Skip to contentSkip to conversation

9p: virtio: make sure 'offs' is initialized in zc_request

PATCH In review Sasha Levin 2 participants 11 messages Aug 13, 2023 – Aug 22, 2023

Subsystems (7)
lore
Share & tools
lore

Sasha Levin

author maintainer
#0 lore raw

PATCH 1/9 9p: virtio: make sure 'offs' is initialized in zc_request

From: Dominique Martinet <[email protected]>

[ Upstream commit 4a73edab69d3a6623f03817fe950a2d9585f80e4 ]

Similarly to the previous patch: offs can be used in handle_rerrors without initializing on small payloads; in this case handle_rerrors will not use it because of the size check, but it doesn't hurt to make sure it is zero to please scan-build.

This fixes the following warning:
net/9p/trans_virtio.c:539:3: warning: 3rd function call argument is an uninitialized value [core.CallAndMessage]

                handle_rerror(req, in_hdr_len, offs, in_pages);
                ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
---
 net/9p/trans_virtio.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/9p/trans_virtio.c +1 -1
index c6a46e8e9eda5..25f5caa57289b 100644
@@ -401,7 +401,7 @@ p9_virtio_zc_request(struct p9_client *client, struct p9_req_t *req,
 	struct page **in_pages = NULL, **out_pages = NULL;
 	struct virtio_chan *chan = client->trans;
 	struct scatterlist *sgs[4];
-	size_t offs;
+	size_t offs = 0;
 	int need_drop = 0;
 
 	p9_debug(P9_DEBUG_TRANS, "virtio request\n");
Signature · 1 line
2.40.1
author maintainer
#1 lore raw

PATCH 2/9 ASoC: da7219: Flush pending AAD IRQ when suspending

From: Dmytro Maluka <[email protected]>

[ Upstream commit 91e292917dad64ab8d1d5ca2ab3069ad9dac6f72 ]

da7219_aad_suspend() disables jack detection, which should prevent generating new interrupts by DA7219 while suspended. However, there is a theoretical possibility that there is a pending interrupt generated just before suspending DA7219 and not handled yet, so the IRQ handler may still run after DA7219 is suspended. To prevent that, wait until the pending IRQ handling is done.

This patch arose as an attempt to fix the following I2C failure occurring sometimes during system suspend or resume:

[  355.876211] i2c_designware i2c_designware.3: Transfer while suspended
[  355.876245] WARNING: CPU: 2 PID: 3576 at drivers/i2c/busses/i2c-designware-master.c:570 i2c_dw_xfer+0x411/0x440
...
[  355.876462] Call Trace:
[  355.876468]  <TASK>
[  355.876475]  ? update_load_avg+0x1b3/0x615
[  355.876484]  __i2c_transfer+0x101/0x1d8
[  355.876494]  i2c_transfer+0x74/0x10d
[  355.876504]  regmap_i2c_read+0x6a/0x9c
[  355.876513]  _regmap_raw_read+0x179/0x223
[  355.876521]  regmap_raw_read+0x1e1/0x28e
[  355.876527]  regmap_bulk_read+0x17d/0x1ba
[  355.876532]  ? __wake_up+0xed/0x1bb
[  355.876542]  da7219_aad_irq_thread+0x54/0x2c9 [snd_soc_da7219 5fb8ebb2179cf2fea29af090f3145d68ed8e2184]
[  355.876556]  irq_thread+0x13c/0x231
[  355.876563]  ? irq_forced_thread_fn+0x5f/0x5f
[  355.876570]  ? irq_thread_fn+0x4d/0x4d
[  355.876576]  kthread+0x13a/0x152
[  355.876581]  ? synchronize_irq+0xc3/0xc3
[  355.876587]  ? kthread_blkcg+0x31/0x31
[  355.876592]  ret_from_fork+0x1f/0x30
[  355.876601]  </TASK>

which indicates that the AAD IRQ handler is unexpectedly running when DA7219 is suspended, and as a result, is trying to read data from DA7219 over I2C and is hitting the I2C driver "Transfer while suspended" failure.

However, with this patch the above failure is still reproducible. So this patch does not fix any real observed issue so far, but at least is useful for confirming that the above issue is not caused by a pending IRQ but rather looks like a DA7219 hardware issue with an IRQ unexpectedly generated after jack detection is already disabled.

---
 sound/soc/codecs/da7219-aad.c | 2 ++
 1 file changed, 2 insertions(+)
sound/soc/codecs/da7219-aad.c +2 -0
index 1d1d10dd92ae2..793c8768f7c44 100644
@@ -854,6 +854,8 @@ void da7219_aad_suspend(struct snd_soc_codec *codec)
 			}
 		}
 	}
+
+	synchronize_irq(da7219_aad->irq);
 }
 
 void da7219_aad_resume(struct snd_soc_codec *codec)
Signature · 1 line
2.40.1
author maintainer
#2 lore raw

PATCH 3/9 ethernet: atheros: fix return value check in atl1c_tso_csum()

In reply to Sasha Levin

From: Yuanjun Gong <[email protected]>

[ Upstream commit 8d01da0a1db237c44c92859ce3612df7af8d3a53 ]

in atl1c_tso_csum, it should check the return value of pskb_trim(), and return an error code if an unexpected value is returned by pskb_trim().

---
 drivers/net/ethernet/atheros/atl1c/atl1c_main.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/net/ethernet/atheros/atl1c/atl1c_main.c +5 -2
index 3615c2a06fdad..6f5c7c1401ce0 100644
@@ -2001,8 +2001,11 @@ static int atl1c_tso_csum(struct atl1c_adapter *adapter,
 			real_len = (((unsigned char *)ip_hdr(skb) - skb->data)
 					+ ntohs(ip_hdr(skb)->tot_len));
 
-			if (real_len < skb->len)
-				pskb_trim(skb, real_len);
+			if (real_len < skb->len) {
+				err = pskb_trim(skb, real_len);
+				if (err)
+					return err;
+			}
 
 			hdr_len = (skb_transport_offset(skb) + tcp_hdrlen(skb));
 			if (unlikely(skb->len == hdr_len)) {
Signature · 1 line
2.40.1
author maintainer
#3 lore raw

PATCH 4/9 vxlan: generalize vxlan_parse_gpe_hdr and remove unused args

In reply to Sasha Levin

From: Jiri Benc <[email protected]>

[ Upstream commit 17a0a64448b568442a101de09575f81ffdc45d15 ]

The vxlan_parse_gpe_hdr function extracts the next protocol value from the GPE header and marks GPE bits as parsed.

In order to be used in the next patch, split the function into protocol extraction and bit marking. The bit marking is meaningful only in vxlan_rcv; move it directly there.

Rename the function to vxlan_parse_gpe_proto to reflect what it now does. Remove unused arguments skb and vxflags. Move the function earlier in the file to allow it to be called from more places in the next patch.

---
 drivers/net/vxlan.c | 58 ++++++++++++++++++++++-----------------------
 1 file changed, 28 insertions(+), 30 deletions(-)
drivers/net/vxlan.c +28 -30
index 31657f15eb071..52aa32d7e4c5c 100644
@@ -542,6 +542,32 @@ static int vxlan_fdb_append(struct vxlan_fdb *f,
 	return 1;
 }
 
+static bool vxlan_parse_gpe_proto(struct vxlanhdr *hdr, __be16 *protocol)
+{
+	struct vxlanhdr_gpe *gpe = (struct vxlanhdr_gpe *)hdr;
+
+	/* Need to have Next Protocol set for interfaces in GPE mode. */
+	if (!gpe->np_applied)
+		return false;
+	/* "The initial version is 0. If a receiver does not support the
+	 * version indicated it MUST drop the packet.
+	 */
+	if (gpe->version != 0)
+		return false;
+	/* "When the O bit is set to 1, the packet is an OAM packet and OAM
+	 * processing MUST occur." However, we don't implement OAM
+	 * processing, thus drop the packet.
+	 */
+	if (gpe->oam_flag)
+		return false;
+
+	*protocol = tun_p_to_eth_p(gpe->next_protocol);
+	if (!*protocol)
+		return false;
+
+	return true;
+}
+
 static struct vxlanhdr *vxlan_gro_remcsum(struct sk_buff *skb,
 					  unsigned int off,
 					  struct vxlanhdr *vh, size_t hdrlen,
@@ -1278,35 +1304,6 @@ static void vxlan_parse_gbp_hdr(struct vxlanhdr *unparsed,
 	unparsed->vx_flags &= ~VXLAN_GBP_USED_BITS;
 }
 
-static bool vxlan_parse_gpe_hdr(struct vxlanhdr *unparsed,
-				__be16 *protocol,
-				struct sk_buff *skb, u32 vxflags)
-{
-	struct vxlanhdr_gpe *gpe = (struct vxlanhdr_gpe *)unparsed;
-
-	/* Need to have Next Protocol set for interfaces in GPE mode. */
-	if (!gpe->np_applied)
-		return false;
-	/* "The initial version is 0. If a receiver does not support the
-	 * version indicated it MUST drop the packet.
-	 */
-	if (gpe->version != 0)
-		return false;
-	/* "When the O bit is set to 1, the packet is an OAM packet and OAM
-	 * processing MUST occur." However, we don't implement OAM
-	 * processing, thus drop the packet.
-	 */
-	if (gpe->oam_flag)
-		return false;
-
-	*protocol = tun_p_to_eth_p(gpe->next_protocol);
-	if (!*protocol)
-		return false;
-
-	unparsed->vx_flags &= ~VXLAN_GPE_USED_BITS;
-	return true;
-}
-
 static bool vxlan_set_mac(struct vxlan_dev *vxlan,
 			  struct vxlan_sock *vs,
 			  struct sk_buff *skb, __be32 vni)
@@ -1408,8 +1405,9 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 	 * used by VXLAN extensions if explicitly requested.
 	 */
 	if (vs->flags & VXLAN_F_GPE) {
-		if (!vxlan_parse_gpe_hdr(&unparsed, &protocol, skb, vs->flags))
+		if (!vxlan_parse_gpe_proto(&unparsed, &protocol))
 			goto drop;
+		unparsed.vx_flags &= ~VXLAN_GPE_USED_BITS;
 		raw_proto = true;
 	}
 
Signature · 1 line
2.40.1
#4 lore raw

Re: [PATCH AUTOSEL 4.14 4/9] vxlan: generalize vxlan_parse_gpe_hdr and remove unused args

Hi!

Quoted 7 lines

From: Jiri Benc <[email protected]>

[ Upstream commit 17a0a64448b568442a101de09575f81ffdc45d15 ]

The vxlan_parse_gpe_hdr function extracts the next protocol value from the GPE header and marks GPE bits as parsed.

Quoted 8 lines

In order to be used in the next patch, split the function into protocol extraction and bit marking. The bit marking is meaningful only in vxlan_rcv; move it directly there.

Rename the function to vxlan_parse_gpe_proto to reflect what it now does. Remove unused arguments skb and vxflags. Move the function earlier in the file to allow it to be called from more places in the next patch.

This seems to be a cleanup/preparation for a patch we don't have queued for 4.14. Please drop.

Best regards,
Pavel

Signature · 2 lines
DENX Software Engineering GmbH,        Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
author maintainer
#5 lore raw

PATCH 5/9 m68k: Fix invalid .section syntax

In reply to Sasha Levin

From: Ben Hutchings <[email protected]>

[ Upstream commit 922a9bd138101e3e5718f0f4d40dba68ef89bb43 ]

gas supports several different forms for .section for ELF targets, including:
.section NAME [, "FLAGS"[, @TYPE[,FLAG_SPECIFIC_ARGUMENTS]]]
and:

    .section "NAME"[, #FLAGS...]

In several places we use a mix of these two forms:
.section NAME, #FLAGS...

A current development snapshot of binutils (2.40.50.20230611) treats this mixed syntax as an error.

Change to consistently use:
.section NAME, "FLAGS"
as is used elsewhere in the kernel.

---
 arch/m68k/fpsp040/skeleton.S       | 4 ++--
 arch/m68k/ifpsp060/os.S            | 4 ++--
 arch/m68k/kernel/relocate_kernel.S | 4 ++--
 3 files changed, 6 insertions(+), 6 deletions(-)
arch/m68k/fpsp040/skeleton.S +2 -2
index a8f41615d94a7..31a9c634c81ed 100644
@@ -499,12 +499,12 @@ in_ea:
 	dbf	%d0,morein
 	rts
 
-	.section .fixup,#alloc,#execinstr
+	.section .fixup,"ax"
 	.even
 1:
 	jbra	fpsp040_die
 
-	.section __ex_table,#alloc
+	.section __ex_table,"a"
 	.align	4
 
 	.long	in_ea,1b
arch/m68k/ifpsp060/os.S +2 -2
index 7a0d6e4280665..89e2ec224ab6c 100644
@@ -379,11 +379,11 @@ _060_real_access:
 
 
 | Execption handling for movs access to illegal memory
-	.section .fixup,#alloc,#execinstr
+	.section .fixup,"ax"
 	.even
 1:	moveq		#-1,%d1
 	rts
-.section __ex_table,#alloc
+.section __ex_table,"a"
 	.align 4
 	.long	dmrbuae,1b
 	.long	dmrwuae,1b
arch/m68k/kernel/relocate_kernel.S +2 -2
index ab0f1e7d46535..f7667079e08e9 100644
@@ -26,7 +26,7 @@ ENTRY(relocate_new_kernel)
 	lea %pc@(.Lcopy),%a4
 2:	addl #0x00000000,%a4		/* virt_to_phys() */
 
-	.section ".m68k_fixup","aw"
+	.section .m68k_fixup,"aw"
 	.long M68K_FIXUP_MEMOFFSET, 2b+2
 	.previous
 
@@ -49,7 +49,7 @@ ENTRY(relocate_new_kernel)
 	lea %pc@(.Lcont040),%a4
 5:	addl #0x00000000,%a4		/* virt_to_phys() */
 
-	.section ".m68k_fixup","aw"
+	.section .m68k_fixup,"aw"
 	.long M68K_FIXUP_MEMOFFSET, 5b+2
 	.previous
 
Signature · 1 line
2.40.1
author maintainer
#6 lore raw

PATCH 6/9 s390/dasd: use correct number of retries for ERP requests

In reply to Sasha Levin

From: Stefan Haberland <[email protected]>

[ Upstream commit acea28a6b74f458defda7417d2217b051ba7d444 ]

If a DASD request fails an error recovery procedure (ERP) request might be built as a copy of the original request to do error recovery.

The ERP request gets a number of retries assigned.
This number is always 256 no matter what other value might have been set for the original request. This is not what is expected when a user specifies a certain amount of retries for the device via sysfs.

Correctly use the number of retries of the original request for ERP requests.

---
 drivers/s390/block/dasd_3990_erp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/s390/block/dasd_3990_erp.c +1 -1
index ee14d8e45c971..6d26343b12f25 100644
@@ -2423,7 +2423,7 @@ static struct dasd_ccw_req *dasd_3990_erp_add_erp(struct dasd_ccw_req *cqr)
 	erp->block    = cqr->block;
 	erp->magic    = cqr->magic;
 	erp->expires  = cqr->expires;
-	erp->retries  = 256;
+	erp->retries  = device->default_retries;
 	erp->buildclk = get_tod_clock();
 	erp->status = DASD_CQR_FILLED;
 
Signature · 1 line
2.40.1
author maintainer
#7 lore raw

PATCH 7/9 fs/nls: make load_nls() take a const parameter

In reply to Sasha Levin

From: Winston Wen <[email protected]>

[ Upstream commit c1ed39ec116272935528ca9b348b8ee79b0791da ]

load_nls() take a char * parameter, use it to find nls module in list or construct the module name to load it.

This change make load_nls() take a const parameter, so we don't need do some cast like this:

        ses->local_nls = load_nls((char *)ctx->local_nls->charset);
---
 fs/nls/nls_base.c   | 4 ++--
 include/linux/nls.h | 2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)
fs/nls/nls_base.c +2 -2
index 52ccd34b1e792..a026dbd3593f6 100644
@@ -272,7 +272,7 @@ int unregister_nls(struct nls_table * nls)
 	return -EINVAL;
 }
 
-static struct nls_table *find_nls(char *charset)
+static struct nls_table *find_nls(const char *charset)
 {
 	struct nls_table *nls;
 	spin_lock(&nls_lock);
@@ -288,7 +288,7 @@ static struct nls_table *find_nls(char *charset)
 	return nls;
 }
 
-struct nls_table *load_nls(char *charset)
+struct nls_table *load_nls(const char *charset)
 {
 	return try_then_request_module(find_nls(charset), "nls_%s", charset);
 }
include/linux/nls.h +1 -1
index 499e486b3722d..e0bf8367b274a 100644
@@ -47,7 +47,7 @@ enum utf16_endian {
 /* nls_base.c */
 extern int __register_nls(struct nls_table *, struct module *);
 extern int unregister_nls(struct nls_table *);
-extern struct nls_table *load_nls(char *);
+extern struct nls_table *load_nls(const char *charset);
 extern void unload_nls(struct nls_table *);
 extern struct nls_table *load_nls_default(void);
 #define register_nls(nls) __register_nls((nls), THIS_MODULE)
Signature · 1 line
2.40.1
#8 lore raw

Re: [PATCH AUTOSEL 4.14 7/9] fs/nls: make load_nls() take a const parameter

Quoted 11 lines · Sasha Levin

From: Winston Wen <[email protected]>

[ Upstream commit c1ed39ec116272935528ca9b348b8ee79b0791da ]

load_nls() take a char * parameter, use it to find nls module in list or construct the module name to load it.

This change make load_nls() take a const parameter, so we don't need do some cast like this:

        ses->local_nls = load_nls((char *)ctx->local_nls->charset);

Nice cleanup, but should not be in stable.

Best regards,
Pavel

Quoted 9 lines
---
 fs/nls/nls_base.c   | 4 ++--
 include/linux/nls.h | 2 +-
Signature · 2 lines
DENX Software Engineering GmbH,        Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
author maintainer
#9 lore raw

PATCH 8/9 ASoc: codecs: ES8316: Fix DMIC config

In reply to Sasha Levin

From: Edgar <[email protected]>

[ Upstream commit d20d35d1ad62c6cca36368c1e8f29335a068659e ]

According to the datasheet, the DMIC config should be changed to { 0, 2 ,3 }

---
 sound/soc/codecs/es8316.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
sound/soc/codecs/es8316.c +1 -1
index 0410f2e5183c3..d35a157bad57f 100644
@@ -140,7 +140,7 @@ static const char * const es8316_dmic_txt[] = {
 		"dmic data at high level",
 		"dmic data at low level",
 };
-static const unsigned int es8316_dmic_values[] = { 0, 1, 2 };
+static const unsigned int es8316_dmic_values[] = { 0, 2, 3 };
 static const struct soc_enum es8316_dmic_src_enum =
 	SOC_VALUE_ENUM_SINGLE(ES8316_ADC_DMIC, 0, 3,
 			      ARRAY_SIZE(es8316_dmic_txt),
Signature · 1 line
2.40.1
author maintainer
#10 lore raw

PATCH 9/9 security: keys: perform capable check only on privileged operations

In reply to Sasha Levin

From: Christian Göttsche <[email protected]>

[ Upstream commit 2d7f105edbb3b2be5ffa4d833abbf9b6965e9ce7 ]

If the current task fails the check for the queried capability via capable(CAP_SYS_ADMIN) LSMs like SELinux generate a denial message. Issuing such denial messages unnecessarily can lead to a policy author granting more privileges to a subject than needed to silence them.

Reorder CAP_SYS_ADMIN checks after the check whether the operation is actually privileged.

---
 security/keys/keyctl.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
security/keys/keyctl.c +8 -3
index 9394d72a77e80..9e52a3e0fc672 100644
@@ -922,14 +922,19 @@ long keyctl_chown_key(key_serial_t id, uid_t user, gid_t group)
 	ret = -EACCES;
 	down_write(&key->sem);
 
-	if (!capable(CAP_SYS_ADMIN)) {
+	{
+		bool is_privileged_op = false;
+
 		/* only the sysadmin can chown a key to some other UID */
 		if (user != (uid_t) -1 && !uid_eq(key->uid, uid))
-			goto error_put;
+			is_privileged_op = true;
 
 		/* only the sysadmin can set the key's GID to a group other
 		 * than one of those that the current process subscribes to */
 		if (group != (gid_t) -1 && !gid_eq(gid, key->gid) && !in_group_p(gid))
+			is_privileged_op = true;
+
+		if (is_privileged_op && !capable(CAP_SYS_ADMIN))
 			goto error_put;
 	}
 
@@ -1029,7 +1034,7 @@ long keyctl_setperm_key(key_serial_t id, key_perm_t perm)
 	down_write(&key->sem);
 
 	/* if we're not the sysadmin, we can only change a key that we own */
-	if (capable(CAP_SYS_ADMIN) || uid_eq(key->uid, current_fsuid())) {
+	if (uid_eq(key->uid, current_fsuid()) || capable(CAP_SYS_ADMIN)) {
 		key->perm = perm;
 		ret = 0;
 	}
Signature · 1 line
2.40.1