September 2024 - Linux-stable-mirror

[PATCH 6.1.y] udf: Fold udf_getblk() into udf_bread()

by Sergey Shtylyov

From: Jan Kara <jack(a)suse.cz> udf_getblk() has a single call site. Fold it there. [Sergey: moved back to using udf_get_block() and buffer_{mapped|new}().] Signed-off-by: Jan Kara <jack(a)suse.cz> Signed-off-by: Sergey Shtylyov <s.shtylyov(a)omp.ru> --- This patch prevents NULL pointer dereference in case sb_getblk() fails... fs/udf/inode.c | 45 +++++++++++++++++++-------------------------- 1 file changed, 19 insertions(+), 26 deletions(-) Index: linux-stable/fs/udf/inode.c =================================================================== --- linux-stable.orig/fs/udf/inode.c +++ linux-stable/fs/udf/inode.c @@ -459,30 +459,6 @@ abort: return err; } -static struct buffer_head *udf_getblk(struct inode *inode, udf_pblk_t block, - int create, int *err) -{ - struct buffer_head *bh; - struct buffer_head dummy; - - dummy.b_state = 0; - dummy.b_blocknr = -1000; - *err = udf_get_block(inode, block, &dummy, create); - if (!*err && buffer_mapped(&dummy)) { - bh = sb_getblk(inode->i_sb, dummy.b_blocknr); - if (buffer_new(&dummy)) { - lock_buffer(bh); - memset(bh->b_data, 0x00, inode->i_sb->s_blocksize); - set_buffer_uptodate(bh); - unlock_buffer(bh); - mark_buffer_dirty_inode(bh, inode); - } - return bh; - } - - return NULL; -} - /* Extend the file with new blocks totaling 'new_block_bytes', * return the number of extents added */ @@ -1198,10 +1174,27 @@ struct buffer_head *udf_bread(struct ino int create, int *err) { struct buffer_head *bh = NULL; + struct buffer_head dummy; + + dummy.b_state = 0; + dummy.b_blocknr = -1000; + *err = udf_get_block(inode, block, &dummy, create); + if (*err || !buffer_mapped(&dummy)) + return NULL; - bh = udf_getblk(inode, block, create, err); - if (!bh) + bh = sb_getblk(inode->i_sb, dummy.b_blocknr); + if (!bh) { + *err = -ENOMEM; return NULL; + } + if (buffer_new(&dummy)) { + lock_buffer(bh); + memset(bh->b_data, 0x00, inode->i_sb->s_blocksize); + set_buffer_uptodate(bh); + unlock_buffer(bh); + mark_buffer_dirty_inode(bh, inode); + return bh; + } if (bh_read(bh, 0) >= 0) return bh;

3 months

1
1
0 0

[PATCH] scsi: megaraid: Remove redundant check in megasas_init_fw()

by George Rurikov

This is cosmetic fix. The memory for 'fusion' (instance->ctrl_context) is allocated in megasas_alloc_ctrl_mem() in a call to megasas_alloc_fusion_context(). A possible allocation error is handled after megasas_alloc_ctrl_mem() with a fail_alloc_dma_buf label transition. Found by Linux Verification Center (linuxtesting.org) with SVACE. Cc: stable(a)vger.kernel.org Signed-off-by: George Rurikov <g.ryurikov(a)securitycode.ru> --- drivers/scsi/megaraid/megaraid_sas_base.c | 111 +++++++++++----------- 1 file changed, 54 insertions(+), 57 deletions(-) diff --git a/drivers/scsi/megaraid/megaraid_sas_base.c b/drivers/scsi/megaraid/megaraid_sas_base.c index 6c79c350a4d5..e4823680e009 100644 --- a/drivers/scsi/megaraid/megaraid_sas_base.c +++ b/drivers/scsi/megaraid/megaraid_sas_base.c @@ -6149,68 +6149,65 @@ static int megasas_init_fw(struct megasas_instance *instance) scratch_pad_1 = megasas_readl (instance, &instance->reg_set->outbound_scratch_pad_1); - /* Check max MSI-X vectors */ - if (fusion) { - if (instance->adapter_type == THUNDERBOLT_SERIES) { - /* Thunderbolt Series*/ - instance->msix_vectors = (scratch_pad_1 - & MR_MAX_REPLY_QUEUES_OFFSET) + 1; - } else { - instance->msix_vectors = ((scratch_pad_1 - & MR_MAX_REPLY_QUEUES_EXT_OFFSET) - >> MR_MAX_REPLY_QUEUES_EXT_OFFSET_SHIFT) + 1; - - /* - * For Invader series, > 8 MSI-x vectors - * supported by FW/HW implies combined - * reply queue mode is enabled. - * For Ventura series, > 16 MSI-x vectors - * supported by FW/HW implies combined - * reply queue mode is enabled. - */ - switch (instance->adapter_type) { - case INVADER_SERIES: - if (instance->msix_vectors > 8) - instance->msix_combined = true; - break; - case AERO_SERIES: - case VENTURA_SERIES: - if (instance->msix_vectors > 16) - instance->msix_combined = true; - break; - } - if (rdpq_enable) - instance->is_rdpq = (scratch_pad_1 & MR_RDPQ_MODE_OFFSET) ? - 1 : 0; + if (instance->adapter_type == THUNDERBOLT_SERIES) { + /* Thunderbolt Series*/ + instance->msix_vectors = (scratch_pad_1 + & MR_MAX_REPLY_QUEUES_OFFSET) + 1; + } else { + instance->msix_vectors = ((scratch_pad_1 + & MR_MAX_REPLY_QUEUES_EXT_OFFSET) + >> MR_MAX_REPLY_QUEUES_EXT_OFFSET_SHIFT) + 1; - if (instance->adapter_type >= INVADER_SERIES && - !instance->msix_combined) { - instance->msix_load_balance = true; - instance->smp_affinity_enable = false; - } + /* + * For Invader series, > 8 MSI-x vectors + * supported by FW/HW implies combined + * reply queue mode is enabled. + * For Ventura series, > 16 MSI-x vectors + * supported by FW/HW implies combined + * reply queue mode is enabled. + */ + switch (instance->adapter_type) { + case INVADER_SERIES: + if (instance->msix_vectors > 8) + instance->msix_combined = true; + break; + case AERO_SERIES: + case VENTURA_SERIES: + if (instance->msix_vectors > 16) + instance->msix_combined = true; + break; + } - /* Save 1-15 reply post index address to local memory - * Index 0 is already saved from reg offset - * MPI2_REPLY_POST_HOST_INDEX_OFFSET - */ - for (loop = 1; loop < MR_MAX_MSIX_REG_ARRAY; loop++) { - instance->reply_post_host_index_addr[loop] = - (u32 __iomem *) - ((u8 __iomem *)instance->reg_set + - MPI2_SUP_REPLY_POST_HOST_INDEX_OFFSET - + (loop * 0x10)); - } + if (rdpq_enable) + instance->is_rdpq = (scratch_pad_1 & MR_RDPQ_MODE_OFFSET) ? + 1 : 0; + + if (instance->adapter_type >= INVADER_SERIES && + !instance->msix_combined) { + instance->msix_load_balance = true; + instance->smp_affinity_enable = false; } - dev_info(&instance->pdev->dev, - "firmware supports msix\t: (%d)", - instance->msix_vectors); - if (msix_vectors) - instance->msix_vectors = min(msix_vectors, - instance->msix_vectors); - } else /* MFI adapters */ - instance->msix_vectors = 1; + /* Save 1-15 reply post index address to local memory + * Index 0 is already saved from reg offset + * MPI2_REPLY_POST_HOST_INDEX_OFFSET + */ + for (loop = 1; loop < MR_MAX_MSIX_REG_ARRAY; loop++) { + instance->reply_post_host_index_addr[loop] = + (u32 __iomem *) + ((u8 __iomem *)instance->reg_set + + MPI2_SUP_REPLY_POST_HOST_INDEX_OFFSET + + (loop * 0x10)); + } + } + + dev_info(&instance->pdev->dev, + "firmware supports msix\t: (%d)", + instance->msix_vectors); + if (msix_vectors) + instance->msix_vectors = min(msix_vectors, + instance->msix_vectors); /* -- 2.34.1 Заявление о конфиденциальности Данное электронное письмо и любые приложения к нему являются конфиденциальными и предназначены исключительно для адресата. Если Вы не являетесь адресатом данного письма, пожалуйста, уведомите немедленно отправителя, не раскрывайте содержание другим лицам, не используйте его в каких-либо целях, не храните и не копируйте информацию любым способом.

3 months

1
0
0 0

[PATCH AUTOSEL 6.10 09/18] spi: spidev: Add an entry for elgin,jg10309-01

by Sasha Levin

From: Fabio Estevam <festevam(a)gmail.com> [ Upstream commit 5f3eee1eef5d0edd23d8ac0974f56283649a1512 ] The rv1108-elgin-r1 board has an LCD controlled via SPI in userspace. The marking on the LCD is JG10309-01. Add the "elgin,jg10309-01" compatible string. Signed-off-by: Fabio Estevam <festevam(a)gmail.com> Reviewed-by: Heiko Stuebner <heiko(a)sntech.de> Link: https://patch.msgid.link/20240828180057.3167190-2-festevam@gmail.com Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- drivers/spi/spidev.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/spi/spidev.c b/drivers/spi/spidev.c index 5304728c68c2..14bf0fa65bef 100644 --- a/drivers/spi/spidev.c +++ b/drivers/spi/spidev.c @@ -731,6 +731,7 @@ static int spidev_of_check(struct device *dev) static const struct of_device_id spidev_dt_ids[] = { { .compatible = "cisco,spi-petra", .data = &spidev_of_check }, { .compatible = "dh,dhcom-board", .data = &spidev_of_check }, + { .compatible = "elgin,jg10309-01", .data = &spidev_of_check }, { .compatible = "lineartechnology,ltc2488", .data = &spidev_of_check }, { .compatible = "lwn,bk4", .data = &spidev_of_check }, { .compatible = "menlo,m53cpld", .data = &spidev_of_check }, -- 2.43.0

3 months

2
2
0 0

[PATCH 5.4.y 1/1] inet: inet_defrag: prevent sk release while still in use

by Saeed Mirzamohammadi

From: Florian Westphal <fw(a)strlen.de> [ Upstream commit 18685451fc4e546fc0e718580d32df3c0e5c8272 ] ip_local_out() and other functions can pass skb->sk as function argument. If the skb is a fragment and reassembly happens before such function call returns, the sk must not be released. This affects skb fragments reassembled via netfilter or similar modules, e.g. openvswitch or ct_act.c, when run as part of tx pipeline. Eric Dumazet made an initial analysis of this bug. Quoting Eric: Calling ip_defrag() in output path is also implying skb_orphan(), which is buggy because output path relies on sk not disappearing. A relevant old patch about the issue was : 8282f27449bf ("inet: frag: Always orphan skbs inside ip_defrag()") [..] net/ipv4/ip_output.c depends on skb->sk being set, and probably to an inet socket, not an arbitrary one. If we orphan the packet in ipvlan, then downstream things like FQ packet scheduler will not work properly. We need to change ip_defrag() to only use skb_orphan() when really needed, ie whenever frag_list is going to be used. Eric suggested to stash sk in fragment queue and made an initial patch. However there is a problem with this: If skb is refragmented again right after, ip_do_fragment() will copy head->sk to the new fragments, and sets up destructor to sock_wfree. IOW, we have no choice but to fix up sk_wmem accouting to reflect the fully reassembled skb, else wmem will underflow. This change moves the orphan down into the core, to last possible moment. As ip_defrag_offset is aliased with sk_buff->sk member, we must move the offset into the FRAG_CB, else skb->sk gets clobbered. This allows to delay the orphaning long enough to learn if the skb has to be queued or if the skb is completing the reasm queue. In the former case, things work as before, skb is orphaned. This is safe because skb gets queued/stolen and won't continue past reasm engine. In the latter case, we will steal the skb->sk reference, reattach it to the head skb, and fix up wmem accouting when inet_frag inflates truesize. Fixes: 7026b1ddb6b8 ("netfilter: Pass socket pointer down through okfn().") Diagnosed-by: Eric Dumazet <edumazet(a)google.com> Reported-by: xingwei lee <xrivendell7(a)gmail.com> Reported-by: yue sun <samsun1006219(a)gmail.com> Reported-by: syzbot+e5167d7144a62715044c(a)syzkaller.appspotmail.com Signed-off-by: Florian Westphal <fw(a)strlen.de> Reviewed-by: Eric Dumazet <edumazet(a)google.com> Link: https://lore.kernel.org/r/20240326101845.30836-1-fw@strlen.de Signed-off-by: Paolo Abeni <pabeni(a)redhat.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> (cherry picked from commit 7d0567842b78390dd9b60f00f1d8f838d540e325) CVE: CVE-2024-26921 Cc: stable(a)vger.kernel.org # 5.4 Signed-off-by: Saeed Mirzamohammadi <saeed.mirzamohammadi(a)oracle.com> --- include/linux/skbuff.h | 5 +- net/core/sock_destructor.h | 12 +++++ net/ipv4/inet_fragment.c | 70 ++++++++++++++++++++----- net/ipv4/ip_fragment.c | 2 +- net/ipv6/netfilter/nf_conntrack_reasm.c | 2 +- 5 files changed, 72 insertions(+), 19 deletions(-) create mode 100644 net/core/sock_destructor.h diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index 29ccc33a1c627..3191d0ffc6e9a 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -704,10 +704,7 @@ struct sk_buff { struct list_head list; }; - union { - struct sock *sk; - int ip_defrag_offset; - }; + struct sock *sk; union { ktime_t tstamp; diff --git a/net/core/sock_destructor.h b/net/core/sock_destructor.h new file mode 100644 index 0000000000000..2f396e6bfba5a --- /dev/null +++ b/net/core/sock_destructor.h @@ -0,0 +1,12 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +#ifndef _NET_CORE_SOCK_DESTRUCTOR_H +#define _NET_CORE_SOCK_DESTRUCTOR_H +#include <net/tcp.h> + +static inline bool is_skb_wmem(const struct sk_buff *skb) +{ + return skb->destructor == sock_wfree || + skb->destructor == __sock_wfree || + (IS_ENABLED(CONFIG_INET) && skb->destructor == tcp_wfree); +} +#endif diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c index e0e8a65d561ec..12ef3cb26676d 100644 --- a/net/ipv4/inet_fragment.c +++ b/net/ipv4/inet_fragment.c @@ -24,6 +24,8 @@ #include <net/ip.h> #include <net/ipv6.h> +#include "../core/sock_destructor.h" + /* Use skb->cb to track consecutive/adjacent fragments coming at * the end of the queue. Nodes in the rb-tree queue will * contain "runs" of one or more adjacent fragments. @@ -39,6 +41,7 @@ struct ipfrag_skb_cb { }; struct sk_buff *next_frag; int frag_run_len; + int ip_defrag_offset; }; #define FRAG_CB(skb) ((struct ipfrag_skb_cb *)((skb)->cb)) @@ -359,12 +362,12 @@ int inet_frag_queue_insert(struct inet_frag_queue *q, struct sk_buff *skb, */ if (!last) fragrun_create(q, skb); /* First fragment. */ - else if (last->ip_defrag_offset + last->len < end) { + else if (FRAG_CB(last)->ip_defrag_offset + last->len < end) { /* This is the common case: skb goes to the end. */ /* Detect and discard overlaps. */ - if (offset < last->ip_defrag_offset + last->len) + if (offset < FRAG_CB(last)->ip_defrag_offset + last->len) return IPFRAG_OVERLAP; - if (offset == last->ip_defrag_offset + last->len) + if (offset == FRAG_CB(last)->ip_defrag_offset + last->len) fragrun_append_to_last(q, skb); else fragrun_create(q, skb); @@ -381,13 +384,13 @@ int inet_frag_queue_insert(struct inet_frag_queue *q, struct sk_buff *skb, parent = *rbn; curr = rb_to_skb(parent); - curr_run_end = curr->ip_defrag_offset + + curr_run_end = FRAG_CB(curr)->ip_defrag_offset + FRAG_CB(curr)->frag_run_len; - if (end <= curr->ip_defrag_offset) + if (end <= FRAG_CB(curr)->ip_defrag_offset) rbn = &parent->rb_left; else if (offset >= curr_run_end) rbn = &parent->rb_right; - else if (offset >= curr->ip_defrag_offset && + else if (offset >= FRAG_CB(curr)->ip_defrag_offset && end <= curr_run_end) return IPFRAG_DUP; else @@ -401,7 +404,7 @@ int inet_frag_queue_insert(struct inet_frag_queue *q, struct sk_buff *skb, rb_insert_color(&skb->rbnode, &q->rb_fragments); } - skb->ip_defrag_offset = offset; + FRAG_CB(skb)->ip_defrag_offset = offset; return IPFRAG_OK; } @@ -411,13 +414,28 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, struct sk_buff *parent) { struct sk_buff *fp, *head = skb_rb_first(&q->rb_fragments); - struct sk_buff **nextp; + void (*destructor)(struct sk_buff *); + unsigned int orig_truesize = 0; + struct sk_buff **nextp = NULL; + struct sock *sk = skb->sk; int delta; + if (sk && is_skb_wmem(skb)) { + /* TX: skb->sk might have been passed as argument to + * dst->output and must remain valid until tx completes. + * + * Move sk to reassembled skb and fix up wmem accounting. + */ + orig_truesize = skb->truesize; + destructor = skb->destructor; + } + if (head != skb) { fp = skb_clone(skb, GFP_ATOMIC); - if (!fp) - return NULL; + if (!fp) { + head = skb; + goto out_restore_sk; + } FRAG_CB(fp)->next_frag = FRAG_CB(skb)->next_frag; if (RB_EMPTY_NODE(&skb->rbnode)) FRAG_CB(parent)->next_frag = fp; @@ -426,6 +444,12 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, &q->rb_fragments); if (q->fragments_tail == skb) q->fragments_tail = fp; + + if (orig_truesize) { + /* prevent skb_morph from releasing sk */ + skb->sk = NULL; + skb->destructor = NULL; + } skb_morph(skb, head); FRAG_CB(skb)->next_frag = FRAG_CB(head)->next_frag; rb_replace_node(&head->rbnode, &skb->rbnode, @@ -433,13 +457,13 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, consume_skb(head); head = skb; } - WARN_ON(head->ip_defrag_offset != 0); + WARN_ON(FRAG_CB(head)->ip_defrag_offset != 0); delta = -head->truesize; /* Head of list must not be cloned. */ if (skb_unclone(head, GFP_ATOMIC)) - return NULL; + goto out_restore_sk; delta += head->truesize; if (delta) @@ -455,7 +479,7 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, clone = alloc_skb(0, GFP_ATOMIC); if (!clone) - return NULL; + goto out_restore_sk; skb_shinfo(clone)->frag_list = skb_shinfo(head)->frag_list; skb_frag_list_init(head); for (i = 0; i < skb_shinfo(head)->nr_frags; i++) @@ -472,6 +496,21 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, nextp = &skb_shinfo(head)->frag_list; } +out_restore_sk: + if (orig_truesize) { + int ts_delta = head->truesize - orig_truesize; + + /* if this reassembled skb is fragmented later, + * fraglist skbs will get skb->sk assigned from head->sk, + * and each frag skb will be released via sock_wfree. + * + * Update sk_wmem_alloc. + */ + head->sk = sk; + head->destructor = destructor; + refcount_add(ts_delta, &sk->sk_wmem_alloc); + } + return nextp; } EXPORT_SYMBOL(inet_frag_reasm_prepare); @@ -479,6 +518,8 @@ EXPORT_SYMBOL(inet_frag_reasm_prepare); void inet_frag_reasm_finish(struct inet_frag_queue *q, struct sk_buff *head, void *reasm_data, bool try_coalesce) { + struct sock *sk = is_skb_wmem(head) ? head->sk : NULL; + const unsigned int head_truesize = head->truesize; struct sk_buff **nextp = (struct sk_buff **)reasm_data; struct rb_node *rbn; struct sk_buff *fp; @@ -541,6 +582,9 @@ void inet_frag_reasm_finish(struct inet_frag_queue *q, struct sk_buff *head, skb_mark_not_on_list(head); head->prev = NULL; head->tstamp = q->stamp; + + if (sk) + refcount_add(sum_truesize - head_truesize, &sk->sk_wmem_alloc); } EXPORT_SYMBOL(inet_frag_reasm_finish); diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c index fad803d2d711e..ec2264adf2a6a 100644 --- a/net/ipv4/ip_fragment.c +++ b/net/ipv4/ip_fragment.c @@ -377,6 +377,7 @@ static int ip_frag_queue(struct ipq *qp, struct sk_buff *skb) } skb_dst_drop(skb); + skb_orphan(skb); return -EINPROGRESS; insert_error: @@ -479,7 +480,6 @@ int ip_defrag(struct net *net, struct sk_buff *skb, u32 user) struct ipq *qp; __IP_INC_STATS(net, IPSTATS_MIB_REASMREQDS); - skb_orphan(skb); /* Lookup (or create) queue header */ qp = ip_find(net, ip_hdr(skb), user, vif); diff --git a/net/ipv6/netfilter/nf_conntrack_reasm.c b/net/ipv6/netfilter/nf_conntrack_reasm.c index fed9666a2f7da..cab68c63ea65e 100644 --- a/net/ipv6/netfilter/nf_conntrack_reasm.c +++ b/net/ipv6/netfilter/nf_conntrack_reasm.c @@ -296,6 +296,7 @@ static int nf_ct_frag6_queue(struct frag_queue *fq, struct sk_buff *skb, } skb_dst_drop(skb); + skb_orphan(skb); return -EINPROGRESS; insert_error: @@ -461,7 +462,6 @@ int nf_ct_frag6_gather(struct net *net, struct sk_buff *skb, u32 user) hdr = ipv6_hdr(skb); fhdr = (struct frag_hdr *)skb_transport_header(skb); - skb_orphan(skb); fq = fq_find(net, fhdr->identification, user, hdr, skb->dev ? skb->dev->ifindex : 0); if (fq == NULL) { -- 2.45.2

3 months

1
0
0 0

[PATCH 5.10.y 1/1] inet: inet_defrag: prevent sk release while still in use

by Saeed Mirzamohammadi

From: Florian Westphal <fw(a)strlen.de> [ Upstream commit 18685451fc4e546fc0e718580d32df3c0e5c8272 ] ip_local_out() and other functions can pass skb->sk as function argument. If the skb is a fragment and reassembly happens before such function call returns, the sk must not be released. This affects skb fragments reassembled via netfilter or similar modules, e.g. openvswitch or ct_act.c, when run as part of tx pipeline. Eric Dumazet made an initial analysis of this bug. Quoting Eric: Calling ip_defrag() in output path is also implying skb_orphan(), which is buggy because output path relies on sk not disappearing. A relevant old patch about the issue was : 8282f27449bf ("inet: frag: Always orphan skbs inside ip_defrag()") [..] net/ipv4/ip_output.c depends on skb->sk being set, and probably to an inet socket, not an arbitrary one. If we orphan the packet in ipvlan, then downstream things like FQ packet scheduler will not work properly. We need to change ip_defrag() to only use skb_orphan() when really needed, ie whenever frag_list is going to be used. Eric suggested to stash sk in fragment queue and made an initial patch. However there is a problem with this: If skb is refragmented again right after, ip_do_fragment() will copy head->sk to the new fragments, and sets up destructor to sock_wfree. IOW, we have no choice but to fix up sk_wmem accouting to reflect the fully reassembled skb, else wmem will underflow. This change moves the orphan down into the core, to last possible moment. As ip_defrag_offset is aliased with sk_buff->sk member, we must move the offset into the FRAG_CB, else skb->sk gets clobbered. This allows to delay the orphaning long enough to learn if the skb has to be queued or if the skb is completing the reasm queue. In the former case, things work as before, skb is orphaned. This is safe because skb gets queued/stolen and won't continue past reasm engine. In the latter case, we will steal the skb->sk reference, reattach it to the head skb, and fix up wmem accouting when inet_frag inflates truesize. Fixes: 7026b1ddb6b8 ("netfilter: Pass socket pointer down through okfn().") Diagnosed-by: Eric Dumazet <edumazet(a)google.com> Reported-by: xingwei lee <xrivendell7(a)gmail.com> Reported-by: yue sun <samsun1006219(a)gmail.com> Reported-by: syzbot+e5167d7144a62715044c(a)syzkaller.appspotmail.com Signed-off-by: Florian Westphal <fw(a)strlen.de> Reviewed-by: Eric Dumazet <edumazet(a)google.com> Link: https://lore.kernel.org/r/20240326101845.30836-1-fw@strlen.de Signed-off-by: Paolo Abeni <pabeni(a)redhat.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> (cherry picked from commit 7d0567842b78390dd9b60f00f1d8f838d540e325) CVE: CVE-2024-26921 Cc: stable(a)vger.kernel.org # 5.10 Signed-off-by: Saeed Mirzamohammadi <saeed.mirzamohammadi(a)oracle.com> --- include/linux/skbuff.h | 5 +- net/core/sock_destructor.h | 12 +++++ net/ipv4/inet_fragment.c | 70 ++++++++++++++++++++----- net/ipv4/ip_fragment.c | 2 +- net/ipv6/netfilter/nf_conntrack_reasm.c | 2 +- 5 files changed, 72 insertions(+), 19 deletions(-) create mode 100644 net/core/sock_destructor.h diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index 31755d496b01d..31ae4b74d4352 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -733,10 +733,7 @@ struct sk_buff { struct list_head list; }; - union { - struct sock *sk; - int ip_defrag_offset; - }; + struct sock *sk; union { ktime_t tstamp; diff --git a/net/core/sock_destructor.h b/net/core/sock_destructor.h new file mode 100644 index 0000000000000..2f396e6bfba5a --- /dev/null +++ b/net/core/sock_destructor.h @@ -0,0 +1,12 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +#ifndef _NET_CORE_SOCK_DESTRUCTOR_H +#define _NET_CORE_SOCK_DESTRUCTOR_H +#include <net/tcp.h> + +static inline bool is_skb_wmem(const struct sk_buff *skb) +{ + return skb->destructor == sock_wfree || + skb->destructor == __sock_wfree || + (IS_ENABLED(CONFIG_INET) && skb->destructor == tcp_wfree); +} +#endif diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c index e0e8a65d561ec..12ef3cb26676d 100644 --- a/net/ipv4/inet_fragment.c +++ b/net/ipv4/inet_fragment.c @@ -24,6 +24,8 @@ #include <net/ip.h> #include <net/ipv6.h> +#include "../core/sock_destructor.h" + /* Use skb->cb to track consecutive/adjacent fragments coming at * the end of the queue. Nodes in the rb-tree queue will * contain "runs" of one or more adjacent fragments. @@ -39,6 +41,7 @@ struct ipfrag_skb_cb { }; struct sk_buff *next_frag; int frag_run_len; + int ip_defrag_offset; }; #define FRAG_CB(skb) ((struct ipfrag_skb_cb *)((skb)->cb)) @@ -359,12 +362,12 @@ int inet_frag_queue_insert(struct inet_frag_queue *q, struct sk_buff *skb, */ if (!last) fragrun_create(q, skb); /* First fragment. */ - else if (last->ip_defrag_offset + last->len < end) { + else if (FRAG_CB(last)->ip_defrag_offset + last->len < end) { /* This is the common case: skb goes to the end. */ /* Detect and discard overlaps. */ - if (offset < last->ip_defrag_offset + last->len) + if (offset < FRAG_CB(last)->ip_defrag_offset + last->len) return IPFRAG_OVERLAP; - if (offset == last->ip_defrag_offset + last->len) + if (offset == FRAG_CB(last)->ip_defrag_offset + last->len) fragrun_append_to_last(q, skb); else fragrun_create(q, skb); @@ -381,13 +384,13 @@ int inet_frag_queue_insert(struct inet_frag_queue *q, struct sk_buff *skb, parent = *rbn; curr = rb_to_skb(parent); - curr_run_end = curr->ip_defrag_offset + + curr_run_end = FRAG_CB(curr)->ip_defrag_offset + FRAG_CB(curr)->frag_run_len; - if (end <= curr->ip_defrag_offset) + if (end <= FRAG_CB(curr)->ip_defrag_offset) rbn = &parent->rb_left; else if (offset >= curr_run_end) rbn = &parent->rb_right; - else if (offset >= curr->ip_defrag_offset && + else if (offset >= FRAG_CB(curr)->ip_defrag_offset && end <= curr_run_end) return IPFRAG_DUP; else @@ -401,7 +404,7 @@ int inet_frag_queue_insert(struct inet_frag_queue *q, struct sk_buff *skb, rb_insert_color(&skb->rbnode, &q->rb_fragments); } - skb->ip_defrag_offset = offset; + FRAG_CB(skb)->ip_defrag_offset = offset; return IPFRAG_OK; } @@ -411,13 +414,28 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, struct sk_buff *parent) { struct sk_buff *fp, *head = skb_rb_first(&q->rb_fragments); - struct sk_buff **nextp; + void (*destructor)(struct sk_buff *); + unsigned int orig_truesize = 0; + struct sk_buff **nextp = NULL; + struct sock *sk = skb->sk; int delta; + if (sk && is_skb_wmem(skb)) { + /* TX: skb->sk might have been passed as argument to + * dst->output and must remain valid until tx completes. + * + * Move sk to reassembled skb and fix up wmem accounting. + */ + orig_truesize = skb->truesize; + destructor = skb->destructor; + } + if (head != skb) { fp = skb_clone(skb, GFP_ATOMIC); - if (!fp) - return NULL; + if (!fp) { + head = skb; + goto out_restore_sk; + } FRAG_CB(fp)->next_frag = FRAG_CB(skb)->next_frag; if (RB_EMPTY_NODE(&skb->rbnode)) FRAG_CB(parent)->next_frag = fp; @@ -426,6 +444,12 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, &q->rb_fragments); if (q->fragments_tail == skb) q->fragments_tail = fp; + + if (orig_truesize) { + /* prevent skb_morph from releasing sk */ + skb->sk = NULL; + skb->destructor = NULL; + } skb_morph(skb, head); FRAG_CB(skb)->next_frag = FRAG_CB(head)->next_frag; rb_replace_node(&head->rbnode, &skb->rbnode, @@ -433,13 +457,13 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, consume_skb(head); head = skb; } - WARN_ON(head->ip_defrag_offset != 0); + WARN_ON(FRAG_CB(head)->ip_defrag_offset != 0); delta = -head->truesize; /* Head of list must not be cloned. */ if (skb_unclone(head, GFP_ATOMIC)) - return NULL; + goto out_restore_sk; delta += head->truesize; if (delta) @@ -455,7 +479,7 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, clone = alloc_skb(0, GFP_ATOMIC); if (!clone) - return NULL; + goto out_restore_sk; skb_shinfo(clone)->frag_list = skb_shinfo(head)->frag_list; skb_frag_list_init(head); for (i = 0; i < skb_shinfo(head)->nr_frags; i++) @@ -472,6 +496,21 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, nextp = &skb_shinfo(head)->frag_list; } +out_restore_sk: + if (orig_truesize) { + int ts_delta = head->truesize - orig_truesize; + + /* if this reassembled skb is fragmented later, + * fraglist skbs will get skb->sk assigned from head->sk, + * and each frag skb will be released via sock_wfree. + * + * Update sk_wmem_alloc. + */ + head->sk = sk; + head->destructor = destructor; + refcount_add(ts_delta, &sk->sk_wmem_alloc); + } + return nextp; } EXPORT_SYMBOL(inet_frag_reasm_prepare); @@ -479,6 +518,8 @@ EXPORT_SYMBOL(inet_frag_reasm_prepare); void inet_frag_reasm_finish(struct inet_frag_queue *q, struct sk_buff *head, void *reasm_data, bool try_coalesce) { + struct sock *sk = is_skb_wmem(head) ? head->sk : NULL; + const unsigned int head_truesize = head->truesize; struct sk_buff **nextp = (struct sk_buff **)reasm_data; struct rb_node *rbn; struct sk_buff *fp; @@ -541,6 +582,9 @@ void inet_frag_reasm_finish(struct inet_frag_queue *q, struct sk_buff *head, skb_mark_not_on_list(head); head->prev = NULL; head->tstamp = q->stamp; + + if (sk) + refcount_add(sum_truesize - head_truesize, &sk->sk_wmem_alloc); } EXPORT_SYMBOL(inet_frag_reasm_finish); diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c index fad803d2d711e..ec2264adf2a6a 100644 --- a/net/ipv4/ip_fragment.c +++ b/net/ipv4/ip_fragment.c @@ -377,6 +377,7 @@ static int ip_frag_queue(struct ipq *qp, struct sk_buff *skb) } skb_dst_drop(skb); + skb_orphan(skb); return -EINPROGRESS; insert_error: @@ -479,7 +480,6 @@ int ip_defrag(struct net *net, struct sk_buff *skb, u32 user) struct ipq *qp; __IP_INC_STATS(net, IPSTATS_MIB_REASMREQDS); - skb_orphan(skb); /* Lookup (or create) queue header */ qp = ip_find(net, ip_hdr(skb), user, vif); diff --git a/net/ipv6/netfilter/nf_conntrack_reasm.c b/net/ipv6/netfilter/nf_conntrack_reasm.c index c129ad334eb39..8c2163f95711c 100644 --- a/net/ipv6/netfilter/nf_conntrack_reasm.c +++ b/net/ipv6/netfilter/nf_conntrack_reasm.c @@ -296,6 +296,7 @@ static int nf_ct_frag6_queue(struct frag_queue *fq, struct sk_buff *skb, } skb_dst_drop(skb); + skb_orphan(skb); return -EINPROGRESS; insert_error: @@ -471,7 +472,6 @@ int nf_ct_frag6_gather(struct net *net, struct sk_buff *skb, u32 user) hdr = ipv6_hdr(skb); fhdr = (struct frag_hdr *)skb_transport_header(skb); - skb_orphan(skb); fq = fq_find(net, fhdr->identification, user, hdr, skb->dev ? skb->dev->ifindex : 0); if (fq == NULL) { -- 2.45.2

3 months

1
0
0 0

[PATCH 5.15.y 1/1] inet: inet_defrag: prevent sk release while still in use

by Saeed Mirzamohammadi

From: Florian Westphal <fw(a)strlen.de> [ Upstream commit 18685451fc4e546fc0e718580d32df3c0e5c8272 ] ip_local_out() and other functions can pass skb->sk as function argument. If the skb is a fragment and reassembly happens before such function call returns, the sk must not be released. This affects skb fragments reassembled via netfilter or similar modules, e.g. openvswitch or ct_act.c, when run as part of tx pipeline. Eric Dumazet made an initial analysis of this bug. Quoting Eric: Calling ip_defrag() in output path is also implying skb_orphan(), which is buggy because output path relies on sk not disappearing. A relevant old patch about the issue was : 8282f27449bf ("inet: frag: Always orphan skbs inside ip_defrag()") [..] net/ipv4/ip_output.c depends on skb->sk being set, and probably to an inet socket, not an arbitrary one. If we orphan the packet in ipvlan, then downstream things like FQ packet scheduler will not work properly. We need to change ip_defrag() to only use skb_orphan() when really needed, ie whenever frag_list is going to be used. Eric suggested to stash sk in fragment queue and made an initial patch. However there is a problem with this: If skb is refragmented again right after, ip_do_fragment() will copy head->sk to the new fragments, and sets up destructor to sock_wfree. IOW, we have no choice but to fix up sk_wmem accouting to reflect the fully reassembled skb, else wmem will underflow. This change moves the orphan down into the core, to last possible moment. As ip_defrag_offset is aliased with sk_buff->sk member, we must move the offset into the FRAG_CB, else skb->sk gets clobbered. This allows to delay the orphaning long enough to learn if the skb has to be queued or if the skb is completing the reasm queue. In the former case, things work as before, skb is orphaned. This is safe because skb gets queued/stolen and won't continue past reasm engine. In the latter case, we will steal the skb->sk reference, reattach it to the head skb, and fix up wmem accouting when inet_frag inflates truesize. Fixes: 7026b1ddb6b8 ("netfilter: Pass socket pointer down through okfn().") Diagnosed-by: Eric Dumazet <edumazet(a)google.com> Reported-by: xingwei lee <xrivendell7(a)gmail.com> Reported-by: yue sun <samsun1006219(a)gmail.com> Reported-by: syzbot+e5167d7144a62715044c(a)syzkaller.appspotmail.com Signed-off-by: Florian Westphal <fw(a)strlen.de> Reviewed-by: Eric Dumazet <edumazet(a)google.com> Link: https://lore.kernel.org/r/20240326101845.30836-1-fw@strlen.de Signed-off-by: Paolo Abeni <pabeni(a)redhat.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> (cherry picked from commit 7d0567842b78390dd9b60f00f1d8f838d540e325) CVE: CVE-2024-26921 Cc: stable(a)vger.kernel.org # 5.15 Signed-off-by: Saeed Mirzamohammadi <saeed.mirzamohammadi(a)oracle.com> --- include/linux/skbuff.h | 7 +-- net/ipv4/inet_fragment.c | 70 ++++++++++++++++++++----- net/ipv4/ip_fragment.c | 2 +- net/ipv6/netfilter/nf_conntrack_reasm.c | 2 +- 4 files changed, 60 insertions(+), 21 deletions(-) diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index b230c422dc3b9..7f52562fac19c 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -660,8 +660,6 @@ typedef unsigned char *sk_buff_data_t; * @rbnode: RB tree node, alternative to next/prev for netem/tcp * @list: queue head * @sk: Socket we are owned by - * @ip_defrag_offset: (aka @sk) alternate use of @sk, used in - * fragmentation management * @dev: Device we arrived on/are leaving by * @dev_scratch: (aka @dev) alternate use of @dev when @dev would be %NULL * @cb: Control buffer. Free for use by every layer. Put private vars here @@ -778,10 +776,7 @@ struct sk_buff { struct list_head list; }; - union { - struct sock *sk; - int ip_defrag_offset; - }; + struct sock *sk; union { ktime_t tstamp; diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c index 341096807100c..7e38170111999 100644 --- a/net/ipv4/inet_fragment.c +++ b/net/ipv4/inet_fragment.c @@ -24,6 +24,8 @@ #include <net/ip.h> #include <net/ipv6.h> +#include "../core/sock_destructor.h" + /* Use skb->cb to track consecutive/adjacent fragments coming at * the end of the queue. Nodes in the rb-tree queue will * contain "runs" of one or more adjacent fragments. @@ -39,6 +41,7 @@ struct ipfrag_skb_cb { }; struct sk_buff *next_frag; int frag_run_len; + int ip_defrag_offset; }; #define FRAG_CB(skb) ((struct ipfrag_skb_cb *)((skb)->cb)) @@ -390,12 +393,12 @@ int inet_frag_queue_insert(struct inet_frag_queue *q, struct sk_buff *skb, */ if (!last) fragrun_create(q, skb); /* First fragment. */ - else if (last->ip_defrag_offset + last->len < end) { + else if (FRAG_CB(last)->ip_defrag_offset + last->len < end) { /* This is the common case: skb goes to the end. */ /* Detect and discard overlaps. */ - if (offset < last->ip_defrag_offset + last->len) + if (offset < FRAG_CB(last)->ip_defrag_offset + last->len) return IPFRAG_OVERLAP; - if (offset == last->ip_defrag_offset + last->len) + if (offset == FRAG_CB(last)->ip_defrag_offset + last->len) fragrun_append_to_last(q, skb); else fragrun_create(q, skb); @@ -412,13 +415,13 @@ int inet_frag_queue_insert(struct inet_frag_queue *q, struct sk_buff *skb, parent = *rbn; curr = rb_to_skb(parent); - curr_run_end = curr->ip_defrag_offset + + curr_run_end = FRAG_CB(curr)->ip_defrag_offset + FRAG_CB(curr)->frag_run_len; - if (end <= curr->ip_defrag_offset) + if (end <= FRAG_CB(curr)->ip_defrag_offset) rbn = &parent->rb_left; else if (offset >= curr_run_end) rbn = &parent->rb_right; - else if (offset >= curr->ip_defrag_offset && + else if (offset >= FRAG_CB(curr)->ip_defrag_offset && end <= curr_run_end) return IPFRAG_DUP; else @@ -432,7 +435,7 @@ int inet_frag_queue_insert(struct inet_frag_queue *q, struct sk_buff *skb, rb_insert_color(&skb->rbnode, &q->rb_fragments); } - skb->ip_defrag_offset = offset; + FRAG_CB(skb)->ip_defrag_offset = offset; return IPFRAG_OK; } @@ -442,13 +445,28 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, struct sk_buff *parent) { struct sk_buff *fp, *head = skb_rb_first(&q->rb_fragments); - struct sk_buff **nextp; + void (*destructor)(struct sk_buff *); + unsigned int orig_truesize = 0; + struct sk_buff **nextp = NULL; + struct sock *sk = skb->sk; int delta; + if (sk && is_skb_wmem(skb)) { + /* TX: skb->sk might have been passed as argument to + * dst->output and must remain valid until tx completes. + * + * Move sk to reassembled skb and fix up wmem accounting. + */ + orig_truesize = skb->truesize; + destructor = skb->destructor; + } + if (head != skb) { fp = skb_clone(skb, GFP_ATOMIC); - if (!fp) - return NULL; + if (!fp) { + head = skb; + goto out_restore_sk; + } FRAG_CB(fp)->next_frag = FRAG_CB(skb)->next_frag; if (RB_EMPTY_NODE(&skb->rbnode)) FRAG_CB(parent)->next_frag = fp; @@ -457,6 +475,12 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, &q->rb_fragments); if (q->fragments_tail == skb) q->fragments_tail = fp; + + if (orig_truesize) { + /* prevent skb_morph from releasing sk */ + skb->sk = NULL; + skb->destructor = NULL; + } skb_morph(skb, head); FRAG_CB(skb)->next_frag = FRAG_CB(head)->next_frag; rb_replace_node(&head->rbnode, &skb->rbnode, @@ -464,13 +488,13 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, consume_skb(head); head = skb; } - WARN_ON(head->ip_defrag_offset != 0); + WARN_ON(FRAG_CB(head)->ip_defrag_offset != 0); delta = -head->truesize; /* Head of list must not be cloned. */ if (skb_unclone(head, GFP_ATOMIC)) - return NULL; + goto out_restore_sk; delta += head->truesize; if (delta) @@ -486,7 +510,7 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, clone = alloc_skb(0, GFP_ATOMIC); if (!clone) - return NULL; + goto out_restore_sk; skb_shinfo(clone)->frag_list = skb_shinfo(head)->frag_list; skb_frag_list_init(head); for (i = 0; i < skb_shinfo(head)->nr_frags; i++) @@ -503,6 +527,21 @@ void *inet_frag_reasm_prepare(struct inet_frag_queue *q, struct sk_buff *skb, nextp = &skb_shinfo(head)->frag_list; } +out_restore_sk: + if (orig_truesize) { + int ts_delta = head->truesize - orig_truesize; + + /* if this reassembled skb is fragmented later, + * fraglist skbs will get skb->sk assigned from head->sk, + * and each frag skb will be released via sock_wfree. + * + * Update sk_wmem_alloc. + */ + head->sk = sk; + head->destructor = destructor; + refcount_add(ts_delta, &sk->sk_wmem_alloc); + } + return nextp; } EXPORT_SYMBOL(inet_frag_reasm_prepare); @@ -510,6 +549,8 @@ EXPORT_SYMBOL(inet_frag_reasm_prepare); void inet_frag_reasm_finish(struct inet_frag_queue *q, struct sk_buff *head, void *reasm_data, bool try_coalesce) { + struct sock *sk = is_skb_wmem(head) ? head->sk : NULL; + const unsigned int head_truesize = head->truesize; struct sk_buff **nextp = (struct sk_buff **)reasm_data; struct rb_node *rbn; struct sk_buff *fp; @@ -572,6 +613,9 @@ void inet_frag_reasm_finish(struct inet_frag_queue *q, struct sk_buff *head, skb_mark_not_on_list(head); head->prev = NULL; head->tstamp = q->stamp; + + if (sk) + refcount_add(sum_truesize - head_truesize, &sk->sk_wmem_alloc); } EXPORT_SYMBOL(inet_frag_reasm_finish); diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c index fad803d2d711e..ec2264adf2a6a 100644 --- a/net/ipv4/ip_fragment.c +++ b/net/ipv4/ip_fragment.c @@ -377,6 +377,7 @@ static int ip_frag_queue(struct ipq *qp, struct sk_buff *skb) } skb_dst_drop(skb); + skb_orphan(skb); return -EINPROGRESS; insert_error: @@ -479,7 +480,6 @@ int ip_defrag(struct net *net, struct sk_buff *skb, u32 user) struct ipq *qp; __IP_INC_STATS(net, IPSTATS_MIB_REASMREQDS); - skb_orphan(skb); /* Lookup (or create) queue header */ qp = ip_find(net, ip_hdr(skb), user, vif); diff --git a/net/ipv6/netfilter/nf_conntrack_reasm.c b/net/ipv6/netfilter/nf_conntrack_reasm.c index 2e5b090d7c89f..0ec5ec5a5b45a 100644 --- a/net/ipv6/netfilter/nf_conntrack_reasm.c +++ b/net/ipv6/netfilter/nf_conntrack_reasm.c @@ -297,6 +297,7 @@ static int nf_ct_frag6_queue(struct frag_queue *fq, struct sk_buff *skb, } skb_dst_drop(skb); + skb_orphan(skb); return -EINPROGRESS; insert_error: @@ -472,7 +473,6 @@ int nf_ct_frag6_gather(struct net *net, struct sk_buff *skb, u32 user) hdr = ipv6_hdr(skb); fhdr = (struct frag_hdr *)skb_transport_header(skb); - skb_orphan(skb); fq = fq_find(net, fhdr->identification, user, hdr, skb->dev ? skb->dev->ifindex : 0); if (fq == NULL) { -- 2.45.2

3 months

1
0
0 0

[PATCH v4 5/5] tpm: flush the auth session only when /dev/tpm0 is open

by Jarkko Sakkinen

Instead of flushing and reloading the auth session for every single transaction, keep the session open unless /dev/tpm0 is used. In practice this means applying TPM2_SA_CONTINUE_SESSION to the session attributes. Flush the session always when /dev/tpm0 is written. Reported-by: Pengyu Ma <mapengyu(a)gmail.com> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219229 Cc: stable(a)vger.kernel.org # v6.10+ Fixes: 7ca110f2679b ("tpm: Address !chip->auth in tpm_buf_append_hmac_session*()") Tested-by: Pengyu Ma <mapengyu(a)gmail.com> Signed-off-by: Jarkko Sakkinen <jarkko(a)kernel.org> --- v4: - Changed as bug. v3: - Refined the commit message. - Removed the conditional for applying TPM2_SA_CONTINUE_SESSION only when /dev/tpm0 is open. It is not required as the auth session is flushed, not saved. v2: - A new patch. --- drivers/char/tpm/tpm-chip.c | 1 + drivers/char/tpm/tpm-dev-common.c | 1 + drivers/char/tpm/tpm-interface.c | 1 + drivers/char/tpm/tpm2-sessions.c | 3 +++ 4 files changed, 6 insertions(+) diff --git a/drivers/char/tpm/tpm-chip.c b/drivers/char/tpm/tpm-chip.c index 0ea00e32f575..7a6bb30d1f32 100644 --- a/drivers/char/tpm/tpm-chip.c +++ b/drivers/char/tpm/tpm-chip.c @@ -680,6 +680,7 @@ void tpm_chip_unregister(struct tpm_chip *chip) rc = tpm_try_get_ops(chip); if (!rc) { if (chip->flags & TPM_CHIP_FLAG_TPM2) { + tpm2_end_auth_session(chip); tpm2_flush_context(chip, chip->null_key); chip->null_key = 0; } diff --git a/drivers/char/tpm/tpm-dev-common.c b/drivers/char/tpm/tpm-dev-common.c index 4eaa8e05c291..a3ed7a99a394 100644 --- a/drivers/char/tpm/tpm-dev-common.c +++ b/drivers/char/tpm/tpm-dev-common.c @@ -29,6 +29,7 @@ static ssize_t tpm_dev_transmit(struct tpm_chip *chip, struct tpm_space *space, #ifdef CONFIG_TCG_TPM2_HMAC if (chip->flags & TPM_CHIP_FLAG_TPM2) { + tpm2_end_auth_session(chip); tpm2_flush_context(chip, chip->null_key); chip->null_key = 0; } diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c index bfa47d48b0f2..2363018fa8fb 100644 --- a/drivers/char/tpm/tpm-interface.c +++ b/drivers/char/tpm/tpm-interface.c @@ -381,6 +381,7 @@ int tpm_pm_suspend(struct device *dev) if (!rc) { if (chip->flags & TPM_CHIP_FLAG_TPM2) { #ifdef CONFIG_TCG_TPM2_HMAC + tpm2_end_auth_session(chip); tpm2_flush_context(chip, chip->null_key); chip->null_key = 0; #endif diff --git a/drivers/char/tpm/tpm2-sessions.c b/drivers/char/tpm/tpm2-sessions.c index 6371e0ee88b0..e9d3a6a9d397 100644 --- a/drivers/char/tpm/tpm2-sessions.c +++ b/drivers/char/tpm/tpm2-sessions.c @@ -333,6 +333,9 @@ void tpm_buf_append_hmac_session(struct tpm_chip *chip, struct tpm_buf *buf, } #ifdef CONFIG_TCG_TPM2_HMAC + /* The first write to /dev/tpm{rm0} will flush the session. */ + attributes |= TPM2_SA_CONTINUE_SESSION; + /* * The Architecture Guide requires us to strip trailing zeros * before computing the HMAC -- 2.46.0

3 months

1
0
0 0

[PATCH v4 4/5] tpm: Allocate chip->auth in tpm2_start_auth_session()

by Jarkko Sakkinen

Move allocation of chip->auth to tpm2_start_auth_session() so that the field can be used as flag to tell whether auth session is active or not. Cc: stable(a)vger.kernel.org # v6.10+ Fixes: 699e3efd6c64 ("tpm: Add HMAC session start and end functions") Signed-off-by: Jarkko Sakkinen <jarkko(a)kernel.org> --- v4: - Change to bug. v3: - No changes. v2: - A new patch. --- drivers/char/tpm/tpm2-sessions.c | 43 +++++++++++++++++++------------- 1 file changed, 25 insertions(+), 18 deletions(-) diff --git a/drivers/char/tpm/tpm2-sessions.c b/drivers/char/tpm/tpm2-sessions.c index 42eb910e9acc..6371e0ee88b0 100644 --- a/drivers/char/tpm/tpm2-sessions.c +++ b/drivers/char/tpm/tpm2-sessions.c @@ -484,7 +484,8 @@ static void tpm2_KDFe(u8 z[EC_PT_SZ], const char *str, u8 *pt_u, u8 *pt_v, sha256_final(&sctx, out); } -static void tpm_buf_append_salt(struct tpm_buf *buf, struct tpm_chip *chip) +static void tpm_buf_append_salt(struct tpm_buf *buf, struct tpm_chip *chip, + struct tpm2_auth *auth) { struct crypto_kpp *kpp; struct kpp_request *req; @@ -543,7 +544,7 @@ static void tpm_buf_append_salt(struct tpm_buf *buf, struct tpm_chip *chip) sg_set_buf(&s[0], chip->null_ec_key_x, EC_PT_SZ); sg_set_buf(&s[1], chip->null_ec_key_y, EC_PT_SZ); kpp_request_set_input(req, s, EC_PT_SZ*2); - sg_init_one(d, chip->auth->salt, EC_PT_SZ); + sg_init_one(d, auth->salt, EC_PT_SZ); kpp_request_set_output(req, d, EC_PT_SZ); crypto_kpp_compute_shared_secret(req); kpp_request_free(req); @@ -554,8 +555,7 @@ static void tpm_buf_append_salt(struct tpm_buf *buf, struct tpm_chip *chip) * This works because KDFe fully consumes the secret before it * writes the salt */ - tpm2_KDFe(chip->auth->salt, "SECRET", x, chip->null_ec_key_x, - chip->auth->salt); + tpm2_KDFe(auth->salt, "SECRET", x, chip->null_ec_key_x, auth->salt); out: crypto_free_kpp(kpp); @@ -854,6 +854,8 @@ int tpm_buf_check_hmac_response(struct tpm_chip *chip, struct tpm_buf *buf, /* manually close the session if it wasn't consumed */ tpm2_flush_context(chip, auth->handle); memzero_explicit(auth, sizeof(*auth)); + kfree(auth); + chip->auth = NULL; } else { /* reset for next use */ auth->session = TPM_HEADER_SIZE; @@ -882,6 +884,8 @@ void tpm2_end_auth_session(struct tpm_chip *chip) tpm2_flush_context(chip, auth->handle); memzero_explicit(auth, sizeof(*auth)); + kfree(auth); + chip->auth = NULL; } EXPORT_SYMBOL(tpm2_end_auth_session); @@ -969,25 +973,29 @@ static int tpm2_load_null(struct tpm_chip *chip, u32 *null_key) */ int tpm2_start_auth_session(struct tpm_chip *chip) { + struct tpm2_auth *auth; struct tpm_buf buf; - struct tpm2_auth *auth = chip->auth; - int rc; u32 null_key; + int rc; - if (!auth) { - dev_warn_once(&chip->dev, "auth session is not active\n"); + if (chip->auth) { + dev_warn_once(&chip->dev, "auth session is active\n"); return 0; } + auth = kzalloc(sizeof(*auth), GFP_KERNEL); + if (!auth) + return -ENOMEM; + rc = tpm2_load_null(chip, &null_key); if (rc) - goto out; + goto err; auth->session = TPM_HEADER_SIZE; rc = tpm_buf_init(&buf, TPM2_ST_NO_SESSIONS, TPM2_CC_START_AUTH_SESS); if (rc) - goto out; + goto err; /* salt key handle */ tpm_buf_append_u32(&buf, null_key); @@ -999,7 +1007,7 @@ int tpm2_start_auth_session(struct tpm_chip *chip) tpm_buf_append(&buf, auth->our_nonce, sizeof(auth->our_nonce)); /* append encrypted salt and squirrel away unencrypted in auth */ - tpm_buf_append_salt(&buf, chip); + tpm_buf_append_salt(&buf, chip, auth); /* session type (HMAC, audit or policy) */ tpm_buf_append_u8(&buf, TPM2_SE_HMAC); @@ -1020,10 +1028,13 @@ int tpm2_start_auth_session(struct tpm_chip *chip) tpm_buf_destroy(&buf); - if (rc) - goto out; + if (rc == TPM2_RC_SUCCESS) { + chip->auth = auth; + return 0; + } - out: +err: + kfree(auth); return rc; } EXPORT_SYMBOL(tpm2_start_auth_session); @@ -1375,10 +1386,6 @@ int tpm2_sessions_init(struct tpm_chip *chip) if (rc) return rc; - chip->auth = kmalloc(sizeof(*chip->auth), GFP_KERNEL); - if (!chip->auth) - return -ENOMEM; - return rc; } #endif /* CONFIG_TCG_TPM2_HMAC */ -- 2.46.0

3 months

1
0
0 0

[PATCH v4 3/5] tpm: flush the null key only when /dev/tpm0 is accessed

by Jarkko Sakkinen

Instead of flushing and reloading the null key for every single auth session, flush it only when: 1. User space needs to access /dev/tpm{rm}0. 2. When going to sleep. 3. When unregistering the chip. This removes the need to load and swap the null key between TPM and regular memory per transaction, when the user space is not using the chip. Cc: stable(a)vger.kernel.org # v6.10+ Fixes: d2add27cf2b8 ("tpm: Add NULL primary creation") Tested-by: Pengyu Ma <mapengyu(a)gmail.com> Signed-off-by: Jarkko Sakkinen <jarkko(a)kernel.org> --- v4: - Changed to bug fix as not having the patch there is a major hit to bootup times. v3: - Unchanged. v2: - Refined the commit message. - Added tested-by from Pengyu Ma <mapengyu(a)gmail.com>. - Removed spurious pr_info() statement. --- drivers/char/tpm/tpm-chip.c | 13 +++++++++++++ drivers/char/tpm/tpm-dev-common.c | 7 +++++++ drivers/char/tpm/tpm-interface.c | 9 +++++++-- drivers/char/tpm/tpm2-cmd.c | 3 +++ drivers/char/tpm/tpm2-sessions.c | 17 ++++++++++++++--- include/linux/tpm.h | 2 ++ 6 files changed, 46 insertions(+), 5 deletions(-) diff --git a/drivers/char/tpm/tpm-chip.c b/drivers/char/tpm/tpm-chip.c index 854546000c92..0ea00e32f575 100644 --- a/drivers/char/tpm/tpm-chip.c +++ b/drivers/char/tpm/tpm-chip.c @@ -674,6 +674,19 @@ EXPORT_SYMBOL_GPL(tpm_chip_register); */ void tpm_chip_unregister(struct tpm_chip *chip) { +#ifdef CONFIG_TCG_TPM2_HMAC + int rc; + + rc = tpm_try_get_ops(chip); + if (!rc) { + if (chip->flags & TPM_CHIP_FLAG_TPM2) { + tpm2_flush_context(chip, chip->null_key); + chip->null_key = 0; + } + tpm_put_ops(chip); + } +#endif + tpm_del_legacy_sysfs(chip); if (tpm_is_hwrng_enabled(chip)) hwrng_unregister(&chip->hwrng); diff --git a/drivers/char/tpm/tpm-dev-common.c b/drivers/char/tpm/tpm-dev-common.c index 30b4c288c1bb..4eaa8e05c291 100644 --- a/drivers/char/tpm/tpm-dev-common.c +++ b/drivers/char/tpm/tpm-dev-common.c @@ -27,6 +27,13 @@ static ssize_t tpm_dev_transmit(struct tpm_chip *chip, struct tpm_space *space, struct tpm_header *header = (void *)buf; ssize_t ret, len; +#ifdef CONFIG_TCG_TPM2_HMAC + if (chip->flags & TPM_CHIP_FLAG_TPM2) { + tpm2_flush_context(chip, chip->null_key); + chip->null_key = 0; + } +#endif + ret = tpm2_prepare_space(chip, space, buf, bufsiz); /* If the command is not implemented by the TPM, synthesize a * response with a TPM2_RC_COMMAND_CODE return for user-space. diff --git a/drivers/char/tpm/tpm-interface.c b/drivers/char/tpm/tpm-interface.c index 5da134f12c9a..bfa47d48b0f2 100644 --- a/drivers/char/tpm/tpm-interface.c +++ b/drivers/char/tpm/tpm-interface.c @@ -379,10 +379,15 @@ int tpm_pm_suspend(struct device *dev) rc = tpm_try_get_ops(chip); if (!rc) { - if (chip->flags & TPM_CHIP_FLAG_TPM2) + if (chip->flags & TPM_CHIP_FLAG_TPM2) { +#ifdef CONFIG_TCG_TPM2_HMAC + tpm2_flush_context(chip, chip->null_key); + chip->null_key = 0; +#endif tpm2_shutdown(chip, TPM2_SU_STATE); - else + } else { rc = tpm1_pm_suspend(chip, tpm_suspend_pcr); + } tpm_put_ops(chip); } diff --git a/drivers/char/tpm/tpm2-cmd.c b/drivers/char/tpm/tpm2-cmd.c index 1e856259219e..aba024cbe7c5 100644 --- a/drivers/char/tpm/tpm2-cmd.c +++ b/drivers/char/tpm/tpm2-cmd.c @@ -364,6 +364,9 @@ void tpm2_flush_context(struct tpm_chip *chip, u32 handle) struct tpm_buf buf; int rc; + if (!handle) + return; + rc = tpm_buf_init(&buf, TPM2_ST_NO_SESSIONS, TPM2_CC_FLUSH_CONTEXT); if (rc) { dev_warn(&chip->dev, "0x%08x was not flushed, out of memory\n", diff --git a/drivers/char/tpm/tpm2-sessions.c b/drivers/char/tpm/tpm2-sessions.c index a62f64e21511..42eb910e9acc 100644 --- a/drivers/char/tpm/tpm2-sessions.c +++ b/drivers/char/tpm/tpm2-sessions.c @@ -920,11 +920,19 @@ static int tpm2_load_null(struct tpm_chip *chip, u32 *null_key) u32 tmp_null_key; int rc; + /* fast path */ + if (chip->null_key) { + *null_key = chip->null_key; + return 0; + } + rc = tpm2_load_context(chip, chip->null_key_context, &offset, &tmp_null_key); if (rc != -EINVAL) { - if (!rc) + if (!rc) { + chip->null_key = tmp_null_key; *null_key = tmp_null_key; + } return rc; } dev_info(&chip->dev, "the null key has been reset\n"); @@ -935,6 +943,7 @@ static int tpm2_load_null(struct tpm_chip *chip, u32 *null_key) /* Return the null key if the name has not been changed: */ if (memcmp(name, chip->null_key_name, sizeof(name)) == 0) { + chip->null_key = tmp_null_key; *null_key = tmp_null_key; return 0; } @@ -1005,7 +1014,6 @@ int tpm2_start_auth_session(struct tpm_chip *chip) tpm_buf_append_u16(&buf, TPM_ALG_SHA256); rc = tpm_transmit_cmd(chip, &buf, 0, "start auth session"); - tpm2_flush_context(chip, null_key); if (rc == TPM2_RC_SUCCESS) rc = tpm2_parse_start_auth_session(auth, &buf); @@ -1337,7 +1345,10 @@ static int tpm2_create_null_primary(struct tpm_chip *chip) rc = tpm2_save_context(chip, null_key, chip->null_key_context, sizeof(chip->null_key_context), &offset); - tpm2_flush_context(chip, null_key); + if (rc) + tpm2_flush_context(chip, null_key); + else + chip->null_key = null_key; } if (rc < 0) diff --git a/include/linux/tpm.h b/include/linux/tpm.h index e93ee8d936a9..4eb39db80e05 100644 --- a/include/linux/tpm.h +++ b/include/linux/tpm.h @@ -205,6 +205,8 @@ struct tpm_chip { #ifdef CONFIG_TCG_TPM2_HMAC /* details for communication security via sessions */ + /* loaded null key */ + u32 null_key; /* saved context for NULL seed */ u8 null_key_context[TPM2_MAX_CONTEXT_SIZE]; /* name of NULL seed */ -- 2.46.0

3 months

1
0
0 0

[PATCH v4 2/5] tpm: Return on tpm2_create_primary() failure in tpm2_load_null()

by Jarkko Sakkinen

tpm2_load_null() ignores the return value of tpm2_create_primary(). Further, it does not heal from the situation when memcmp() returns zero. Address this by returning on failure and saving the null key if there was no detected interference in the bus. Cc: stable(a)vger.kernel.org # v6.10+ Fixes: eb24c9788cd9 ("tpm: disable the TPM if NULL name changes") Signed-off-by: Jarkko Sakkinen <jarkko(a)kernel.org> --- v3: - Update log messages. Previously the log message incorrectly stated on load failure that integrity check had been failed, even tho the check is done *after* the load operation. v2: - Refined the commit message. - Reverted tpm2_create_primary() changes. They are not required if tmp_null_key is used as the parameter. --- drivers/char/tpm/tpm2-sessions.c | 38 +++++++++++++++++--------------- 1 file changed, 20 insertions(+), 18 deletions(-) diff --git a/drivers/char/tpm/tpm2-sessions.c b/drivers/char/tpm/tpm2-sessions.c index 795f4c7c6adb..a62f64e21511 100644 --- a/drivers/char/tpm/tpm2-sessions.c +++ b/drivers/char/tpm/tpm2-sessions.c @@ -915,32 +915,34 @@ static int tpm2_parse_start_auth_session(struct tpm2_auth *auth, static int tpm2_load_null(struct tpm_chip *chip, u32 *null_key) { - int rc; unsigned int offset = 0; /* dummy offset for null seed context */ u8 name[SHA256_DIGEST_SIZE + 2]; + u32 tmp_null_key; + int rc; rc = tpm2_load_context(chip, chip->null_key_context, &offset, - null_key); - if (rc != -EINVAL) + &tmp_null_key); + if (rc != -EINVAL) { + if (!rc) + *null_key = tmp_null_key; return rc; + } + dev_info(&chip->dev, "the null key has been reset\n"); - /* an integrity failure may mean the TPM has been reset */ - dev_err(&chip->dev, "NULL key integrity failure!\n"); - /* check the null name against what we know */ - tpm2_create_primary(chip, TPM2_RH_NULL, NULL, name); - if (memcmp(name, chip->null_key_name, sizeof(name)) == 0) - /* name unchanged, assume transient integrity failure */ + rc = tpm2_create_primary(chip, TPM2_RH_NULL, &tmp_null_key, name); + if (rc) return rc; - /* - * Fatal TPM failure: the NULL seed has actually changed, so - * the TPM must have been illegally reset. All in-kernel TPM - * operations will fail because the NULL primary can't be - * loaded to salt the sessions, but disable the TPM anyway so - * userspace programmes can't be compromised by it. - */ - dev_err(&chip->dev, "NULL name has changed, disabling TPM due to interference\n"); - chip->flags |= TPM_CHIP_FLAG_DISABLE; + /* Return the null key if the name has not been changed: */ + if (memcmp(name, chip->null_key_name, sizeof(name)) == 0) { + *null_key = tmp_null_key; + return 0; + } + + /* Deduce from the name change TPM interference: */ + dev_err(&chip->dev, "the null key integrity check failedh\n"); + tpm2_flush_context(chip, tmp_null_key); + chip->flags |= TPM_CHIP_FLAG_DISABLE; return rc; } -- 2.46.0

3 months

1
0
0 0

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror September 2024