _OSC definition in the ACPI specification (in particular, Section 6.2.11.1.3 "Sequence of _OSC Calls" that contradicts other perts of the _OSC definition).

---
 drivers/acpi/bus.c |   27 +++++++++++++++++++--------
 1 file changed, 19 insertions(+), 8 deletions(-)
drivers/acpi/bus.c +19 -8
@@ -332,21 +332,32 @@ static void acpi_bus_osc_negotiate_platf
 	if (ACPI_FAILURE(acpi_run_osc(handle, &context)))
 		return;
 
-	kfree(context.ret.pointer);
+	capbuf_ret = context.ret.pointer;
+	if (context.ret.length <= OSC_SUPPORT_DWORD) {
+		kfree(context.ret.pointer);
+		return;
+	}
 
-	/* Now run _OSC again with query flag clear */
+	/*
+	 * Now run _OSC again with query flag clear and with the caps
+	 * supported by both the OS and the platform.
+	 */
 	capbuf[OSC_QUERY_DWORD] = 0;
+	capbuf[OSC_SUPPORT_DWORD] = capbuf_ret[OSC_SUPPORT_DWORD];
+	kfree(context.ret.pointer);
 
 	if (ACPI_FAILURE(acpi_run_osc(handle, &context)))
 		return;
 
 	capbuf_ret = context.ret.pointer;
-	osc_sb_apei_support_acked =
-		capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_APEI_SUPPORT;
-	osc_pc_lpi_support_confirmed =
-		capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_PCLPI_SUPPORT;
-	osc_sb_native_usb4_support_confirmed =
-		capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_NATIVE_USB4_SUPPORT;
+	if (context.ret.length > OSC_SUPPORT_DWORD) {
+		osc_sb_apei_support_acked =
+			capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_APEI_SUPPORT;
+		osc_pc_lpi_support_confirmed =
+			capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_PCLPI_SUPPORT;
+		osc_sb_native_usb4_support_confirmed =
+			capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_NATIVE_USB4_SUPPORT;
+	}
 
 	kfree(context.ret.pointer);
 }
author maintainer
#131 lore raw

PATCH 131/1126 ACPI: properties: Consistently return -ENOENT if there are no more references

In reply to Greg Kroah-Hartman

From: Sakari Ailus <[email protected]>

commit babc92da5928f81af951663fc436997352e02d3a upstream.

__acpi_node_get_property_reference() is documented to return -ENOENT if the caller requests a property reference at an index that does not exist, not -EINVAL which it actually does.

Fix this by returning -ENOENT consistenly, independently of whether the property value is a plain reference or a package.

Fixes c343bc2ce2c6 ("ACPI: properties: Align return codes of __acpi_node_get_property_reference()")
Cc 4.14+ <[email protected]> # 4.14+
---
 drivers/acpi/property.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/acpi/property.c +1 -1
@@ -685,7 +685,7 @@ int __acpi_node_get_property_reference(c
 	 */
 	if (obj->type == ACPI_TYPE_LOCAL_REFERENCE) {
 		if (index)
-			return -EINVAL;
+			return -ENOENT;
 
 		device = acpi_fetch_acpi_dev(obj->reference.handle);
 		if (!device)
author maintainer
#132 lore raw

PATCH 132/1126 coredump: Also dump first pages of non-executable ELF libraries

In reply to Greg Kroah-Hartman

From: Jann Horn <[email protected]>

commit 84158b7f6a0624b81800b4e7c90f7fb7fdecf66c upstream.

When I rewrote the VMA dumping logic for coredumps, I changed it to recognize ELF library mappings based on the file being executable instead of the mapping having an ELF header. But turns out, distros ship many ELF libraries as non-executable, so the heuristic goes wrong...

Restore the old behavior where FILTER(ELF_HEADERS) dumps the first page of any offset-0 readable mapping that starts with the ELF magic.

This fix is technically layer-breaking a bit, because it checks for something ELF-specific in fs/coredump.c; but since we probably want to share this between standard ELF and FDPIC ELF anyway, I guess it's fine? And this also keeps the change small for backporting.

Fixes 429a22e776a2 ("coredump: rework elf/elf_fdpic vma_dump_size() into common helper")
---
 fs/coredump.c |   39 ++++++++++++++++++++++++++++++++++-----
 1 file changed, 34 insertions(+), 5 deletions(-)
fs/coredump.c +34 -5
@@ -42,6 +42,7 @@
 #include <linux/path.h>
 #include <linux/timekeeping.h>
 #include <linux/sysctl.h>
+#include <linux/elf.h>
 
 #include <linux/uaccess.h>
 #include <asm/mmu_context.h>
@@ -980,6 +981,8 @@ static bool always_dump_vma(struct vm_ar
 	return false;
 }
 
+#define DUMP_SIZE_MAYBE_ELFHDR_PLACEHOLDER 1
+
 /*
  * Decide how much of @vma's contents should be included in a core dump.
  */
@@ -1039,9 +1042,20 @@ static unsigned long vma_dump_size(struc
 	 * dump the first page to aid in determining what was mapped here.
 	 */
 	if (FILTER(ELF_HEADERS) &&
-	    vma->vm_pgoff == 0 && (vma->vm_flags & VM_READ) &&
-	    (READ_ONCE(file_inode(vma->vm_file)->i_mode) & 0111) != 0)
-		return PAGE_SIZE;
+	    vma->vm_pgoff == 0 && (vma->vm_flags & VM_READ)) {
+		if ((READ_ONCE(file_inode(vma->vm_file)->i_mode) & 0111) != 0)
+			return PAGE_SIZE;
+
+		/*
+		 * ELF libraries aren't always executable.
+		 * We'll want to check whether the mapping starts with the ELF
+		 * magic, but not now - we're holding the mmap lock,
+		 * so copy_from_user() doesn't work here.
+		 * Use a placeholder instead, and fix it up later in
+		 * dump_vma_snapshot().
+		 */
+		return DUMP_SIZE_MAYBE_ELFHDR_PLACEHOLDER;
+	}
 
 #undef	FILTER
 
@@ -1116,8 +1130,6 @@ int dump_vma_snapshot(struct coredump_pa
 		m->end = vma->vm_end;
 		m->flags = vma->vm_flags;
 		m->dump_size = vma_dump_size(vma, cprm->mm_flags);
-
-		vma_data_size += m->dump_size;
 	}
 
 	mmap_write_unlock(mm);
@@ -1127,6 +1139,23 @@ int dump_vma_snapshot(struct coredump_pa
 		return -EFAULT;
 	}
 
+	for (i = 0; i < *vma_count; i++) {
+		struct core_vma_metadata *m = (*vma_meta) + i;
+
+		if (m->dump_size == DUMP_SIZE_MAYBE_ELFHDR_PLACEHOLDER) {
+			char elfmag[SELFMAG];
+
+			if (copy_from_user(elfmag, (void __user *)m->start, SELFMAG) ||
+					memcmp(elfmag, ELFMAG, SELFMAG) != 0) {
+				m->dump_size = 0;
+			} else {
+				m->dump_size = PAGE_SIZE;
+			}
+		}
+
+		vma_data_size += m->dump_size;
+	}
+
 	*vma_data_size_ptr = vma_data_size;
 	return 0;
 }
author maintainer
#133 lore raw

PATCH 133/1126 ext4: fix ext4_fc_stats trace point

In reply to Greg Kroah-Hartman

From: Ritesh Harjani <[email protected]>

commit 7af1974af0a9ba8a8ed2e3e947d87dd4d9a78d27 upstream.

ftrace's __print_symbolic() requires that any enum values used in the symbol to string translation table be wrapped in a TRACE_DEFINE_ENUM so that the enum value can be decoded from the ftrace ring buffer by user space tooling.

This patch also fixes few other problems found in this trace point. e.g. dereferencing structures in TP_printk which should not be done at any cost.

Also to avoid checkpatch warnings, this patch removes those whitespaces/tab stops issues.

---
 include/trace/events/ext4.h |   80 +++++++++++++++++++++++++++-----------------
 1 file changed, 50 insertions(+), 30 deletions(-)
include/trace/events/ext4.h +50 -30
@@ -95,6 +95,17 @@ TRACE_DEFINE_ENUM(ES_REFERENCED_B);
 	{ FALLOC_FL_COLLAPSE_RANGE,	"COLLAPSE_RANGE"},	\
 	{ FALLOC_FL_ZERO_RANGE,		"ZERO_RANGE"})
 
+TRACE_DEFINE_ENUM(EXT4_FC_REASON_XATTR);
+TRACE_DEFINE_ENUM(EXT4_FC_REASON_CROSS_RENAME);
+TRACE_DEFINE_ENUM(EXT4_FC_REASON_JOURNAL_FLAG_CHANGE);
+TRACE_DEFINE_ENUM(EXT4_FC_REASON_NOMEM);
+TRACE_DEFINE_ENUM(EXT4_FC_REASON_SWAP_BOOT);
+TRACE_DEFINE_ENUM(EXT4_FC_REASON_RESIZE);
+TRACE_DEFINE_ENUM(EXT4_FC_REASON_RENAME_DIR);
+TRACE_DEFINE_ENUM(EXT4_FC_REASON_FALLOC_RANGE);
+TRACE_DEFINE_ENUM(EXT4_FC_REASON_INODE_JOURNAL_DATA);
+TRACE_DEFINE_ENUM(EXT4_FC_REASON_MAX);
+
 #define show_fc_reason(reason)						\
 	__print_symbolic(reason,					\
 		{ EXT4_FC_REASON_XATTR,		"XATTR"},		\
@@ -2723,41 +2734,50 @@ TRACE_EVENT(ext4_fc_commit_stop,
 
 #define FC_REASON_NAME_STAT(reason)					\
 	show_fc_reason(reason),						\
-	__entry->sbi->s_fc_stats.fc_ineligible_reason_count[reason]
+	__entry->fc_ineligible_rc[reason]
 
 TRACE_EVENT(ext4_fc_stats,
-	    TP_PROTO(struct super_block *sb),
+	TP_PROTO(struct super_block *sb),
+
+	TP_ARGS(sb),
+
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__array(unsigned int, fc_ineligible_rc, EXT4_FC_REASON_MAX)
+		__field(unsigned long, fc_commits)
+		__field(unsigned long, fc_ineligible_commits)
+		__field(unsigned long, fc_numblks)
+	),
 
-	    TP_ARGS(sb),
+	TP_fast_assign(
+		int i;
 
-	    TP_STRUCT__entry(
-		    __field(dev_t, dev)
-		    __field(struct ext4_sb_info *, sbi)
-		    __field(int, count)
-		    ),
-
-	    TP_fast_assign(
-		    __entry->dev = sb->s_dev;
-		    __entry->sbi = EXT4_SB(sb);
-		    ),
-
-	    TP_printk("dev %d:%d fc ineligible reasons:\n"
-		      "%s:%d, %s:%d, %s:%d, %s:%d, %s:%d, %s:%d, %s:%d, %s:%d, %s:%d; "
-		      "num_commits:%ld, ineligible: %ld, numblks: %ld",
-		      MAJOR(__entry->dev), MINOR(__entry->dev),
-		      FC_REASON_NAME_STAT(EXT4_FC_REASON_XATTR),
-		      FC_REASON_NAME_STAT(EXT4_FC_REASON_CROSS_RENAME),
-		      FC_REASON_NAME_STAT(EXT4_FC_REASON_JOURNAL_FLAG_CHANGE),
-		      FC_REASON_NAME_STAT(EXT4_FC_REASON_NOMEM),
-		      FC_REASON_NAME_STAT(EXT4_FC_REASON_SWAP_BOOT),
-		      FC_REASON_NAME_STAT(EXT4_FC_REASON_RESIZE),
-		      FC_REASON_NAME_STAT(EXT4_FC_REASON_RENAME_DIR),
-		      FC_REASON_NAME_STAT(EXT4_FC_REASON_FALLOC_RANGE),
-		      FC_REASON_NAME_STAT(EXT4_FC_REASON_INODE_JOURNAL_DATA),
-		      __entry->sbi->s_fc_stats.fc_num_commits,
-		      __entry->sbi->s_fc_stats.fc_ineligible_commits,
-		      __entry->sbi->s_fc_stats.fc_numblks)
+		__entry->dev = sb->s_dev;
+		for (i = 0; i < EXT4_FC_REASON_MAX; i++) {
+			__entry->fc_ineligible_rc[i] =
+				EXT4_SB(sb)->s_fc_stats.fc_ineligible_reason_count[i];
+		}
+		__entry->fc_commits = EXT4_SB(sb)->s_fc_stats.fc_num_commits;
+		__entry->fc_ineligible_commits =
+			EXT4_SB(sb)->s_fc_stats.fc_ineligible_commits;
+		__entry->fc_numblks = EXT4_SB(sb)->s_fc_stats.fc_numblks;
+	),
 
+	TP_printk("dev %d,%d fc ineligible reasons:\n"
+		  "%s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u, %s:%u "
+		  "num_commits:%lu, ineligible: %lu, numblks: %lu",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  FC_REASON_NAME_STAT(EXT4_FC_REASON_XATTR),
+		  FC_REASON_NAME_STAT(EXT4_FC_REASON_CROSS_RENAME),
+		  FC_REASON_NAME_STAT(EXT4_FC_REASON_JOURNAL_FLAG_CHANGE),
+		  FC_REASON_NAME_STAT(EXT4_FC_REASON_NOMEM),
+		  FC_REASON_NAME_STAT(EXT4_FC_REASON_SWAP_BOOT),
+		  FC_REASON_NAME_STAT(EXT4_FC_REASON_RESIZE),
+		  FC_REASON_NAME_STAT(EXT4_FC_REASON_RENAME_DIR),
+		  FC_REASON_NAME_STAT(EXT4_FC_REASON_FALLOC_RANGE),
+		  FC_REASON_NAME_STAT(EXT4_FC_REASON_INODE_JOURNAL_DATA),
+		  __entry->fc_commits, __entry->fc_ineligible_commits,
+		  __entry->fc_numblks)
 );
 
 #define DEFINE_TRACE_DENTRY_EVENT(__type)				\
author maintainer
#134 lore raw

PATCH 134/1126 ext4: fix fs corruption when tring to remove a non-empty directory with IO error

In reply to Greg Kroah-Hartman

From: Ye Bin <[email protected]>

commit 7aab5c84a0f6ec2290e2ba4a6b245178b1bf949a upstream.

We inject IO error when rmdir non empty direcory, then got issue as follows:
step1: mkfs.ext4 -F /dev/sda
step2: mount /dev/sda test
step3: cd test
step4: mkdir -p 1/2
step5: rmdir 1

	[  110.920551] ext4_empty_dir: inject fault
	[  110.921926] EXT4-fs warning (device sda): ext4_rmdir:3113: inode #12:
	comm rmdir: empty directory '1' has too many links (3)

step6: cd ..
step7: umount test
step8: fsck.ext4 -f /dev/sda
e2fsck 1.42.9 (28-Dec-2013)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Entry '..' in .../??? (13) has deleted/unused inode 12. Clear<y>? yes Pass 3: Checking directory connectivity
Unconnected directory inode 13 (...)
Connect to /lost+found<y>? yes
Pass 4: Checking reference counts
Inode 13 ref count is 3, should be 2. Fix<y>? yes
Pass 5: Checking group summary information

/dev/sda: ***** FILE SYSTEM WAS MODIFIED *****
/dev/sda: 12/131072 files (0.0% non-contiguous), 26157/524288 blocks

ext4_rmdir

	if (!ext4_empty_dir(inode))
		goto end_rmdir;
ext4_empty_dir
	bh = ext4_read_dirblock(inode, 0, DIRENT_HTREE);
	if (IS_ERR(bh))
		return true;

Now if read directory block failed, 'ext4_empty_dir' will return true, assume directory is empty. Obviously, it will lead to above issue.
To solve this issue, if read directory block failed 'ext4_empty_dir' just return false. To avoid making things worse when file system is already corrupted, 'ext4_empty_dir' also return false.

---
 fs/ext4/inline.c |    9 ++++-----
 fs/ext4/namei.c  |   10 +++++-----
 2 files changed, 9 insertions(+), 10 deletions(-)
fs/ext4/inline.c +4 -5
@@ -1783,19 +1783,20 @@ bool empty_inline_dir(struct inode *dir,
 	void *inline_pos;
 	unsigned int offset;
 	struct ext4_dir_entry_2 *de;
-	bool ret = true;
+	bool ret = false;
 
 	err = ext4_get_inode_loc(dir, &iloc);
 	if (err) {
 		EXT4_ERROR_INODE_ERR(dir, -err,
 				     "error %d getting inode %lu block",
 				     err, dir->i_ino);
-		return true;
+		return false;
 	}
 
 	down_read(&EXT4_I(dir)->xattr_sem);
 	if (!ext4_has_inline_data(dir)) {
 		*has_inline_data = 0;
+		ret = true;
 		goto out;
 	}
 
@@ -1804,7 +1805,6 @@ bool empty_inline_dir(struct inode *dir,
 		ext4_warning(dir->i_sb,
 			     "bad inline directory (dir #%lu) - no `..'",
 			     dir->i_ino);
-		ret = true;
 		goto out;
 	}
 
@@ -1823,16 +1823,15 @@ bool empty_inline_dir(struct inode *dir,
 				     dir->i_ino, le32_to_cpu(de->inode),
 				     le16_to_cpu(de->rec_len), de->name_len,
 				     inline_size);
-			ret = true;
 			goto out;
 		}
 		if (le32_to_cpu(de->inode)) {
-			ret = false;
 			goto out;
 		}
 		offset += ext4_rec_len_from_disk(de->rec_len, inline_size);
 	}
 
+	ret = true;
 out:
 	up_read(&EXT4_I(dir)->xattr_sem);
 	brelse(iloc.bh);
fs/ext4/namei.c +5 -5
@@ -2997,14 +2997,14 @@ bool ext4_empty_dir(struct inode *inode)
 	if (inode->i_size < ext4_dir_rec_len(1, NULL) +
 					ext4_dir_rec_len(2, NULL)) {
 		EXT4_ERROR_INODE(inode, "invalid size");
-		return true;
+		return false;
 	}
 	/* The first directory block must not be a hole,
 	 * so treat it as DIRENT_HTREE
 	 */
 	bh = ext4_read_dirblock(inode, 0, DIRENT_HTREE);
 	if (IS_ERR(bh))
-		return true;
+		return false;
 
 	de = (struct ext4_dir_entry_2 *) bh->b_data;
 	if (ext4_check_dir_entry(inode, NULL, de, bh, bh->b_data, bh->b_size,
@@ -3012,7 +3012,7 @@ bool ext4_empty_dir(struct inode *inode)
 	    le32_to_cpu(de->inode) != inode->i_ino || strcmp(".", de->name)) {
 		ext4_warning_inode(inode, "directory missing '.'");
 		brelse(bh);
-		return true;
+		return false;
 	}
 	offset = ext4_rec_len_from_disk(de->rec_len, sb->s_blocksize);
 	de = ext4_next_entry(de, sb->s_blocksize);
@@ -3021,7 +3021,7 @@ bool ext4_empty_dir(struct inode *inode)
 	    le32_to_cpu(de->inode) == 0 || strcmp("..", de->name)) {
 		ext4_warning_inode(inode, "directory missing '..'");
 		brelse(bh);
-		return true;
+		return false;
 	}
 	offset += ext4_rec_len_from_disk(de->rec_len, sb->s_blocksize);
 	while (offset < inode->i_size) {
@@ -3035,7 +3035,7 @@ bool ext4_empty_dir(struct inode *inode)
 				continue;
 			}
 			if (IS_ERR(bh))
-				return true;
+				return false;
 		}
 		de = (struct ext4_dir_entry_2 *) (bh->b_data +
 					(offset & (sb->s_blocksize - 1)));
author maintainer
#135 lore raw

PATCH 135/1126 ext4: make mb_optimize_scan option work with set/unset mount cmd

In reply to Greg Kroah-Hartman

From: Ojaswin Mujoo <[email protected]>

commit 27b38686a3bb601db48901dbc4e2fc5d77ffa2c1 upstream.

After moving to the new mount API, mb_optimize_scan mount option handling was not working as expected due to the parsed value always being overwritten by default. Refactor and fix this to the expected behavior described below:

*  mb_optimize_scan=1 - On
*  mb_optimize_scan=0 - Off
*  mb_optimize_scan not passed - On if no. of BGs > threshold else off
*  Remounts retain previous value unless we explicitly pass the option

with a new value

---
 fs/ext4/super.c |   24 ++++++++++++++----------
 1 file changed, 14 insertions(+), 10 deletions(-)
fs/ext4/super.c +14 -10
@@ -2021,12 +2021,12 @@ static int ext4_set_test_dummy_encryptio
 #define EXT4_SPEC_s_commit_interval		(1 << 16)
 #define EXT4_SPEC_s_fc_debug_max_replay		(1 << 17)
 #define EXT4_SPEC_s_sb_block			(1 << 18)
+#define EXT4_SPEC_mb_optimize_scan		(1 << 19)
 
 struct ext4_fs_context {
 	char		*s_qf_names[EXT4_MAXQUOTAS];
 	char		*test_dummy_enc_arg;
 	int		s_jquota_fmt;	/* Format of quota to use */
-	int		mb_optimize_scan;
 #ifdef CONFIG_EXT4_DEBUG
 	int s_fc_debug_max_replay;
 #endif
@@ -2451,12 +2451,17 @@ static int ext4_parse_param(struct fs_co
 			ctx_clear_mount_opt(ctx, m->mount_opt);
 		return 0;
 	case Opt_mb_optimize_scan:
-		if (result.int_32 != 0 && result.int_32 != 1) {
+		if (result.int_32 == 1) {
+			ctx_set_mount_opt2(ctx, EXT4_MOUNT2_MB_OPTIMIZE_SCAN);
+			ctx->spec |= EXT4_SPEC_mb_optimize_scan;
+		} else if (result.int_32 == 0) {
+			ctx_clear_mount_opt2(ctx, EXT4_MOUNT2_MB_OPTIMIZE_SCAN);
+			ctx->spec |= EXT4_SPEC_mb_optimize_scan;
+		} else {
 			ext4_msg(NULL, KERN_WARNING,
 				 "mb_optimize_scan should be set to 0 or 1.");
 			return -EINVAL;
 		}
-		ctx->mb_optimize_scan = result.int_32;
 		return 0;
 	}
 
@@ -4369,7 +4374,6 @@ static int __ext4_fill_super(struct fs_c
 
 	/* Set defaults for the variables that will be set during parsing */
 	ctx->journal_ioprio = DEFAULT_JOURNAL_IOPRIO;
-	ctx->mb_optimize_scan = DEFAULT_MB_OPTIMIZE_SCAN;
 
 	sbi->s_inode_readahead_blks = EXT4_DEF_INODE_READAHEAD_BLKS;
 	sbi->s_sectors_written_start =
@@ -5320,12 +5324,12 @@ no_journal:
 	 * turned off by passing "mb_optimize_scan=0". This can also be
 	 * turned on forcefully by passing "mb_optimize_scan=1".
 	 */
-	if (ctx->mb_optimize_scan == 1)
-		set_opt2(sb, MB_OPTIMIZE_SCAN);
-	else if (ctx->mb_optimize_scan == 0)
-		clear_opt2(sb, MB_OPTIMIZE_SCAN);
-	else if (sbi->s_groups_count >= MB_DEFAULT_LINEAR_SCAN_THRESHOLD)
-		set_opt2(sb, MB_OPTIMIZE_SCAN);
+	if (!(ctx->spec & EXT4_SPEC_mb_optimize_scan)) {
+		if (sbi->s_groups_count >= MB_DEFAULT_LINEAR_SCAN_THRESHOLD)
+			set_opt2(sb, MB_OPTIMIZE_SCAN);
+		else
+			clear_opt2(sb, MB_OPTIMIZE_SCAN);
+	}
 
 	err = ext4_mb_init(sb);
 	if (err) {
author maintainer
#136 lore raw

PATCH 136/1126 ext4: make mb_optimize_scan performance mount option work with extents

In reply to Greg Kroah-Hartman

From: Ojaswin Mujoo <[email protected]>

commit 077d0c2c78df6f7260cdd015a991327efa44d8ad upstream.

Currently mb_optimize_scan scan feature which improves filesystem performance heavily (when FS is fragmented), seems to be not working with files with extents (ext4 by default has files with extents).

This patch fixes that and makes mb_optimize_scan feature work for files with extents.

Below are some performance numbers obtained when allocating a 10M and 100M file with and w/o this patch on a filesytem with no 1M contiguous block.

<perf numbers>
===============
Workload: dd if=/dev/urandom of=test conv=fsync bs=1M count=10/100

Time taken
=====================================================

no.     Size   without-patch     with-patch    Diff(%)
1       10M      0m8.401s         0m5.623s     33.06%
2       100M     1m40.465s        1m14.737s    25.6%

<debug stats>
=============
w/o patch:
mballoc:

    reqs: 17056
    success: 11407
    groups_scanned: 13643
    cr0_stats:
            hits: 37
            groups_considered: 9472
            useless_loops: 36
            bad_suggestions: 0
    cr1_stats:
            hits: 11418
            groups_considered: 908560
            useless_loops: 1894
            bad_suggestions: 0
    cr2_stats:
            hits: 1873
            groups_considered: 6913
            useless_loops: 21
    cr3_stats:
            hits: 21
            groups_considered: 5040
            useless_loops: 21
    extents_scanned: 417364
            goal_hits: 3707
            2^n_hits: 37
            breaks: 1873
            lost: 0
    buddies_generated: 239/240
    buddies_time_used: 651080
    preallocated: 705
    discarded: 478

with patch:
mballoc:

    reqs: 12768
    success: 11305
    groups_scanned: 12768
    cr0_stats:
            hits: 1
            groups_considered: 18
            useless_loops: 0
            bad_suggestions: 0
    cr1_stats:
            hits: 5829
            groups_considered: 50626
            useless_loops: 0
            bad_suggestions: 0
    cr2_stats:
            hits: 6938
            groups_considered: 580363
            useless_loops: 0
    cr3_stats:
            hits: 0
            groups_considered: 0
            useless_loops: 0
    extents_scanned: 309059
            goal_hits: 0
            2^n_hits: 1
            breaks: 1463
            lost: 0
    buddies_generated: 239/240
    buddies_time_used: 791392
    preallocated: 673
    discarded: 446
---
 fs/ext4/mballoc.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ext4/mballoc.c +1 -1
@@ -1000,7 +1000,7 @@ static inline int should_optimize_scan(s
 		return 0;
 	if (ac->ac_criteria >= 2)
 		return 0;
-	if (ext4_test_inode_flag(ac->ac_inode, EXT4_INODE_EXTENTS))
+	if (!ext4_test_inode_flag(ac->ac_inode, EXT4_INODE_EXTENTS))
 		return 0;
 	return 1;
 }
author maintainer
#137 lore raw

PATCH 137/1126 samples/landlock: Fix path_list memory leak

In reply to Greg Kroah-Hartman

From: Tom Rix <[email protected]>

commit 66b513b7c64a7290c1fbb88e657f7cece992e131 upstream.

Clang static analysis reports this error

sandboxer.c:134:8: warning: Potential leak of memory
pointed to by 'path_list'

        ret = 0;
              ^

path_list is allocated in parse_path() but never freed.

---
 samples/landlock/sandboxer.c |    1 +
 1 file changed, 1 insertion(+)
samples/landlock/sandboxer.c +1 -0
@@ -134,6 +134,7 @@ static int populate_ruleset(
 	ret = 0;
 
 out_free_name:
+	free(path_list);
 	free(env_path_name);
 	return ret;
 }
author maintainer
#138 lore raw

PATCH 138/1126 landlock: Use square brackets around "landlock-ruleset"

In reply to Greg Kroah-Hartman

From: Christian Brauner <[email protected]>

commit aea0b9f2486da8497f35c7114b764bf55e17c7ea upstream.

Make the name of the anon inode fd "[landlock-ruleset]" instead of "landlock-ruleset". This is minor but most anon inode fds already carry square brackets around their name:

    [eventfd]
    [eventpoll]
    [fanotify]
    [fscontext]
    [io_uring]
    [pidfd]
    [signalfd]
    [timerfd]
    [userfaultfd]

For the sake of consistency lets do the same for the landlock-ruleset anon inode fd that comes with landlock. We did the same in 1cdc415f1083 ("uapi, fsopen: use square brackets around "fscontext" [ver #2]") for the new mount api.

---
 security/landlock/syscalls.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
security/landlock/syscalls.c +1 -1
@@ -192,7 +192,7 @@ SYSCALL_DEFINE3(landlock_create_ruleset,
 		return PTR_ERR(ruleset);
 
 	/* Creates anonymous FD referring to the ruleset. */
-	ruleset_fd = anon_inode_getfd("landlock-ruleset", &ruleset_fops,
+	ruleset_fd = anon_inode_getfd("[landlock-ruleset]", &ruleset_fops,
 			ruleset, O_RDWR | O_CLOEXEC);
 	if (ruleset_fd < 0)
 		landlock_put_ruleset(ruleset);
author maintainer
#139 lore raw

PATCH 139/1126 mailbox: tegra-hsp: Flush whole channel

In reply to Greg Kroah-Hartman

From: Pekka Pessi <[email protected]>

commit 60de2d2dc284e0dd1c2c897d08625bde24ef3454 upstream.

The txdone can re-fill the mailbox. Keep polling the mailbox during the flush until all the messages have been delivered.

This fixes an issue with the Tegra Combined UART (TCU) where output can get truncated under high traffic load.

Fixes 91b1b1c3da8a ("mailbox: tegra-hsp: Add support for shared mailboxes")
---
 drivers/mailbox/tegra-hsp.c |    5 +++++
 1 file changed, 5 insertions(+)
drivers/mailbox/tegra-hsp.c +5 -0
@@ -412,6 +412,11 @@ static int tegra_hsp_mailbox_flush(struc
 		value = tegra_hsp_channel_readl(ch, HSP_SM_SHRD_MBOX);
 		if ((value & HSP_SM_SHRD_MBOX_FULL) == 0) {
 			mbox_chan_txdone(chan, 0);
+
+			/* Wait until channel is empty */
+			if (chan->active_req != NULL)
+				continue;
+
 			return 0;
 		}
 
author maintainer
#140 lore raw

PATCH 140/1126 btrfs: zoned: put block group after final usage

In reply to Greg Kroah-Hartman

From: Nikolay Borisov <[email protected]>

commit d3e29967079c522ce1c5cab0e9fab2c280b977eb upstream.

It's counter-intuitive (and wrong) to put the block group _before_ the final usage in submit_eb_page. Fix it by re-ordering the call to btrfs_put_block_group after its final reference. Also fix a minor typo in 'implies'

Fixes be1a1d7a5d24 ("btrfs: zoned: finish fully written block group")
---
 fs/btrfs/extent_io.c |    7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
fs/btrfs/extent_io.c +4 -3
@@ -4780,11 +4780,12 @@ static int submit_eb_page(struct page *p
 		return ret;
 	}
 	if (cache) {
-		/* Impiles write in zoned mode */
-		btrfs_put_block_group(cache);
-		/* Mark the last eb in a block group */
+		/*
+		 * Implies write in zoned mode. Mark the last eb in a block group.
+		 */
 		if (cache->seq_zone && eb->start + eb->len == cache->zone_capacity)
 			set_bit(EXTENT_BUFFER_ZONE_FINISH, &eb->bflags);
+		btrfs_put_block_group(cache);
 	}
 	ret = write_one_eb(eb, wbc, epd);
 	free_extent_buffer(eb);
author maintainer
#141 lore raw

PATCH 141/1126 block: fix rq-qos breakage from skipping rq_qos_done_bio()

In reply to Greg Kroah-Hartman

From: Tejun Heo <[email protected]>

commit aa1b46dcdc7baaf5fec0be25782ef24b26aa209e upstream.

a647a524a467 ("block: don't call rq_qos_ops->done_bio if the bio isn't tracked") made bio_endio() skip rq_qos_done_bio() if BIO_TRACKED is not set. While this fixed a potential oops, it also broke blk-iocost by skipping the done_bio callback for merged bios.

Before, whether a bio goes through rq_qos_throttle() or rq_qos_merge(), rq_qos_done_bio() would be called on the bio on completion with BIO_TRACKED distinguishing the former from the latter. rq_qos_done_bio() is not called for bios which wenth through rq_qos_merge(). This royally confuses blk-iocost as the merged bios never finish and are considered perpetually in-flight.

One reliably reproducible failure mode is an intermediate cgroup geting stuck active preventing its children from being activated due to the leaf-only rule, leading to loss of control. The following is from resctl-bench protection scenario which emulates isolating a web server like workload from a memory bomb run on an iocost configuration which should yield a reasonable level of protection.

  # cat /sys/block/nvme2n1/device/model
  Samsung SSD 970 PRO 512GB
  # cat /sys/fs/cgroup/io.cost.model
  259:0 ctrl=user model=linear rbps=834913556 rseqiops=93622 rrandiops=102913 wbps=618985353 wseqiops=72325 wrandiops=71025
  # cat /sys/fs/cgroup/io.cost.qos
  259:0 enable=1 ctrl=user rpct=95.00 rlat=18776 wpct=95.00 wlat=8897 min=60.00 max=100.00
  # resctl-bench -m 29.6G -r out.json run protection::scenario=mem-hog,loops=1
  ...
  Memory Hog Summary
  ==================

IO Latency: R p50=242u:336u/2.5m p90=794u:1.4m/7.5m p99=2.7m:8.0m/62.5m max=8.0m:36.4m/350m

              W p50=221u:323u/1.5m p90=709u:1.2m/5.5m p99=1.5m:2.5m/9.5m max=6.9m:35.9m/350m

Isolation and Request Latency Impact Distributions:

min p01 p05 p10 p25 p50 p75 p90 p95 p99 max mean stdev
isol% 15.90 15.90 15.90 40.05 57.24 59.07 60.01 74.63 74.63 90.35 90.35 58.12 15.82
lat-imp% 0 0 0 0 0 4.55 14.68 15.54 233.5 548.1 548.1 53.88 143.6

Result: isol=58.12:15.82% lat_imp=53.88%:143.6 work_csv=100.0% missing=3.96%

The isolation result of 58.12% is close to what this device would show without any IO control.

Fix it by introducing a new flag BIO_QOS_MERGED to mark merged bios and calling rq_qos_done_bio() on them too. For consistency and clarity, rename BIO_TRACKED to BIO_QOS_THROTTLED. The flag checks are moved into rq_qos_done_bio() so that it's next to the code paths that set the flags.

With the patch applied, the above same benchmark shows:

  # resctl-bench -m 29.6G -r out.json run protection::scenario=mem-hog,loops=1
  ...
  Memory Hog Summary
  ==================

IO Latency: R p50=123u:84.4u/985u p90=322u:256u/2.5m p99=1.6m:1.4m/9.5m max=11.1m:36.0m/350m

              W p50=429u:274u/995u p90=1.7m:1.3m/4.5m p99=3.4m:2.7m/11.5m max=7.9m:5.9m/26.5m

Isolation and Request Latency Impact Distributions:

min p01 p05 p10 p25 p50 p75 p90 p95 p99 max mean stdev
isol% 84.91 84.91 89.51 90.73 92.31 94.49 96.36 98.04 98.71 100.0 100.0 94.42 2.81
lat-imp% 0 0 0 0 0 2.81 5.73 11.11 13.92 17.53 22.61 4.10 4.68

Result: isol=94.42:2.81% lat_imp=4.10%:4.68 work_csv=58.34% missing=0%

Fixes a647a524a467 ("block: don't call rq_qos_ops->done_bio if the bio isn't tracked")
Cc [email protected] # v5.15+
Cc Ming Lei <[email protected]>
Cc Yu Kuai <[email protected]>
---
 block/bio.c               |    3 +--
 block/blk-iolatency.c     |    2 +-
 block/blk-rq-qos.h        |   20 +++++++++++---------
 include/linux/blk_types.h |    3 ++-
 4 files changed, 15 insertions(+), 13 deletions(-)
block/bio.c +1 -2
@@ -1486,8 +1486,7 @@ again:
 	if (!bio_integrity_endio(bio))
 		return;
 
-	if (bio->bi_bdev && bio_flagged(bio, BIO_TRACKED))
-		rq_qos_done_bio(bdev_get_queue(bio->bi_bdev), bio);
+	rq_qos_done_bio(bio);
 
 	if (bio->bi_bdev && bio_flagged(bio, BIO_TRACE_COMPLETION)) {
 		trace_block_bio_complete(bdev_get_queue(bio->bi_bdev), bio);
block/blk-iolatency.c +1 -1
@@ -598,7 +598,7 @@ static void blkcg_iolatency_done_bio(str
 	int inflight = 0;
 
 	blkg = bio->bi_blkg;
-	if (!blkg || !bio_flagged(bio, BIO_TRACKED))
+	if (!blkg || !bio_flagged(bio, BIO_QOS_THROTTLED))
 		return;
 
 	iolat = blkg_to_lat(bio->bi_blkg);
block/blk-rq-qos.h +11 -9
@@ -177,20 +177,20 @@ static inline void rq_qos_requeue(struct
 		__rq_qos_requeue(q->rq_qos, rq);
 }
 
-static inline void rq_qos_done_bio(struct request_queue *q, struct bio *bio)
+static inline void rq_qos_done_bio(struct bio *bio)
 {
-	if (q->rq_qos)
-		__rq_qos_done_bio(q->rq_qos, bio);
+	if (bio->bi_bdev && (bio_flagged(bio, BIO_QOS_THROTTLED) ||
+			     bio_flagged(bio, BIO_QOS_MERGED))) {
+		struct request_queue *q = bdev_get_queue(bio->bi_bdev);
+		if (q->rq_qos)
+			__rq_qos_done_bio(q->rq_qos, bio);
+	}
 }
 
 static inline void rq_qos_throttle(struct request_queue *q, struct bio *bio)
 {
-	/*
-	 * BIO_TRACKED lets controllers know that a bio went through the
-	 * normal rq_qos path.
-	 */
 	if (q->rq_qos) {
-		bio_set_flag(bio, BIO_TRACKED);
+		bio_set_flag(bio, BIO_QOS_THROTTLED);
 		__rq_qos_throttle(q->rq_qos, bio);
 	}
 }
@@ -205,8 +205,10 @@ static inline void rq_qos_track(struct r
 static inline void rq_qos_merge(struct request_queue *q, struct request *rq,
 				struct bio *bio)
 {
-	if (q->rq_qos)
+	if (q->rq_qos) {
+		bio_set_flag(bio, BIO_QOS_MERGED);
 		__rq_qos_merge(q->rq_qos, rq, bio);
+	}
 }
 
 static inline void rq_qos_queue_depth_changed(struct request_queue *q)
include/linux/blk_types.h +2 -1
@@ -317,7 +317,8 @@ enum {
 	BIO_TRACE_COMPLETION,	/* bio_endio() should trace the final completion
 				 * of this bio. */
 	BIO_CGROUP_ACCT,	/* has been accounted to a cgroup */
-	BIO_TRACKED,		/* set if bio goes through the rq_qos path */
+	BIO_QOS_THROTTLED,	/* bio went through rq_qos throttle path */
+	BIO_QOS_MERGED,		/* but went through rq_qos merge path */
 	BIO_REMAPPED,
 	BIO_ZONE_WRITE_LOCKED,	/* Owns a zoned device zone write lock */
 	BIO_PERCPU_CACHE,	/* can participate in per-cpu alloc cache */
author maintainer
#142 lore raw

PATCH 142/1126 block: limit request dispatch loop duration

In reply to Greg Kroah-Hartman

From: Shin'ichiro Kawasaki <[email protected]>

commit 572299f03afd676dd4e20669cdaf5ed0fe1379d4 upstream.

When IO requests are made continuously and the target block device handles requests faster than request arrival, the request dispatch loop keeps on repeating to dispatch the arriving requests very long time, more than a minute. Since the loop runs as a workqueue worker task, the very long loop duration triggers workqueue watchdog timeout and BUG [1].

To avoid the very long loop duration, break the loop periodically. When opportunity to dispatch requests still exists, check need_resched(). If need_resched() returns true, the dispatch loop already consumed its time slice, then reschedule the dispatch work and break the loop. With heavy IO load, need_resched() does not return true for 20~30 seconds. To cover such case, check time spent in the dispatch loop with jiffies. If more than 1 second is spent, reschedule the dispatch work and break the loop.

[1]

[  609.691437] BUG: workqueue lockup - pool cpus=10 node=1 flags=0x0 nice=-20 stuck for 35s!
[  609.701820] Showing busy workqueues and worker pools:
[  609.707915] workqueue events: flags=0x0
[  609.712615]   pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/256 refcnt=2
[  609.712626]     pending: drm_fb_helper_damage_work [drm_kms_helper]
[  609.712687] workqueue events_freezable: flags=0x4
[  609.732943]   pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/256 refcnt=2
[  609.732952]     pending: pci_pme_list_scan
[  609.732968] workqueue events_power_efficient: flags=0x80
[  609.751947]   pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/256 refcnt=2
[  609.751955]     pending: neigh_managed_work
[  609.752018] workqueue kblockd: flags=0x18
[  609.769480]   pwq 21: cpus=10 node=1 flags=0x0 nice=-20 active=3/256 refcnt=4
[  609.769488]     in-flight: 1020:blk_mq_run_work_fn
[  609.769498]     pending: blk_mq_timeout_work, blk_mq_run_work_fn
[  609.769744] pool 21: cpus=10 node=1 flags=0x0 nice=-20 hung=35s workers=2 idle: 67
[  639.899730] BUG: workqueue lockup - pool cpus=10 node=1 flags=0x0 nice=-20 stuck for 66s!
[  639.909513] Showing busy workqueues and worker pools:
[  639.915404] workqueue events: flags=0x0
[  639.920197]   pwq 0: cpus=0 node=0 flags=0x0 nice=0 active=1/256 refcnt=2
[  639.920215]     pending: drm_fb_helper_damage_work [drm_kms_helper]
[  639.920365] workqueue kblockd: flags=0x18
[  639.939932]   pwq 21: cpus=10 node=1 flags=0x0 nice=-20 active=3/256 refcnt=4
[  639.939942]     in-flight: 1020:blk_mq_run_work_fn
[  639.939955]     pending: blk_mq_timeout_work, blk_mq_run_work_fn
[  639.940212] pool 21: cpus=10 node=1 flags=0x0 nice=-20 hung=66s workers=2 idle: 67
---
 block/blk-mq-sched.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
block/blk-mq-sched.c +8 -1
@@ -180,11 +180,18 @@ static int __blk_mq_do_dispatch_sched(st
 
 static int blk_mq_do_dispatch_sched(struct blk_mq_hw_ctx *hctx)
 {
+	unsigned long end = jiffies + HZ;
 	int ret;
 
 	do {
 		ret = __blk_mq_do_dispatch_sched(hctx);
-	} while (ret == 1);
+		if (ret != 1)
+			break;
+		if (need_resched() || time_is_before_jiffies(end)) {
+			blk_mq_delay_run_hw_queue(hctx, 0);
+			break;
+		}
+	} while (1);
 
 	return ret;
 }
author maintainer
#143 lore raw

PATCH 143/1126 block: dont merge across cgroup boundaries if blkcg is enabled

In reply to Greg Kroah-Hartman

From: Tejun Heo <[email protected]>

commit 6b2b04590b51aa4cf395fcd185ce439cab5961dc upstream.

blk-iocost and iolatency are cgroup aware rq-qos policies but they didn't disable merges across different cgroups. This obviously can lead to accounting and control errors but more importantly to priority inversions - e.g. an IO which belongs to a higher priority cgroup or IO class may end up getting throttled incorrectly because it gets merged to an IO issued from a low priority cgroup.

Fix it by adding blk_cgroup_mergeable() which is called from merge paths and rejects cross-cgroup and cross-issue_as_root merges.

Fixes d70675121546 ("block: introduce blk-iolatency io controller")
Cc [email protected] # v4.19+
Cc Josef Bacik <[email protected]>
---
 block/blk-merge.c          |   11 +++++++++++
 include/linux/blk-cgroup.h |   17 +++++++++++++++++
 2 files changed, 28 insertions(+)
block/blk-merge.c +11 -0
@@ -9,6 +9,7 @@
 #include <linux/blk-integrity.h>
 #include <linux/scatterlist.h>
 #include <linux/part_stat.h>
+#include <linux/blk-cgroup.h>
 
 #include <trace/events/block.h>
 
@@ -600,6 +601,9 @@ static inline unsigned int blk_rq_get_ma
 static inline int ll_new_hw_segment(struct request *req, struct bio *bio,
 		unsigned int nr_phys_segs)
 {
+	if (!blk_cgroup_mergeable(req, bio))
+		goto no_merge;
+
 	if (blk_integrity_merge_bio(req->q, req, bio) == false)
 		goto no_merge;
 
@@ -696,6 +700,9 @@ static int ll_merge_requests_fn(struct r
 	if (total_phys_segments > blk_rq_get_max_segments(req))
 		return 0;
 
+	if (!blk_cgroup_mergeable(req, next->bio))
+		return 0;
+
 	if (blk_integrity_merge_rq(q, req, next) == false)
 		return 0;
 
@@ -904,6 +911,10 @@ bool blk_rq_merge_ok(struct request *rq,
 	if (bio_data_dir(bio) != rq_data_dir(rq))
 		return false;
 
+	/* don't merge across cgroup boundaries */
+	if (!blk_cgroup_mergeable(rq, bio))
+		return false;
+
 	/* only merge integrity protected bio into ditto rq */
 	if (blk_integrity_merge_bio(rq->q, rq, bio) == false)
 		return false;
include/linux/blk-cgroup.h +17 -0
@@ -24,6 +24,7 @@
 #include <linux/atomic.h>
 #include <linux/kthread.h>
 #include <linux/fs.h>
+#include <linux/blk-mq.h>
 
 /* percpu_counter batch for blkg_[rw]stats, per-cpu drift doesn't matter */
 #define BLKG_STAT_CPU_BATCH	(INT_MAX / 2)
@@ -604,6 +605,21 @@ static inline void blkcg_clear_delay(str
 		atomic_dec(&blkg->blkcg->css.cgroup->congestion_count);
 }
 
+/**
+ * blk_cgroup_mergeable - Determine whether to allow or disallow merges
+ * @rq: request to merge into
+ * @bio: bio to merge
+ *
+ * @bio and @rq should belong to the same cgroup and their issue_as_root should
+ * match. The latter is necessary as we don't want to throttle e.g. a metadata
+ * update because it happens to be next to a regular IO.
+ */
+static inline bool blk_cgroup_mergeable(struct request *rq, struct bio *bio)
+{
+	return rq->bio->bi_blkg == bio->bi_blkg &&
+		bio_issue_as_root_blkg(rq->bio) == bio_issue_as_root_blkg(bio);
+}
+
 void blk_cgroup_bio_start(struct bio *bio);
 void blkcg_add_delay(struct blkcg_gq *blkg, u64 now, u64 delta);
 void blkcg_schedule_throttle(struct request_queue *q, bool use_memdelay);
@@ -659,6 +675,7 @@ static inline void blkg_put(struct blkcg
 static inline bool blkcg_punt_bio_submit(struct bio *bio) { return false; }
 static inline void blkcg_bio_issue_init(struct bio *bio) { }
 static inline void blk_cgroup_bio_start(struct bio *bio) { }
+static inline bool blk_cgroup_mergeable(struct request *rq, struct bio *bio) { return true; }
 
 #define blk_queue_for_each_rl(rl, q)	\
 	for ((rl) = &(q)->root_rl; (rl); (rl) = NULL)
author maintainer
#144 lore raw

PATCH 144/1126 drm/edid: check basic audio support on CEA extension block

In reply to Greg Kroah-Hartman

From: Cooper Chiou <[email protected]>

commit 5662abf6e21338be6d085d6375d3732ac6147fd2 upstream.

Tag code stored in bit7:5 for CTA block byte[3] is not the same as CEA extension block definition. Only check CEA block has basic audio support.

v3: update commit message.

---
 drivers/gpu/drm/drm_edid.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/drm_edid.c +2 -1
@@ -4848,7 +4848,8 @@ bool drm_detect_monitor_audio(struct edi
 	if (!edid_ext)
 		goto end;
 
-	has_audio = ((edid_ext[3] & EDID_BASIC_AUDIO) != 0);
+	has_audio = (edid_ext[0] == CEA_EXT &&
+		    (edid_ext[3] & EDID_BASIC_AUDIO) != 0);
 
 	if (has_audio) {
 		DRM_DEBUG_KMS("Monitor has basic audio support\n");
author maintainer
#145 lore raw

PATCH 145/1126 fbdev: Hot-unplug firmware fb devices on forced removal

In reply to Greg Kroah-Hartman

From: Thomas Zimmermann <[email protected]>

commit 27599aacbaefcbf2af7b06b0029459bbf682000d upstream.

Hot-unplug all firmware-framebuffer devices as part of removing them via remove_conflicting_framebuffers() et al. Releases all memory regions to be acquired by native drivers.

Firmware, such as EFI, install a framebuffer while posting the computer. After removing the firmware-framebuffer device from fbdev, a native driver takes over the hardware and the firmware framebuffer becomes invalid.

Firmware-framebuffer drivers, specifically simplefb, don't release their device from Linux' device hierarchy. It still owns the firmware framebuffer and blocks the native drivers from loading. This has been observed in the vmwgfx driver. [1]

Initiating a device removal (i.e., hot unplug) as part of remove_conflicting_framebuffers() removes the underlying device and returns the memory range to the system.

[1] https://lore.kernel.org/dri-devel/[email protected]/

v2:
* rename variable 'dev' to 'device' (Javier)

---
 drivers/video/fbdev/core/fbmem.c |   29 ++++++++++++++++++++++++++---
 include/linux/fb.h               |    1 +
 2 files changed, 27 insertions(+), 3 deletions(-)
drivers/video/fbdev/core/fbmem.c +26 -3
@@ -25,6 +25,7 @@
 #include <linux/init.h>
 #include <linux/linux_logo.h>
 #include <linux/proc_fs.h>
+#include <linux/platform_device.h>
 #include <linux/seq_file.h>
 #include <linux/console.h>
 #include <linux/kmod.h>
@@ -1559,18 +1560,36 @@ static void do_remove_conflicting_frameb
 	/* check all firmware fbs and kick off if the base addr overlaps */
 	for_each_registered_fb(i) {
 		struct apertures_struct *gen_aper;
+		struct device *device;
 
 		if (!(registered_fb[i]->flags & FBINFO_MISC_FIRMWARE))
 			continue;
 
 		gen_aper = registered_fb[i]->apertures;
+		device = registered_fb[i]->device;
 		if (fb_do_apertures_overlap(gen_aper, a) ||
 			(primary && gen_aper && gen_aper->count &&
 			 gen_aper->ranges[0].base == VGA_FB_PHYS)) {
 
 			printk(KERN_INFO "fb%d: switching to %s from %s\n",
 			       i, name, registered_fb[i]->fix.id);
-			do_unregister_framebuffer(registered_fb[i]);
+
+			/*
+			 * If we kick-out a firmware driver, we also want to remove
+			 * the underlying platform device, such as simple-framebuffer,
+			 * VESA, EFI, etc. A native driver will then be able to
+			 * allocate the memory range.
+			 *
+			 * If it's not a platform device, at least print a warning. A
+			 * fix would add code to remove the device from the system.
+			 */
+			if (dev_is_platform(device)) {
+				registered_fb[i]->forced_out = true;
+				platform_device_unregister(to_platform_device(device));
+			} else {
+				pr_warn("fb%d: cannot remove device\n", i);
+				do_unregister_framebuffer(registered_fb[i]);
+			}
 		}
 	}
 }
@@ -1900,9 +1919,13 @@ EXPORT_SYMBOL(register_framebuffer);
 void
 unregister_framebuffer(struct fb_info *fb_info)
 {
-	mutex_lock(&registration_lock);
+	bool forced_out = fb_info->forced_out;
+
+	if (!forced_out)
+		mutex_lock(&registration_lock);
 	do_unregister_framebuffer(fb_info);
-	mutex_unlock(&registration_lock);
+	if (!forced_out)
+		mutex_unlock(&registration_lock);
 }
 EXPORT_SYMBOL(unregister_framebuffer);
 
include/linux/fb.h +1 -0
@@ -502,6 +502,7 @@ struct fb_info {
 	} *apertures;
 
 	bool skip_vt_switch; /* no VT switch on suspend/resume required */
+	bool forced_out; /* set when being removed by another driver */
 };
 
 static inline struct apertures_struct *alloc_apertures(unsigned int max_num) {
author maintainer
#146 lore raw

PATCH 146/1126 video: fbdev: sm712fb: Fix crash in smtcfb_read()

In reply to Greg Kroah-Hartman

From: Helge Deller <[email protected]>

commit bd771cf5c4254511cc4abb88f3dab3bd58bdf8e8 upstream.

Zheyu Ma reported this crash in the sm712fb driver when reading three bytes from the framebuffer:

 BUG: unable to handle page fault for address: ffffc90001ffffff
 RIP: 0010:smtcfb_read+0x230/0x3e0
 Call Trace:
  vfs_read+0x198/0xa00
  ? do_sys_openat2+0x27d/0x350
  ? __fget_light+0x54/0x340
  ksys_read+0xce/0x190
  do_syscall_64+0x43/0x90

Fix it by removing the open-coded endianess fixup-code and by moving the pointer post decrement out the fb_readl() function.

---
 drivers/video/fbdev/sm712fb.c |   25 +++++++------------------
 1 file changed, 7 insertions(+), 18 deletions(-)
drivers/video/fbdev/sm712fb.c +7 -18
@@ -1047,7 +1047,7 @@ static ssize_t smtcfb_read(struct fb_inf
 	if (count + p > total_size)
 		count = total_size - p;
 
-	buffer = kmalloc((count > PAGE_SIZE) ? PAGE_SIZE : count, GFP_KERNEL);
+	buffer = kmalloc(PAGE_SIZE, GFP_KERNEL);
 	if (!buffer)
 		return -ENOMEM;
 
@@ -1059,25 +1059,14 @@ static ssize_t smtcfb_read(struct fb_inf
 	while (count) {
 		c = (count > PAGE_SIZE) ? PAGE_SIZE : count;
 		dst = buffer;
-		for (i = c >> 2; i--;) {
-			*dst = fb_readl(src++);
-			*dst = big_swap(*dst);
+		for (i = (c + 3) >> 2; i--;) {
+			u32 val;
+
+			val = fb_readl(src);
+			*dst = big_swap(val);
+			src++;
 			dst++;
 		}
-		if (c & 3) {
-			u8 *dst8 = (u8 *)dst;
-			u8 __iomem *src8 = (u8 __iomem *)src;
-
-			for (i = c & 3; i--;) {
-				if (i & 1) {
-					*dst8++ = fb_readb(++src8);
-				} else {
-					*dst8++ = fb_readb(--src8);
-					src8 += 2;
-				}
-			}
-			src = (u32 __iomem *)src8;
-		}
 
 		if (copy_to_user(buf, buffer, c)) {
 			err = -EFAULT;
author maintainer
#147 lore raw

PATCH 147/1126 video: fbdev: atari: Atari 2 bpp (STe) palette bugfix

In reply to Greg Kroah-Hartman

From: Michael Schmitz <[email protected]>

commit c8be5edbd36ceed2ff3d6b8f8e40643c3f396ea3 upstream.

The code to set the shifter STe palette registers has a long standing operator precedence bug, manifesting as colors set on a 2 bits per pixel frame buffer coming up with a distinctive blue tint.

Add parentheses around the calculation of the per-color palette data before shifting those into their respective bit field position.

This bug goes back a long way (2.4 days at the very least) so there won't be a Fixes: tag.

Tested on ARAnyM as well on Falcon030 hardware.

---
 drivers/video/fbdev/atafb.c |   12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/video/fbdev/atafb.c +6 -6
@@ -1683,9 +1683,9 @@ static int falcon_setcolreg(unsigned int
 			   ((blue & 0xfc00) >> 8));
 	if (regno < 16) {
 		shifter_tt.color_reg[regno] =
-			(((red & 0xe000) >> 13) | ((red & 0x1000) >> 12) << 8) |
-			(((green & 0xe000) >> 13) | ((green & 0x1000) >> 12) << 4) |
-			((blue & 0xe000) >> 13) | ((blue & 0x1000) >> 12);
+			((((red & 0xe000) >> 13)   | ((red & 0x1000) >> 12)) << 8)   |
+			((((green & 0xe000) >> 13) | ((green & 0x1000) >> 12)) << 4) |
+			   ((blue & 0xe000) >> 13) | ((blue & 0x1000) >> 12);
 		((u32 *)info->pseudo_palette)[regno] = ((red & 0xf800) |
 						       ((green & 0xfc00) >> 5) |
 						       ((blue & 0xf800) >> 11));
@@ -1971,9 +1971,9 @@ static int stste_setcolreg(unsigned int
 	green >>= 12;
 	if (ATARIHW_PRESENT(EXTD_SHIFTER))
 		shifter_tt.color_reg[regno] =
-			(((red & 0xe) >> 1) | ((red & 1) << 3) << 8) |
-			(((green & 0xe) >> 1) | ((green & 1) << 3) << 4) |
-			((blue & 0xe) >> 1) | ((blue & 1) << 3);
+			((((red & 0xe)   >> 1) | ((red & 1)   << 3)) << 8) |
+			((((green & 0xe) >> 1) | ((green & 1) << 3)) << 4) |
+			  ((blue & 0xe)  >> 1) | ((blue & 1)  << 3);
 	else
 		shifter_tt.color_reg[regno] =
 			((red & 0xe) << 7) |
author maintainer
#148 lore raw

PATCH 148/1126 rfkill: make new event layout opt-in

In reply to Greg Kroah-Hartman

From: Johannes Berg <[email protected]>

commit 54f586a9153201c6cff55e1f561990c78bd99aa7 upstream.

Again new complaints surfaced that we had broken the ABI here, although previously all the userspace tools had agreed that it was their mistake and fixed it. Yet now there are cases (e.g. RHEL) that want to run old userspace with newer kernels, and thus are broken.

Since this is a bit of a whack-a-mole thing, change the whole extensibility scheme of rfkill to no longer just rely on the message lengths, but instead require userspace to opt in via a new ioctl to a given maximum event size that it is willing to understand.

By default, set that to RFKILL_EVENT_SIZE_V1 (8), so that the behaviour for userspace not calling the ioctl will look as if it's just running on an older kernel.

---
 include/uapi/linux/rfkill.h |   14 +++++++++++-
 net/rfkill/core.c           |   48 +++++++++++++++++++++++++++++++-------------
 2 files changed, 46 insertions(+), 16 deletions(-)
include/uapi/linux/rfkill.h +12 -2
@@ -159,8 +159,16 @@ struct rfkill_event_ext {
  * old behaviour for all userspace, unless it explicitly opts in to the
  * rules outlined here by using the new &struct rfkill_event_ext.
  *
- * Userspace using &struct rfkill_event_ext must adhere to the following
- * rules
+ * Additionally, some other userspace (bluez, g-s-d) was reading with a
+ * large size but as streaming reads rather than message-based, or with
+ * too strict checks for the returned size. So eventually, we completely
+ * reverted this, and extended messages need to be opted in to by using
+ * an ioctl:
+ *
+ *  ioctl(fd, RFKILL_IOCTL_MAX_SIZE, sizeof(struct rfkill_event_ext));
+ *
+ * Userspace using &struct rfkill_event_ext and the ioctl must adhere to
+ * the following rules:
  *
  * 1. accept short writes, optionally using them to detect that it's
  *    running on an older kernel;
@@ -175,6 +183,8 @@ struct rfkill_event_ext {
 #define RFKILL_IOC_MAGIC	'R'
 #define RFKILL_IOC_NOINPUT	1
 #define RFKILL_IOCTL_NOINPUT	_IO(RFKILL_IOC_MAGIC, RFKILL_IOC_NOINPUT)
+#define RFKILL_IOC_MAX_SIZE	2
+#define RFKILL_IOCTL_MAX_SIZE	_IOW(RFKILL_IOC_MAGIC, RFKILL_IOC_EXT_SIZE, __u32)
 
 /* and that's all userspace gets */
 
net/rfkill/core.c +34 -14

105 lines not on this page · raw message

author maintainer
#149 lore raw

PATCH 149/1126 ARM: dts: at91: sama7g5: Remove unused properties in i2c nodes

In reply to Greg Kroah-Hartman

From: Tudor Ambarus <[email protected]>

commit cbb92a7717d2e1c512b7e81c6b22c7298b58a881 upstream.

The "atmel,use-dma-rx", "atmel,use-dma-rx" dt properties are not used by the i2c-at91 driver, nor they are defined in the bindings file, thus remove them.

---
 arch/arm/boot/dts/sama7g5.dtsi |    6 ------
 1 file changed, 6 deletions(-)
arch/arm/boot/dts/sama7g5.dtsi +0 -6
@@ -382,8 +382,6 @@
 				dmas = <&dma0 AT91_XDMAC_DT_PERID(7)>,
 					<&dma0 AT91_XDMAC_DT_PERID(8)>;
 				dma-names = "rx", "tx";
-				atmel,use-dma-rx;
-				atmel,use-dma-tx;
 				status = "disabled";
 			};
 		};
@@ -558,8 +556,6 @@
 				dmas = <&dma0 AT91_XDMAC_DT_PERID(21)>,
 					<&dma0 AT91_XDMAC_DT_PERID(22)>;
 				dma-names = "rx", "tx";
-				atmel,use-dma-rx;
-				atmel,use-dma-tx;
 				status = "disabled";
 			};
 		};
@@ -584,8 +580,6 @@
 				dmas = <&dma0 AT91_XDMAC_DT_PERID(23)>,
 					<&dma0 AT91_XDMAC_DT_PERID(24)>;
 				dma-names = "rx", "tx";
-				atmel,use-dma-rx;
-				atmel,use-dma-tx;
 				status = "disabled";
 			};
 		};
author maintainer
#150 lore raw

PATCH 150/1126 ARM: dts: at91: sama5d2: Fix PMERRLOC resource size

In reply to Greg Kroah-Hartman

From: Tudor Ambarus <[email protected]>

commit 0fb578a529ac7aca326a9fa475b4a6f58a756fda upstream.

PMERRLOC resource size was set to 0x100, which resulted in HSMC_ERRLOCx register being truncated to offset x = 21, causing error correction to fail if more than 22 bit errors and if 24 or 32 bit error correction was supported.

---
 arch/arm/boot/dts/sama5d2.dtsi |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/boot/dts/sama5d2.dtsi +1 -1
@@ -415,7 +415,7 @@
 				pmecc: ecc-engine@f8014070 {
 					compatible = "atmel,sama5d2-pmecc";
 					reg = <0xf8014070 0x490>,
-					      <0xf8014500 0x100>;
+					      <0xf8014500 0x200>;
 				};
 			};
 
author maintainer
#151 lore raw

PATCH 151/1126 ARM: dts: exynos: fix UART3 pins configuration in Exynos5250

In reply to Greg Kroah-Hartman

From: Krzysztof Kozlowski <[email protected]>

commit 372d7027fed43c8570018e124cf78b89523a1f8e upstream.

The gpa1-4 pin was put twice in UART3 pin configuration of Exynos5250, instead of proper pin gpa1-5.

Fixes f8bfe2b050f3 ("ARM: dts: add pin state information in client nodes for Exynos5 platforms")
---
 arch/arm/boot/dts/exynos5250-pinctrl.dtsi |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/boot/dts/exynos5250-pinctrl.dtsi +1 -1
@@ -260,7 +260,7 @@
 	};
 
 	uart3_data: uart3-data {
-		samsung,pins = "gpa1-4", "gpa1-4";
+		samsung,pins = "gpa1-4", "gpa1-5";
 		samsung,pin-function = <EXYNOS_PIN_FUNC_2>;
 		samsung,pin-pud = <EXYNOS_PIN_PULL_NONE>;
 		samsung,pin-drv = <EXYNOS4_PIN_DRV_LV1>;
author maintainer
#152 lore raw

PATCH 152/1126 ARM: dts: exynos: add missing HDMI supplies on SMDK5250

In reply to Greg Kroah-Hartman

From: Krzysztof Kozlowski <[email protected]>

commit 60a9914cb2061ba612a3f14f6ad329912b486360 upstream.

Add required VDD supplies to HDMI block on SMDK5250. Without them, the HDMI driver won't probe. Because of lack of schematics, use same supplies as on Arndale 5250 board (voltage matches).

---
 arch/arm/boot/dts/exynos5250-smdk5250.dts |    3 +++
 1 file changed, 3 insertions(+)
arch/arm/boot/dts/exynos5250-smdk5250.dts +3 -0
@@ -118,6 +118,9 @@
 	status = "okay";
 	ddc = <&i2c_2>;
 	hpd-gpios = <&gpx3 7 GPIO_ACTIVE_HIGH>;
+	vdd-supply = <&ldo8_reg>;
+	vdd_osc-supply = <&ldo10_reg>;
+	vdd_pll-supply = <&ldo8_reg>;
 };
 
 &i2c_0 {
author maintainer
#153 lore raw

PATCH 153/1126 ARM: dts: exynos: add missing HDMI supplies on SMDK5420

In reply to Greg Kroah-Hartman

From: Krzysztof Kozlowski <[email protected]>

commit 453a24ded415f7fce0499c6b0a2c7b28f84911f2 upstream.

Add required VDD supplies to HDMI block on SMDK5420. Without them, the HDMI driver won't probe. Because of lack of schematics, use same supplies as on Arndale Octa and Odroid XU3 boards (voltage matches).

---
 arch/arm/boot/dts/exynos5420-smdk5420.dts |    3 +++
 1 file changed, 3 insertions(+)
arch/arm/boot/dts/exynos5420-smdk5420.dts +3 -0
@@ -124,6 +124,9 @@
 	hpd-gpios = <&gpx3 7 GPIO_ACTIVE_HIGH>;
 	pinctrl-names = "default";
 	pinctrl-0 = <&hdmi_hpd_irq>;
+	vdd-supply = <&ldo6_reg>;
+	vdd_osc-supply = <&ldo7_reg>;
+	vdd_pll-supply = <&ldo6_reg>;
 };
 
 &hsi2c_4 {
author maintainer
#154 lore raw

PATCH 154/1126 mgag200 fix memmapsl configuration in GCTL6 register

In reply to Greg Kroah-Hartman

From: Jocelyn Falempe <[email protected]>

commit 028a73e10705af1ffd51f2537460f616dc58680e upstream.

On some servers with MGA G200_SE_A (rev 42), booting with Legacy BIOS, the hardware hangs when using kdump and kexec into the kdump kernel. This happens when the uncompress code tries to write "Decompressing Linux" to the VGA Console.

It can be reproduced by writing to the VGA console (0xB8000) after booting to graphic mode, it generates the following error:

kernel:NMI: PCI system error (SERR) for reason a0 on CPU 0. kernel:Dazed and confused, but trying to continue

The root cause is the configuration of the MGA GCTL6 register

According to the GCTL6 register documentation:

bit 0 is gcgrmode:
0: Enables alpha mode, and the character generator addressing system is
activated.
1: Enables graphics mode, and the character addressing system is not
used.

bit 1 is chainodd even:
0: The A0 signal of the memory address bus is used during system memory
addressing.
1: Allows A0 to be replaced by either the A16 signal of the system
address (ifmemmapsl is ‘00’), or by the hpgoddev (MISC<5>, odd/even page select) field, described on page 3-294).

bit 3-2 are memmapsl:

    Memory map select bits 1 and 0. VGA.
    These bits select where the video memory is mapped, as shown below:
        00 => A0000h - BFFFFh
        01 => A0000h - AFFFFh
        10 => B0000h - B7FFFh
        11 => B8000h - BFFFFh

bit 7-4 are reserved.

Current code set it to 0x05 => memmapsl to b01 => 0xa0000 (graphic mode) But on x86, the VGA console is at 0xb8000 (text mode)
In arch/x86/boot/compressed/misc.c debug strings are written to 0xb8000 As the driver doesn't use this mapping at 0xa0000, it is safe to set it to 0xb8000 instead, to avoid kernel hang on G200_SE_A rev42, with kexec/kdump.

Thus changing the value 0x05 to 0x0d

---
 drivers/gpu/drm/mgag200/mgag200_mode.c |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/gpu/drm/mgag200/mgag200_mode.c +4 -1
@@ -529,7 +529,10 @@ static void mgag200_set_format_regs(stru
 	WREG_GFX(3, 0x00);
 	WREG_GFX(4, 0x00);
 	WREG_GFX(5, 0x40);
-	WREG_GFX(6, 0x05);
+	/* GCTL6 should be 0x05, but we configure memmapsl to 0xb8000 (text mode),
+	 * so that it doesn't hang when running kexec/kdump on G200_SE rev42.
+	 */
+	WREG_GFX(6, 0x0d);
 	WREG_GFX(7, 0x0f);
 	WREG_GFX(8, 0x0f);
 
author maintainer
#155 lore raw

PATCH 155/1126 carl9170: fix missing bit-wise or operator for tx_params

In reply to Greg Kroah-Hartman

From: Colin Ian King <[email protected]>

commit 02a95374b5eebdbd3b6413fd7ddec151d2ea75a1 upstream.

Currently tx_params is being re-assigned with a new value and the previous setting IEEE80211_HT_MCS_TX_RX_DIFF is being overwritten. The assignment operator is incorrect, the original intent was to bit-wise or the value in. Fix this by replacing the = operator with |= instead.

Kudos to Christian Lamparter for suggesting the correct fix.

---
 drivers/net/wireless/ath/carl9170/main.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/wireless/ath/carl9170/main.c +1 -1
@@ -1914,7 +1914,7 @@ static int carl9170_parse_eeprom(struct
 		WARN_ON(!(tx_streams >= 1 && tx_streams <=
 			IEEE80211_HT_MCS_TX_MAX_STREAMS));
 
-		tx_params = (tx_streams - 1) <<
+		tx_params |= (tx_streams - 1) <<
 			    IEEE80211_HT_MCS_TX_MAX_STREAMS_SHIFT;
 
 		carl9170_band_2GHz.ht_cap.mcs.tx_params |= tx_params;
author maintainer
#156 lore raw

PATCH 156/1126 pstore: Dont use semaphores in always-atomic-context code

In reply to Greg Kroah-Hartman

From: Jann Horn <[email protected]>

commit 8126b1c73108bc691f5643df19071a59a69d0bc6 upstream.

pstore_dump() is *always* invoked in atomic context (nowadays in an RCU read-side critical section, before that under a spinlock). It doesn't make sense to try to use semaphores here.

This is mostly a revert of commit ea84b580b955 ("pstore: Convert buf_lock to semaphore"), except that two parts aren't restored back exactly as they were:

- keep the lock initialization in pstore_register
- in efi_pstore_write(), always set the "block" flag to false
- omit "is_locked", that was unnecessary since
commit 959217c84c27 ("pstore: Actually give up during locking failure")
- fix the bailout message

The actual problem that the buggy commit was trying to address may have been that the use of preemptible() in efi_pstore_write() was wrong - it only looks at preempt_count() and the state of IRQs, but __rcu_read_lock() doesn't touch either of those under CONFIG_PREEMPT_RCU. (Sidenote: CONFIG_PREEMPT_RCU means that the scheduler can preempt tasks in RCU read-side critical sections, but you're not allowed to actively block/reschedule.)

Lockdep probably never caught the problem because it's very rare that you actually hit the contended case, so lockdep always just sees the down_trylock(), not the down_interruptible(), and so it can't tell that there's a problem.

---
 drivers/firmware/efi/efi-pstore.c |    2 +-
 fs/pstore/platform.c              |   38 ++++++++++++++++++--------------------
 include/linux/pstore.h            |    6 +++---
 3 files changed, 22 insertions(+), 24 deletions(-)
drivers/firmware/efi/efi-pstore.c +1 -1
@@ -266,7 +266,7 @@ static int efi_pstore_write(struct pstor
 		efi_name[i] = name[i];
 
 	ret = efivar_entry_set_safe(efi_name, vendor, PSTORE_EFI_ATTRIBUTES,
-			      preemptible(), record->size, record->psi->buf);
+			      false, record->size, record->psi->buf);
 
 	if (record->reason == KMSG_DUMP_OOPS && try_module_get(THIS_MODULE))
 		if (!schedule_work(&efivar_work))
fs/pstore/platform.c +18 -20
@@ -143,21 +143,22 @@ static void pstore_timer_kick(void)
 	mod_timer(&pstore_timer, jiffies + msecs_to_jiffies(pstore_update_ms));
 }
 
-/*
- * Should pstore_dump() wait for a concurrent pstore_dump()? If
- * not, the current pstore_dump() will report a failure to dump
- * and return.
- */
-static bool pstore_cannot_wait(enum kmsg_dump_reason reason)
+static bool pstore_cannot_block_path(enum kmsg_dump_reason reason)
 {
-	/* In NMI path, pstore shouldn't block regardless of reason. */
+	/*
+	 * In case of NMI path, pstore shouldn't be blocked
+	 * regardless of reason.
+	 */
 	if (in_nmi())
 		return true;
 
 	switch (reason) {
 	/* In panic case, other cpus are stopped by smp_send_stop(). */
 	case KMSG_DUMP_PANIC:
-	/* Emergency restart shouldn't be blocked. */
+	/*
+	 * Emergency restart shouldn't be blocked by spinning on
+	 * pstore_info::buf_lock.
+	 */
 	case KMSG_DUMP_EMERG:
 		return true;
 	default:
@@ -389,21 +390,19 @@ static void pstore_dump(struct kmsg_dump
 	unsigned long	total = 0;
 	const char	*why;
 	unsigned int	part = 1;
+	unsigned long	flags = 0;
 	int		ret;
 
 	why = kmsg_dump_reason_str(reason);
 
-	if (down_trylock(&psinfo->buf_lock)) {
-		/* Failed to acquire lock: give up if we cannot wait. */
-		if (pstore_cannot_wait(reason)) {
-			pr_err("dump skipped in %s path: may corrupt error record\n",
-				in_nmi() ? "NMI" : why);
-			return;
-		}
-		if (down_interruptible(&psinfo->buf_lock)) {
-			pr_err("could not grab semaphore?!\n");
+	if (pstore_cannot_block_path(reason)) {
+		if (!spin_trylock_irqsave(&psinfo->buf_lock, flags)) {
+			pr_err("dump skipped in %s path because of concurrent dump\n",
+					in_nmi() ? "NMI" : why);
 			return;
 		}
+	} else {
+		spin_lock_irqsave(&psinfo->buf_lock, flags);
 	}
 
 	kmsg_dump_rewind(&iter);
@@ -467,8 +466,7 @@ static void pstore_dump(struct kmsg_dump
 		total += record.size;
 		part++;
 	}
-
-	up(&psinfo->buf_lock);
+	spin_unlock_irqrestore(&psinfo->buf_lock, flags);
 }
 
 static struct kmsg_dumper pstore_dumper = {
@@ -594,7 +592,7 @@ int pstore_register(struct pstore_info *
 		psi->write_user = pstore_write_user_compat;
 	psinfo = psi;
 	mutex_init(&psinfo->read_mutex);
-	sema_init(&psinfo->buf_lock, 1);
+	spin_lock_init(&psinfo->buf_lock);
 
 	if (psi->flags & PSTORE_FLAGS_DMESG)
 		allocate_buf_for_compression();
include/linux/pstore.h +3 -3
@@ -14,7 +14,7 @@
 #include <linux/errno.h>
 #include <linux/kmsg_dump.h>
 #include <linux/mutex.h>
-#include <linux/semaphore.h>
+#include <linux/spinlock.h>
 #include <linux/time.h>
 #include <linux/types.h>
 
@@ -87,7 +87,7 @@ struct pstore_record {
  * @owner:	module which is responsible for this backend driver
  * @name:	name of the backend driver
  *
- * @buf_lock:	semaphore to serialize access to @buf
+ * @buf_lock:	spinlock to serialize access to @buf
  * @buf:	preallocated crash dump buffer
  * @bufsize:	size of @buf available for crash dump bytes (must match
  *		smallest number of bytes available for writing to a
@@ -178,7 +178,7 @@ struct pstore_info {
 	struct module	*owner;
 	const char	*name;
 
-	struct semaphore buf_lock;
+	spinlock_t	buf_lock;
 	char		*buf;
 	size_t		bufsize;
 
author maintainer
#157 lore raw

PATCH 157/1126 thermal: int340x: Increase bitmap size

In reply to Greg Kroah-Hartman

From: Srinivas Pandruvada <[email protected]>

commit 668f69a5f863b877bc3ae129efe9a80b6f055141 upstream.

The number of policies are 10, so can't be supported by the bitmap size of u8.

Even though there are no platfoms with these many policies, but for correctness increase to u32.

Fixes 16fc8eca1975 ("thermal/int340x_thermal: Add additional UUIDs")
Cc 5.1+ <[email protected]> # 5.1+
---
 drivers/thermal/intel/int340x_thermal/int3400_thermal.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/thermal/intel/int340x_thermal/int3400_thermal.c +1 -1
@@ -53,7 +53,7 @@ struct int3400_thermal_priv {
 	struct art *arts;
 	int trt_count;
 	struct trt *trts;
-	u8 uuid_bitmap;
+	u32 uuid_bitmap;
 	int rel_misc_dev_res;
 	int current_uuid_index;
 	char *data_vault;
author maintainer
#158 lore raw

PATCH 158/1126 lib/raid6/test: fix multiple definition linking error

In reply to Greg Kroah-Hartman

From: Dirk Müller <[email protected]>

commit a5359ddd052860bacf957e65fe819c63e974b3a6 upstream.

GCC 10+ defaults to -fno-common, which enforces proper declaration of external references using "extern". without this change a link would fail with:

lib/raid6/test/algos.c:28: multiple definition of `raid6_call'; lib/raid6/test/test.c:22: first defined here

the pq.h header that is included already includes an extern declaration so we can just remove the redundant one here.

---
 lib/raid6/test/test.c |    1 -
 1 file changed, 1 deletion(-)
lib/raid6/test/test.c +0 -1
@@ -19,7 +19,6 @@
 #define NDISKS		16	/* Including P and Q */
 
 const char raid6_empty_zero_page[PAGE_SIZE] __attribute__((aligned(PAGE_SIZE)));
-struct raid6_calls raid6_call;
 
 char *dataptrs[NDISKS];
 char data[NDISKS][PAGE_SIZE] __attribute__((aligned(PAGE_SIZE)));
author maintainer
#159 lore raw

PATCH 159/1126 exec: Force single empty string when argv is empty

In reply to Greg Kroah-Hartman

From: Kees Cook <[email protected]>

commit dcd46d897adb70d63e025f175a00a89797d31a43 upstream.

Quoting[1] Ariadne Conill:

"In several other operating systems, it is a hard requirement that the second argument to execve(2) be the name of a program, thus prohibiting a scenario where argc < 1. POSIX 2017 also recommends this behaviour, but it is not an explicit requirement[2]:

The argument arg0 should point to a filename string that is associated with the process being started by one of the exec functions.
...
Interestingly, Michael Kerrisk opened an issue about this in 2008[3], but there was no consensus to support fixing this issue then. Hopefully now that CVE-2021-4034 shows practical exploitative use[4] of this bug in a shellcode, we can reconsider.

This issue is being tracked in the KSPP issue tracker[5]."

While the initial code searches[6][7] turned up what appeared to be mostly corner case tests, trying to that just reject argv == NULL (or an immediately terminated pointer list) quickly started tripping[8] existing userspace programs.

The next best approach is forcing a single empty string into argv and adjusting argc to match. The number of programs depending on argc == 0 seems a smaller set than those calling execve with a NULL argv.

Account for the additional stack space in bprm_stack_limits(). Inject an empty string when argc == 0 (and set argc = 1). Warn about the case so userspace has some notice about the change:

process './argc0' launched './argc0' with NULL argv: empty string added

Additionally WARN() and reject NULL argv usage for kernel threads.

[1] https://lore.kernel.org/lkml/[email protected]/
[2] https://pubs.opengroup.org/onlinepubs/9699919799/functions/exec.html
[3] https://bugzilla.kernel.org/show_bug.cgi?id=8408
[4] https://www.qualys.com/2022/01/25/cve-2021-4034/pwnkit.txt
[5] https://github.com/KSPP/linux/issues/176
[6] https://codesearch.debian.net/search?q=execve%5C+*%5C%28%5B%5E%2C%5D%2B%2C+*NULL&literal=0
[7] https://codesearch.debian.net/search?q=execlp%3F%5Cs*%5C%28%5B%5E%2C%5D%2B%2C%5Cs*NULL&literal=0
[8] https://lore.kernel.org/lkml/20220131144352.GE16385@xsang-OptiPlex-9020/

---
 fs/exec.c |   26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)
fs/exec.c +25 -1
@@ -495,8 +495,14 @@ static int bprm_stack_limits(struct linu
 	 * the stack. They aren't stored until much later when we can't
 	 * signal to the parent that the child has run out of stack space.
 	 * Instead, calculate it here so it's possible to fail gracefully.
+	 *
+	 * In the case of argc = 0, make sure there is space for adding a
+	 * empty string (which will bump argc to 1), to ensure confused
+	 * userspace programs don't start processing from argv[1], thinking
+	 * argc can never be 0, to keep them from walking envp by accident.
+	 * See do_execveat_common().
 	 */
-	ptr_size = (bprm->argc + bprm->envc) * sizeof(void *);
+	ptr_size = (max(bprm->argc, 1) + bprm->envc) * sizeof(void *);
 	if (limit <= ptr_size)
 		return -E2BIG;
 	limit -= ptr_size;
@@ -1897,6 +1903,9 @@ static int do_execveat_common(int fd, st
 	}
 
 	retval = count(argv, MAX_ARG_STRINGS);
+	if (retval == 0)
+		pr_warn_once("process '%s' launched '%s' with NULL argv: empty string added\n",
+			     current->comm, bprm->filename);
 	if (retval < 0)
 		goto out_free;
 	bprm->argc = retval;
@@ -1923,6 +1932,19 @@ static int do_execveat_common(int fd, st
 	if (retval < 0)
 		goto out_free;
 
+	/*
+	 * When argv is empty, add an empty string ("") as argv[0] to
+	 * ensure confused userspace programs that start processing
+	 * from argv[1] won't end up walking envp. See also
+	 * bprm_stack_limits().
+	 */
+	if (bprm->argc == 0) {
+		retval = copy_string_kernel("", bprm);
+		if (retval < 0)
+			goto out_free;
+		bprm->argc = 1;
+	}
+
 	retval = bprm_execve(bprm, fd, filename, flags);
 out_free:
 	free_bprm(bprm);
@@ -1951,6 +1973,8 @@ int kernel_execve(const char *kernel_fil
 	}
 
 	retval = count_strings_kernel(argv);
+	if (WARN_ON_ONCE(retval == 0))
+		retval = -EINVAL;
 	if (retval < 0)
 		goto out_free;
 	bprm->argc = retval;
author maintainer
#160 lore raw

PATCH 160/1126 crypto: rsa-pkcs1pad - only allow with rsa

In reply to Greg Kroah-Hartman

From: Eric Biggers <[email protected]>

commit 9b30430ea356f237945e52f8a3a42158877bd5a9 upstream.

The pkcs1pad template can be instantiated with an arbitrary akcipher algorithm, which doesn't make sense; it is specifically an RSA padding scheme. Make it check that the underlying algorithm really is RSA.

Fixes 3d5b1ecdea6f ("crypto: rsa - RSA padding algorithm")
Cc <[email protected]> # v4.5+
---
 crypto/rsa-pkcs1pad.c |    5 +++++
 1 file changed, 5 insertions(+)
crypto/rsa-pkcs1pad.c +5 -0
@@ -621,6 +621,11 @@ static int pkcs1pad_create(struct crypto
 
 	rsa_alg = crypto_spawn_akcipher_alg(&ctx->spawn);
 
+	if (strcmp(rsa_alg->base.cra_name, "rsa") != 0) {
+		err = -EINVAL;
+		goto err_free_inst;
+	}
+
 	err = -ENAMETOOLONG;
 	hash_name = crypto_attr_alg_name(tb[2]);
 	if (IS_ERR(hash_name)) {
author maintainer
#161 lore raw

PATCH 161/1126 crypto: rsa-pkcs1pad - correctly get hash from source scatterlist

In reply to Greg Kroah-Hartman

From: Eric Biggers <[email protected]>

commit e316f7179be22912281ce6331d96d7c121fb2b17 upstream.

Commit c7381b012872 ("crypto: akcipher - new verify API for public key algorithms") changed akcipher_alg::verify to take in both the signature and the actual hash and do the signature verification, rather than just return the hash expected by the signature as was the case before. To do this, it implemented a hack where the signature and hash are concatenated with each other in one scatterlist.

Obviously, for this to work correctly, akcipher_alg::verify needs to correctly extract the two items from the scatterlist it is given. Unfortunately, it doesn't correctly extract the hash in the case where the signature is longer than the RSA key size, as it assumes that the signature's length is equal to the RSA key size. This causes a prefix of the hash, or even the entire hash, to be taken from the *signature*.

(Note, the case of a signature longer than the RSA key size should not be allowed in the first place; a separate patch will fix that.)

It is unclear whether the resulting scheme has any useful security properties.

Fix this by correctly extracting the hash from the scatterlist.

Fixes c7381b012872 ("crypto: akcipher - new verify API for public key algorithms")
Cc <[email protected]> # v5.2+
---
 crypto/rsa-pkcs1pad.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
crypto/rsa-pkcs1pad.c +1 -1
@@ -495,7 +495,7 @@ static int pkcs1pad_verify_complete(stru
 			   sg_nents_for_len(req->src,
 					    req->src_len + req->dst_len),
 			   req_ctx->out_buf + ctx->key_size,
-			   req->dst_len, ctx->key_size);
+			   req->dst_len, req->src_len);
 	/* Do the actual verification step. */
 	if (memcmp(req_ctx->out_buf + ctx->key_size, out_buf + pos,
 		   req->dst_len) != 0)
author maintainer
#162 lore raw

PATCH 162/1126 crypto: rsa-pkcs1pad - restore signature length check

In reply to Greg Kroah-Hartman

From: Eric Biggers <[email protected]>

commit d3481accd974541e6a5d6a1fb588924a3519c36e upstream.

RSA PKCS#1 v1.5 signatures are required to be the same length as the RSA key size. RFC8017 specifically requires the verifier to check this
(https://datatracker.ietf.org/doc/html/rfc8017#section-8.2.2).

Commit a49de377e051 ("crypto: Add hash param to pkcs1pad") changed the kernel to allow longer signatures, but didn't explain this part of the change; it seems to be unrelated to the rest of the commit.

Revert this change, since it doesn't appear to be correct.

We can be pretty sure that no one is relying on overly-long signatures (which would have to be front-padded with zeroes) being supported, given that they would have been broken since commit c7381b012872 ("crypto: akcipher - new verify API for public key algorithms").

Fixes a49de377e051 ("crypto: Add hash param to pkcs1pad")
Cc <[email protected]> # v4.6+
Cc Tadeusz Struk <[email protected]>
---
 crypto/rsa-pkcs1pad.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
crypto/rsa-pkcs1pad.c +1 -1
@@ -538,7 +538,7 @@ static int pkcs1pad_verify(struct akciph
 
 	if (WARN_ON(req->dst) ||
 	    WARN_ON(!req->dst_len) ||
-	    !ctx->key_size || req->src_len < ctx->key_size)
+	    !ctx->key_size || req->src_len != ctx->key_size)
 		return -EINVAL;
 
 	req_ctx->out_buf = kmalloc(ctx->key_size + req->dst_len, GFP_KERNEL);
author maintainer
#163 lore raw

PATCH 163/1126 crypto: rsa-pkcs1pad - fix buffer overread in pkcs1pad_verify_complete()

In reply to Greg Kroah-Hartman

From: Eric Biggers <[email protected]>

commit a24611ea356c7f3f0ec926da11b9482ac1f414fd upstream.

Before checking whether the expected digest_info is present, we need to check that there are enough bytes remaining.

Fixes a49de377e051 ("crypto: Add hash param to pkcs1pad")
Cc <[email protected]> # v4.6+
Cc Tadeusz Struk <[email protected]>
---
 crypto/rsa-pkcs1pad.c |    2 ++
 1 file changed, 2 insertions(+)
crypto/rsa-pkcs1pad.c +2 -0
@@ -476,6 +476,8 @@ static int pkcs1pad_verify_complete(stru
 	pos++;
 
 	if (digest_info) {
+		if (digest_info->size > dst_len - pos)
+			goto done;
 		if (crypto_memneq(out_buf + pos, digest_info->data,
 				  digest_info->size))
 			goto done;
author maintainer
#164 lore raw

PATCH 164/1126 bcache: fixup multiple threads crash

In reply to Greg Kroah-Hartman

From: Mingzhe Zou <[email protected]>

commit 887554ab96588de2917b6c8c73e552da082e5368 upstream.

When multiple threads to check btree nodes in parallel, the main thread wait for all threads to stop or CACHE_SET_IO_DISABLE flag:

wait_event_interruptible(check_state->wait,

                         atomic_read(&check_state->started) == 0 ||
                         test_bit(CACHE_SET_IO_DISABLE, &c->flags));

However, the bch_btree_node_read and bch_btree_node_read_done maybe call bch_cache_set_error, then the CACHE_SET_IO_DISABLE will be set. If the flag already set, the main thread return error. At the same time, maybe some threads still running and read NULL pointer, the kernel will crash.

This patch change the event wait condition, the main thread must wait for all threads to stop.

Fixes 8e7102273f597 ("bcache: make bch_btree_check() to be multithreaded")
---
 drivers/md/bcache/btree.c     |    6 ++++--
 drivers/md/bcache/writeback.c |    6 ++++--
 2 files changed, 8 insertions(+), 4 deletions(-)
drivers/md/bcache/btree.c +4 -2
@@ -2060,9 +2060,11 @@ int bch_btree_check(struct cache_set *c)
 		}
 	}
 
+	/*
+	 * Must wait for all threads to stop.
+	 */
 	wait_event_interruptible(check_state->wait,
-				 atomic_read(&check_state->started) == 0 ||
-				  test_bit(CACHE_SET_IO_DISABLE, &c->flags));
+				 atomic_read(&check_state->started) == 0);
 
 	for (i = 0; i < check_state->total_threads; i++) {
 		if (check_state->infos[i].result) {
drivers/md/bcache/writeback.c +4 -2
@@ -998,9 +998,11 @@ void bch_sectors_dirty_init(struct bcach
 		}
 	}
 
+	/*
+	 * Must wait for all threads to stop.
+	 */
 	wait_event_interruptible(state->wait,
-		 atomic_read(&state->started) == 0 ||
-		 test_bit(CACHE_SET_IO_DISABLE, &c->flags));
+		 atomic_read(&state->started) == 0);
 
 out:
 	kfree(state);
author maintainer
#165 lore raw

PATCH 165/1126 PM: domains: Fix sleep-in-atomic bug caused by genpd_debug_remove()

In reply to Greg Kroah-Hartman

From: Shawn Guo <[email protected]>

commit f6bfe8b5b2c2a5ac8bd2fc7bca3706e6c3fc26d8 upstream.

When a genpd with GENPD_FLAG_IRQ_SAFE gets removed, the following sleep-in-atomic bug will be seen, as genpd_debug_remove() will be called with a spinlock being held.

[    0.029183] BUG: sleeping function called from invalid context at kernel/locking/rwsem.c:1460
[    0.029204] in_atomic(): 1, irqs_disabled(): 128, non_block: 0, pid: 1, name: swapper/0
[    0.029219] preempt_count: 1, expected: 0
[    0.029230] CPU: 1 PID: 1 Comm: swapper/0 Not tainted 5.17.0-rc4+ #489
[    0.029245] Hardware name: Thundercomm TurboX CM2290 (DT)
[    0.029256] Call trace:
[    0.029265]  dump_backtrace.part.0+0xbc/0xd0
[    0.029285]  show_stack+0x3c/0xa0
[    0.029298]  dump_stack_lvl+0x7c/0xa0
[    0.029311]  dump_stack+0x18/0x34
[    0.029323]  __might_resched+0x10c/0x13c
[    0.029338]  __might_sleep+0x4c/0x80
[    0.029351]  down_read+0x24/0xd0
[    0.029363]  lookup_one_len_unlocked+0x9c/0xcc
[    0.029379]  lookup_positive_unlocked+0x10/0x50
[    0.029392]  debugfs_lookup+0x68/0xac
[    0.029406]  genpd_remove.part.0+0x12c/0x1b4
[    0.029419]  of_genpd_remove_last+0xa8/0xd4
[    0.029434]  psci_cpuidle_domain_probe+0x174/0x53c
[    0.029449]  platform_probe+0x68/0xe0
[    0.029462]  really_probe+0x190/0x430
[    0.029473]  __driver_probe_device+0x90/0x18c
[    0.029485]  driver_probe_device+0x40/0xe0
[    0.029497]  __driver_attach+0xf4/0x1d0
[    0.029508]  bus_for_each_dev+0x70/0xd0
[    0.029523]  driver_attach+0x24/0x30
[    0.029534]  bus_add_driver+0x164/0x22c
[    0.029545]  driver_register+0x78/0x130
[    0.029556]  __platform_driver_register+0x28/0x34
[    0.029569]  psci_idle_init_domains+0x1c/0x28
[    0.029583]  do_one_initcall+0x50/0x1b0
[    0.029595]  kernel_init_freeable+0x214/0x280
[    0.029609]  kernel_init+0x2c/0x13c
[    0.029622]  ret_from_fork+0x10/0x20

It doesn't seem necessary to call genpd_debug_remove() with the lock, so move it out from locking to fix the problem.

Fixes 718072ceb211 ("PM: domains: create debugfs nodes when adding power domains")
Cc 5.11+ <[email protected]> # 5.11+
---
 drivers/base/power/domain.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/base/power/domain.c +1 -1
@@ -2058,9 +2058,9 @@ static int genpd_remove(struct generic_p
 		kfree(link);
 	}
 
-	genpd_debug_remove(genpd);
 	list_del(&genpd->gpd_list_node);
 	genpd_unlock(genpd);
+	genpd_debug_remove(genpd);
 	cancel_work_sync(&genpd->power_off_work);
 	if (genpd_is_cpu_domain(genpd))
 		free_cpumask_var(genpd->cpus);
author maintainer
#166 lore raw

PATCH 166/1126 DEC: Limit PMAX memory probing to R3k systems

In reply to Greg Kroah-Hartman

From: Maciej W. Rozycki <[email protected]>

commit 244eae91a94c6dab82b3232967d10eeb9dfa21c6 upstream.

Recent tightening of the opcode table in binutils so as to consistently disallow the assembly or disassembly of CP0 instructions not supported by the processor architecture chosen has caused a regression like below:

arch/mips/dec/prom/locore.S: Assembler messages:
arch/mips/dec/prom/locore.S:29: Error: opcode not supported on this processor: r4600 (mips3) `rfe'

in a piece of code used to probe for memory with PMAX DECstation models, which have non-REX firmware. Those computers always have an R2000 CPU and consequently the exception handler used in memory probing uses the RFE instruction, which those processors use.

While adding 64-bit support this code was correctly excluded for 64-bit configurations, however it should have also been excluded for irrelevant 32-bit configurations. Do this now then, and only enable PMAX memory probing for R3k systems.

---
 arch/mips/dec/prom/Makefile      |    2 +-
 arch/mips/include/asm/dec/prom.h |   15 +++++----------
 2 files changed, 6 insertions(+), 11 deletions(-)
arch/mips/dec/prom/Makefile +1 -1
@@ -6,4 +6,4 @@
 
 lib-y			+= init.o memory.o cmdline.o identify.o console.o
 
-lib-$(CONFIG_32BIT)	+= locore.o
+lib-$(CONFIG_CPU_R3000)	+= locore.o
arch/mips/include/asm/dec/prom.h +5 -10
@@ -43,16 +43,11 @@
  */
 #define REX_PROM_MAGIC		0x30464354
 
-#ifdef CONFIG_64BIT
-
-#define prom_is_rex(magic)	1	/* KN04 and KN05 are REX PROMs.  */
-
-#else /* !CONFIG_64BIT */
-
-#define prom_is_rex(magic)	((magic) == REX_PROM_MAGIC)
-
-#endif /* !CONFIG_64BIT */
-
+/* KN04 and KN05 are REX PROMs, so only do the check for R3k systems.  */
+static inline bool prom_is_rex(u32 magic)
+{
+	return !IS_ENABLED(CONFIG_CPU_R3000) || magic == REX_PROM_MAGIC;
+}
 
 /*
  * 3MIN/MAXINE PROM entry points for DS5000/1xx's, DS5000/xx's and
author maintainer
#167 lore raw

PATCH 167/1126 media: gpio-ir-tx: fix transmit with long spaces on Orange Pi PC

In reply to Greg Kroah-Hartman

From: Sean Young <[email protected]>

commit 5ad05ecad4326ddaa26a83ba2233a67be24c1aaa upstream.

Calling udelay for than 1000us does not always yield the correct results.

---
 drivers/media/rc/gpio-ir-tx.c |   28 +++++++++++++++++++++-------
 1 file changed, 21 insertions(+), 7 deletions(-)
drivers/media/rc/gpio-ir-tx.c +21 -7
@@ -48,11 +48,29 @@ static int gpio_ir_tx_set_carrier(struct
 	return 0;
 }
 
+static void delay_until(ktime_t until)
+{
+	/*
+	 * delta should never exceed 0.5 seconds (IR_MAX_DURATION) and on
+	 * m68k ndelay(s64) does not compile; so use s32 rather than s64.
+	 */
+	s32 delta;
+
+	while (true) {
+		delta = ktime_us_delta(until, ktime_get());
+		if (delta <= 0)
+			return;
+
+		/* udelay more than 1ms may not work */
+		delta = min(delta, 1000);
+		udelay(delta);
+	}
+}
+
 static void gpio_ir_tx_unmodulated(struct gpio_ir *gpio_ir, uint *txbuf,
 				   uint count)
 {
 	ktime_t edge;
-	s32 delta;
 	int i;
 
 	local_irq_disable();
@@ -63,9 +81,7 @@ static void gpio_ir_tx_unmodulated(struc
 		gpiod_set_value(gpio_ir->gpio, !(i % 2));
 
 		edge = ktime_add_us(edge, txbuf[i]);
-		delta = ktime_us_delta(edge, ktime_get());
-		if (delta > 0)
-			udelay(delta);
+		delay_until(edge);
 	}
 
 	gpiod_set_value(gpio_ir->gpio, 0);
@@ -97,9 +113,7 @@ static void gpio_ir_tx_modulated(struct
 		if (i % 2) {
 			// space
 			edge = ktime_add_us(edge, txbuf[i]);
-			delta = ktime_us_delta(edge, ktime_get());
-			if (delta > 0)
-				udelay(delta);
+			delay_until(edge);
 		} else {
 			// pulse
 			ktime_t last = ktime_add_us(edge, txbuf[i]);
author maintainer
#168 lore raw

PATCH 168/1126 media: omap3isp: Use struct_group() for memcpy() region

In reply to Greg Kroah-Hartman

From: Kees Cook <[email protected]>

commit d4568fc8525897e683983806f813be1ae9eedaed upstream.

In preparation for FORTIFY_SOURCE performing compile-time and run-time field bounds checking for memcpy(), memmove(), and memset(), avoid intentionally writing across neighboring fields. Wrap the target region in struct_group(). This additionally fixes a theoretical misalignment of the copy (since the size of "buf" changes between 64-bit and 32-bit, but this is likely never built for 64-bit).

FWIW, I think this code is totally broken on 64-bit (which appears to not be a "real" build configuration): it would either always fail (with an uninitialized data->buf_size) or would cause corruption in userspace due to the copy_to_user() in the call path against an uninitialized data->buf value:

omap3isp_stat_request_statistics_time32(...)

    struct omap3isp_stat_data data64;
    ...
    omap3isp_stat_request_statistics(stat, &data64);

int omap3isp_stat_request_statistics(struct ispstat *stat,

                                     struct omap3isp_stat_data *data)
    ...
    buf = isp_stat_buf_get(stat, data);

static struct ispstat_buffer *isp_stat_buf_get(struct ispstat *stat,
struct omap3isp_stat_data *data)
...

    if (buf->buf_size > data->buf_size) {
            ...
            return ERR_PTR(-EINVAL);
    }
    ...
    rval = copy_to_user(data->buf,
                        buf->virt_addr,
                        buf->buf_size);

Regardless, additionally initialize data64 to be zero-filled to avoid undefined behavior.

Cc Arnd Bergmann <[email protected]>
Fixes 378e3f81cb56 ("media: omap3isp: support 64-bit version of omap3isp_stat_data")
---
 drivers/media/platform/omap3isp/ispstat.c |    5 +++--
 include/uapi/linux/omap3isp.h             |   21 +++++++++++++--------
 2 files changed, 16 insertions(+), 10 deletions(-)
drivers/media/platform/omap3isp/ispstat.c +3 -2
@@ -512,7 +512,7 @@ int omap3isp_stat_request_statistics(str
 int omap3isp_stat_request_statistics_time32(struct ispstat *stat,
 					struct omap3isp_stat_data_time32 *data)
 {
-	struct omap3isp_stat_data data64;
+	struct omap3isp_stat_data data64 = { };
 	int ret;
 
 	ret = omap3isp_stat_request_statistics(stat, &data64);
@@ -521,7 +521,8 @@ int omap3isp_stat_request_statistics_tim
 
 	data->ts.tv_sec = data64.ts.tv_sec;
 	data->ts.tv_usec = data64.ts.tv_usec;
-	memcpy(&data->buf, &data64.buf, sizeof(*data) - sizeof(data->ts));
+	data->buf = (uintptr_t)data64.buf;
+	memcpy(&data->frame, &data64.frame, sizeof(data->frame));
 
 	return 0;
 }
include/uapi/linux/omap3isp.h +13 -8
@@ -162,6 +162,7 @@ struct omap3isp_h3a_aewb_config {
  * struct omap3isp_stat_data - Statistic data sent to or received from user
  * @ts: Timestamp of returned framestats.
  * @buf: Pointer to pass to user.
+ * @buf_size: Size of buffer.
  * @frame_number: Frame number of requested stats.
  * @cur_frame: Current frame number being processed.
  * @config_counter: Number of the configuration associated with the data.
@@ -176,10 +177,12 @@ struct omap3isp_stat_data {
 	struct timeval ts;
 #endif
 	void __user *buf;
-	__u32 buf_size;
-	__u16 frame_number;
-	__u16 cur_frame;
-	__u16 config_counter;
+	__struct_group(/* no tag */, frame, /* no attrs */,
+		__u32 buf_size;
+		__u16 frame_number;
+		__u16 cur_frame;
+		__u16 config_counter;
+	);
 };
 
 #ifdef __KERNEL__
@@ -189,10 +192,12 @@ struct omap3isp_stat_data_time32 {
 		__s32	tv_usec;
 	} ts;
 	__u32 buf;
-	__u32 buf_size;
-	__u16 frame_number;
-	__u16 cur_frame;
-	__u16 config_counter;
+	__struct_group(/* no tag */, frame, /* no attrs */,
+		__u32 buf_size;
+		__u16 frame_number;
+		__u16 cur_frame;
+		__u16 config_counter;
+	);
 };
 #endif
 
author maintainer
#169 lore raw

PATCH 169/1126 media: venus: vdec: fixed possible memory leak issue

In reply to Greg Kroah-Hartman

From: Ameer Hamza <[email protected]>

commit 8403fdd775858a7bf04868d43daea0acbe49ddfc upstream.

The venus_helper_alloc_dpb_bufs() implementation allows an early return on an error path when checking the id from ida_alloc_min() which would not release the earlier buffer allocation.

Move the direct kfree() from the error checking of dma_alloc_attrs() to the common fail path to ensure that allocations are released on all error paths in this function.

Addresses-Coverity: 1494120 ("Resource leak")

Fixes 40d87aafee29 ("media: venus: vdec: decoded picture buffer handling during reconfig sequence")
---
 drivers/media/platform/qcom/venus/helpers.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/media/platform/qcom/venus/helpers.c +1 -1
@@ -189,7 +189,6 @@ int venus_helper_alloc_dpb_bufs(struct v
 		buf->va = dma_alloc_attrs(dev, buf->size, &buf->da, GFP_KERNEL,
 					  buf->attrs);
 		if (!buf->va) {
-			kfree(buf);
 			ret = -ENOMEM;
 			goto fail;
 		}
@@ -209,6 +208,7 @@ int venus_helper_alloc_dpb_bufs(struct v
 	return 0;
 
 fail:
+	kfree(buf);
 	venus_helper_free_dpb_bufs(inst);
 	return ret;
 }
author maintainer
#170 lore raw

PATCH 170/1126 media: venus: hfi_cmds: List HDR10 property as unsupported for v1 and v3

In reply to Greg Kroah-Hartman

From: Stanimir Varbanov <[email protected]>

commit 22beb839f48d841ec75974872863dc253d37c21c upstream.

The HFI_PROPERTY_PARAM_VENC_HDR10_PQ_SEI HFI property is not supported on Venus v1 and v3.

Fixes 9172652d72f8 ("media: venus: venc: Add support for CLL and Mastering display controls")
---
 drivers/media/platform/qcom/venus/hfi_cmds.c |    2 ++
 1 file changed, 2 insertions(+)
drivers/media/platform/qcom/venus/hfi_cmds.c +2 -0
@@ -1054,6 +1054,8 @@ static int pkt_session_set_property_1x(s
 		pkt->shdr.hdr.size += sizeof(u32) + sizeof(*info);
 		break;
 	}
+	case HFI_PROPERTY_PARAM_VENC_HDR10_PQ_SEI:
+		return -ENOTSUPP;
 
 	/* FOLLOWING PROPERTIES ARE NOT IMPLEMENTED IN CORE YET */
 	case HFI_PROPERTY_CONFIG_BUFFER_REQUIREMENTS:
author maintainer
#171 lore raw

PATCH 171/1126 media: venus: venc: Fix h264 8x8 transform control

In reply to Greg Kroah-Hartman

From: Stanimir Varbanov <[email protected]>

commit 61b3317dd424a3488b6754d7ff8301944d9d17d7 upstream.

During encoder driver open controls are initialized via a call to v4l2_ctrl_handler_setup which returns EINVAL error for V4L2_CID_MPEG_VIDEO_H264_8X8_TRANSFORM v4l2 control. The control default value is disabled and because of firmware limitations 8x8 transform cannot be disabled for the supported HIGH and CONSTRAINED_HIGH profiles.

To fix the issue change the control default value to enabled (this is fine because the firmware enables 8x8 transform for high and constrained_high profiles by default). Also, correct the checking of profile ids in s_ctrl from hfi to v4l2 ids.

Fixes bfee75f73c37 ("media: venus: venc: add support for V4L2_CID_MPEG_VIDEO_H264_8X8_TRANSFORM control")
---
 drivers/media/platform/qcom/venus/venc.c       |    4 ++--
 drivers/media/platform/qcom/venus/venc_ctrls.c |    6 +++---
 2 files changed, 5 insertions(+), 5 deletions(-)
drivers/media/platform/qcom/venus/venc.c +2 -2
@@ -662,8 +662,8 @@ static int venc_set_properties(struct ve
 
 		ptype = HFI_PROPERTY_PARAM_VENC_H264_TRANSFORM_8X8;
 		h264_transform.enable_type = 0;
-		if (ctr->profile.h264 == HFI_H264_PROFILE_HIGH ||
-		    ctr->profile.h264 == HFI_H264_PROFILE_CONSTRAINED_HIGH)
+		if (ctr->profile.h264 == V4L2_MPEG_VIDEO_H264_PROFILE_HIGH ||
+		    ctr->profile.h264 == V4L2_MPEG_VIDEO_H264_PROFILE_CONSTRAINED_HIGH)
 			h264_transform.enable_type = ctr->h264_8x8_transform;
 
 		ret = hfi_session_set_property(inst, ptype, &h264_transform);
drivers/media/platform/qcom/venus/venc_ctrls.c +3 -3
@@ -320,8 +320,8 @@ static int venc_op_s_ctrl(struct v4l2_ct
 		ctr->intra_refresh_period = ctrl->val;
 		break;
 	case V4L2_CID_MPEG_VIDEO_H264_8X8_TRANSFORM:
-		if (ctr->profile.h264 != HFI_H264_PROFILE_HIGH &&
-		    ctr->profile.h264 != HFI_H264_PROFILE_CONSTRAINED_HIGH)
+		if (ctr->profile.h264 != V4L2_MPEG_VIDEO_H264_PROFILE_HIGH &&
+		    ctr->profile.h264 != V4L2_MPEG_VIDEO_H264_PROFILE_CONSTRAINED_HIGH)
 			return -EINVAL;
 
 		/*
@@ -457,7 +457,7 @@ int venc_ctrl_init(struct venus_inst *in
 			  V4L2_CID_MPEG_VIDEO_H264_I_FRAME_MIN_QP, 1, 51, 1, 1);
 
 	v4l2_ctrl_new_std(&inst->ctrl_handler, &venc_ctrl_ops,
-			  V4L2_CID_MPEG_VIDEO_H264_8X8_TRANSFORM, 0, 1, 1, 0);
+			  V4L2_CID_MPEG_VIDEO_H264_8X8_TRANSFORM, 0, 1, 1, 1);
 
 	v4l2_ctrl_new_std(&inst->ctrl_handler, &venc_ctrl_ops,
 			  V4L2_CID_MPEG_VIDEO_H264_P_FRAME_MIN_QP, 1, 51, 1, 1);
author maintainer
#172 lore raw

PATCH 172/1126 media: davinci: vpif: fix unbalanced runtime PM get

In reply to Greg Kroah-Hartman

From: Johan Hovold <[email protected]>

commit 4a321de239213300a714fa0353a5f1272d381a44 upstream.

Make sure to balance the runtime PM usage counter on driver unbind.

Fixes 407ccc65bfd2 ("[media] davinci: vpif: add pm_runtime support")
Cc Lad, Prabhakar <[email protected]>
---
 drivers/media/platform/davinci/vpif.c |    1 +
 1 file changed, 1 insertion(+)
drivers/media/platform/davinci/vpif.c +1 -0
@@ -495,6 +495,7 @@ static int vpif_probe(struct platform_de
 
 static int vpif_remove(struct platform_device *pdev)
 {
+	pm_runtime_put(&pdev->dev);
 	pm_runtime_disable(&pdev->dev);
 	return 0;
 }
author maintainer
#173 lore raw

PATCH 173/1126 media: davinci: vpif: fix unbalanced runtime PM enable

In reply to Greg Kroah-Hartman

From: Johan Hovold <[email protected]>

commit d42b3ad105b5d3481f6a56bc789aa2b27aa09325 upstream.

Make sure to disable runtime PM before returning on probe errors.

Fixes 479f7a118105 ("[media] davinci: vpif: adaptions for DT support")
Cc Kevin Hilman <[email protected]>
---
 drivers/media/platform/davinci/vpif.c |   11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)
drivers/media/platform/davinci/vpif.c +9 -2
@@ -428,6 +428,7 @@ static int vpif_probe(struct platform_de
 	static struct resource *res_irq;
 	struct platform_device *pdev_capture, *pdev_display;
 	struct device_node *endpoint = NULL;
+	int ret;
 
 	vpif_base = devm_platform_ioremap_resource(pdev, 0);
 	if (IS_ERR(vpif_base))
@@ -456,8 +457,8 @@ static int vpif_probe(struct platform_de
 	res_irq = platform_get_resource(pdev, IORESOURCE_IRQ, 0);
 	if (!res_irq) {
 		dev_warn(&pdev->dev, "Missing IRQ resource.\n");
-		pm_runtime_put(&pdev->dev);
-		return -EINVAL;
+		ret = -EINVAL;
+		goto err_put_rpm;
 	}
 
 	pdev_capture = devm_kzalloc(&pdev->dev, sizeof(*pdev_capture),
@@ -491,6 +492,12 @@ static int vpif_probe(struct platform_de
 	}
 
 	return 0;
+
+err_put_rpm:
+	pm_runtime_put(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+
+	return ret;
 }
 
 static int vpif_remove(struct platform_device *pdev)
author maintainer
#174 lore raw

PATCH 174/1126 media: davinci: vpif: fix use-after-free on driver unbind

In reply to Greg Kroah-Hartman

From: Johan Hovold <[email protected]>

commit 43acb728bbc40169d2e2425e84a80068270974be upstream.

The driver allocates and registers two platform device structures during probe, but the devices were never deregistered on driver unbind.

This results in a use-after-free on driver unbind as the device structures were allocated using devres and would be freed by driver core when remove() returns.

Fix this by adding the missing deregistration calls to the remove() callback and failing probe on registration errors.

Note that the platform device structures must be freed using a proper release callback to avoid leaking associated resources like device names.

Fixes 479f7a118105 ("[media] davinci: vpif: adaptions for DT support")
Cc Kevin Hilman <[email protected]>
---
 drivers/media/platform/davinci/vpif.c |   97 ++++++++++++++++++++++++----------
 1 file changed, 71 insertions(+), 26 deletions(-)
drivers/media/platform/davinci/vpif.c +71 -26
@@ -41,6 +41,11 @@ MODULE_ALIAS("platform:" VPIF_DRIVER_NAM
 #define VPIF_CH2_MAX_MODES	15
 #define VPIF_CH3_MAX_MODES	2
 
+struct vpif_data {
+	struct platform_device *capture;
+	struct platform_device *display;
+};
+
 DEFINE_SPINLOCK(vpif_lock);
 EXPORT_SYMBOL_GPL(vpif_lock);
 
@@ -423,17 +428,31 @@ int vpif_channel_getfid(u8 channel_id)
 }
 EXPORT_SYMBOL(vpif_channel_getfid);
 
+static void vpif_pdev_release(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+
+	kfree(pdev);
+}
+
 static int vpif_probe(struct platform_device *pdev)
 {
 	static struct resource *res_irq;
 	struct platform_device *pdev_capture, *pdev_display;
 	struct device_node *endpoint = NULL;
+	struct vpif_data *data;
 	int ret;
 
 	vpif_base = devm_platform_ioremap_resource(pdev, 0);
 	if (IS_ERR(vpif_base))
 		return PTR_ERR(vpif_base);
 
+	data = kzalloc(sizeof(*data), GFP_KERNEL);
+	if (!data)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, data);
+
 	pm_runtime_enable(&pdev->dev);
 	pm_runtime_get(&pdev->dev);
 
@@ -461,49 +480,75 @@ static int vpif_probe(struct platform_de
 		goto err_put_rpm;
 	}
 
-	pdev_capture = devm_kzalloc(&pdev->dev, sizeof(*pdev_capture),
-				    GFP_KERNEL);
-	if (pdev_capture) {
-		pdev_capture->name = "vpif_capture";
-		pdev_capture->id = -1;
-		pdev_capture->resource = res_irq;
-		pdev_capture->num_resources = 1;
-		pdev_capture->dev.dma_mask = pdev->dev.dma_mask;
-		pdev_capture->dev.coherent_dma_mask = pdev->dev.coherent_dma_mask;
-		pdev_capture->dev.parent = &pdev->dev;
-		platform_device_register(pdev_capture);
-	} else {
-		dev_warn(&pdev->dev, "Unable to allocate memory for pdev_capture.\n");
+	pdev_capture = kzalloc(sizeof(*pdev_capture), GFP_KERNEL);
+	if (!pdev_capture) {
+		ret = -ENOMEM;
+		goto err_put_rpm;
 	}
 
-	pdev_display = devm_kzalloc(&pdev->dev, sizeof(*pdev_display),
-				    GFP_KERNEL);
-	if (pdev_display) {
-		pdev_display->name = "vpif_display";
-		pdev_display->id = -1;
-		pdev_display->resource = res_irq;
-		pdev_display->num_resources = 1;
-		pdev_display->dev.dma_mask = pdev->dev.dma_mask;
-		pdev_display->dev.coherent_dma_mask = pdev->dev.coherent_dma_mask;
-		pdev_display->dev.parent = &pdev->dev;
-		platform_device_register(pdev_display);
-	} else {
-		dev_warn(&pdev->dev, "Unable to allocate memory for pdev_display.\n");
+	pdev_capture->name = "vpif_capture";
+	pdev_capture->id = -1;
+	pdev_capture->resource = res_irq;
+	pdev_capture->num_resources = 1;
+	pdev_capture->dev.dma_mask = pdev->dev.dma_mask;
+	pdev_capture->dev.coherent_dma_mask = pdev->dev.coherent_dma_mask;
+	pdev_capture->dev.parent = &pdev->dev;
+	pdev_capture->dev.release = vpif_pdev_release;
+
+	ret = platform_device_register(pdev_capture);
+	if (ret)
+		goto err_put_pdev_capture;
+
+	pdev_display = kzalloc(sizeof(*pdev_display), GFP_KERNEL);
+	if (!pdev_display) {
+		ret = -ENOMEM;
+		goto err_put_pdev_capture;
 	}
 
+	pdev_display->name = "vpif_display";
+	pdev_display->id = -1;
+	pdev_display->resource = res_irq;
+	pdev_display->num_resources = 1;
+	pdev_display->dev.dma_mask = pdev->dev.dma_mask;
+	pdev_display->dev.coherent_dma_mask = pdev->dev.coherent_dma_mask;
+	pdev_display->dev.parent = &pdev->dev;
+	pdev_display->dev.release = vpif_pdev_release;
+
+	ret = platform_device_register(pdev_display);
+	if (ret)
+		goto err_put_pdev_display;
+
+	data->capture = pdev_capture;
+	data->display = pdev_display;
+
 	return 0;
 
+err_put_pdev_display:
+	platform_device_put(pdev_display);
+err_put_pdev_capture:
+	platform_device_put(pdev_capture);
 err_put_rpm:
 	pm_runtime_put(&pdev->dev);
 	pm_runtime_disable(&pdev->dev);
+	kfree(data);
 
 	return ret;
 }
 
 static int vpif_remove(struct platform_device *pdev)
 {
+	struct vpif_data *data = platform_get_drvdata(pdev);
+
+	if (data->capture)
+		platform_device_unregister(data->capture);
+	if (data->display)
+		platform_device_unregister(data->display);
+
 	pm_runtime_put(&pdev->dev);
 	pm_runtime_disable(&pdev->dev);
+
+	kfree(data);
+
 	return 0;
 }
 
author maintainer
#175 lore raw

PATCH 175/1126 mips: Always permit to build u-boot images

In reply to Greg Kroah-Hartman

From: Paul Cercueil <[email protected]>

commit 34275ac292ae141ebc1296b72f005f71b25998a7 upstream.

The platforms where the kernel should be loaded above 0x8000.0000 do not support loading u-boot images, that doesn't mean that we shouldn't be able to generate them.

Additionally, since commit 79876cc1d7b8 ("MIPS: new Kconfig option ZBOOT_LOAD_ADDRESS"), the $(zload-y) variable was no longer hardcoded, which made it impossible to use the uzImage.bin target.

Fixes 79876cc1d7b8 ("MIPS: new Kconfig option ZBOOT_LOAD_ADDRESS")
---
 arch/mips/Makefile |    4 ----
 1 file changed, 4 deletions(-)
arch/mips/Makefile +0 -4
@@ -340,14 +340,12 @@ drivers-$(CONFIG_PM)	+= arch/mips/power/
 boot-y			:= vmlinux.bin
 boot-y			+= vmlinux.ecoff
 boot-y			+= vmlinux.srec
-ifeq ($(shell expr $(load-y) \< 0xffffffff80000000 2> /dev/null), 0)
 boot-y			+= uImage
 boot-y			+= uImage.bin
 boot-y			+= uImage.bz2
 boot-y			+= uImage.gz
 boot-y			+= uImage.lzma
 boot-y			+= uImage.lzo
-endif
 boot-y			+= vmlinux.itb
 boot-y			+= vmlinux.gz.itb
 boot-y			+= vmlinux.bz2.itb
@@ -359,9 +357,7 @@ bootz-y			:= vmlinuz
 bootz-y			+= vmlinuz.bin
 bootz-y			+= vmlinuz.ecoff
 bootz-y			+= vmlinuz.srec
-ifeq ($(shell expr $(zload-y) \< 0xffffffff80000000 2> /dev/null), 0)
 bootz-y			+= uzImage.bin
-endif
 bootz-y			+= vmlinuz.itb
 
 #
author maintainer
#176 lore raw

PATCH 176/1126 btrfs: zoned: mark relocation as writing

In reply to Greg Kroah-Hartman

From: Naohiro Aota <[email protected]>

commit ca5e4ea0beaec8bc674121838bf8614c089effb9 upstream.

There is a hung_task issue with running generic/068 on an SMR device. The hang occurs while a process is trying to thaw the filesystem. The process is trying to take sb->s_umount to thaw the FS. The lock is held by fsstress, which calls btrfs_sync_fs() and is waiting for an ordered extent to finish. However, as the FS is frozen, the ordered extents never finish.

Having an ordered extent while the FS is frozen is the root cause of the hang. The ordered extent is initiated from btrfs_relocate_chunk() which is called from btrfs_reclaim_bgs_work().

This commit adds sb_*_write() around btrfs_relocate_chunk() call site. For the usual "btrfs balance" command, we already call it with mnt_want_file() in btrfs_ioctl_balance().

---
 fs/btrfs/block-group.c |    8 +++++++-
 fs/btrfs/volumes.c     |    3 +++
 2 files changed, 10 insertions(+), 1 deletion(-)
fs/btrfs/block-group.c +7 -1
@@ -1522,8 +1522,12 @@ void btrfs_reclaim_bgs_work(struct work_
 	if (!test_bit(BTRFS_FS_OPEN, &fs_info->flags))
 		return;
 
-	if (!btrfs_exclop_start(fs_info, BTRFS_EXCLOP_BALANCE))
+	sb_start_write(fs_info->sb);
+
+	if (!btrfs_exclop_start(fs_info, BTRFS_EXCLOP_BALANCE)) {
+		sb_end_write(fs_info->sb);
 		return;
+	}
 
 	/*
 	 * Long running balances can keep us blocked here for eternity, so
@@ -1531,6 +1535,7 @@ void btrfs_reclaim_bgs_work(struct work_
 	 */
 	if (!mutex_trylock(&fs_info->reclaim_bgs_lock)) {
 		btrfs_exclop_finish(fs_info);
+		sb_end_write(fs_info->sb);
 		return;
 	}
 
@@ -1605,6 +1610,7 @@ next:
 	spin_unlock(&fs_info->unused_bgs_lock);
 	mutex_unlock(&fs_info->reclaim_bgs_lock);
 	btrfs_exclop_finish(fs_info);
+	sb_end_write(fs_info->sb);
 }
 
 void btrfs_reclaim_bgs(struct btrfs_fs_info *fs_info)
fs/btrfs/volumes.c +3 -0
@@ -8299,10 +8299,12 @@ static int relocating_repair_kthread(voi
 	target = cache->start;
 	btrfs_put_block_group(cache);
 
+	sb_start_write(fs_info->sb);
 	if (!btrfs_exclop_start(fs_info, BTRFS_EXCLOP_BALANCE)) {
 		btrfs_info(fs_info,
 			   "zoned: skip relocating block group %llu to repair: EBUSY",
 			   target);
+		sb_end_write(fs_info->sb);
 		return -EBUSY;
 	}
 
@@ -8330,6 +8332,7 @@ out:
 		btrfs_put_block_group(cache);
 	mutex_unlock(&fs_info->reclaim_bgs_lock);
 	btrfs_exclop_finish(fs_info);
+	sb_end_write(fs_info->sb);
 
 	return ret;
 }
author maintainer
#177 lore raw

PATCH 177/1126 btrfs: extend locking to all space_info members accesses

In reply to Greg Kroah-Hartman

From: Niels Dossche <[email protected]>

commit 06bae876634ebf837ba70ea3de532b288326103d upstream.

bytes_pinned is always accessed under space_info->lock, except in btrfs_preempt_reclaim_metadata_space, however the other members are accessed under that lock. The reserved member of the rsv's are also partially accessed under a lock and partially not. Move all these accesses into the same lock to ensure consistency.

This could potentially race and lead to a flush instead of a commit but it's not a big problem as it's only for preemptive flush.

---
 fs/btrfs/space-info.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
fs/btrfs/space-info.c +2 -1
@@ -1061,7 +1061,6 @@ static void btrfs_preempt_reclaim_metada
 			trans_rsv->reserved;
 		if (block_rsv_size < space_info->bytes_may_use)
 			delalloc_size = space_info->bytes_may_use - block_rsv_size;
-		spin_unlock(&space_info->lock);
 
 		/*
 		 * We don't want to include the global_rsv in our calculation,
@@ -1092,6 +1091,8 @@ static void btrfs_preempt_reclaim_metada
 			flush = FLUSH_DELAYED_REFS_NR;
 		}
 
+		spin_unlock(&space_info->lock);
+
 		/*
 		 * We don't want to reclaim everything, just a portion, so scale
 		 * down the to_reclaim by 1/4.  If it takes us down to 0,
author maintainer
#178 lore raw

PATCH 178/1126 btrfs: verify the tranisd of the to-be-written dirty extent buffer

In reply to Greg Kroah-Hartman

From: Qu Wenruo <[email protected]>

commit 3777369ff1518b579560611a0d0c33f930154f64 upstream.

[BUG]
There is a bug report that a bitflip in the transid part of an extent buffer makes btrfs to reject certain tree blocks:

BTRFS error (device dm-0): parent transid verify failed on 1382301696 wanted 262166 found 22

[CAUSE]
Note the failed transid check, hex(262166) = 0x40016, while hex(22) = 0x16.

It's an obvious bitflip.

Furthermore, the reporter also confirmed the bitflip is from the hardware, so it's a real hardware caused bitflip, and such problem can not be detected by the existing tree-checker framework.

As tree-checker can only verify the content inside one tree block, while generation of a tree block can only be verified against its parent.

So such problem remain undetected.

[FIX]
Although tree-checker can not verify it at write-time, we still have a quick (but not the most accurate) way to catch such obvious corruption.

Function csum_one_extent_buffer() is called before we submit metadata write.

Thus it means, all the extent buffer passed in should be dirty tree blocks, and should be newer than last committed transaction.

Using that we can catch the above bitflip.

Although it's not a perfect solution, as if the corrupted generation is higher than the correct value, we have no way to catch it at all.

---
 fs/btrfs/disk-io.c |   26 ++++++++++++++++++++------
 1 file changed, 20 insertions(+), 6 deletions(-)
fs/btrfs/disk-io.c +20 -6
@@ -441,17 +441,31 @@ static int csum_one_extent_buffer(struct
 	else
 		ret = btrfs_check_leaf_full(eb);
 
-	if (ret < 0) {
-		btrfs_print_tree(eb, 0);
+	if (ret < 0)
+		goto error;
+
+	/*
+	 * Also check the generation, the eb reached here must be newer than
+	 * last committed. Or something seriously wrong happened.
+	 */
+	if (unlikely(btrfs_header_generation(eb) <= fs_info->last_trans_committed)) {
+		ret = -EUCLEAN;
 		btrfs_err(fs_info,
-			"block=%llu write time tree block corruption detected",
-			eb->start);
-		WARN_ON(IS_ENABLED(CONFIG_BTRFS_DEBUG));
-		return ret;
+			"block=%llu bad generation, have %llu expect > %llu",
+			  eb->start, btrfs_header_generation(eb),
+			  fs_info->last_trans_committed);
+		goto error;
 	}
 	write_extent_buffer(eb, result, 0, fs_info->csum_size);
 
 	return 0;
+
+error:
+	btrfs_print_tree(eb, 0);
+	btrfs_err(fs_info, "block=%llu write time tree block corruption detected",
+		  eb->start);
+	WARN_ON(IS_ENABLED(CONFIG_BTRFS_DEBUG));
+	return ret;
 }
 
 /* Checksum all dirty extent buffers in one bio_vec */
author maintainer
#179 lore raw

PATCH 179/1126 xtensa: define update_mmu_tlb function

In reply to Greg Kroah-Hartman

From: Max Filippov <[email protected]>

commit 1c4664faa38923330d478f046dc743a00c1e2dec upstream.

Before the commit f9ce0be71d1f ("mm: Cleanup faultaround and finish_fault() codepaths") there was a call to update_mmu_cache in alloc_set_pte that used to invalidate TLB entry caching invalid PTE that caused a page fault. That commit removed that call so now invalid TLB entry survives causing repetitive page faults on the CPU that took the initial fault until that TLB entry is occasionally evicted. This issue is spotted by the xtensa TLB sanity checker.

Fix this issue by defining update_mmu_tlb function that flushes TLB entry for the faulting address.

---
 arch/xtensa/include/asm/pgtable.h |    4 ++++
 arch/xtensa/mm/tlb.c              |    6 ++++++
 2 files changed, 10 insertions(+)
arch/xtensa/include/asm/pgtable.h +4 -0
@@ -411,6 +411,10 @@ extern  void update_mmu_cache(struct vm_
 
 typedef pte_t *pte_addr_t;
 
+void update_mmu_tlb(struct vm_area_struct *vma,
+		    unsigned long address, pte_t *ptep);
+#define __HAVE_ARCH_UPDATE_MMU_TLB
+
 #endif /* !defined (__ASSEMBLY__) */
 
 #define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_YOUNG
arch/xtensa/mm/tlb.c +6 -0
@@ -162,6 +162,12 @@ void local_flush_tlb_kernel_range(unsign
 	}
 }
 
+void update_mmu_tlb(struct vm_area_struct *vma,
+		    unsigned long address, pte_t *ptep)
+{
+	local_flush_tlb_page(vma, address);
+}
+
 #ifdef CONFIG_DEBUG_TLB_SANITY
 
 static unsigned get_pte_for_vaddr(unsigned vaddr)
author maintainer
#180 lore raw

PATCH 180/1126 xtensa: fix stop_machine_cpuslocked call in patch_text

In reply to Greg Kroah-Hartman

From: Max Filippov <[email protected]>

commit f406f2d03e07afc199dd8cf501f361dde6be8a69 upstream.

patch_text must invoke patch_text_stop_machine on all online CPUs, but it calls stop_machine_cpuslocked with NULL cpumask. As a result only one CPU runs patch_text_stop_machine potentially leaving stale icache entries on other CPUs. Fix that by calling stop_machine_cpuslocked with cpu_online_mask as the last argument.

Fixes 64711f9a47d4 ("xtensa: implement jump_label support")
---
 arch/xtensa/kernel/jump_label.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/xtensa/kernel/jump_label.c +1 -1
@@ -61,7 +61,7 @@ static void patch_text(unsigned long add
 			.data = data,
 		};
 		stop_machine_cpuslocked(patch_text_stop_machine,
-					&patch, NULL);
+					&patch, cpu_online_mask);
 	} else {
 		unsigned long flags;
 
author maintainer
#181 lore raw

PATCH 181/1126 xtensa: fix xtensa_wsr always writing 0

In reply to Greg Kroah-Hartman

From: Max Filippov <[email protected]>

commit a3d0245c58f962ee99d4440ea0eaf45fb7f5a5cc upstream.

The commit cad6fade6e78 ("xtensa: clean up WSR*/RSR*/get_sr/set_sr") replaced 'WSR' macro in the function xtensa_wsr with 'xtensa_set_sr', but variable 'v' in the xtensa_set_sr body shadowed the argument 'v' passed to it, resulting in wrong value written to debug registers.

Fix that by removing intermediate variable from the xtensa_set_sr macro body.

Fixes cad6fade6e78 ("xtensa: clean up WSR*/RSR*/get_sr/set_sr")
---
 arch/xtensa/include/asm/processor.h |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
arch/xtensa/include/asm/processor.h +2 -2
@@ -246,8 +246,8 @@ extern unsigned long __get_wchan(struct
 
 #define xtensa_set_sr(x, sr) \
 	({ \
-	 unsigned int v = (unsigned int)(x); \
-	 __asm__ __volatile__ ("wsr %0, "__stringify(sr) :: "a"(v)); \
+	 __asm__ __volatile__ ("wsr %0, "__stringify(sr) :: \
+			       "a"((unsigned int)(x))); \
 	 })
 
 #define xtensa_get_sr(sr) \
author maintainer
#182 lore raw

PATCH 182/1126 KVM: s390x: fix SCK locking

In reply to Greg Kroah-Hartman

From: Claudio Imbrenda <[email protected]>

commit c0573ba5c5a2244dc02060b1f374d4593c1d20b7 upstream.

When handling the SCK instruction, the kvm lock is taken, even though the vcpu lock is already being held. The normal locking order is kvm lock first and then vcpu lock. This is can (and in some circumstances does) lead to deadlocks.

The function kvm_s390_set_tod_clock is called both by the SCK handler and by some IOCTLs to set the clock. The IOCTLs will not hold the vcpu lock, so they can safely take the kvm lock. The SCK handler holds the vcpu lock, but will also somehow need to acquire the kvm lock without relinquishing the vcpu lock.

The solution is to factor out the code to set the clock, and provide two wrappers. One is called like the original function and does the locking, the other is called kvm_s390_try_set_tod_clock and uses trylock to try to acquire the kvm lock. This new wrapper is then used in the SCK handler. If locking fails, -EAGAIN is returned, which is eventually propagated to userspace, thus also freeing the vcpu lock and allowing for forward progress.

This is not the most efficient or elegant way to solve this issue, but the SCK instruction is deprecated and its performance is not critical.

The goal of this patch is just to provide a simple but correct way to fix the bug.

---
 arch/s390/kvm/kvm-s390.c |   19 ++++++++++++++++---
 arch/s390/kvm/kvm-s390.h |    4 ++--
 arch/s390/kvm/priv.c     |   15 ++++++++++++++-
 3 files changed, 32 insertions(+), 6 deletions(-)
arch/s390/kvm/kvm-s390.c +16 -3
@@ -3869,14 +3869,12 @@ retry:
 	return 0;
 }
 
-void kvm_s390_set_tod_clock(struct kvm *kvm,
-			    const struct kvm_s390_vm_tod_clock *gtod)
+static void __kvm_s390_set_tod_clock(struct kvm *kvm, const struct kvm_s390_vm_tod_clock *gtod)
 {
 	struct kvm_vcpu *vcpu;
 	union tod_clock clk;
 	unsigned long i;
 
-	mutex_lock(&kvm->lock);
 	preempt_disable();
 
 	store_tod_clock_ext(&clk);
@@ -3897,7 +3895,22 @@ void kvm_s390_set_tod_clock(struct kvm *
 
 	kvm_s390_vcpu_unblock_all(kvm);
 	preempt_enable();
+}
+
+void kvm_s390_set_tod_clock(struct kvm *kvm, const struct kvm_s390_vm_tod_clock *gtod)
+{
+	mutex_lock(&kvm->lock);
+	__kvm_s390_set_tod_clock(kvm, gtod);
+	mutex_unlock(&kvm->lock);
+}
+
+int kvm_s390_try_set_tod_clock(struct kvm *kvm, const struct kvm_s390_vm_tod_clock *gtod)
+{
+	if (!mutex_trylock(&kvm->lock))
+		return 0;
+	__kvm_s390_set_tod_clock(kvm, gtod);
 	mutex_unlock(&kvm->lock);
+	return 1;
 }
 
 /**
arch/s390/kvm/kvm-s390.h +2 -2
@@ -349,8 +349,8 @@ int kvm_s390_handle_sigp(struct kvm_vcpu
 int kvm_s390_handle_sigp_pei(struct kvm_vcpu *vcpu);
 
 /* implemented in kvm-s390.c */
-void kvm_s390_set_tod_clock(struct kvm *kvm,
-			    const struct kvm_s390_vm_tod_clock *gtod);
+void kvm_s390_set_tod_clock(struct kvm *kvm, const struct kvm_s390_vm_tod_clock *gtod);
+int kvm_s390_try_set_tod_clock(struct kvm *kvm, const struct kvm_s390_vm_tod_clock *gtod);
 long kvm_arch_fault_in_page(struct kvm_vcpu *vcpu, gpa_t gpa, int writable);
 int kvm_s390_store_status_unloaded(struct kvm_vcpu *vcpu, unsigned long addr);
 int kvm_s390_vcpu_store_status(struct kvm_vcpu *vcpu, unsigned long addr);
arch/s390/kvm/priv.c +14 -1
@@ -102,7 +102,20 @@ static int handle_set_clock(struct kvm_v
 		return kvm_s390_inject_prog_cond(vcpu, rc);
 
 	VCPU_EVENT(vcpu, 3, "SCK: setting guest TOD to 0x%llx", gtod.tod);
-	kvm_s390_set_tod_clock(vcpu->kvm, &gtod);
+	/*
+	 * To set the TOD clock the kvm lock must be taken, but the vcpu lock
+	 * is already held in handle_set_clock. The usual lock order is the
+	 * opposite.  As SCK is deprecated and should not be used in several
+	 * cases, for example when the multiple epoch facility or TOD clock
+	 * steering facility is installed (see Principles of Operation),  a
+	 * slow path can be used.  If the lock can not be taken via try_lock,
+	 * the instruction will be retried via -EAGAIN at a later point in
+	 * time.
+	 */
+	if (!kvm_s390_try_set_tod_clock(vcpu->kvm, &gtod)) {
+		kvm_s390_retry_instr(vcpu);
+		return -EAGAIN;
+	}
 
 	kvm_s390_set_psw_cc(vcpu, 0);
 	return 0;
author maintainer
#183 lore raw

PATCH 183/1126 drm/syncobj: flatten dma_fence_chains on transfer

In reply to Greg Kroah-Hartman

From: Christian König <[email protected]>

commit 721255b52700b320c4ae2e23d57f7d9ad1db50b9 upstream.

It is illegal to add a dma_fence_chain as timeline point. Flatten out the fences into a dma_fence_array instead.

---
 drivers/gpu/drm/drm_syncobj.c |   61 ++++++++++++++++++++++++++++++++++++++----
 1 file changed, 56 insertions(+), 5 deletions(-)
drivers/gpu/drm/drm_syncobj.c +56 -5
@@ -853,12 +853,57 @@ drm_syncobj_fd_to_handle_ioctl(struct dr
 					&args->handle);
 }
 
+
+/*
+ * Try to flatten a dma_fence_chain into a dma_fence_array so that it can be
+ * added as timeline fence to a chain again.
+ */
+static int drm_syncobj_flatten_chain(struct dma_fence **f)
+{
+	struct dma_fence_chain *chain = to_dma_fence_chain(*f);
+	struct dma_fence *tmp, **fences;
+	struct dma_fence_array *array;
+	unsigned int count;
+
+	if (!chain)
+		return 0;
+
+	count = 0;
+	dma_fence_chain_for_each(tmp, &chain->base)
+		++count;
+
+	fences = kmalloc_array(count, sizeof(*fences), GFP_KERNEL);
+	if (!fences)
+		return -ENOMEM;
+
+	count = 0;
+	dma_fence_chain_for_each(tmp, &chain->base)
+		fences[count++] = dma_fence_get(tmp);
+
+	array = dma_fence_array_create(count, fences,
+				       dma_fence_context_alloc(1),
+				       1, false);
+	if (!array)
+		goto free_fences;
+
+	dma_fence_put(*f);
+	*f = &array->base;
+	return 0;
+
+free_fences:
+	while (count--)
+		dma_fence_put(fences[count]);
+
+	kfree(fences);
+	return -ENOMEM;
+}
+
 static int drm_syncobj_transfer_to_timeline(struct drm_file *file_private,
 					    struct drm_syncobj_transfer *args)
 {
 	struct drm_syncobj *timeline_syncobj = NULL;
-	struct dma_fence *fence;
 	struct dma_fence_chain *chain;
+	struct dma_fence *fence;
 	int ret;
 
 	timeline_syncobj = drm_syncobj_find(file_private, args->dst_handle);
@@ -869,16 +914,22 @@ static int drm_syncobj_transfer_to_timel
 				     args->src_point, args->flags,
 				     &fence);
 	if (ret)
-		goto err;
+		goto err_put_timeline;
+
+	ret = drm_syncobj_flatten_chain(&fence);
+	if (ret)
+		goto err_free_fence;
+
 	chain = dma_fence_chain_alloc();
 	if (!chain) {
 		ret = -ENOMEM;
-		goto err1;
+		goto err_free_fence;
 	}
+
 	drm_syncobj_add_point(timeline_syncobj, chain, fence, args->dst_point);
-err1:
+err_free_fence:
 	dma_fence_put(fence);
-err:
+err_put_timeline:
 	drm_syncobj_put(timeline_syncobj);
 
 	return ret;
author maintainer
#184 lore raw

PATCH 184/1126 drm/nouveau/backlight: Fix LVDS backlight detection on some laptops

In reply to Greg Kroah-Hartman

From: Lyude Paul <[email protected]>

commit 6b0076540faffd47f5a899bf12f3528c4f0e726b upstream.

It seems that some laptops will report having both an eDP and LVDS connector, even though only the LVDS connector is actually hooked up. This can lead to issues with backlight registration if the eDP connector ends up getting registered before the LVDS connector, as the backlight device will then be registered to the eDP connector instead of the LVDS connector.

So, fix this by only registering the backlight on connectors that are reported as being connected.

Fixes 6eca310e8924 ("drm/nouveau/kms/nv50-: Add basic DPCD backlight support for nouveau")

Bugzilla: https://gitlab.freedesktop.org/drm/nouveau/-/issues/137

---
 drivers/gpu/drm/nouveau/nouveau_backlight.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/nouveau/nouveau_backlight.c +2 -1
@@ -294,7 +294,8 @@ nv50_backlight_init(struct nouveau_backl
 	struct nouveau_drm *drm = nouveau_drm(nv_encoder->base.base.dev);
 	struct nvif_object *device = &drm->client.device.object;
 
-	if (!nvif_rd32(device, NV50_PDISP_SOR_PWM_CTL(ffs(nv_encoder->dcb->or) - 1)))
+	if (!nvif_rd32(device, NV50_PDISP_SOR_PWM_CTL(ffs(nv_encoder->dcb->or) - 1)) ||
+	    nv_conn->base.status != connector_status_connected)
 		return -ENODEV;
 
 	if (nv_conn->type == DCB_CONNECTOR_eDP) {
author maintainer
#185 lore raw

PATCH 185/1126 drm/nouveau/backlight: Just set all backlight types as RAW

In reply to Greg Kroah-Hartman

From: Lyude Paul <[email protected]>

commit b21a142fd2055d8276169efcc95b624ff908a341 upstream.

Currently we can get a warning on systems with eDP backlights like so:

nv_backlight: invalid backlight type

  WARNING: CPU: 4 PID: 454 at drivers/video/backlight/backlight.c:420
    backlight_device_register+0x226/0x250

This happens as a result of us not filling out props.type for the eDP backlight, even though we do it for all other backlight types.

Since nothing in our driver uses anything but BACKLIGHT_RAW, let's take the props\.type assignments out of the codepaths for individual backlight types and just set it unconditionally to prevent this from happening again.

Fixes 6eca310e8924 ("drm/nouveau/kms/nv50-: Add basic DPCD backlight support for nouveau")
Cc <[email protected]> # v5.15+
---
 drivers/gpu/drm/nouveau/nouveau_backlight.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/gpu/drm/nouveau/nouveau_backlight.c +1 -2
@@ -101,7 +101,6 @@ nv40_backlight_init(struct nouveau_encod
 	if (!(nvif_rd32(device, NV40_PMC_BACKLIGHT) & NV40_PMC_BACKLIGHT_MASK))
 		return -ENODEV;
 
-	props->type = BACKLIGHT_RAW;
 	props->max_brightness = 31;
 	*ops = &nv40_bl_ops;
 	return 0;
@@ -343,7 +342,6 @@ nv50_backlight_init(struct nouveau_backl
 	else
 		*ops = &nva3_bl_ops;
 
-	props->type = BACKLIGHT_RAW;
 	props->max_brightness = 100;
 
 	return 0;
@@ -411,6 +409,7 @@ nouveau_backlight_init(struct drm_connec
 		goto fail_alloc;
 	}
 
+	props.type = BACKLIGHT_RAW;
 	bl->dev = backlight_device_register(backlight_name, connector->kdev,
 					    nv_encoder, ops, &props);
 	if (IS_ERR(bl->dev)) {
author maintainer
#186 lore raw

PATCH 186/1126 drm/fb-helper: Mark screen buffers in system memory with FBINFO_VIRTFB

In reply to Greg Kroah-Hartman

From: Thomas Zimmermann <[email protected]>

commit cd9f7f7ac5932129fe81b4c7559cfcb226ec7c5c upstream.

Mark screen buffers in system memory with FBINFO_VIRTFB. Otherwise, fbdev deferred I/O marks mmap'ed areas of system memory with VM_IO. (There's an inverse relationship between the two flags.)

For shadow buffers, also set the FBINFO_READS_FAST hint.

v3:
* change FB_ to FBINFO_ in commit description
v2:
* updated commit description (Daniel)
* added Fixes tag

Fixes d536540f304c ("drm/fb-helper: Add generic fbdev emulation .fb_probe function")
Cc <[email protected]> # v4.19+
---
 drivers/gpu/drm/drm_fb_helper.c |    9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/gpu/drm/drm_fb_helper.c +6 -3
@@ -2346,6 +2346,7 @@ static int drm_fb_helper_generic_probe(s
 	fbi->fbops = &drm_fbdev_fb_ops;
 	fbi->screen_size = sizes->surface_height * fb->pitches[0];
 	fbi->fix.smem_len = fbi->screen_size;
+	fbi->flags = FBINFO_DEFAULT;
 
 	drm_fb_helper_fill_info(fbi, fb_helper, sizes);
 
@@ -2353,19 +2354,21 @@ static int drm_fb_helper_generic_probe(s
 		fbi->screen_buffer = vzalloc(fbi->screen_size);
 		if (!fbi->screen_buffer)
 			return -ENOMEM;
+		fbi->flags |= FBINFO_VIRTFB | FBINFO_READS_FAST;
 
 		fbi->fbdefio = &drm_fbdev_defio;
-
 		fb_deferred_io_init(fbi);
 	} else {
 		/* buffer is mapped for HW framebuffer */
 		ret = drm_client_buffer_vmap(fb_helper->buffer, &map);
 		if (ret)
 			return ret;
-		if (map.is_iomem)
+		if (map.is_iomem) {
 			fbi->screen_base = map.vaddr_iomem;
-		else
+		} else {
 			fbi->screen_buffer = map.vaddr;
+			fbi->flags |= FBINFO_VIRTFB;
+		}
 
 		/*
 		 * Shamelessly leak the physical address to user-space. As
author maintainer
#187 lore raw

PATCH 187/1126 brcmfmac: firmware: Allocate space for default boardrev in nvram

In reply to Greg Kroah-Hartman

From: Hector Martin <[email protected]>

commit d19d8e3ba256f81ea4a27209dbbd1f0a00ef1903 upstream.

If boardrev is missing from the NVRAM we add a default one, but this might need more space in the output buffer than was allocated. Ensure we have enough padding for this in the buffer.

---
 drivers/net/wireless/broadcom/brcm80211/brcmfmac/firmware.c |    2 ++
 1 file changed, 2 insertions(+)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/firmware.c +2 -0
@@ -207,6 +207,8 @@ static int brcmf_init_nvram_parser(struc
 		size = BRCMF_FW_MAX_NVRAM_SIZE;
 	else
 		size = data_len;
+	/* Add space for properties we may add */
+	size += strlen(BRCMF_FW_DEFAULT_BOARDREV) + 1;
 	/* Alloc for extra 0 byte + roundup by 4 + length field */
 	size += 1 + 3 + sizeof(u32);
 	nvp->nvram = kzalloc(size, GFP_KERNEL);
author maintainer
#188 lore raw

PATCH 188/1126 brcmfmac: pcie: Release firmwares in the brcmf_pcie_setup error path

In reply to Greg Kroah-Hartman

From: Hector Martin <[email protected]>

commit 5e90f0f3ead014867dade7a22f93958119f5efab upstream.

This avoids leaking memory if brcmf_chip_get_raminfo fails. Note that the CLM blob is released in the device remove path.

---
 drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c |    2 ++
 1 file changed, 2 insertions(+)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c +2 -0
@@ -1777,6 +1777,8 @@ static void brcmf_pcie_setup(struct devi
 	ret = brcmf_chip_get_raminfo(devinfo->ci);
 	if (ret) {
 		brcmf_err(bus, "Failed to get RAM info\n");
+		release_firmware(fw);
+		brcmf_fw_nvram_free(nvram);
 		goto fail;
 	}
 
author maintainer
#189 lore raw

PATCH 189/1126 brcmfmac: pcie: Declare missing firmware files in pcie.c

In reply to Greg Kroah-Hartman

From: Hector Martin <[email protected]>

commit 6d766d8cb505ec1fae63da8faef4fc5712c3d794 upstream.

Move one of the declarations from sdio.c to pcie.c, since it makes no sense in the former (SDIO support is optional), and add missing ones.

---
 drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c |    7 +++++++
 drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c |    1 -
 2 files changed, 7 insertions(+), 1 deletion(-)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c +7 -0
@@ -59,6 +59,13 @@ BRCMF_FW_DEF(4366B, "brcmfmac4366b-pcie"
 BRCMF_FW_DEF(4366C, "brcmfmac4366c-pcie");
 BRCMF_FW_DEF(4371, "brcmfmac4371-pcie");
 
+/* firmware config files */
+MODULE_FIRMWARE(BRCMF_FW_DEFAULT_PATH "brcmfmac*-pcie.txt");
+MODULE_FIRMWARE(BRCMF_FW_DEFAULT_PATH "brcmfmac*-pcie.*.txt");
+
+/* per-board firmware binaries */
+MODULE_FIRMWARE(BRCMF_FW_DEFAULT_PATH "brcmfmac*-pcie.*.bin");
+
 static const struct brcmf_firmware_mapping brcmf_pcie_fwnames[] = {
 	BRCMF_FW_ENTRY(BRCM_CC_43602_CHIP_ID, 0xFFFFFFFF, 43602),
 	BRCMF_FW_ENTRY(BRCM_CC_43465_CHIP_ID, 0xFFFFFFF0, 4366C),
drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c +0 -1
@@ -629,7 +629,6 @@ BRCMF_FW_CLM_DEF(43752, "brcmfmac43752-s
 
 /* firmware config files */
 MODULE_FIRMWARE(BRCMF_FW_DEFAULT_PATH "brcmfmac*-sdio.*.txt");
-MODULE_FIRMWARE(BRCMF_FW_DEFAULT_PATH "brcmfmac*-pcie.*.txt");
 
 /* per-board firmware binaries */
 MODULE_FIRMWARE(BRCMF_FW_DEFAULT_PATH "brcmfmac*-sdio.*.bin");
author maintainer
#190 lore raw

PATCH 190/1126 brcmfmac: pcie: Replace brcmf_pcie_copy_mem_todev with memcpy_toio

In reply to Greg Kroah-Hartman

From: Hector Martin <[email protected]>

commit 9466987f246758eb7e9071ae58005253f631271e upstream.

The alignment check was wrong (e.g. & 4 instead of & 3), and the logic was also inefficient if the length was not a multiple of 4, since it would needlessly fall back to copying the entire buffer bytewise.

We already have a perfectly good memcpy_toio function, so just call that instead of rolling our own copy logic here. brcmf_pcie_init_ringbuffers was already using it anyway.

---
 drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c |   48 +---------------
 1 file changed, 4 insertions(+), 44 deletions(-)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c +4 -44
@@ -12,6 +12,7 @@
 #include <linux/interrupt.h>
 #include <linux/bcma/bcma.h>
 #include <linux/sched.h>
+#include <linux/io.h>
 #include <asm/unaligned.h>
 
 #include <soc.h>
@@ -455,47 +456,6 @@ brcmf_pcie_write_ram32(struct brcmf_pcie
 
 
 static void
-brcmf_pcie_copy_mem_todev(struct brcmf_pciedev_info *devinfo, u32 mem_offset,
-			  void *srcaddr, u32 len)
-{
-	void __iomem *address = devinfo->tcm + mem_offset;
-	__le32 *src32;
-	__le16 *src16;
-	u8 *src8;
-
-	if (((ulong)address & 4) || ((ulong)srcaddr & 4) || (len & 4)) {
-		if (((ulong)address & 2) || ((ulong)srcaddr & 2) || (len & 2)) {
-			src8 = (u8 *)srcaddr;
-			while (len) {
-				iowrite8(*src8, address);
-				address++;
-				src8++;
-				len--;
-			}
-		} else {
-			len = len / 2;
-			src16 = (__le16 *)srcaddr;
-			while (len) {
-				iowrite16(le16_to_cpu(*src16), address);
-				address += 2;
-				src16++;
-				len--;
-			}
-		}
-	} else {
-		len = len / 4;
-		src32 = (__le32 *)srcaddr;
-		while (len) {
-			iowrite32(le32_to_cpu(*src32), address);
-			address += 4;
-			src32++;
-			len--;
-		}
-	}
-}
-
-
-static void
 brcmf_pcie_copy_dev_tomem(struct brcmf_pciedev_info *devinfo, u32 mem_offset,
 			  void *dstaddr, u32 len)
 {
@@ -1570,8 +1530,8 @@ static int brcmf_pcie_download_fw_nvram(
 		return err;
 
 	brcmf_dbg(PCIE, "Download FW %s\n", devinfo->fw_name);
-	brcmf_pcie_copy_mem_todev(devinfo, devinfo->ci->rambase,
-				  (void *)fw->data, fw->size);
+	memcpy_toio(devinfo->tcm + devinfo->ci->rambase,
+		    (void *)fw->data, fw->size);
 
 	resetintr = get_unaligned_le32(fw->data);
 	release_firmware(fw);
@@ -1585,7 +1545,7 @@ static int brcmf_pcie_download_fw_nvram(
 		brcmf_dbg(PCIE, "Download NVRAM %s\n", devinfo->nvram_name);
 		address = devinfo->ci->rambase + devinfo->ci->ramsize -
 			  nvram_len;
-		brcmf_pcie_copy_mem_todev(devinfo, address, nvram, nvram_len);
+		memcpy_toio(devinfo->tcm + address, nvram, nvram_len);
 		brcmf_fw_nvram_free(nvram);
 	} else {
 		brcmf_dbg(PCIE, "No matching NVRAM file found %s\n",
author maintainer
#191 lore raw

PATCH 191/1126 brcmfmac: pcie: Fix crashes due to early IRQs

In reply to Greg Kroah-Hartman

From: Hector Martin <[email protected]>

commit b50255c83b914defd61a57fbc81d452334b63f4c upstream.

The driver was enabling IRQs before the message processing was initialized. This could cause IRQs to come in too early and crash the driver. Instead, move the IRQ enable and hostready to a bus preinit function, at which point everything is properly initialized.

---
 drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c |   16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/pcie.c +13 -3
@@ -1315,6 +1315,18 @@ static void brcmf_pcie_down(struct devic
 {
 }
 
+static int brcmf_pcie_preinit(struct device *dev)
+{
+	struct brcmf_bus *bus_if = dev_get_drvdata(dev);
+	struct brcmf_pciedev *buspub = bus_if->bus_priv.pcie;
+
+	brcmf_dbg(PCIE, "Enter\n");
+
+	brcmf_pcie_intr_enable(buspub->devinfo);
+	brcmf_pcie_hostready(buspub->devinfo);
+
+	return 0;
+}
 
 static int brcmf_pcie_tx(struct device *dev, struct sk_buff *skb)
 {
@@ -1423,6 +1435,7 @@ static int brcmf_pcie_reset(struct devic
 }
 
 static const struct brcmf_bus_ops brcmf_pcie_bus_ops = {
+	.preinit = brcmf_pcie_preinit,
 	.txdata = brcmf_pcie_tx,
 	.stop = brcmf_pcie_down,
 	.txctl = brcmf_pcie_tx_ctlpkt,
@@ -1795,9 +1808,6 @@ static void brcmf_pcie_setup(struct devi
 
 	init_waitqueue_head(&devinfo->mbdata_resp_wait);
 
-	brcmf_pcie_intr_enable(devinfo);
-	brcmf_pcie_hostready(devinfo);
-
 	ret = brcmf_attach(&devinfo->pdev->dev);
 	if (ret)
 		goto fail;
author maintainer
#192 lore raw

PATCH 192/1126 drm/i915/opregion: check port number bounds for SWSCI display power state

In reply to Greg Kroah-Hartman

From: Jani Nikula <[email protected]>

commit 24a644ebbfd3b13cda702f98907f9dd123e34bf9 upstream.

The mapping from enum port to whatever port numbering scheme is used by the SWSCI Display Power State Notification is odd, and the memory of it has faded. In any case, the parameter only has space for ports numbered [0..4], and UBSAN reports bit shift beyond it when the platform has port F or more.

Since the SWSCI functionality is supposed to be obsolete for new platforms (i.e. ones that might have port F or more), just bail out early if the mapped and mangled port number is beyond what the Display Power State Notification can support.

---
 drivers/gpu/drm/i915/display/intel_opregion.c |   15 +++++++++++++++
 1 file changed, 15 insertions(+)
drivers/gpu/drm/i915/display/intel_opregion.c +15 -0
@@ -375,6 +375,21 @@ int intel_opregion_notify_encoder(struct
 		return -EINVAL;
 	}
 
+	/*
+	 * The port numbering and mapping here is bizarre. The now-obsolete
+	 * swsci spec supports ports numbered [0..4]. Port E is handled as a
+	 * special case, but port F and beyond are not. The functionality is
+	 * supposed to be obsolete for new platforms. Just bail out if the port
+	 * number is out of bounds after mapping.
+	 */
+	if (port > 4) {
+		drm_dbg_kms(&dev_priv->drm,
+			    "[ENCODER:%d:%s] port %c (index %u) out of bounds for display power state notification\n",
+			    intel_encoder->base.base.id, intel_encoder->base.name,
+			    port_name(intel_encoder->port), port);
+		return -EINVAL;
+	}
+
 	if (!enable)
 		parm |= 4 << 8;
 
author maintainer
#193 lore raw

PATCH 193/1126 drm/i915/gem: add missing boundary check in vm_access

In reply to Greg Kroah-Hartman

From: Mastan Katragadda <[email protected]>

commit 3886a86e7e6cc6ce2ce93c440fecd8f42aed0ce7 upstream.

A missing bounds check in vm_access() can lead to an out-of-bounds read or write in the adjacent memory area, since the len attribute is not validated before the memcpy later in the function, potentially hitting:

[  183.637831] BUG: unable to handle page fault for address: ffffc90000c86000
[  183.637934] #PF: supervisor read access in kernel mode
[  183.637997] #PF: error_code(0x0000) - not-present page
[  183.638059] PGD 100000067 P4D 100000067 PUD 100258067 PMD 106341067 PTE 0
[  183.638144] Oops: 0000 [#2] PREEMPT SMP NOPTI
[  183.638201] CPU: 3 PID: 1790 Comm: poc Tainted: G      D           5.17.0-rc6-ci-drm-11296+ #1
[  183.638298] Hardware name: Intel Corporation CoffeeLake Client Platform/CoffeeLake H DDR4 RVP, BIOS CNLSFWR1.R00.X208.B00.1905301319 05/30/2019
[  183.638430] RIP: 0010:memcpy_erms+0x6/0x10
[  183.640213] RSP: 0018:ffffc90001763d48 EFLAGS: 00010246
[  183.641117] RAX: ffff888109c14000 RBX: ffff888111bece40 RCX: 0000000000000ffc
[  183.642029] RDX: 0000000000001000 RSI: ffffc90000c86000 RDI: ffff888109c14004
[  183.642946] RBP: 0000000000000ffc R08: 800000000000016b R09: 0000000000000000
[  183.643848] R10: ffffc90000c85000 R11: 0000000000000048 R12: 0000000000001000
[  183.644742] R13: ffff888111bed190 R14: ffff888109c14000 R15: 0000000000001000
[  183.645653] FS:  00007fe5ef807540(0000) GS:ffff88845b380000(0000) knlGS:0000000000000000
[  183.646570] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  183.647481] CR2: ffffc90000c86000 CR3: 000000010ff02006 CR4: 00000000003706e0
[  183.648384] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[  183.649271] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[  183.650142] Call Trace:
[  183.650988]  <TASK>
[  183.651793]  vm_access+0x1f0/0x2a0 [i915]
[  183.652726]  __access_remote_vm+0x224/0x380
[  183.653561]  mem_rw.isra.0+0xf9/0x190
[  183.654402]  vfs_read+0x9d/0x1b0
[  183.655238]  ksys_read+0x63/0xe0
[  183.656065]  do_syscall_64+0x38/0xc0
[  183.656882]  entry_SYSCALL_64_after_hwframe+0x44/0xae
[  183.657663] RIP: 0033:0x7fe5ef725142
[  183.659351] RSP: 002b:00007ffe1e81c7e8 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
[  183.660227] RAX: ffffffffffffffda RBX: 0000557055dfb780 RCX: 00007fe5ef725142
[  183.661104] RDX: 0000000000001000 RSI: 00007ffe1e81d880 RDI: 0000000000000005
[  183.661972] RBP: 00007ffe1e81e890 R08: 0000000000000030 R09: 0000000000000046
[  183.662832] R10: 0000557055dfc2e0 R11: 0000000000000246 R12: 0000557055dfb1c0
[  183.663691] R13: 00007ffe1e81e980 R14: 0000000000000000 R15: 0000000000000000

Changes since v1:
- Updated if condition with range_overflows_t [Chris Wilson]

Fixes 9f909e215fea ("drm/i915: Implement vm_ops->access for gdb access into mmaps")
Cc Chris Wilson <[email protected]>
Cc Jon Bloomfield <[email protected]>
Cc Sudeep Dutt <[email protected]>
Cc <[email protected]> # v5.8+
[mauld: tidy up the commit message and add Cc: stable]
(cherry picked from commit 661412e301e2ca86799aa4f400d1cf0bd38c57c6)
---
 drivers/gpu/drm/i915/gem/i915_gem_mman.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/gem/i915_gem_mman.c +1 -1
@@ -439,7 +439,7 @@ vm_access(struct vm_area_struct *area, u
 		return -EACCES;
 
 	addr -= area->vm_start;
-	if (addr >= obj->base.size)
+	if (range_overflows_t(u64, addr, len, obj->base.size))
 		return -EINVAL;
 
 	i915_gem_ww_ctx_init(&ww, true);
author maintainer
#194 lore raw

PATCH 194/1126 PCI: imx6: Allow to probe when dw_pcie_wait_for_link() fails

In reply to Greg Kroah-Hartman

From: Fabio Estevam <[email protected]>

commit f81f095e87715e198471f4653952fe5e3f824874 upstream.

The intention of commit 886a9c134755 ("PCI: dwc: Move link handling into common code") was to standardize the behavior of link down as explained in its commit log:

"The behavior for a link down was inconsistent as some drivers would fail probe in that case while others succeed. Let's standardize this to succeed as there are usecases where devices (and the link) appear later even without hotplug. For example, a reconfigured FPGA device."

The pci-imx6 still fails to probe when the link is not present, which causes the following warning:

imx6q-pcie 8ffc000.pcie: Phy link never came up
imx6q-pcie: probe of 8ffc000.pcie failed with error -110

------------[ cut here ]------------
WARNING: CPU: 0 PID: 30 at drivers/regulator/core.c:2257 _regulator_put.part.0+0x1b8/0x1dc
Modules linked in:
CPU: 0 PID: 30 Comm: kworker/u2:2 Not tainted 5.15.0-next-20211103 #1
Hardware name: Freescale i.MX6 SoloX (Device Tree)
Workqueue: events_unbound async_run_entry_fn
[<c0111730>] (unwind_backtrace) from [<c010bb74>] (show_stack+0x10/0x14)
[<c010bb74>] (show_stack) from [<c0f90290>] (dump_stack_lvl+0x58/0x70)
[<c0f90290>] (dump_stack_lvl) from [<c012631c>] (__warn+0xd4/0x154)
[<c012631c>] (__warn) from [<c0f87b00>] (warn_slowpath_fmt+0x74/0xa8)
[<c0f87b00>] (warn_slowpath_fmt) from [<c076b4bc>] (_regulator_put.part.0+0x1b8/0x1dc)
[<c076b4bc>] (_regulator_put.part.0) from [<c076b574>] (regulator_put+0x2c/0x3c)
[<c076b574>] (regulator_put) from [<c08c3740>] (release_nodes+0x50/0x178)

Fix this problem by ignoring the dw_pcie_wait_for_link() error like it is done on the other dwc drivers.

Tested on imx6sx-sdb and imx6q-sabresd boards.

---
 drivers/pci/controller/dwc/pci-imx6.c |   10 ++--------
 1 file changed, 2 insertions(+), 8 deletions(-)
drivers/pci/controller/dwc/pci-imx6.c +2 -8
@@ -809,9 +809,7 @@ static int imx6_pcie_start_link(struct d
 	/* Start LTSSM. */
 	imx6_pcie_ltssm_enable(dev);
 
-	ret = dw_pcie_wait_for_link(pci);
-	if (ret)
-		goto err_reset_phy;
+	dw_pcie_wait_for_link(pci);
 
 	if (pci->link_gen == 2) {
 		/* Allow Gen2 mode after the link is up. */
@@ -847,11 +845,7 @@ static int imx6_pcie_start_link(struct d
 		}
 
 		/* Make sure link training is finished as well! */
-		ret = dw_pcie_wait_for_link(pci);
-		if (ret) {
-			dev_err(dev, "Failed to bring link up!\n");
-			goto err_reset_phy;
-		}
+		dw_pcie_wait_for_link(pci);
 	} else {
 		dev_info(dev, "Link: Gen2 disabled\n");
 	}
author maintainer
#195 lore raw

PATCH 195/1126 PCI: pciehp: Clear cmd_busy bit in polling mode

In reply to Greg Kroah-Hartman

From: Liguang Zhang <[email protected]>

commit 92912b175178c7e895f5e5e9f1e30ac30319162b upstream.

Writes to a Downstream Port's Slot Control register are PCIe hotplug "commands." If the Port supports Command Completed events, software must wait for a command to complete before writing to Slot Control again.

pcie_do_write_cmd() sets ctrl->cmd_busy when it writes to Slot Control. If software notification is enabled, i.e., PCI_EXP_SLTCTL_HPIE and PCI_EXP_SLTCTL_CCIE are set, ctrl->cmd_busy is cleared by pciehp_isr().

But when software notification is disabled, as it is when pcie_init() powers off an empty slot, pcie_wait_cmd() uses pcie_poll_cmd() to poll for command completion, and it neglects to clear ctrl->cmd_busy, which leads to spurious timeouts:

pcieport 0000:00:03.0: pciehp: Timeout on hotplug command 0x01c0 (issued 2264 msec ago) pcieport 0000:00:03.0: pciehp: Timeout on hotplug command 0x05c0 (issued 2288 msec ago)

Clear ctrl->cmd_busy in pcie_poll_cmd() when it detects a Command Completed event (PCI_EXP_SLTSTA_CC).

---
 drivers/pci/hotplug/pciehp_hpc.c |    2 ++
 1 file changed, 2 insertions(+)
drivers/pci/hotplug/pciehp_hpc.c +2 -0
@@ -98,6 +98,8 @@ static int pcie_poll_cmd(struct controll
 		if (slot_status & PCI_EXP_SLTSTA_CC) {
 			pcie_capability_write_word(pdev, PCI_EXP_SLTSTA,
 						   PCI_EXP_SLTSTA_CC);
+			ctrl->cmd_busy = 0;
+			smp_mb();
 			return 1;
 		}
 		msleep(10);
author maintainer
#196 lore raw

PATCH 196/1126 PCI: xgene: Revert "PCI: xgene: Use inbound resources for setup"

In reply to Greg Kroah-Hartman

From: Marc Zyngier <[email protected]>

commit 1874b6d7ab1bdc900e8398026350313ac29caddb upstream.

Commit 6dce5aa59e0b ("PCI: xgene: Use inbound resources for setup") killed PCIe on my XGene-1 box (a Mustang board). The machine itself is still alive, but half of its storage (over NVMe) is gone, and the NVMe driver just times out.

Note that this machine boots with a device tree provided by the UEFI firmware (2016 vintage), which could well be non conformant with the spec, hence the breakage.

With the patch reverted, the box boots 5.17-rc8 with flying colors.

Fixes 6dce5aa59e0b ("PCI: xgene: Use inbound resources for setup")
Cc Rob Herring <[email protected]>
Cc Toan Le <[email protected]>
Cc Lorenzo Pieralisi <[email protected]>
Cc Krzysztof Wilczyński <[email protected]>
Cc Bjorn Helgaas <[email protected]>
Cc Stéphane Graber <[email protected]>
Cc dann frazier <[email protected]>
---
 drivers/pci/controller/pci-xgene.c |   33 ++++++++++++++++++++++-----------
 1 file changed, 22 insertions(+), 11 deletions(-)
drivers/pci/controller/pci-xgene.c +22 -11
@@ -479,28 +479,27 @@ static int xgene_pcie_select_ib_reg(u8 *
 }
 
 static void xgene_pcie_setup_ib_reg(struct xgene_pcie *port,
-				    struct resource_entry *entry,
-				    u8 *ib_reg_mask)
+				    struct of_pci_range *range, u8 *ib_reg_mask)
 {
 	void __iomem *cfg_base = port->cfg_base;
 	struct device *dev = port->dev;
 	void __iomem *bar_addr;
 	u32 pim_reg;
-	u64 cpu_addr = entry->res->start;
-	u64 pci_addr = cpu_addr - entry->offset;
-	u64 size = resource_size(entry->res);
+	u64 cpu_addr = range->cpu_addr;
+	u64 pci_addr = range->pci_addr;
+	u64 size = range->size;
 	u64 mask = ~(size - 1) | EN_REG;
 	u32 flags = PCI_BASE_ADDRESS_MEM_TYPE_64;
 	u32 bar_low;
 	int region;
 
-	region = xgene_pcie_select_ib_reg(ib_reg_mask, size);
+	region = xgene_pcie_select_ib_reg(ib_reg_mask, range->size);
 	if (region < 0) {
 		dev_warn(dev, "invalid pcie dma-range config\n");
 		return;
 	}
 
-	if (entry->res->flags & IORESOURCE_PREFETCH)
+	if (range->flags & IORESOURCE_PREFETCH)
 		flags |= PCI_BASE_ADDRESS_MEM_PREFETCH;
 
 	bar_low = pcie_bar_low_val((u32)cpu_addr, flags);
@@ -531,13 +530,25 @@ static void xgene_pcie_setup_ib_reg(stru
 
 static int xgene_pcie_parse_map_dma_ranges(struct xgene_pcie *port)
 {
-	struct pci_host_bridge *bridge = pci_host_bridge_from_priv(port);
-	struct resource_entry *entry;
+	struct device_node *np = port->node;
+	struct of_pci_range range;
+	struct of_pci_range_parser parser;
+	struct device *dev = port->dev;
 	u8 ib_reg_mask = 0;
 
-	resource_list_for_each_entry(entry, &bridge->dma_ranges)
-		xgene_pcie_setup_ib_reg(port, entry, &ib_reg_mask);
+	if (of_pci_dma_range_parser_init(&parser, np)) {
+		dev_err(dev, "missing dma-ranges property\n");
+		return -EINVAL;
+	}
 
+	/* Get the dma-ranges from DT */
+	for_each_of_pci_range(&parser, &range) {
+		u64 end = range.cpu_addr + range.size - 1;
+
+		dev_dbg(dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
+			range.flags, range.cpu_addr, end, range.pci_addr);
+		xgene_pcie_setup_ib_reg(port, &range, &ib_reg_mask);
+	}
 	return 0;
 }
 
author maintainer
#197 lore raw

PATCH 197/1126 PCI: xgene: Revert "PCI: xgene: Fix IB window setup"

In reply to Greg Kroah-Hartman

From: Marc Zyngier <[email protected]>

commit 825da4e9cec68713fbb02dc6f71fe1bf65fe8050 upstream.

Commit c7a75d07827a ("PCI: xgene: Fix IB window setup") tried to fix the damages that 6dce5aa59e0b ("PCI: xgene: Use inbound resources for setup") caused, but actually didn't improve anything for some plarforms (at least Mustang and m400 are still broken).

Given that 6dce5aa59e0b has been reverted, revert this patch as well, restoring the PCIe support on XGene to its pre-5.5, working state.

Fixes c7a75d07827a ("PCI: xgene: Fix IB window setup")
Cc Rob Herring <[email protected]>
Cc Toan Le <[email protected]>
Cc Lorenzo Pieralisi <[email protected]>
Cc Krzysztof Wilczyński <[email protected]>
Cc Bjorn Helgaas <[email protected]>
Cc Stéphane Graber <[email protected]>
Cc dann frazier <[email protected]>
---
 drivers/pci/controller/pci-xgene.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/pci/controller/pci-xgene.c +1 -1
@@ -465,7 +465,7 @@ static int xgene_pcie_select_ib_reg(u8 *
 		return 1;
 	}
 
-	if ((size > SZ_1K) && (size < SZ_4G) && !(*ib_reg_mask & (1 << 0))) {
+	if ((size > SZ_1K) && (size < SZ_1T) && !(*ib_reg_mask & (1 << 0))) {
 		*ib_reg_mask |= (1 << 0);
 		return 0;
 	}
author maintainer
#198 lore raw

PATCH 198/1126 regulator: qcom_smd: fix for_each_child.cocci warnings

In reply to Greg Kroah-Hartman

From: kernel test robot <[email protected]>

[ Upstream commit 6390d42c21efff0b4c10956a38e341f4e84ecd3d ]

drivers/regulator/qcom_smd-regulator.c:1318:1-33: WARNING: Function "for_each_available_child_of_node" should have of_node_put() before return around line 1321.

Semantic patch information:
False positives can be due to function calls within the for_each loop that may encapsulate an of_node_put.

Generated by: scripts/coccinelle/iterators/for_each_child.cocci

---
 drivers/regulator/qcom_smd-regulator.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/regulator/qcom_smd-regulator.c +3 -1
index 9fc666107a06..8490aa8eecb1 100644
@@ -1317,8 +1317,10 @@ static int rpm_reg_probe(struct platform_device *pdev)
 
 	for_each_available_child_of_node(dev->of_node, node) {
 		vreg = devm_kzalloc(&pdev->dev, sizeof(*vreg), GFP_KERNEL);
-		if (!vreg)
+		if (!vreg) {
+			of_node_put(node);
 			return -ENOMEM;
+		}
 
 		ret = rpm_regulator_init_vreg(vreg, dev, node, rpm, vreg_data);
 
Signature · 1 line
2.34.1
author maintainer
#199 lore raw

PATCH 199/1126 selinux: access superblock_security_struct in LSM blob way

In reply to Greg Kroah-Hartman

From: GONG, Ruiqi <[email protected]>

[ Upstream commit 0266c25e7c2821181b610595df42cbca6bc93cb8 ]

LSM blob has been involved for superblock's security struct. So fix the remaining direct access to sb->s_security by using the LSM blob mechanism.

Fixes 08abe46b2cfc ("selinux: fall back to SECURITY_FS_USE_GENFS if no xattr support")
Fixes 69c4a42d72eb ("lsm,selinux: add new hook to compare new mount to an existing mount")
---
 security/selinux/hooks.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
security/selinux/hooks.c +2 -2
index 5b6895e4fc29..a0243bae8423 100644
@@ -479,7 +479,7 @@ static int selinux_is_sblabel_mnt(struct super_block *sb)
 
 static int sb_check_xattr_support(struct super_block *sb)
 {
-	struct superblock_security_struct *sbsec = sb->s_security;
+	struct superblock_security_struct *sbsec = selinux_superblock(sb);
 	struct dentry *root = sb->s_root;
 	struct inode *root_inode = d_backing_inode(root);
 	u32 sid;
@@ -2647,7 +2647,7 @@ static int selinux_sb_eat_lsm_opts(char *options, void **mnt_opts)
 static int selinux_sb_mnt_opts_compat(struct super_block *sb, void *mnt_opts)
 {
 	struct selinux_mnt_opts *opts = mnt_opts;
-	struct superblock_security_struct *sbsec = sb->s_security;
+	struct superblock_security_struct *sbsec = selinux_superblock(sb);
 	u32 sid;
 	int rc;
 
Signature · 1 line
2.34.1
author maintainer
#200 lore raw

PATCH 200/1126 selinux: check return value of sel_make_avc_files

In reply to Greg Kroah-Hartman

From: Christian Göttsche <[email protected]>

[ Upstream commit bcb62828e3e8c813b6613db6eb7fd9657db248fc ]

sel_make_avc_files() might fail and return a negative errno value on memory allocation failures. Re-add the check of the return value, dropped in 66f8e2f03c02 ("selinux: sidtab reverse lookup hash table").

Reported by clang-analyzer:

security/selinux/selinuxfs.c:2129:2: warning: Value stored to
'ret' is never read [deadcode.DeadStores]
ret = sel_make_avc_files(dentry);
^ ~~~~~~~~~~~~~~~~~~~~~~~~~~

Fixes 66f8e2f03c02 ("selinux: sidtab reverse lookup hash table")
[PM: description line wrapping, added proper commit ref]
---
 security/selinux/selinuxfs.c | 2 ++
 1 file changed, 2 insertions(+)
security/selinux/selinuxfs.c +2 -0
index e4cd7cb856f3..f2f6203e0fff 100644
@@ -2127,6 +2127,8 @@ static int sel_fill_super(struct super_block *sb, struct fs_context *fc)
 	}
 
 	ret = sel_make_avc_files(dentry);
+	if (ret)
+		goto err;
 
 	dentry = sel_make_dir(sb->s_root, "ss", &fsi->last_ino);
 	if (IS_ERR(dentry)) {
Signature · 1 line
2.34.1
author maintainer
#201 lore raw

PATCH 201/1126 crypto: ccp - Ensure psp_ret is always initd in __sev_platform_init_locked()

In reply to Greg Kroah-Hartman

From: Peter Gonda <[email protected]>

[ Upstream commit 1e1ec11d3ec3134e05d4710f4dee5f9bd05e828d ]

Initialize psp_ret inside of __sev_platform_init_locked() because there are many failure paths with PSP initialization that do not set __sev_do_cmd_locked().

Fixes e423b9d75e77: ("crypto: ccp - Move SEV_INIT retry for corrupted data")
---
 drivers/crypto/ccp/sev-dev.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/crypto/ccp/sev-dev.c +1 -1
index 8fd774a10edc..6ab93dfd478a 100644
@@ -413,7 +413,7 @@ static int __sev_platform_init_locked(int *error)
 {
 	struct psp_device *psp = psp_master;
 	struct sev_device *sev;
-	int rc, psp_ret;
+	int rc, psp_ret = -1;
 	int (*init_function)(int *error);
 
 	if (!psp || !psp->sev_data)
Signature · 1 line
2.34.1
author maintainer
#202 lore raw

PATCH 202/1126 crypto: qat - fix a signedness bug in get_service_enabled()

In reply to Greg Kroah-Hartman

From: Dan Carpenter <[email protected]>

[ Upstream commit 844318dfd31f7c99f6cebbdba5d6f8392c4c115c ]

The "ret" variable needs to be signed or there is an error message which will not be printed correctly.

Fixes 0cec19c761e5 ("crypto: qat - add support for compression for 4xxx")
---
 drivers/crypto/qat/qat_4xxx/adf_4xxx_hw_data.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/crypto/qat/qat_4xxx/adf_4xxx_hw_data.c +1 -1
index 6d10edc40aca..68d39c833332 100644
@@ -52,7 +52,7 @@ static const char *const dev_cfg_services[] = {
 static int get_service_enabled(struct adf_accel_dev *accel_dev)
 {
 	char services[ADF_CFG_MAX_VAL_LEN_IN_BYTES] = {0};
-	u32 ret;
+	int ret;
 
 	ret = adf_cfg_get_param_value(accel_dev, ADF_GENERAL_SEC,
 				      ADF_SERVICES_ENABLED, services);
Signature · 1 line
2.34.1
author maintainer
#203 lore raw

PATCH 203/1126 hwrng: cavium - HW_RANDOM_CAVIUM should depend on ARCH_THUNDER

In reply to Greg Kroah-Hartman

From: Geert Uytterhoeven <[email protected]>

[ Upstream commit ab7d88549e2f7ae116afd303f32e1950cb790a1d ]

The Cavium ThunderX Random Number Generator is only present on Cavium ThunderX SoCs, and not available as an independent PCIe endpoint. Hence add a dependency on ARCH_THUNDER, to prevent asking the user about this driver when configuring a kernel without Cavium Thunder SoC support.

Fixes cc2f1908c6b8f625 ("hwrng: cavium - Add Cavium HWRNG driver for ThunderX SoC.")
---
 drivers/char/hw_random/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/char/hw_random/Kconfig +1 -1
index 9704963f9d50..a087156a5818 100644
@@ -401,7 +401,7 @@ config HW_RANDOM_MESON
 
 config HW_RANDOM_CAVIUM
 	tristate "Cavium ThunderX Random Number Generator support"
-	depends on HW_RANDOM && PCI && ARM64
+	depends on HW_RANDOM && PCI && ARCH_THUNDER
 	default HW_RANDOM
 	help
 	  This driver provides kernel-side support for the Random Number
Signature · 1 line
2.34.1
author maintainer
#204 lore raw

PATCH 204/1126 crypto: sun8i-ss - really disable hash on A80

In reply to Greg Kroah-Hartman

From: Corentin Labbe <[email protected]>

[ Upstream commit 881fc7fba6c3e7d77d608b9a50b01a89d5e0c61b ]

When adding hashes support to sun8i-ss, I have added them only on A83T. But I forgot that 0 is a valid algorithm ID, so hashes are enabled on A80 but with an incorrect ID.
Anyway, even with correct IDs, hashes do not work on A80 and I cannot find why.
So let's disable all of them on A80.

Fixes d9b45418a917 ("crypto: sun8i-ss - support hash algorithms")
---
 drivers/crypto/allwinner/sun8i-ss/sun8i-ss-core.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/crypto/allwinner/sun8i-ss/sun8i-ss-core.c +2 -0
index 80e89066dbd1..319fe3279a71 100644
@@ -30,6 +30,8 @@
 static const struct ss_variant ss_a80_variant = {
 	.alg_cipher = { SS_ALG_AES, SS_ALG_DES, SS_ALG_3DES,
 	},
+	.alg_hash = { SS_ID_NOTSUPP, SS_ID_NOTSUPP, SS_ID_NOTSUPP, SS_ID_NOTSUPP,
+	},
 	.op_mode = { SS_OP_ECB, SS_OP_CBC,
 	},
 	.ss_clks = {
Signature · 1 line
2.34.1
author maintainer
#205 lore raw

PATCH 205/1126 crypto: kdf - Select hmac in addition to sha256

In reply to Greg Kroah-Hartman

From: Herbert Xu <[email protected]>

[ Upstream commit a88592cc27efd4ed0ceba79016eb4a3ddb90e05e ]

In addition to sha256 we must also enable hmac for the kdf self-test to work.

Fixes 304b4acee2f0 ("crypto: kdf - select SHA-256 required...")
Fixes 026a733e6659 ("crypto: kdf - add SP800-108 counter key...")
---
 crypto/Kconfig | 1 +
 1 file changed, 1 insertion(+)
crypto/Kconfig +1 -0
index 442765219c37..2cca54c59fec 100644
@@ -1847,6 +1847,7 @@ config CRYPTO_JITTERENTROPY
 
 config CRYPTO_KDF800108_CTR
 	tristate
+	select CRYPTO_HMAC
 	select CRYPTO_SHA256
 
 config CRYPTO_USER_API
Signature · 1 line
2.34.1
author maintainer
#206 lore raw

PATCH 206/1126 crypto: qat - fix access to PFVF interrupt registers for GEN4

In reply to Greg Kroah-Hartman

From: Giovanni Cabiddu <[email protected]>

[ Upstream commit 642a7d49c249f04007e68c124a148847471dd476 ]

The logic that detects, enables and disables pfvf interrupts was expecting a single CSR per VF. Instead, the source and mask register are two registers with a bit per VF.
Due to this, the driver is reading and setting reserved CSRs and not masking the correct source of interrupts.

Fix the access to the source and mask register for QAT GEN4 devices by removing the outer loop in adf_gen4_get_vf2pf_sources(), adf_gen4_enable_vf2pf_interrupts() and adf_gen4_disable_vf2pf_interrupts() and changing the helper macros ADF_4XXX_VM2PF_SOU and ADF_4XXX_VM2PF_MSK.

Fixes a9dc0d966605 ("crypto: qat - add PFVF support to the GEN4 host driver")
---
 drivers/crypto/qat/qat_common/adf_gen4_pfvf.c | 42 ++++---------------
 1 file changed, 9 insertions(+), 33 deletions(-)
drivers/crypto/qat/qat_common/adf_gen4_pfvf.c +9 -33
index 8efbedf63bc8..3b3ea849c5e5 100644
@@ -9,15 +9,12 @@
 #include "adf_pfvf_pf_proto.h"
 #include "adf_pfvf_utils.h"
 
-#define ADF_4XXX_MAX_NUM_VFS		16
-
 #define ADF_4XXX_PF2VM_OFFSET(i)	(0x40B010 + ((i) * 0x20))
 #define ADF_4XXX_VM2PF_OFFSET(i)	(0x40B014 + ((i) * 0x20))
 
 /* VF2PF interrupt source registers */
-#define ADF_4XXX_VM2PF_SOU(i)		(0x41A180 + ((i) * 4))
-#define ADF_4XXX_VM2PF_MSK(i)		(0x41A1C0 + ((i) * 4))
-#define ADF_4XXX_VM2PF_INT_EN_MSK	BIT(0)
+#define ADF_4XXX_VM2PF_SOU		0x41A180
+#define ADF_4XXX_VM2PF_MSK		0x41A1C0
 
 #define ADF_PFVF_GEN4_MSGTYPE_SHIFT	2
 #define ADF_PFVF_GEN4_MSGTYPE_MASK	0x3F
@@ -41,51 +38,30 @@ static u32 adf_gen4_pf_get_vf2pf_offset(u32 i)
 
 static u32 adf_gen4_get_vf2pf_sources(void __iomem *pmisc_addr)
 {
-	int i;
 	u32 sou, mask;
-	int num_csrs = ADF_4XXX_MAX_NUM_VFS;
-	u32 vf_mask = 0;
 
-	for (i = 0; i < num_csrs; i++) {
-		sou = ADF_CSR_RD(pmisc_addr, ADF_4XXX_VM2PF_SOU(i));
-		mask = ADF_CSR_RD(pmisc_addr, ADF_4XXX_VM2PF_MSK(i));
-		sou &= ~mask;
-		vf_mask |= sou << i;
-	}
+	sou = ADF_CSR_RD(pmisc_addr, ADF_4XXX_VM2PF_SOU);
+	mask = ADF_CSR_RD(pmisc_addr, ADF_4XXX_VM2PF_MSK);
 
-	return vf_mask;
+	return sou &= ~mask;
 }
 
 static void adf_gen4_enable_vf2pf_interrupts(void __iomem *pmisc_addr,
 					     u32 vf_mask)
 {
-	int num_csrs = ADF_4XXX_MAX_NUM_VFS;
-	unsigned long mask = vf_mask;
 	unsigned int val;
-	int i;
-
-	for_each_set_bit(i, &mask, num_csrs) {
-		unsigned int offset = ADF_4XXX_VM2PF_MSK(i);
 
-		val = ADF_CSR_RD(pmisc_addr, offset) & ~ADF_4XXX_VM2PF_INT_EN_MSK;
-		ADF_CSR_WR(pmisc_addr, offset, val);
-	}
+	val = ADF_CSR_RD(pmisc_addr, ADF_4XXX_VM2PF_MSK) & ~vf_mask;
+	ADF_CSR_WR(pmisc_addr, ADF_4XXX_VM2PF_MSK, val);
 }
 
 static void adf_gen4_disable_vf2pf_interrupts(void __iomem *pmisc_addr,
 					      u32 vf_mask)
 {
-	int num_csrs = ADF_4XXX_MAX_NUM_VFS;
-	unsigned long mask = vf_mask;
 	unsigned int val;
-	int i;
-
-	for_each_set_bit(i, &mask, num_csrs) {
-		unsigned int offset = ADF_4XXX_VM2PF_MSK(i);
 
-		val = ADF_CSR_RD(pmisc_addr, offset) | ADF_4XXX_VM2PF_INT_EN_MSK;
-		ADF_CSR_WR(pmisc_addr, offset, val);
-	}
+	val = ADF_CSR_RD(pmisc_addr, ADF_4XXX_VM2PF_MSK) | vf_mask;
+	ADF_CSR_WR(pmisc_addr, ADF_4XXX_VM2PF_MSK, val);
 }
 
 static int adf_gen4_pfvf_send(struct adf_accel_dev *accel_dev,
Signature · 1 line
2.34.1
author maintainer
#207 lore raw

PATCH 207/1126 crypto: authenc - Fix sleep in atomic context in decrypt_tail

In reply to Greg Kroah-Hartman

From: Herbert Xu <[email protected]>

[ Upstream commit 66eae850333d639fc278d6f915c6fc01499ea893 ]

The function crypto_authenc_decrypt_tail discards its flags argument and always relies on the flags from the original request when starting its sub-request.

This is clearly wrong as it may cause the SLEEPABLE flag to be set when it shouldn't.

Fixes 92d95ba91772 ("crypto: authenc - Convert to new AEAD interface")
---
 crypto/authenc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
crypto/authenc.c +1 -1
index 670bf1a01d00..17f674a7cdff 100644
@@ -253,7 +253,7 @@ static int crypto_authenc_decrypt_tail(struct aead_request *req,
 		dst = scatterwalk_ffwd(areq_ctx->dst, req->dst, req->assoclen);
 
 	skcipher_request_set_tfm(skreq, ctx->enc);
-	skcipher_request_set_callback(skreq, aead_request_flags(req),
+	skcipher_request_set_callback(skreq, flags,
 				      req->base.complete, req->base.data);
 	skcipher_request_set_crypt(skreq, src, dst,
 				   req->cryptlen - authsize, req->iv);
Signature · 1 line
2.34.1
author maintainer
#208 lore raw

PATCH 208/1126 crypto: octeontx2 - select CONFIG_NET_DEVLINK

In reply to Greg Kroah-Hartman

From: Shijith Thotton <[email protected]>

[ Upstream commit 85872d1a6f38d133133784c8027d25d1c5328f4f ]

OcteonTX2 CPT driver will fail to link without devlink support.

aarch64-linux-gnu-ld: otx2_cpt_devlink.o: in function `otx2_cpt_dl_egrp_delete':
otx2_cpt_devlink.c:18: undefined reference to `devlink_priv'
aarch64-linux-gnu-ld: otx2_cpt_devlink.o: in function `otx2_cpt_dl_egrp_create':
otx2_cpt_devlink.c:9: undefined reference to `devlink_priv'
aarch64-linux-gnu-ld: otx2_cpt_devlink.o: in function `otx2_cpt_dl_uc_info':
otx2_cpt_devlink.c:27: undefined reference to `devlink_priv'

Fixes fed8f4d5f946 ("crypto: octeontx2 - parameters for custom engine groups")
---
 drivers/crypto/marvell/Kconfig | 1 +
 1 file changed, 1 insertion(+)
drivers/crypto/marvell/Kconfig +1 -0
index 9125199f1702..a48591af12d0 100644
@@ -47,6 +47,7 @@ config CRYPTO_DEV_OCTEONTX2_CPT
 	select CRYPTO_SKCIPHER
 	select CRYPTO_HASH
 	select CRYPTO_AEAD
+	select NET_DEVLINK
 	help
 		This driver allows you to utilize the Marvell Cryptographic
 		Accelerator Unit(CPT) found in OcteonTX2 series of processors.
Signature · 1 line
2.34.1
author maintainer
#209 lore raw

PATCH 209/1126 crypto: mxs-dcp - Fix scatterlist processing

In reply to Greg Kroah-Hartman

From: Tomas Paukrt <[email protected]>

[ Upstream commit 28e9b6d8199a3f124682b143800c2dacdc3d70dd ]

This patch fixes a bug in scatterlist processing that may cause incorrect AES block encryption/decryption.

Fixes 2e6d793e1bf0 ("crypto: mxs-dcp - Use sg_mapping_iter to copy data")
---
 drivers/crypto/mxs-dcp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/crypto/mxs-dcp.c +1 -1
index d19e5ffb5104..d6f9e2fe863d 100644
@@ -331,7 +331,7 @@ static int mxs_dcp_aes_block_crypt(struct crypto_async_request *arq)
 		memset(key + AES_KEYSIZE_128, 0, AES_KEYSIZE_128);
 	}
 
-	for_each_sg(req->src, src, sg_nents(src), i) {
+	for_each_sg(req->src, src, sg_nents(req->src), i) {
 		src_buf = sg_virt(src);
 		len = sg_dma_len(src);
 		tlen += len;
Signature · 1 line
2.34.1
author maintainer
#210 lore raw

PATCH 210/1126 selinux: Fix selinux_sb_mnt_opts_compat()

In reply to Greg Kroah-Hartman

From: Scott Mayhew <[email protected]>

[ Upstream commit b8b87fd954b4b1bdd2d739c8f50bf685351a1a94 ]

selinux_sb_mnt_opts_compat() is called under the sb_lock spinlock and shouldn't be performing any memory allocations. Fix this by parsing the sids at the same time we're chopping up the security mount options string and then using the pre-parsed sids when doing the comparison.

Fixes cc274ae7763d ("selinux: fix sleeping function called from invalid context")
Fixes 69c4a42d72eb ("lsm,selinux: add new hook to compare new mount to an existing mount")
---
 security/selinux/hooks.c | 75 ++++++++++++++++++++++------------------
 1 file changed, 41 insertions(+), 34 deletions(-)
security/selinux/hooks.c +41 -34
index a0243bae8423..a7306208649a 100644
@@ -342,6 +342,10 @@ static void inode_free_security(struct inode *inode)
 
 struct selinux_mnt_opts {
 	const char *fscontext, *context, *rootcontext, *defcontext;
+	u32 fscontext_sid;
+	u32 context_sid;
+	u32 rootcontext_sid;
+	u32 defcontext_sid;
 };
 
 static void selinux_free_mnt_opts(void *mnt_opts)
@@ -598,15 +602,14 @@ static int bad_option(struct superblock_security_struct *sbsec, char flag,
 	return 0;
 }
 
-static int parse_sid(struct super_block *sb, const char *s, u32 *sid,
-		     gfp_t gfp)
+static int parse_sid(struct super_block *sb, const char *s, u32 *sid)
 {
 	int rc = security_context_str_to_sid(&selinux_state, s,
-					     sid, gfp);
+					     sid, GFP_KERNEL);
 	if (rc)
 		pr_warn("SELinux: security_context_str_to_sid"
 		       "(%s) failed for (dev %s, type %s) errno=%d\n",
-		       s, sb->s_id, sb->s_type->name, rc);
+		       s, sb ? sb->s_id : "?", sb ? sb->s_type->name : "?", rc);
 	return rc;
 }
 
@@ -673,8 +676,7 @@ static int selinux_set_mnt_opts(struct super_block *sb,
 	 */
 	if (opts) {
 		if (opts->fscontext) {
-			rc = parse_sid(sb, opts->fscontext, &fscontext_sid,
-					GFP_KERNEL);
+			rc = parse_sid(sb, opts->fscontext, &fscontext_sid);
 			if (rc)
 				goto out;
 			if (bad_option(sbsec, FSCONTEXT_MNT, sbsec->sid,
@@ -683,8 +685,7 @@ static int selinux_set_mnt_opts(struct super_block *sb,
 			sbsec->flags |= FSCONTEXT_MNT;
 		}
 		if (opts->context) {
-			rc = parse_sid(sb, opts->context, &context_sid,
-					GFP_KERNEL);
+			rc = parse_sid(sb, opts->context, &context_sid);
 			if (rc)
 				goto out;
 			if (bad_option(sbsec, CONTEXT_MNT, sbsec->mntpoint_sid,
@@ -693,8 +694,7 @@ static int selinux_set_mnt_opts(struct super_block *sb,
 			sbsec->flags |= CONTEXT_MNT;
 		}
 		if (opts->rootcontext) {
-			rc = parse_sid(sb, opts->rootcontext, &rootcontext_sid,
-					GFP_KERNEL);
+			rc = parse_sid(sb, opts->rootcontext, &rootcontext_sid);
 			if (rc)
 				goto out;
 			if (bad_option(sbsec, ROOTCONTEXT_MNT, root_isec->sid,
@@ -703,8 +703,7 @@ static int selinux_set_mnt_opts(struct super_block *sb,
 			sbsec->flags |= ROOTCONTEXT_MNT;
 		}
 		if (opts->defcontext) {
-			rc = parse_sid(sb, opts->defcontext, &defcontext_sid,
-					GFP_KERNEL);
+			rc = parse_sid(sb, opts->defcontext, &defcontext_sid);
 			if (rc)
 				goto out;
 			if (bad_option(sbsec, DEFCONTEXT_MNT, sbsec->def_sid,
@@ -996,21 +995,29 @@ static int selinux_add_opt(int token, const char *s, void **mnt_opts)
 		if (opts->context || opts->defcontext)
 			goto err;
 		opts->context = s;
+		if (selinux_initialized(&selinux_state))
+			parse_sid(NULL, s, &opts->context_sid);
 		break;
 	case Opt_fscontext:
 		if (opts->fscontext)
 			goto err;
 		opts->fscontext = s;
+		if (selinux_initialized(&selinux_state))
+			parse_sid(NULL, s, &opts->fscontext_sid);
 		break;
 	case Opt_rootcontext:
 		if (opts->rootcontext)
 			goto err;
 		opts->rootcontext = s;
+		if (selinux_initialized(&selinux_state))
+			parse_sid(NULL, s, &opts->rootcontext_sid);
 		break;
 	case Opt_defcontext:
 		if (opts->context || opts->defcontext)
 			goto err;
 		opts->defcontext = s;
+		if (selinux_initialized(&selinux_state))
+			parse_sid(NULL, s, &opts->defcontext_sid);
 		break;
 	}
 
@@ -2648,8 +2655,6 @@ static int selinux_sb_mnt_opts_compat(struct super_block *sb, void *mnt_opts)
 {
 	struct selinux_mnt_opts *opts = mnt_opts;
 	struct superblock_security_struct *sbsec = selinux_superblock(sb);
-	u32 sid;
-	int rc;
 
 	/*
 	 * Superblock not initialized (i.e. no options) - reject if any
@@ -2666,34 +2671,36 @@ static int selinux_sb_mnt_opts_compat(struct super_block *sb, void *mnt_opts)
 		return (sbsec->flags & SE_MNTMASK) ? 1 : 0;
 
 	if (opts->fscontext) {
-		rc = parse_sid(sb, opts->fscontext, &sid, GFP_NOWAIT);
-		if (rc)
+		if (opts->fscontext_sid == SECSID_NULL)
 			return 1;
-		if (bad_option(sbsec, FSCONTEXT_MNT, sbsec->sid, sid))
+		else if (bad_option(sbsec, FSCONTEXT_MNT, sbsec->sid,
+				       opts->fscontext_sid))
 			return 1;
 	}
 	if (opts->context) {
-		rc = parse_sid(sb, opts->context, &sid, GFP_NOWAIT);
-		if (rc)
+		if (opts->context_sid == SECSID_NULL)
 			return 1;
-		if (bad_option(sbsec, CONTEXT_MNT, sbsec->mntpoint_sid, sid))
+		else if (bad_option(sbsec, CONTEXT_MNT, sbsec->mntpoint_sid,
+				       opts->context_sid))
 			return 1;
 	}
 	if (opts->rootcontext) {
-		struct inode_security_struct *root_isec;
-
-		root_isec = backing_inode_security(sb->s_root);
-		rc = parse_sid(sb, opts->rootcontext, &sid, GFP_NOWAIT);
-		if (rc)
-			return 1;
-		if (bad_option(sbsec, ROOTCONTEXT_MNT, root_isec->sid, sid))
+		if (opts->rootcontext_sid == SECSID_NULL)
 			return 1;
+		else {
+			struct inode_security_struct *root_isec;
+
+			root_isec = backing_inode_security(sb->s_root);
+			if (bad_option(sbsec, ROOTCONTEXT_MNT, root_isec->sid,
+				       opts->rootcontext_sid))
+				return 1;
+		}
 	}
 	if (opts->defcontext) {
-		rc = parse_sid(sb, opts->defcontext, &sid, GFP_NOWAIT);
-		if (rc)
+		if (opts->defcontext_sid == SECSID_NULL)
 			return 1;
-		if (bad_option(sbsec, DEFCONTEXT_MNT, sbsec->def_sid, sid))
+		else if (bad_option(sbsec, DEFCONTEXT_MNT, sbsec->def_sid,
+				       opts->defcontext_sid))
 			return 1;
 	}
 	return 0;
@@ -2713,14 +2720,14 @@ static int selinux_sb_remount(struct super_block *sb, void *mnt_opts)
 		return 0;
 
 	if (opts->fscontext) {
-		rc = parse_sid(sb, opts->fscontext, &sid, GFP_KERNEL);
+		rc = parse_sid(sb, opts->fscontext, &sid);
 		if (rc)
 			return rc;
 		if (bad_option(sbsec, FSCONTEXT_MNT, sbsec->sid, sid))
 			goto out_bad_option;
 	}
 	if (opts->context) {
-		rc = parse_sid(sb, opts->context, &sid, GFP_KERNEL);
+		rc = parse_sid(sb, opts->context, &sid);
 		if (rc)
 			return rc;
 		if (bad_option(sbsec, CONTEXT_MNT, sbsec->mntpoint_sid, sid))
@@ -2729,14 +2736,14 @@ static int selinux_sb_remount(struct super_block *sb, void *mnt_opts)
 	if (opts->rootcontext) {
 		struct inode_security_struct *root_isec;
 		root_isec = backing_inode_security(sb->s_root);
-		rc = parse_sid(sb, opts->rootcontext, &sid, GFP_KERNEL);
+		rc = parse_sid(sb, opts->rootcontext, &sid);
 		if (rc)
 			return rc;
 		if (bad_option(sbsec, ROOTCONTEXT_MNT, root_isec->sid, sid))
 			goto out_bad_option;
 	}
 	if (opts->defcontext) {
-		rc = parse_sid(sb, opts->defcontext, &sid, GFP_KERNEL);
+		rc = parse_sid(sb, opts->defcontext, &sid);
 		if (rc)
 			return rc;
 		if (bad_option(sbsec, DEFCONTEXT_MNT, sbsec->def_sid, sid))
Signature · 1 line
2.34.1
author maintainer
#211 lore raw

PATCH 211/1126 thermal: int340x: Check for NULL after calling kmemdup()

In reply to Greg Kroah-Hartman

From: Jiasheng Jiang <[email protected]>

[ Upstream commit 38b16d6cfe54c820848bcfc999bc5e8a7da1cefb ]

As the potential failure of the allocation, kmemdup() may return NULL.

Then, 'bin_attr_data_vault.private' will be NULL, but 'bin_attr_data_vault.size' is not 0, which is not consistent.

Therefore, it is better to check the return value of kmemdup() to avoid the confusion.

Fixes 0ba13c763aac ("thermal/int340x_thermal: Export GDDV")
[ rjw: Subject and changelog edits ]
---
 drivers/thermal/intel/int340x_thermal/int3400_thermal.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/thermal/intel/int340x_thermal/int3400_thermal.c +5 -0
index 3e253b767b49..a0b599100106 100644
@@ -468,6 +468,11 @@ static void int3400_setup_gddv(struct int3400_thermal_priv *priv)
 	priv->data_vault = kmemdup(obj->package.elements[0].buffer.pointer,
 				   obj->package.elements[0].buffer.length,
 				   GFP_KERNEL);
+	if (!priv->data_vault) {
+		kfree(buffer.pointer);
+		return;
+	}
+
 	bin_attr_data_vault.private = priv->data_vault;
 	bin_attr_data_vault.size = obj->package.elements[0].buffer.length;
 	kfree(buffer.pointer);
Signature · 1 line
2.34.1
author maintainer
#212 lore raw

PATCH 212/1126 crypto: octeontx2 - remove CONFIG_DM_CRYPT check

In reply to Greg Kroah-Hartman

From: Shijith Thotton <[email protected]>

[ Upstream commit 2d841af23ae8f398c85dd1ff2dc24b5ec8ba4569 ]

No issues were found while using the driver with dm-crypt enabled. So CONFIG_DM_CRYPT check in the driver can be removed.

This also fixes the NULL pointer dereference in driver release if CONFIG_DM_CRYPT is enabled.

...

Unable to handle kernel NULL pointer dereference at virtual address 0000000000000008
...
Call trace:
 crypto_unregister_alg+0x68/0xfc
 crypto_unregister_skciphers+0x44/0x60
 otx2_cpt_crypto_exit+0x100/0x1a0
 otx2_cptvf_remove+0xf8/0x200
 pci_device_remove+0x3c/0xd4
 __device_release_driver+0x188/0x234
 device_release_driver+0x2c/0x4c

...

Fixes 6f03f0e8b6c8 ("crypto: octeontx2 - register with linux crypto framework")
---
 .../crypto/marvell/octeontx2/otx2_cptvf_algs.c  | 17 +++++++----------
 1 file changed, 7 insertions(+), 10 deletions(-)
drivers/crypto/marvell/octeontx2/otx2_cptvf_algs.c +7 -10
index 2748a3327e39..f8f8542ce3e4 100644
@@ -1634,16 +1634,13 @@ static inline int cpt_register_algs(void)
 {
 	int i, err = 0;
 
-	if (!IS_ENABLED(CONFIG_DM_CRYPT)) {
-		for (i = 0; i < ARRAY_SIZE(otx2_cpt_skciphers); i++)
-			otx2_cpt_skciphers[i].base.cra_flags &=
-							~CRYPTO_ALG_DEAD;
-
-		err = crypto_register_skciphers(otx2_cpt_skciphers,
-						ARRAY_SIZE(otx2_cpt_skciphers));
-		if (err)
-			return err;
-	}
+	for (i = 0; i < ARRAY_SIZE(otx2_cpt_skciphers); i++)
+		otx2_cpt_skciphers[i].base.cra_flags &= ~CRYPTO_ALG_DEAD;
+
+	err = crypto_register_skciphers(otx2_cpt_skciphers,
+					ARRAY_SIZE(otx2_cpt_skciphers));
+	if (err)
+		return err;
 
 	for (i = 0; i < ARRAY_SIZE(otx2_cpt_aeads); i++)
 		otx2_cpt_aeads[i].base.cra_flags &= ~CRYPTO_ALG_DEAD;
Signature · 1 line
2.34.1
author maintainer
#213 lore raw

PATCH 213/1126 spi: tegra114: Add missing IRQ check in tegra_spi_probe

In reply to Greg Kroah-Hartman

From: Miaoqian Lin <[email protected]>

[ Upstream commit 4f92724d4b92c024e721063f520d66e11ca4b54b ]

This func misses checking for platform_get_irq()'s call and may passes the negative error codes to request_threaded_irq(), which takes unsigned IRQ #, causing it to fail with -EINVAL, overriding an original error code. Stop calling request_threaded_irq() with invalid IRQ #s.

---
 drivers/spi/spi-tegra114.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/spi/spi-tegra114.c +4 -0
index e9de1d958bbd..8f345247a8c3 100644
@@ -1352,6 +1352,10 @@ static int tegra_spi_probe(struct platform_device *pdev)
 	tspi->phys = r->start;
 
 	spi_irq = platform_get_irq(pdev, 0);
+	if (spi_irq < 0) {
+		ret = spi_irq;
+		goto exit_free_master;
+	}
 	tspi->irq = spi_irq;
 
 	tspi->clk = devm_clk_get(&pdev->dev, "spi");
Signature · 1 line
2.34.1
author maintainer
#214 lore raw

PATCH 214/1126 spi: tegra210-quad: Fix missin IRQ check in tegra_qspi_probe

In reply to Greg Kroah-Hartman

From: Miaoqian Lin <[email protected]>

[ Upstream commit 47c3e06ed95aa9b74932dbc6b23b544f644faf84 ]

This func misses checking for platform_get_irq()'s call and may passes the negative error codes to request_threaded_irq(), which takes unsigned IRQ #, causing it to fail with -EINVAL, overriding an original error code. Stop calling request_threaded_irq() with invalid IRQ #s.

Fixes 921fc1838fb0 ("spi: tegra210-quad: Add support for Tegra210 QSPI controller")
---
 drivers/spi/spi-tegra210-quad.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/spi/spi-tegra210-quad.c +2 -0
index ce1bdb4767ea..cb00ac2fc7d8 100644
@@ -1240,6 +1240,8 @@ static int tegra_qspi_probe(struct platform_device *pdev)
 
 	tqspi->phys = r->start;
 	qspi_irq = platform_get_irq(pdev, 0);
+	if (qspi_irq < 0)
+		return qspi_irq;
 	tqspi->irq = qspi_irq;
 
 	tqspi->clk = devm_clk_get(&pdev->dev, "qspi");
Signature · 1 line
2.34.1
author maintainer
#215 lore raw

PATCH 215/1126 perf: MARVELL_CN10K_TAD_PMU should depend on ARCH_THUNDER

In reply to Greg Kroah-Hartman

From: Geert Uytterhoeven <[email protected]>

[ Upstream commit e564518b0727c8960942d8b1452703bbabf1a5ec ]

The Marvell CN10K Last-Level cache Tag-and-data Units (LLC-TAD) performance monitor is only present on Marvell CN10K SoCs. Hence add a dependency on ARCH_THUNDER, to prevent asking the user about this driver when configuring a kernel without Cavium Thunder (incl. Marvell CN10K) SoC support.

---
 drivers/perf/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/perf/Kconfig +1 -1
index e1a0c44bc686..7d6ffdf44a41 100644
@@ -141,7 +141,7 @@ config ARM_DMC620_PMU
 
 config MARVELL_CN10K_TAD_PMU
 	tristate "Marvell CN10K LLC-TAD PMU"
-	depends on ARM64 || (COMPILE_TEST && 64BIT)
+	depends on ARCH_THUNDER || (COMPILE_TEST && 64BIT)
 	help
 	  Provides support for Last-Level cache Tag-and-data Units (LLC-TAD)
 	  performance monitors on CN10K family silicons.
Signature · 1 line
2.34.1
author maintainer
#216 lore raw

PATCH 216/1126 selftests/sgx: Fix NULL-pointer-dereference upon early test failure

In reply to Greg Kroah-Hartman

From: Reinette Chatre <[email protected]>

[ Upstream commit 2d03861e0d1d1ee81efc59338101cdd86a7474f6 ]

== Background ==

The SGX selftests track parts of the enclave binaries in an array:
encl->segment_tbl[]. That array is dynamically allocated early (but not first) in the test's lifetime. The array is referenced at the end of the test in encl_delete().

== Problem ==

encl->segment_tbl[] can be NULL if the test fails before its allocation. That leads to a NULL-pointer-dereference in encl_delete(). This is triggered during early failures of the selftest like if the enclave binary ("test_encl.elf") is deleted.

== Solution ==

Ensure encl->segment_tbl[] is valid before attempting to access its members. The offset with which it is accessed, encl->nr_segments, is initialized before encl->segment_tbl[] and thus considered valid to use after the encl->segment_tbl[] check succeeds.

---
 tools/testing/selftests/sgx/load.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
tools/testing/selftests/sgx/load.c +5 -4
index 9d4322c946e2..006b464c8fc9 100644
@@ -21,7 +21,7 @@
 
 void encl_delete(struct encl *encl)
 {
-	struct encl_segment *heap_seg = &encl->segment_tbl[encl->nr_segments - 1];
+	struct encl_segment *heap_seg;
 
 	if (encl->encl_base)
 		munmap((void *)encl->encl_base, encl->encl_size);
@@ -32,10 +32,11 @@ void encl_delete(struct encl *encl)
 	if (encl->fd)
 		close(encl->fd);
 
-	munmap(heap_seg->src, heap_seg->size);
-
-	if (encl->segment_tbl)
+	if (encl->segment_tbl) {
+		heap_seg = &encl->segment_tbl[encl->nr_segments - 1];
+		munmap(heap_seg->src, heap_seg->size);
 		free(encl->segment_tbl);
+	}
 
 	memset(encl, 0, sizeof(*encl));
 }
Signature · 1 line
2.34.1
author maintainer
#217 lore raw

PATCH 217/1126 selftests/sgx: Do not attempt enclave build without valid enclave

In reply to Greg Kroah-Hartman

From: Reinette Chatre <[email protected]>

[ Upstream commit fff36bcbfde1126f6b81cb8ee12a58aada17ca29 ]

It is not possible to build an enclave if it was not possible to load the binary from which it should be constructed. Do not attempt to make further progress but instead return with failure. A "return false" from setup_test_encl() is expected to trip an ASSERT_TRUE() and abort the rest of the test.

---
 tools/testing/selftests/sgx/main.c | 1 +
 1 file changed, 1 insertion(+)
tools/testing/selftests/sgx/main.c +1 -0
index 370c4995f7c4..a7cd2c3e6f7e 100644
@@ -147,6 +147,7 @@ static bool setup_test_encl(unsigned long heap_size, struct encl *encl,
 	if (!encl_load("test_encl.elf", encl, heap_size)) {
 		encl_delete(encl);
 		TH_LOG("Failed to load the test enclave.\n");
+		return false;
 	}
 
 	if (!encl_measure(encl))
Signature · 1 line
2.34.1
author maintainer
#218 lore raw

PATCH 218/1126 selftests/sgx: Ensure enclave data available during debug print

In reply to Greg Kroah-Hartman

From: Reinette Chatre <[email protected]>

[ Upstream commit 2db703fc3b15e7ef68c82eca613a3c00d43d70af ]

In support of debugging the SGX tests print details from the enclave and its memory mappings if any failure is encountered during enclave loading.

When a failure is encountered no data is printed because the printing of the data is preceded by cleanup of the data.

Move the data cleanup after the data print.

---
 tools/testing/selftests/sgx/main.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
tools/testing/selftests/sgx/main.c +2 -2
index a7cd2c3e6f7e..b0bd95a4730d 100644
@@ -186,8 +186,6 @@ static bool setup_test_encl(unsigned long heap_size, struct encl *encl,
 	return true;
 
 err:
-	encl_delete(encl);
-
 	for (i = 0; i < encl->nr_segments; i++) {
 		seg = &encl->segment_tbl[i];
 
@@ -208,6 +206,8 @@ static bool setup_test_encl(unsigned long heap_size, struct encl *encl,
 
 	TH_LOG("Failed to initialize the test enclave.\n");
 
+	encl_delete(encl);
+
 	return false;
 }
 
Signature · 1 line
2.34.1
author maintainer
#219 lore raw

PATCH 219/1126 stack: Constrain and fix stack offset randomization with Clang builds

In reply to Greg Kroah-Hartman

From: Marco Elver <[email protected]>

[ Upstream commit efa90c11f62e6b7252fb75efe2787056872a627c ]

All supported versions of Clang perform auto-init of __builtin_alloca() when stack auto-init is on (CONFIG_INIT_STACK_ALL_{ZERO,PATTERN}).

add_random_kstack_offset() uses __builtin_alloca() to add a stack offset. This means, when CONFIG_INIT_STACK_ALL_{ZERO,PATTERN} is enabled, add_random_kstack_offset() will auto-init that unused portion of the stack used to add an offset.

There are several problems with this:

1. These offsets can be as large as 1023 bytes. Performing memset() on them isn't exactly cheap, and this is done on every syscall entry.

2. Architectures adding add_random_kstack_offset() to syscall entry implemented in C require them to be 'noinstr' (e.g. see x86 and s390). The potential problem here is that a call to memset may occur, which is not noinstr.

A x86_64 defconfig kernel with Clang 11 and CONFIG_VMLINUX_VALIDATION shows:

 | vmlinux.o: warning: objtool: do_syscall_64()+0x9d: call to memset() leaves .noinstr.text section
 | vmlinux.o: warning: objtool: do_int80_syscall_32()+0xab: call to memset() leaves .noinstr.text section
 | vmlinux.o: warning: objtool: __do_fast_syscall_32()+0xe2: call to memset() leaves .noinstr.text section
 | vmlinux.o: warning: objtool: fixup_bad_iret()+0x2f: call to memset() leaves .noinstr.text section

Clang 14 (unreleased) will introduce a way to skip alloca initialization via __builtin_alloca_uninitialized() (https://reviews.llvm.org/D115440).

Constrain RANDOMIZE_KSTACK_OFFSET to only be enabled if no stack auto-init is enabled, the compiler is GCC, or Clang is version 14+. Use __builtin_alloca_uninitialized() if the compiler provides it, as is done by Clang 14.

---
 arch/Kconfig                     |  1 +
 include/linux/randomize_kstack.h | 16 ++++++++++++++--
 2 files changed, 15 insertions(+), 2 deletions(-)
arch/Kconfig +1 -0
index 678a80713b21..5e88237f84d2 100644
@@ -1162,6 +1162,7 @@ config HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
 config RANDOMIZE_KSTACK_OFFSET_DEFAULT
 	bool "Randomize kernel stack offset on syscall entry"
 	depends on HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
+	depends on INIT_STACK_NONE || !CC_IS_CLANG || CLANG_VERSION >= 140000
 	help
 	  The kernel stack offset can be randomized (after pt_regs) by
 	  roughly 5 bits of entropy, frustrating memory corruption
include/linux/randomize_kstack.h +14 -2
index bebc911161b6..d373f1bcbf7c 100644
@@ -16,8 +16,20 @@ DECLARE_PER_CPU(u32, kstack_offset);
  * alignment. Also, since this use is being explicitly masked to a max of
  * 10 bits, stack-clash style attacks are unlikely. For more details see
  * "VLAs" in Documentation/process/deprecated.rst
+ *
+ * The normal __builtin_alloca() is initialized with INIT_STACK_ALL (currently
+ * only with Clang and not GCC). Initializing the unused area on each syscall
+ * entry is expensive, and generating an implicit call to memset() may also be
+ * problematic (such as in noinstr functions). Therefore, if the compiler
+ * supports it (which it should if it initializes allocas), always use the
+ * "uninitialized" variant of the builtin.
  */
-void *__builtin_alloca(size_t size);
+#if __has_builtin(__builtin_alloca_uninitialized)
+#define __kstack_alloca __builtin_alloca_uninitialized
+#else
+#define __kstack_alloca __builtin_alloca
+#endif
+
 /*
  * Use, at most, 10 bits of entropy. We explicitly cap this to keep the
  * "VLA" from being unbounded (see above). 10 bits leaves enough room for
@@ -36,7 +48,7 @@ void *__builtin_alloca(size_t size);
 	if (static_branch_maybe(CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT,	\
 				&randomize_kstack_offset)) {		\
 		u32 offset = raw_cpu_read(kstack_offset);		\
-		u8 *ptr = __builtin_alloca(KSTACK_OFFSET_MAX(offset));	\
+		u8 *ptr = __kstack_alloca(KSTACK_OFFSET_MAX(offset));	\
 		/* Keep allocation even after "ptr" loses scope. */	\
 		asm volatile("" :: "r"(ptr) : "memory");		\
 	}								\
Signature · 1 line
2.34.1
author maintainer
#220 lore raw

PATCH 220/1126 arm64/mm: avoid fixmap race condition when create pud mapping

In reply to Greg Kroah-Hartman

From: Jianyong Wu <[email protected]>

[ Upstream commit ee017ee353506fcec58e481673e4331ff198a80e ]

The 'fixmap' is a global resource and is used recursively by create pud mapping(), leading to a potential race condition in the presence of a concurrent call to alloc_init_pud():

kernel_init thread virtio-mem workqueue thread
================== ===========================

alloc_init_pud(...) alloc_init_pud(...)
pudp = pud_set_fixmap_offset(...) pudp = pud_set_fixmap_offset(...)
READ_ONCE(*pudp)
pud_clear_fixmap(...)

                                            READ_ONCE(*pudp) // CRASH!

As kernel may sleep during creating pud mapping, introduce a mutex lock to serialise use of the fixmap entries by alloc_init_pud(). However, there is no need for locking in early boot stage and it doesn't work well with KASLR enabled when early boot. So, enable lock when system_state doesn't equal to "SYSTEM_BOOTING".

Fixes f4710445458c ("arm64: mm: use fixmap when creating page tables")
---
 arch/arm64/mm/mmu.c | 9 +++++++++
 1 file changed, 9 insertions(+)
arch/arm64/mm/mmu.c +9 -0
index 044d2021c20c..37b8230cda6a 100644
@@ -63,6 +63,7 @@ static pmd_t bm_pmd[PTRS_PER_PMD] __page_aligned_bss __maybe_unused;
 static pud_t bm_pud[PTRS_PER_PUD] __page_aligned_bss __maybe_unused;
 
 static DEFINE_SPINLOCK(swapper_pgdir_lock);
+static DEFINE_MUTEX(fixmap_lock);
 
 void set_swapper_pgd(pgd_t *pgdp, pgd_t pgd)
 {
@@ -329,6 +330,12 @@ static void alloc_init_pud(pgd_t *pgdp, unsigned long addr, unsigned long end,
 	}
 	BUG_ON(p4d_bad(p4d));
 
+	/*
+	 * No need for locking during early boot. And it doesn't work as
+	 * expected with KASLR enabled.
+	 */
+	if (system_state != SYSTEM_BOOTING)
+		mutex_lock(&fixmap_lock);
 	pudp = pud_set_fixmap_offset(p4dp, addr);
 	do {
 		pud_t old_pud = READ_ONCE(*pudp);
@@ -359,6 +366,8 @@ static void alloc_init_pud(pgd_t *pgdp, unsigned long addr, unsigned long end,
 	} while (pudp++, addr = next, addr != end);
 
 	pud_clear_fixmap();
+	if (system_state != SYSTEM_BOOTING)
+		mutex_unlock(&fixmap_lock);
 }
 
 static void __create_pgd_mapping(pgd_t *pgdir, phys_addr_t phys,
Signature · 1 line
2.34.1
author maintainer
#221 lore raw

PATCH 221/1126 security: add sctp_assoc_established hook

In reply to Greg Kroah-Hartman

From: Ondrej Mosnacek <[email protected]>

[ Upstream commit 5e50f5d4ff31e95599d695df1f0a4e7d2d6fef99 ]

security_sctp_assoc_established() is added to replace security_inet_conn_established() called in sctp_sf_do_5_1E_ca(), so that asoc can be accessed in security subsystem and save the peer secid to asoc->peer_secid.

Fixes 72e89f50084c ("security: Add support for SCTP security hooks")
---
 Documentation/security/SCTP.rst | 22 ++++++++++------------
 include/linux/lsm_hook_defs.h   |  2 ++
 include/linux/lsm_hooks.h       |  5 +++++
 include/linux/security.h        |  8 ++++++++
 net/sctp/sm_statefuns.c         |  8 +++++---
 security/security.c             |  7 +++++++
 6 files changed, 37 insertions(+), 15 deletions(-)
Documentation/security/SCTP.rst +10 -12
index d5fd6ccc3dcb..406cc68b8808 100644
@@ -15,10 +15,7 @@ For security module support, three SCTP specific hooks have been implemented::
     security_sctp_assoc_request()
     security_sctp_bind_connect()
     security_sctp_sk_clone()
-
-Also the following security hook has been utilised::
-
-    security_inet_conn_established()
+    security_sctp_assoc_established()
 
 The usage of these hooks are described below with the SELinux implementation
 described in the `SCTP SELinux Support`_ chapter.
@@ -122,11 +119,12 @@ calls **sctp_peeloff**\(3).
     @newsk - pointer to new sock structure.
 
 
-security_inet_conn_established()
+security_sctp_assoc_established()
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-Called when a COOKIE ACK is received::
+Called when a COOKIE ACK is received, and the peer secid will be
+saved into ``@asoc->peer_secid`` for client::
 
-    @sk  - pointer to sock structure.
+    @asoc - pointer to sctp association structure.
     @skb - pointer to skbuff of the COOKIE ACK packet.
 
 
@@ -134,7 +132,7 @@ Security Hooks used for Association Establishment
 -------------------------------------------------
 
 The following diagram shows the use of ``security_sctp_bind_connect()``,
-``security_sctp_assoc_request()``, ``security_inet_conn_established()`` when
+``security_sctp_assoc_request()``, ``security_sctp_assoc_established()`` when
 establishing an association.
 ::
 
@@ -172,7 +170,7 @@ establishing an association.
           <------------------------------------------- COOKIE ACK
           |                                               |
     sctp_sf_do_5_1E_ca                                    |
- Call security_inet_conn_established()                    |
+ Call security_sctp_assoc_established()                   |
  to set the peer label.                                   |
           |                                               |
           |                               If SCTP_SOCKET_TCP or peeled off
@@ -198,7 +196,7 @@ hooks with the SELinux specifics expanded below::
     security_sctp_assoc_request()
     security_sctp_bind_connect()
     security_sctp_sk_clone()
-    security_inet_conn_established()
+    security_sctp_assoc_established()
 
 
 security_sctp_assoc_request()
@@ -271,12 +269,12 @@ sockets sid and peer sid to that contained in the ``@asoc sid`` and
     @newsk - pointer to new sock structure.
 
 
-security_inet_conn_established()
+security_sctp_assoc_established()
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 Called when a COOKIE ACK is received where it sets the connection's peer sid
 to that in ``@skb``::
 
-    @sk  - pointer to sock structure.
+    @asoc - pointer to sctp association structure.
     @skb - pointer to skbuff of the COOKIE ACK packet.
 
 
include/linux/lsm_hook_defs.h +2 -0
index 819ec92dc2a8..db924fe379c9 100644
@@ -332,6 +332,8 @@ LSM_HOOK(int, 0, sctp_bind_connect, struct sock *sk, int optname,
 	 struct sockaddr *address, int addrlen)
 LSM_HOOK(void, LSM_RET_VOID, sctp_sk_clone, struct sctp_association *asoc,
 	 struct sock *sk, struct sock *newsk)
+LSM_HOOK(int, 0, sctp_assoc_established, struct sctp_association *asoc,
+	 struct sk_buff *skb)
 #endif /* CONFIG_SECURITY_NETWORK */
 
 #ifdef CONFIG_SECURITY_INFINIBAND
include/linux/lsm_hooks.h +5 -0
index 3bf5c658bc44..419b5febc3ca 100644
@@ -1046,6 +1046,11 @@
  *	@asoc pointer to current sctp association structure.
  *	@sk pointer to current sock structure.
  *	@newsk pointer to new sock structure.
+ * @sctp_assoc_established:
+ *	Passes the @asoc and @chunk->skb of the association COOKIE_ACK packet
+ *	to the security module.
+ *	@asoc pointer to sctp association structure.
+ *	@skb pointer to skbuff of association packet.
  *
  * Security hooks for Infiniband
  *
include/linux/security.h +8 -0
index 6d72772182c8..25b3ef71f495 100644
@@ -1422,6 +1422,8 @@ int security_sctp_bind_connect(struct sock *sk, int optname,
 			       struct sockaddr *address, int addrlen);
 void security_sctp_sk_clone(struct sctp_association *asoc, struct sock *sk,
 			    struct sock *newsk);
+int security_sctp_assoc_established(struct sctp_association *asoc,
+				    struct sk_buff *skb);
 
 #else	/* CONFIG_SECURITY_NETWORK */
 static inline int security_unix_stream_connect(struct sock *sock,
@@ -1641,6 +1643,12 @@ static inline void security_sctp_sk_clone(struct sctp_association *asoc,
 					  struct sock *newsk)
 {
 }
+
+static inline int security_sctp_assoc_established(struct sctp_association *asoc,
+						  struct sk_buff *skb)
+{
+	return 0;
+}
 #endif	/* CONFIG_SECURITY_NETWORK */
 
 #ifdef CONFIG_SECURITY_INFINIBAND
net/sctp/sm_statefuns.c +5 -3
index cc544a97c4af..7f342bc12735 100644
@@ -930,6 +930,11 @@ enum sctp_disposition sctp_sf_do_5_1E_ca(struct net *net,
 	if (!sctp_vtag_verify(chunk, asoc))
 		return sctp_sf_pdiscard(net, ep, asoc, type, arg, commands);
 
+	/* Set peer label for connection. */
+	if (security_sctp_assoc_established((struct sctp_association *)asoc,
+					    chunk->skb))
+		return sctp_sf_pdiscard(net, ep, asoc, type, arg, commands);
+
 	/* Verify that the chunk length for the COOKIE-ACK is OK.
 	 * If we don't do this, any bundled chunks may be junked.
 	 */
@@ -945,9 +950,6 @@ enum sctp_disposition sctp_sf_do_5_1E_ca(struct net *net,
 	 */
 	sctp_add_cmd_sf(commands, SCTP_CMD_INIT_COUNTER_RESET, SCTP_NULL());
 
-	/* Set peer label for connection. */
-	security_inet_conn_established(ep->base.sk, chunk->skb);
-
 	/* RFC 2960 5.1 Normal Establishment of an Association
 	 *
 	 * E) Upon reception of the COOKIE ACK, endpoint "A" will move
security/security.c +7 -0
index 22261d79f333..e9526f005f09 100644
@@ -2391,6 +2391,13 @@ void security_sctp_sk_clone(struct sctp_association *asoc, struct sock *sk,
 }
 EXPORT_SYMBOL(security_sctp_sk_clone);
 
+int security_sctp_assoc_established(struct sctp_association *asoc,
+				    struct sk_buff *skb)
+{
+	return call_int_hook(sctp_assoc_established, 0, asoc, skb);
+}
+EXPORT_SYMBOL(security_sctp_assoc_established);
+
 #endif	/* CONFIG_SECURITY_NETWORK */
 
 #ifdef CONFIG_SECURITY_INFINIBAND
Signature · 1 line
2.34.1
author maintainer
#222 lore raw

PATCH 222/1126 security: implement sctp_assoc_established hook in selinux

In reply to Greg Kroah-Hartman

From: Ondrej Mosnacek <[email protected]>

[ Upstream commit 3eb8eaf2ca3e98d4f6e52bed6148ee8fe3069a3d ]

Do this by extracting the peer labeling per-association logic from selinux_sctp_assoc_request() into a new helper selinux_sctp_process_new_assoc() and use this helper in both selinux_sctp_assoc_request() and selinux_sctp_assoc_established(). This ensures that the peer labeling behavior as documented in Documentation/security/SCTP.rst is applied both on the client and server side:
"""
An SCTP socket will only have one peer label assigned to it. This will be assigned during the establishment of the first association. Any further associations on this socket will have their packet peer label compared to the sockets peer label, and only if they are different will the ``association`` permission be validated. This is validated by checking the socket peer sid against the received packets peer sid to determine whether the association should be allowed or denied.
"""

At the same time, it also ensures that the peer label of the association is set to the correct value, such that if it is peeled off into a new socket, the socket's peer label will then be set to the association's peer label, same as it already works on the server side.

While selinux_inet_conn_established() (which we are replacing by selinux_sctp_assoc_established() for SCTP) only deals with assigning a peer label to the connection (socket), in case of SCTP we need to also copy the (local) socket label to the association, so that selinux_sctp_sk_clone() can then pick it up for the new socket in case of SCTP peeloff.

Careful readers will notice that the selinux_sctp_process_new_assoc() helper also includes the "IPv4 packet received over an IPv6 socket" check, even though it hadn't been in selinux_sctp_assoc_request() before. While such check is not necessary in selinux_inet_conn_request() (because struct request_sock's family field is already set according to the skb's family), here it is needed, as we don't have request_sock and we take the initial family from the socket. In selinux_sctp_assoc_established() it is similarly needed as well (and also selinux_inet_conn_established() already has it).

Fixes 72e89f50084c ("security: Add support for SCTP security hooks")
---
 security/selinux/hooks.c | 90 +++++++++++++++++++++++++++++-----------
 1 file changed, 66 insertions(+), 24 deletions(-)
security/selinux/hooks.c +66 -24
index a7306208649a..b55f5efd3e0d 100644
@@ -5306,37 +5306,38 @@ static void selinux_sock_graft(struct sock *sk, struct socket *parent)
 	sksec->sclass = isec->sclass;
 }
 
-/* Called whenever SCTP receives an INIT chunk. This happens when an incoming
- * connect(2), sctp_connectx(3) or sctp_sendmsg(3) (with no association
- * already present).
+/*
+ * Determines peer_secid for the asoc and updates socket's peer label
+ * if it's the first association on the socket.
  */
-static int selinux_sctp_assoc_request(struct sctp_association *asoc,
-				      struct sk_buff *skb)
+static int selinux_sctp_process_new_assoc(struct sctp_association *asoc,
+					  struct sk_buff *skb)
 {
-	struct sk_security_struct *sksec = asoc->base.sk->sk_security;
+	struct sock *sk = asoc->base.sk;
+	u16 family = sk->sk_family;
+	struct sk_security_struct *sksec = sk->sk_security;
 	struct common_audit_data ad;
 	struct lsm_network_audit net = {0,};
-	u8 peerlbl_active;
-	u32 peer_sid = SECINITSID_UNLABELED;
-	u32 conn_sid;
-	int err = 0;
+	int err;
 
-	if (!selinux_policycap_extsockclass())
-		return 0;
+	/* handle mapped IPv4 packets arriving via IPv6 sockets */
+	if (family == PF_INET6 && skb->protocol == htons(ETH_P_IP))
+		family = PF_INET;
 
-	peerlbl_active = selinux_peerlbl_enabled();
+	if (selinux_peerlbl_enabled()) {
+		asoc->peer_secid = SECSID_NULL;
 
-	if (peerlbl_active) {
 		/* This will return peer_sid = SECSID_NULL if there are
 		 * no peer labels, see security_net_peersid_resolve().
 		 */
-		err = selinux_skb_peerlbl_sid(skb, asoc->base.sk->sk_family,
-					      &peer_sid);
+		err = selinux_skb_peerlbl_sid(skb, family, &asoc->peer_secid);
 		if (err)
 			return err;
 
-		if (peer_sid == SECSID_NULL)
-			peer_sid = SECINITSID_UNLABELED;
+		if (asoc->peer_secid == SECSID_NULL)
+			asoc->peer_secid = SECINITSID_UNLABELED;
+	} else {
+		asoc->peer_secid = SECINITSID_UNLABELED;
 	}
 
 	if (sksec->sctp_assoc_state == SCTP_ASSOC_UNSET) {
@@ -5347,8 +5348,8 @@ static int selinux_sctp_assoc_request(struct sctp_association *asoc,
 		 * then it is approved by policy and used as the primary
 		 * peer SID for getpeercon(3).
 		 */
-		sksec->peer_sid = peer_sid;
-	} else if  (sksec->peer_sid != peer_sid) {
+		sksec->peer_sid = asoc->peer_secid;
+	} else if (sksec->peer_sid != asoc->peer_secid) {
 		/* Other association peer SIDs are checked to enforce
 		 * consistency among the peer SIDs.
 		 */
@@ -5356,11 +5357,32 @@ static int selinux_sctp_assoc_request(struct sctp_association *asoc,
 		ad.u.net = &net;
 		ad.u.net->sk = asoc->base.sk;
 		err = avc_has_perm(&selinux_state,
-				   sksec->peer_sid, peer_sid, sksec->sclass,
-				   SCTP_SOCKET__ASSOCIATION, &ad);
+				   sksec->peer_sid, asoc->peer_secid,
+				   sksec->sclass, SCTP_SOCKET__ASSOCIATION,
+				   &ad);
 		if (err)
 			return err;
 	}
+	return 0;
+}
+
+/* Called whenever SCTP receives an INIT or COOKIE ECHO chunk. This
+ * happens on an incoming connect(2), sctp_connectx(3) or
+ * sctp_sendmsg(3) (with no association already present).
+ */
+static int selinux_sctp_assoc_request(struct sctp_association *asoc,
+				      struct sk_buff *skb)
+{
+	struct sk_security_struct *sksec = asoc->base.sk->sk_security;
+	u32 conn_sid;
+	int err;
+
+	if (!selinux_policycap_extsockclass())
+		return 0;
+
+	err = selinux_sctp_process_new_assoc(asoc, skb);
+	if (err)
+		return err;
 
 	/* Compute the MLS component for the connection and store
 	 * the information in asoc. This will be used by SCTP TCP type
@@ -5368,17 +5390,36 @@ static int selinux_sctp_assoc_request(struct sctp_association *asoc,
 	 * socket to be generated. selinux_sctp_sk_clone() will then
 	 * plug this into the new socket.
 	 */
-	err = selinux_conn_sid(sksec->sid, peer_sid, &conn_sid);
+	err = selinux_conn_sid(sksec->sid, asoc->peer_secid, &conn_sid);
 	if (err)
 		return err;
 
 	asoc->secid = conn_sid;
-	asoc->peer_secid = peer_sid;
 
 	/* Set any NetLabel labels including CIPSO/CALIPSO options. */
 	return selinux_netlbl_sctp_assoc_request(asoc, skb);
 }
 
+/* Called when SCTP receives a COOKIE ACK chunk as the final
+ * response to an association request (initited by us).
+ */
+static int selinux_sctp_assoc_established(struct sctp_association *asoc,
+					  struct sk_buff *skb)
+{
+	struct sk_security_struct *sksec = asoc->base.sk->sk_security;
+
+	if (!selinux_policycap_extsockclass())
+		return 0;
+
+	/* Inherit secid from the parent socket - this will be picked up
+	 * by selinux_sctp_sk_clone() if the association gets peeled off
+	 * into a new socket.
+	 */
+	asoc->secid = sksec->sid;
+
+	return selinux_sctp_process_new_assoc(asoc, skb);
+}
+
 /* Check if sctp IPv4/IPv6 addresses are valid for binding or connecting
  * based on their @optname.
  */
@@ -7199,6 +7240,7 @@ static struct security_hook_list selinux_hooks[] __lsm_ro_after_init = {
 	LSM_HOOK_INIT(sctp_assoc_request, selinux_sctp_assoc_request),
 	LSM_HOOK_INIT(sctp_sk_clone, selinux_sctp_sk_clone),
 	LSM_HOOK_INIT(sctp_bind_connect, selinux_sctp_bind_connect),
+	LSM_HOOK_INIT(sctp_assoc_established, selinux_sctp_assoc_established),
 	LSM_HOOK_INIT(inet_conn_request, selinux_inet_conn_request),
 	LSM_HOOK_INIT(inet_csk_clone, selinux_inet_csk_clone),
 	LSM_HOOK_INIT(inet_conn_established, selinux_inet_conn_established),
Signature · 1 line
2.34.1
author maintainer
#223 lore raw

PATCH 223/1126 blk-cgroup: set blkg iostat after percpu stat aggregation

In reply to Greg Kroah-Hartman

From: Chengming Zhou <[email protected]>

[ Upstream commit f122d103b564e5fb7c82de902c6f8f6cbdf50ec9 ]

Don't need to do blkg_iostat_set for top blkg iostat on each CPU, so move it after percpu stat aggregation.

Fixes ef45fe470e1e ("blk-cgroup: show global disk stats in root cgroup io.stat")
---
 block/blk-cgroup.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
block/blk-cgroup.c +5 -5
index 650f7e27989f..87a1c0c3fa40 100644
@@ -857,11 +857,11 @@ static void blkcg_fill_root_iostats(void)
 			blk_queue_root_blkg(bdev_get_queue(bdev));
 		struct blkg_iostat tmp;
 		int cpu;
+		unsigned long flags;
 
 		memset(&tmp, 0, sizeof(tmp));
 		for_each_possible_cpu(cpu) {
 			struct disk_stats *cpu_dkstats;
-			unsigned long flags;
 
 			cpu_dkstats = per_cpu_ptr(bdev->bd_stats, cpu);
 			tmp.ios[BLKG_IOSTAT_READ] +=
@@ -877,11 +877,11 @@ static void blkcg_fill_root_iostats(void)
 				cpu_dkstats->sectors[STAT_WRITE] << 9;
 			tmp.bytes[BLKG_IOSTAT_DISCARD] +=
 				cpu_dkstats->sectors[STAT_DISCARD] << 9;
-
-			flags = u64_stats_update_begin_irqsave(&blkg->iostat.sync);
-			blkg_iostat_set(&blkg->iostat.cur, &tmp);
-			u64_stats_update_end_irqrestore(&blkg->iostat.sync, flags);
 		}
+
+		flags = u64_stats_update_begin_irqsave(&blkg->iostat.sync);
+		blkg_iostat_set(&blkg->iostat.cur, &tmp);
+		u64_stats_update_end_irqrestore(&blkg->iostat.sync, flags);
 	}
 }
 
Signature · 1 line
2.34.1
author maintainer
#224 lore raw

PATCH 224/1126 selftests/x86: Add validity check and allow field splitting

In reply to Greg Kroah-Hartman

From: Muhammad Usama Anjum <[email protected]>

[ Upstream commit b06e15ebd5bfb670f93c7f11a29b8299c1178bc6 ]

Add check to test if CC has a string. CC can have multiple sub-strings like "ccache gcc". Erorr pops up if it is treated as single string and double quotes are used around it. This can be fixed by removing the quotes and not treating CC as a single string.

Fixes e9886ace222e ("selftests, x86: Rework x86 target architecture detection")
---
 tools/testing/selftests/x86/check_cc.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/x86/check_cc.sh +1 -1
index 3e2089c8cf54..8c669c0d662e 100755
@@ -7,7 +7,7 @@ CC="$1"
 TESTPROG="$2"
 shift 2
 
-if "$CC" -o /dev/null "$TESTPROG" -O0 "$@" 2>/dev/null; then
+if [ -n "$CC" ] && $CC -o /dev/null "$TESTPROG" -O0 "$@" 2>/dev/null; then
     echo 1
 else
     echo 0
Signature · 1 line
2.34.1
author maintainer
#225 lore raw

PATCH 225/1126 selftests/sgx: Treat CC as one argument

In reply to Greg Kroah-Hartman

From: Muhammad Usama Anjum <[email protected]>

[ Upstream commit 6170abb21e2380477080b25145da9747ad467d3d ]

CC can have multiple sub-strings like "ccache gcc". For check_cc.sh, CC needs to be treated like one argument. Put double quotes around it to make CC one string and hence one argument.

---
 tools/testing/selftests/sgx/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/sgx/Makefile +1 -1
index 2956584e1e37..75af864e07b6 100644
@@ -4,7 +4,7 @@ include ../lib.mk
 
 .PHONY: all clean
 
-CAN_BUILD_X86_64 := $(shell ../x86/check_cc.sh $(CC) \
+CAN_BUILD_X86_64 := $(shell ../x86/check_cc.sh "$(CC)" \
 			    ../x86/trivial_64bit_program.c)
 
 ifndef OBJCOPY
Signature · 1 line
2.34.1
author maintainer
#226 lore raw

PATCH 226/1126 crypto: rockchip - ECB does not need IV

In reply to Greg Kroah-Hartman

From: Corentin Labbe <[email protected]>

[ Upstream commit 973d74e93820d99d8ea203882631c76edab699c9 ]

When loading rockchip crypto module, testmgr complains that ivsize of ecb-des3-ede-rk is not the same than generic implementation.
In fact ECB does not use an IV.

Fixes ce0183cb6464b ("crypto: rockchip - switch to skcipher API")
---
 drivers/crypto/rockchip/rk3288_crypto_skcipher.c | 1 -
 1 file changed, 1 deletion(-)
drivers/crypto/rockchip/rk3288_crypto_skcipher.c +0 -1
index 1cece1a7d3f0..5bbf0d2722e1 100644
@@ -506,7 +506,6 @@ struct rk_crypto_tmp rk_ecb_des3_ede_alg = {
 		.exit			= rk_ablk_exit_tfm,
 		.min_keysize		= DES3_EDE_KEY_SIZE,
 		.max_keysize		= DES3_EDE_KEY_SIZE,
-		.ivsize			= DES_BLOCK_SIZE,
 		.setkey			= rk_tdes_setkey,
 		.encrypt		= rk_des3_ede_ecb_encrypt,
 		.decrypt		= rk_des3_ede_ecb_decrypt,
Signature · 1 line
2.34.1
author maintainer
#227 lore raw

PATCH 227/1126 block: update io_ticks when io hang

In reply to Greg Kroah-Hartman

From: Zhang Wensheng <[email protected]>

[ Upstream commit 86d7331299fda7634b11c1b7c911432679d525a5 ]

When the inflight IOs are slow and no new IOs are issued, we expect iostat could manifest the IO hang problem. However after commit 5b18b5a73760 ("block: delete part_round_stats and switch to less precise counting"), io_tick and time_in_queue will not be updated until the end of IO, and the avgqu-sz and %util columns of iostat will be zero.

Because it has using stat.nsecs accumulation to express time_in_queue which is not suitable to change, and may %util will express the status better when io hang occur. To fix io_ticks, we use update_io_ticks and inflight to update io_ticks when diskstats_show and part_stat_show been called.

Fixes 5b18b5a73760 ("block: delete part_round_stats and switch to less precise counting")
---
 block/genhd.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
block/genhd.c +12 -2
index 9eca1f7d35c9..6fa8b2068cc2 100644
@@ -927,12 +927,17 @@ ssize_t part_stat_show(struct device *dev,
 	struct disk_stats stat;
 	unsigned int inflight;
 
-	part_stat_read_all(bdev, &stat);
 	if (queue_is_mq(q))
 		inflight = blk_mq_in_flight(q, bdev);
 	else
 		inflight = part_in_flight(bdev);
 
+	if (inflight) {
+		part_stat_lock();
+		update_io_ticks(bdev, jiffies, true);
+		part_stat_unlock();
+	}
+	part_stat_read_all(bdev, &stat);
 	return sprintf(buf,
 		"%8lu %8lu %8llu %8u "
 		"%8lu %8lu %8llu %8u "
@@ -1188,12 +1193,17 @@ static int diskstats_show(struct seq_file *seqf, void *v)
 	xa_for_each(&gp->part_tbl, idx, hd) {
 		if (bdev_is_partition(hd) && !bdev_nr_sectors(hd))
 			continue;
-		part_stat_read_all(hd, &stat);
 		if (queue_is_mq(gp->queue))
 			inflight = blk_mq_in_flight(gp->queue, hd);
 		else
 			inflight = part_in_flight(hd);
 
+		if (inflight) {
+			part_stat_lock();
+			update_io_ticks(hd, jiffies, true);
+			part_stat_unlock();
+		}
+		part_stat_read_all(hd, &stat);
 		seq_printf(seqf, "%4d %7d %pg "
 			   "%lu %lu %lu %u "
 			   "%lu %lu %lu %u "
Signature · 1 line
2.34.1
author maintainer
#228 lore raw

PATCH 228/1126 audit: log AUDIT_TIME_* records only from rules

In reply to Greg Kroah-Hartman

From: Richard Guy Briggs <[email protected]>

[ Upstream commit 272ceeaea355214b301530e262a0df8600bfca95 ]

AUDIT_TIME_* events are generated when there are syscall rules present that are not related to time keeping. This will produce noisy log entries that could flood the logs and hide events we really care about.

Rather than immediately produce the AUDIT_TIME_* records, store the data in the context and log it at syscall exit time respecting the filter rules.

Note: This eats the audit_buffer, unlike any others in show_special().

Please see https://bugzilla.redhat.com/show_bug.cgi?id=1991919

Fixes 7e8eda734d30 ("ntp: Audit NTP parameters adjustment")
Fixes 2d87a0674bd6 ("timekeeping: Audit clock adjustments")
[PM: fixed style/whitespace issues]
---
 kernel/audit.h   |  4 +++
 kernel/auditsc.c | 87 +++++++++++++++++++++++++++++++++++++-----------
 2 files changed, 71 insertions(+), 20 deletions(-)
kernel/audit.h +4 -0
index c4498090a5bd..58b66543b4d5 100644
@@ -201,6 +201,10 @@ struct audit_context {
 		struct {
 			char			*name;
 		} module;
+		struct {
+			struct audit_ntp_data	ntp_data;
+			struct timespec64	tk_injoffset;
+		} time;
 	};
 	int fds[2];
 	struct audit_proctitle proctitle;
kernel/auditsc.c +67 -20

110 lines not on this page · raw message

Signature · 1 line
2.34.1
author maintainer
#229 lore raw

PATCH 229/1126 EVM: fix the evm= __setup handler return value

In reply to Greg Kroah-Hartman

From: Randy Dunlap <[email protected]>

[ Upstream commit f2544f5e6c691679d56bb38637d2f347075b36fa ]

__setup() handlers should return 1 if the parameter is handled. Returning 0 causes the entire string to be added to init's environment strings (limited to 32 strings), unnecessarily polluting it.

Using the documented string "evm=fix" causes an Unknown parameter message:
Unknown kernel command line parameters "BOOT_IMAGE=/boot/bzImage-517rc5 evm=fix", will be passed to user space.

and that string is added to init's environment string space:
Run /sbin/init as init process

    with arguments:
     /sbin/init
    with environment:
     HOME=/
     TERM=linux
     BOOT_IMAGE=/boot/bzImage-517rc5
     evm=fix

With this change, using "evm=fix" acts as expected and an invalid option ("evm=evm") causes a warning to be printed:
evm: invalid "evm" mode
but init's environment is not polluted with this string, as expected.

Fixes 7102ebcd65c1 ("evm: permit only valid security.evm xattrs to be updated")
Link lore.kernel.org/r/[email protected]
---
 security/integrity/evm/evm_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
security/integrity/evm/evm_main.c +1 -1
index 08f907382c61..7d87772f0ce6 100644
@@ -86,7 +86,7 @@ static int __init evm_set_fixmode(char *str)
 	else
 		pr_err("invalid \"%s\" mode", str);
 
-	return 0;
+	return 1;
 }
 __setup("evm=", evm_set_fixmode);
 
Signature · 1 line
2.34.1
author maintainer
#230 lore raw

PATCH 230/1126 crypto: ccree - dont attempt 0 len DMA mappings

In reply to Greg Kroah-Hartman

From: Gilad Ben-Yossef <[email protected]>

[ Upstream commit 1fb37b5692c915edcc2448a6b37255738c7c77e0 ]

Refuse to try mapping zero bytes as this may cause a fault on some configurations / platforms and it seems the prev. attempt is not enough and we need to be more explicit.

Fixes ce0fc6db38de ("crypto: ccree - protect against empty or NULL

scatterlists")

---
 drivers/crypto/ccree/cc_buffer_mgr.c | 7 +++++++
 1 file changed, 7 insertions(+)
drivers/crypto/ccree/cc_buffer_mgr.c +7 -0
index a5e041d9d2cf..11e0278c8631 100644
@@ -258,6 +258,13 @@ static int cc_map_sg(struct device *dev, struct scatterlist *sg,
 {
 	int ret = 0;
 
+	if (!nbytes) {
+		*mapped_nents = 0;
+		*lbytes = 0;
+		*nents = 0;
+		return 0;
+	}
+
 	*nents = cc_get_sgl_nents(dev, sg, nbytes, lbytes);
 	if (*nents > max_sg_nents) {
 		*nents = 0;
Signature · 1 line
2.34.1
author maintainer
#231 lore raw

PATCH 231/1126 crypto: hisilicon/sec - fix the aead software fallback for engine

In reply to Greg Kroah-Hartman

From: Kai Ye <[email protected]>

[ Upstream commit 0a2a464f863187f97e96ebc6384c052cafd4a54c ]

Due to the subreq pointer misuse the private context memory. The aead soft crypto occasionally casues the OS panic as setting the 64K page. Here is fix it.

Fixes 6c46a3297bea ("crypto: hisilicon/sec - add fallback tfm...")
---
 drivers/crypto/hisilicon/sec2/sec_crypto.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)
drivers/crypto/hisilicon/sec2/sec_crypto.c +13 -3
index 6a45bd23b363..090920ed50c8 100644
@@ -2284,9 +2284,10 @@ static int sec_aead_soft_crypto(struct sec_ctx *ctx,
 				struct aead_request *aead_req,
 				bool encrypt)
 {
-	struct aead_request *subreq = aead_request_ctx(aead_req);
 	struct sec_auth_ctx *a_ctx = &ctx->a_ctx;
 	struct device *dev = ctx->dev;
+	struct aead_request *subreq;
+	int ret;
 
 	/* Kunpeng920 aead mode not support input 0 size */
 	if (!a_ctx->fallback_aead_tfm) {
@@ -2294,6 +2295,10 @@ static int sec_aead_soft_crypto(struct sec_ctx *ctx,
 		return -EINVAL;
 	}
 
+	subreq = aead_request_alloc(a_ctx->fallback_aead_tfm, GFP_KERNEL);
+	if (!subreq)
+		return -ENOMEM;
+
 	aead_request_set_tfm(subreq, a_ctx->fallback_aead_tfm);
 	aead_request_set_callback(subreq, aead_req->base.flags,
 				  aead_req->base.complete, aead_req->base.data);
@@ -2301,8 +2306,13 @@ static int sec_aead_soft_crypto(struct sec_ctx *ctx,
 			       aead_req->cryptlen, aead_req->iv);
 	aead_request_set_ad(subreq, aead_req->assoclen);
 
-	return encrypt ? crypto_aead_encrypt(subreq) :
-		   crypto_aead_decrypt(subreq);
+	if (encrypt)
+		ret = crypto_aead_encrypt(subreq);
+	else
+		ret = crypto_aead_decrypt(subreq);
+	aead_request_free(subreq);
+
+	return ret;
 }
 
 static int sec_aead_crypto(struct aead_request *a_req, bool encrypt)
Signature · 1 line
2.34.1
author maintainer
#232 lore raw

PATCH 232/1126 spi: pxa2xx-pci: Balance reference count for PCI DMA device

In reply to Greg Kroah-Hartman

From: Andy Shevchenko <[email protected]>

[ Upstream commit 609d7ffdc42199a0ec949db057e3b4be6745d6c5 ]

The pci_get_slot() increases its reference count, the caller must decrement the reference count by calling pci_dev_put().

Fixes 743485ea3bee ("spi: pxa2xx-pci: Do a specific setup in a separate function")
Fixes 25014521603f ("spi: pxa2xx-pci: Enable DMA for Intel Merrifield")
---
 drivers/spi/spi-pxa2xx-pci.c | 17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)
drivers/spi/spi-pxa2xx-pci.c +16 -1
index 2e134eb4bd2c..6502fda6243e 100644
@@ -76,14 +76,23 @@ static bool lpss_dma_filter(struct dma_chan *chan, void *param)
 	return true;
 }
 
+static void lpss_dma_put_device(void *dma_dev)
+{
+	pci_dev_put(dma_dev);
+}
+
 static int lpss_spi_setup(struct pci_dev *dev, struct pxa_spi_info *c)
 {
 	struct pci_dev *dma_dev;
+	int ret;
 
 	c->num_chipselect = 1;
 	c->max_clk_rate = 50000000;
 
 	dma_dev = pci_get_slot(dev->bus, PCI_DEVFN(PCI_SLOT(dev->devfn), 0));
+	ret = devm_add_action_or_reset(&dev->dev, lpss_dma_put_device, dma_dev);
+	if (ret)
+		return ret;
 
 	if (c->tx_param) {
 		struct dw_dma_slave *slave = c->tx_param;
@@ -107,8 +116,9 @@ static int lpss_spi_setup(struct pci_dev *dev, struct pxa_spi_info *c)
 
 static int mrfld_spi_setup(struct pci_dev *dev, struct pxa_spi_info *c)
 {
-	struct pci_dev *dma_dev = pci_get_slot(dev->bus, PCI_DEVFN(21, 0));
 	struct dw_dma_slave *tx, *rx;
+	struct pci_dev *dma_dev;
+	int ret;
 
 	switch (PCI_FUNC(dev->devfn)) {
 	case 0:
@@ -133,6 +143,11 @@ static int mrfld_spi_setup(struct pci_dev *dev, struct pxa_spi_info *c)
 		return -ENODEV;
 	}
 
+	dma_dev = pci_get_slot(dev->bus, PCI_DEVFN(21, 0));
+	ret = devm_add_action_or_reset(&dev->dev, lpss_dma_put_device, dma_dev);
+	if (ret)
+		return ret;
+
 	tx = c->tx_param;
 	tx->dma_dev = &dma_dev->dev;
 
Signature · 1 line
2.34.1
author maintainer
#233 lore raw

PATCH 233/1126 hwmon: (pmbus) Add mutex to regulator ops

In reply to Greg Kroah-Hartman

From: Patrick Rudolph <[email protected]>

[ Upstream commit 686d303ee6301261b422ea51e64833d7909a2c36 ]

On PMBUS devices with multiple pages, the regulator ops need to be protected with the update mutex. This prevents accidentally changing the page in a separate thread while operating on the PMBUS_OPERATION register.

Tested on Infineon xdpe11280 while a separate thread polls for sensor data.

---
 drivers/hwmon/pmbus/pmbus_core.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)
drivers/hwmon/pmbus/pmbus_core.c +13 -3
index ac2fbee1ba9c..b1386a4df4cc 100644
@@ -2391,10 +2391,14 @@ static int pmbus_regulator_is_enabled(struct regulator_dev *rdev)
 {
 	struct device *dev = rdev_get_dev(rdev);
 	struct i2c_client *client = to_i2c_client(dev->parent);
+	struct pmbus_data *data = i2c_get_clientdata(client);
 	u8 page = rdev_get_id(rdev);
 	int ret;
 
+	mutex_lock(&data->update_lock);
 	ret = pmbus_read_byte_data(client, page, PMBUS_OPERATION);
+	mutex_unlock(&data->update_lock);
+
 	if (ret < 0)
 		return ret;
 
@@ -2405,11 +2409,17 @@ static int _pmbus_regulator_on_off(struct regulator_dev *rdev, bool enable)
 {
 	struct device *dev = rdev_get_dev(rdev);
 	struct i2c_client *client = to_i2c_client(dev->parent);
+	struct pmbus_data *data = i2c_get_clientdata(client);
 	u8 page = rdev_get_id(rdev);
+	int ret;
 
-	return pmbus_update_byte_data(client, page, PMBUS_OPERATION,
-				      PB_OPERATION_CONTROL_ON,
-				      enable ? PB_OPERATION_CONTROL_ON : 0);
+	mutex_lock(&data->update_lock);
+	ret = pmbus_update_byte_data(client, page, PMBUS_OPERATION,
+				     PB_OPERATION_CONTROL_ON,
+				     enable ? PB_OPERATION_CONTROL_ON : 0);
+	mutex_unlock(&data->update_lock);
+
+	return ret;
 }
 
 static int pmbus_regulator_enable(struct regulator_dev *rdev)
Signature · 1 line
2.34.1
author maintainer
#234 lore raw

PATCH 234/1126 hwmon: (sch56xx-common) Replace WDOG_ACTIVE with WDOG_HW_RUNNING

In reply to Greg Kroah-Hartman

From: Armin Wolf <[email protected]>

[ Upstream commit 647d6f09bea7dacf4cdb6d4ea7e3051883955297 ]

If the watchdog was already enabled by the BIOS after booting, the watchdog infrastructure needs to regularly send keepalives to prevent a unexpected reset.
WDOG_ACTIVE only serves as an status indicator for userspace, we want to use WDOG_HW_RUNNING instead.

Since my Fujitsu Esprimo P720 does not support the watchdog, this change is compile-tested only.

---
 drivers/hwmon/sch56xx-common.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/hwmon/sch56xx-common.c +1 -1
index 40cdadad35e5..f85eede6d766 100644
@@ -422,7 +422,7 @@ void sch56xx_watchdog_register(struct device *parent, u16 addr, u32 revision,
 	data->wddev.max_timeout = 255 * 60;
 	watchdog_set_nowayout(&data->wddev, nowayout);
 	if (output_enable & SCH56XX_WDOG_OUTPUT_ENABLE)
-		set_bit(WDOG_ACTIVE, &data->wddev.status);
+		set_bit(WDOG_HW_RUNNING, &data->wddev.status);
 
 	/* Since the watchdog uses a downcounter there is no register to read
 	   the BIOS set timeout from (if any was set at all) ->
Signature · 1 line
2.34.1
author maintainer
#235 lore raw

PATCH 235/1126 nvme: cleanup __nvme_check_ids

In reply to Greg Kroah-Hartman

From: Christoph Hellwig <[email protected]>

[ Upstream commit fd8099e7918cd2df39ef306dd1d1af7178a15b81 ]

Pass the actual nvme_ns_ids used for the comparison instead of the ns_head that isn't needed and use a more descriptive function name.

---
 drivers/nvme/host/core.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)
drivers/nvme/host/core.c +4 -5
index fd4720d37cc0..f921c917f4b0 100644
@@ -3598,16 +3598,15 @@ static struct nvme_ns_head *nvme_find_ns_head(struct nvme_subsystem *subsys,
 	return NULL;
 }
 
-static int __nvme_check_ids(struct nvme_subsystem *subsys,
-		struct nvme_ns_head *new)
+static int nvme_subsys_check_duplicate_ids(struct nvme_subsystem *subsys,
+		struct nvme_ns_ids *ids)
 {
 	struct nvme_ns_head *h;
 
 	lockdep_assert_held(&subsys->lock);
 
 	list_for_each_entry(h, &subsys->nsheads, entry) {
-		if (nvme_ns_ids_valid(&new->ids) &&
-		    nvme_ns_ids_equal(&new->ids, &h->ids))
+		if (nvme_ns_ids_valid(ids) && nvme_ns_ids_equal(ids, &h->ids))
 			return -EINVAL;
 	}
 
@@ -3706,7 +3705,7 @@ static struct nvme_ns_head *nvme_alloc_ns_head(struct nvme_ctrl *ctrl,
 	head->ids = *ids;
 	kref_init(&head->ref);
 
-	ret = __nvme_check_ids(ctrl->subsys, head);
+	ret = nvme_subsys_check_duplicate_ids(ctrl->subsys, &head->ids);
 	if (ret) {
 		dev_err(ctrl->device,
 			"duplicate IDs for nsid %d\n", nsid);
Signature · 1 line
2.34.1
author maintainer
#236 lore raw

PATCH 236/1126 nvme: fix the check for duplicate unique identifiers

In reply to Greg Kroah-Hartman

From: Christoph Hellwig <[email protected]>

[ Upstream commit e2724cb9f0c406b8fb66efd3aa9e8b3edfd8d5c8 ]

nvme_subsys_check_duplicate_ids should needs to return an error if any of the identifiers matches, not just if all of them match. But it does not need to and should not look at the CSI value for this sanity check.

Rewrite the logic to be separate from nvme_ns_ids_equal and optimize it by reducing duplicate checks for non-present identifiers.

Fixes ed754e5deeb1 ("nvme: track shared namespaces")
---
 drivers/nvme/host/core.c | 18 ++++++++++--------
 1 file changed, 10 insertions(+), 8 deletions(-)
drivers/nvme/host/core.c +10 -8
index f921c917f4b0..7cf81bb5d211 100644
@@ -1683,13 +1683,6 @@ static void nvme_config_discard(struct gendisk *disk, struct nvme_ns *ns)
 		blk_queue_max_write_zeroes_sectors(queue, UINT_MAX);
 }
 
-static bool nvme_ns_ids_valid(struct nvme_ns_ids *ids)
-{
-	return !uuid_is_null(&ids->uuid) ||
-		memchr_inv(ids->nguid, 0, sizeof(ids->nguid)) ||
-		memchr_inv(ids->eui64, 0, sizeof(ids->eui64));
-}
-
 static bool nvme_ns_ids_equal(struct nvme_ns_ids *a, struct nvme_ns_ids *b)
 {
 	return uuid_equal(&a->uuid, &b->uuid) &&
@@ -3601,12 +3594,21 @@ static struct nvme_ns_head *nvme_find_ns_head(struct nvme_subsystem *subsys,
 static int nvme_subsys_check_duplicate_ids(struct nvme_subsystem *subsys,
 		struct nvme_ns_ids *ids)
 {
+	bool has_uuid = !uuid_is_null(&ids->uuid);
+	bool has_nguid = memchr_inv(ids->nguid, 0, sizeof(ids->nguid));
+	bool has_eui64 = memchr_inv(ids->eui64, 0, sizeof(ids->eui64));
 	struct nvme_ns_head *h;
 
 	lockdep_assert_held(&subsys->lock);
 
 	list_for_each_entry(h, &subsys->nsheads, entry) {
-		if (nvme_ns_ids_valid(ids) && nvme_ns_ids_equal(ids, &h->ids))
+		if (has_uuid && uuid_equal(&ids->uuid, &h->ids.uuid))
+			return -EINVAL;
+		if (has_nguid &&
+		    memcmp(&ids->nguid, &h->ids.nguid, sizeof(ids->nguid)) == 0)
+			return -EINVAL;
+		if (has_eui64 &&
+		    memcmp(&ids->eui64, &h->ids.eui64, sizeof(ids->eui64)) == 0)
 			return -EINVAL;
 	}
 
Signature · 1 line
2.34.1