Linux-stable-mirror

linux-stable-mirror@lists.linaro.org

340 participants
124247 discussions

FAILED: patch "[PATCH] arm64: dts: qcom: sdm845: Fix slimbam num-channels/ees" failed to apply to 5.15-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.15-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.15.y git checkout FETCH_HEAD git cherry-pick -x 316294bb6695a43a9181973ecd4e6fb3e576a9f7 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025101653-gray-unbroken-db61@gregkh' --subject-prefix 'PATCH 5.15.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 316294bb6695a43a9181973ecd4e6fb3e576a9f7 Mon Sep 17 00:00:00 2001 From: Stephan Gerhold <stephan.gerhold(a)linaro.org> Date: Thu, 21 Aug 2025 10:15:09 +0200 Subject: [PATCH] arm64: dts: qcom: sdm845: Fix slimbam num-channels/ees Reading the hardware registers of the &slimbam on RB3 reveals that the BAM supports only 23 pipes (channels) and supports 4 EEs instead of 2. This hasn't caused problems so far since nothing is using the extra channels, but attempting to use them would lead to crashes. The bam_dma driver might warn in the future if the num-channels in the DT are wrong, so correct the properties in the DT to avoid future regressions. Cc: stable(a)vger.kernel.org Fixes: 27ca1de07dc3 ("arm64: dts: qcom: sdm845: add slimbus nodes") Signed-off-by: Stephan Gerhold <stephan.gerhold(a)linaro.org> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov(a)oss.qualcomm.com> Link: https://lore.kernel.org/r/20250821-sdm845-slimbam-channels-v1-1-498f7d46b9e… Signed-off-by: Bjorn Andersson <andersson(a)kernel.org> diff --git a/arch/arm64/boot/dts/qcom/sdm845.dtsi b/arch/arm64/boot/dts/qcom/sdm845.dtsi index 828b55cb6baf..02536114edb8 100644 --- a/arch/arm64/boot/dts/qcom/sdm845.dtsi +++ b/arch/arm64/boot/dts/qcom/sdm845.dtsi @@ -5396,11 +5396,11 @@ slimbam: dma-controller@17184000 { compatible = "qcom,bam-v1.7.4", "qcom,bam-v1.7.0"; qcom,controlled-remotely; reg = <0 0x17184000 0 0x2a000>; - num-channels = <31>; + num-channels = <23>; interrupts = <GIC_SPI 164 IRQ_TYPE_LEVEL_HIGH>; #dma-cells = <1>; qcom,ee = <1>; - qcom,num-ees = <2>; + qcom,num-ees = <4>; iommus = <&apps_smmu 0x1806 0x0>; };

2 months, 4 weeks

FAILED: patch "[PATCH] ACPI: property: Do not pass NULL handles to" failed to apply to 6.1-stable tree

by gregkh＠linuxfoundation.org

2 months, 4 weeks

[PATCH 6.6.y 0/5] s390/bpf: Tail call counter fixes

by Ilya Leoshkevich

Hi, As discussed in [1], here are the backports of the s390 BPF tail counter fixes. Commits 1-3 are NFC prerequisites, commits 4-5 are fixes themselves. [1] https://lore.kernel.org/stable/CA+G9fYsdErtgqKuyPfFhMS9haGKavBVCHQnipv2EeXM… Best regards, Ilya Ilya Leoshkevich (5): s390/bpf: Change seen_reg to a mask s390/bpf: Centralize frame offset calculations s390/bpf: Describe the frame using a struct instead of constants s390/bpf: Write back tail call counter for BPF_PSEUDO_CALL s390/bpf: Write back tail call counter for BPF_TRAMP_F_CALL_ORIG arch/s390/net/bpf_jit.h | 55 ------------ arch/s390/net/bpf_jit_comp.c | 161 +++++++++++++++++++++-------------- 2 files changed, 97 insertions(+), 119 deletions(-) delete mode 100644 arch/s390/net/bpf_jit.h -- 2.51.0

2 months, 4 weeks

FAILED: patch "[PATCH] arm64: kprobes: call set_memory_rox() for kprobe page" failed to apply to 6.6-stable tree

by gregkh＠linuxfoundation.org

2 months, 4 weeks

FAILED: patch "[PATCH] arm64: dts: qcom: qcs615: add missing dt property in QUP SEs" failed to apply to 6.17-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.17-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.17.y git checkout FETCH_HEAD git cherry-pick -x 6a5e9b9738a32229e2673d4eccfcbfe2ef3a1ab4 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025101637-delegate-oversold-c71e@gregkh' --subject-prefix 'PATCH 6.17.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 6a5e9b9738a32229e2673d4eccfcbfe2ef3a1ab4 Mon Sep 17 00:00:00 2001 From: Viken Dadhaniya <viken.dadhaniya(a)oss.qualcomm.com> Date: Mon, 30 Jun 2025 12:13:38 +0530 Subject: [PATCH] arm64: dts: qcom: qcs615: add missing dt property in QUP SEs Add the missing required-opps and operating-points-v2 properties to several I2C, SPI, and UART nodes in the QUP SEs. Fixes: f6746dc9e379 ("arm64: dts: qcom: qcs615: Add QUPv3 configuration") Cc: stable(a)vger.kernel.org Signed-off-by: Viken Dadhaniya <viken.dadhaniya(a)oss.qualcomm.com> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov(a)oss.qualcomm.com> Link: https://lore.kernel.org/r/20250630064338.2487409-1-viken.dadhaniya@oss.qual… Signed-off-by: Bjorn Andersson <andersson(a)kernel.org> diff --git a/arch/arm64/boot/dts/qcom/sm6150.dtsi b/arch/arm64/boot/dts/qcom/sm6150.dtsi index bfbb21035492..e033b53f0f0f 100644 --- a/arch/arm64/boot/dts/qcom/sm6150.dtsi +++ b/arch/arm64/boot/dts/qcom/sm6150.dtsi @@ -631,6 +631,7 @@ &mc_virt SLAVE_EBI1 QCOM_ICC_TAG_ALWAYS>, interconnect-names = "qup-core", "qup-config"; power-domains = <&rpmhpd RPMHPD_CX>; + operating-points-v2 = <&qup_opp_table>; status = "disabled"; }; @@ -654,6 +655,7 @@ &config_noc SLAVE_QUP_0 QCOM_ICC_TAG_ALWAYS>, "qup-config", "qup-memory"; power-domains = <&rpmhpd RPMHPD_CX>; + required-opps = <&rpmhpd_opp_low_svs>; dmas = <&gpi_dma0 0 1 QCOM_GPI_I2C>, <&gpi_dma0 1 1 QCOM_GPI_I2C>; dma-names = "tx", @@ -681,6 +683,7 @@ &config_noc SLAVE_QUP_0 QCOM_ICC_TAG_ALWAYS>, "qup-config", "qup-memory"; power-domains = <&rpmhpd RPMHPD_CX>; + required-opps = <&rpmhpd_opp_low_svs>; dmas = <&gpi_dma0 0 2 QCOM_GPI_I2C>, <&gpi_dma0 1 2 QCOM_GPI_I2C>; dma-names = "tx", @@ -703,6 +706,7 @@ &mc_virt SLAVE_EBI1 QCOM_ICC_TAG_ALWAYS>, interconnect-names = "qup-core", "qup-config"; power-domains = <&rpmhpd RPMHPD_CX>; + operating-points-v2 = <&qup_opp_table>; dmas = <&gpi_dma0 0 2 QCOM_GPI_SPI>, <&gpi_dma0 1 2 QCOM_GPI_SPI>; dma-names = "tx", @@ -728,6 +732,7 @@ &mc_virt SLAVE_EBI1 QCOM_ICC_TAG_ALWAYS>, interconnect-names = "qup-core", "qup-config"; power-domains = <&rpmhpd RPMHPD_CX>; + operating-points-v2 = <&qup_opp_table>; status = "disabled"; }; @@ -751,6 +756,7 @@ &config_noc SLAVE_QUP_0 QCOM_ICC_TAG_ALWAYS>, "qup-config", "qup-memory"; power-domains = <&rpmhpd RPMHPD_CX>; + required-opps = <&rpmhpd_opp_low_svs>; dmas = <&gpi_dma0 0 3 QCOM_GPI_I2C>, <&gpi_dma0 1 3 QCOM_GPI_I2C>; dma-names = "tx",

2 months, 4 weeks

FAILED: patch "[PATCH] ACPI: property: Do not pass NULL handles to" failed to apply to 6.6-stable tree

by gregkh＠linuxfoundation.org

2 months, 4 weeks

FAILED: patch "[PATCH] ACPI: property: Do not pass NULL handles to" failed to apply to 6.12-stable tree

by gregkh＠linuxfoundation.org

2 months, 4 weeks

FAILED: patch "[PATCH] ipmi: Rework user message limit handling" failed to apply to 6.1-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.1-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.1.y git checkout FETCH_HEAD git cherry-pick -x b52da4054ee0bf9ecb44996f2c83236ff50b3812 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025101647-unsteady-antitrust-ef9e@gregkh' --subject-prefix 'PATCH 6.1.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From b52da4054ee0bf9ecb44996f2c83236ff50b3812 Mon Sep 17 00:00:00 2001 From: Corey Minyard <corey(a)minyard.net> Date: Fri, 5 Sep 2025 11:33:39 -0500 Subject: [PATCH] ipmi: Rework user message limit handling The limit on the number of user messages had a number of issues, improper counting in some cases and a use after free. Restructure how this is all done to handle more in the receive message allocation routine, so all refcouting and user message limit counts are done in that routine. It's a lot cleaner and safer. Reported-by: Gilles BULOZ <gilles.buloz(a)kontron.com> Closes: https://lore.kernel.org/lkml/aLsw6G0GyqfpKs2S@mail.minyard.net/ Fixes: 8e76741c3d8b ("ipmi: Add a limit on the number of users that may use IPMI") Cc: <stable(a)vger.kernel.org> # 4.19 Signed-off-by: Corey Minyard <corey(a)minyard.net> Tested-by: Gilles BULOZ <gilles.buloz(a)kontron.com> diff --git a/drivers/char/ipmi/ipmi_msghandler.c b/drivers/char/ipmi/ipmi_msghandler.c index b78cc359534d..d2fbf2203bd2 100644 --- a/drivers/char/ipmi/ipmi_msghandler.c +++ b/drivers/char/ipmi/ipmi_msghandler.c @@ -38,7 +38,9 @@ #define IPMI_DRIVER_VERSION "39.2" -static struct ipmi_recv_msg *ipmi_alloc_recv_msg(void); +static struct ipmi_recv_msg *ipmi_alloc_recv_msg(struct ipmi_user *user); +static void ipmi_set_recv_msg_user(struct ipmi_recv_msg *msg, + struct ipmi_user *user); static int ipmi_init_msghandler(void); static void smi_work(struct work_struct *t); static void handle_new_recv_msgs(struct ipmi_smi *intf); @@ -955,7 +957,6 @@ static int deliver_response(struct ipmi_smi *intf, struct ipmi_recv_msg *msg) * risk. At this moment, simply skip it in that case. */ ipmi_free_recv_msg(msg); - atomic_dec(&msg->user->nr_msgs); } else { /* * Deliver it in smi_work. The message will hold a @@ -1611,8 +1612,7 @@ int ipmi_set_gets_events(struct ipmi_user *user, bool val) } list_for_each_entry_safe(msg, msg2, &msgs, link) { - msg->user = user; - kref_get(&user->refcount); + ipmi_set_recv_msg_user(msg, user); deliver_local_response(intf, msg); } } @@ -2277,22 +2277,15 @@ static int i_ipmi_request(struct ipmi_user *user, int run_to_completion = READ_ONCE(intf->run_to_completion); int rv = 0; - if (user) { - if (atomic_add_return(1, &user->nr_msgs) > max_msgs_per_user) { - /* Decrement will happen at the end of the routine. */ - rv = -EBUSY; - goto out; - } - } - - if (supplied_recv) + if (supplied_recv) { recv_msg = supplied_recv; - else { - recv_msg = ipmi_alloc_recv_msg(); - if (recv_msg == NULL) { - rv = -ENOMEM; - goto out; - } + recv_msg->user = user; + if (user) + atomic_inc(&user->nr_msgs); + } else { + recv_msg = ipmi_alloc_recv_msg(user); + if (IS_ERR(recv_msg)) + return PTR_ERR(recv_msg); } recv_msg->user_msg_data = user_msg_data; @@ -2303,8 +2296,7 @@ static int i_ipmi_request(struct ipmi_user *user, if (smi_msg == NULL) { if (!supplied_recv) ipmi_free_recv_msg(recv_msg); - rv = -ENOMEM; - goto out; + return -ENOMEM; } } @@ -2315,10 +2307,6 @@ static int i_ipmi_request(struct ipmi_user *user, goto out_err; } - recv_msg->user = user; - if (user) - /* The put happens when the message is freed. */ - kref_get(&user->refcount); recv_msg->msgid = msgid; /* * Store the message to send in the receive message so timeout @@ -2347,8 +2335,10 @@ static int i_ipmi_request(struct ipmi_user *user, if (rv) { out_err: - ipmi_free_smi_msg(smi_msg); - ipmi_free_recv_msg(recv_msg); + if (!supplied_smi) + ipmi_free_smi_msg(smi_msg); + if (!supplied_recv) + ipmi_free_recv_msg(recv_msg); } else { dev_dbg(intf->si_dev, "Send: %*ph\n", smi_msg->data_size, smi_msg->data); @@ -2358,9 +2348,6 @@ static int i_ipmi_request(struct ipmi_user *user, if (!run_to_completion) mutex_unlock(&intf->users_mutex); -out: - if (rv && user) - atomic_dec(&user->nr_msgs); return rv; } @@ -3851,7 +3838,7 @@ static int handle_ipmb_get_msg_cmd(struct ipmi_smi *intf, unsigned char chan; struct ipmi_user *user = NULL; struct ipmi_ipmb_addr *ipmb_addr; - struct ipmi_recv_msg *recv_msg; + struct ipmi_recv_msg *recv_msg = NULL; if (msg->rsp_size < 10) { /* Message not big enough, just ignore it. */ @@ -3872,9 +3859,8 @@ static int handle_ipmb_get_msg_cmd(struct ipmi_smi *intf, rcvr = find_cmd_rcvr(intf, netfn, cmd, chan); if (rcvr) { user = rcvr->user; - kref_get(&user->refcount); - } else - user = NULL; + recv_msg = ipmi_alloc_recv_msg(user); + } rcu_read_unlock(); if (user == NULL) { @@ -3904,47 +3890,41 @@ static int handle_ipmb_get_msg_cmd(struct ipmi_smi *intf, * causes it to not be freed or queued. */ rv = -1; + } else if (!IS_ERR(recv_msg)) { + /* Extract the source address from the data. */ + ipmb_addr = (struct ipmi_ipmb_addr *) &recv_msg->addr; + ipmb_addr->addr_type = IPMI_IPMB_ADDR_TYPE; + ipmb_addr->slave_addr = msg->rsp[6]; + ipmb_addr->lun = msg->rsp[7] & 3; + ipmb_addr->channel = msg->rsp[3] & 0xf; + + /* + * Extract the rest of the message information + * from the IPMB header. + */ + recv_msg->recv_type = IPMI_CMD_RECV_TYPE; + recv_msg->msgid = msg->rsp[7] >> 2; + recv_msg->msg.netfn = msg->rsp[4] >> 2; + recv_msg->msg.cmd = msg->rsp[8]; + recv_msg->msg.data = recv_msg->msg_data; + + /* + * We chop off 10, not 9 bytes because the checksum + * at the end also needs to be removed. + */ + recv_msg->msg.data_len = msg->rsp_size - 10; + memcpy(recv_msg->msg_data, &msg->rsp[9], + msg->rsp_size - 10); + if (deliver_response(intf, recv_msg)) + ipmi_inc_stat(intf, unhandled_commands); + else + ipmi_inc_stat(intf, handled_commands); } else { - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { - /* - * We couldn't allocate memory for the - * message, so requeue it for handling - * later. - */ - rv = 1; - kref_put(&user->refcount, free_ipmi_user); - } else { - /* Extract the source address from the data. */ - ipmb_addr = (struct ipmi_ipmb_addr *) &recv_msg->addr; - ipmb_addr->addr_type = IPMI_IPMB_ADDR_TYPE; - ipmb_addr->slave_addr = msg->rsp[6]; - ipmb_addr->lun = msg->rsp[7] & 3; - ipmb_addr->channel = msg->rsp[3] & 0xf; - - /* - * Extract the rest of the message information - * from the IPMB header. - */ - recv_msg->user = user; - recv_msg->recv_type = IPMI_CMD_RECV_TYPE; - recv_msg->msgid = msg->rsp[7] >> 2; - recv_msg->msg.netfn = msg->rsp[4] >> 2; - recv_msg->msg.cmd = msg->rsp[8]; - recv_msg->msg.data = recv_msg->msg_data; - - /* - * We chop off 10, not 9 bytes because the checksum - * at the end also needs to be removed. - */ - recv_msg->msg.data_len = msg->rsp_size - 10; - memcpy(recv_msg->msg_data, &msg->rsp[9], - msg->rsp_size - 10); - if (deliver_response(intf, recv_msg)) - ipmi_inc_stat(intf, unhandled_commands); - else - ipmi_inc_stat(intf, handled_commands); - } + /* + * We couldn't allocate memory for the message, so + * requeue it for handling later. + */ + rv = 1; } return rv; @@ -3957,7 +3937,7 @@ static int handle_ipmb_direct_rcv_cmd(struct ipmi_smi *intf, int rv = 0; struct ipmi_user *user = NULL; struct ipmi_ipmb_direct_addr *daddr; - struct ipmi_recv_msg *recv_msg; + struct ipmi_recv_msg *recv_msg = NULL; unsigned char netfn = msg->rsp[0] >> 2; unsigned char cmd = msg->rsp[3]; @@ -3966,9 +3946,8 @@ static int handle_ipmb_direct_rcv_cmd(struct ipmi_smi *intf, rcvr = find_cmd_rcvr(intf, netfn, cmd, 0); if (rcvr) { user = rcvr->user; - kref_get(&user->refcount); - } else - user = NULL; + recv_msg = ipmi_alloc_recv_msg(user); + } rcu_read_unlock(); if (user == NULL) { @@ -3990,44 +3969,38 @@ static int handle_ipmb_direct_rcv_cmd(struct ipmi_smi *intf, * causes it to not be freed or queued. */ rv = -1; + } else if (!IS_ERR(recv_msg)) { + /* Extract the source address from the data. */ + daddr = (struct ipmi_ipmb_direct_addr *)&recv_msg->addr; + daddr->addr_type = IPMI_IPMB_DIRECT_ADDR_TYPE; + daddr->channel = 0; + daddr->slave_addr = msg->rsp[1]; + daddr->rs_lun = msg->rsp[0] & 3; + daddr->rq_lun = msg->rsp[2] & 3; + + /* + * Extract the rest of the message information + * from the IPMB header. + */ + recv_msg->recv_type = IPMI_CMD_RECV_TYPE; + recv_msg->msgid = (msg->rsp[2] >> 2); + recv_msg->msg.netfn = msg->rsp[0] >> 2; + recv_msg->msg.cmd = msg->rsp[3]; + recv_msg->msg.data = recv_msg->msg_data; + + recv_msg->msg.data_len = msg->rsp_size - 4; + memcpy(recv_msg->msg_data, msg->rsp + 4, + msg->rsp_size - 4); + if (deliver_response(intf, recv_msg)) + ipmi_inc_stat(intf, unhandled_commands); + else + ipmi_inc_stat(intf, handled_commands); } else { - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { - /* - * We couldn't allocate memory for the - * message, so requeue it for handling - * later. - */ - rv = 1; - kref_put(&user->refcount, free_ipmi_user); - } else { - /* Extract the source address from the data. */ - daddr = (struct ipmi_ipmb_direct_addr *)&recv_msg->addr; - daddr->addr_type = IPMI_IPMB_DIRECT_ADDR_TYPE; - daddr->channel = 0; - daddr->slave_addr = msg->rsp[1]; - daddr->rs_lun = msg->rsp[0] & 3; - daddr->rq_lun = msg->rsp[2] & 3; - - /* - * Extract the rest of the message information - * from the IPMB header. - */ - recv_msg->user = user; - recv_msg->recv_type = IPMI_CMD_RECV_TYPE; - recv_msg->msgid = (msg->rsp[2] >> 2); - recv_msg->msg.netfn = msg->rsp[0] >> 2; - recv_msg->msg.cmd = msg->rsp[3]; - recv_msg->msg.data = recv_msg->msg_data; - - recv_msg->msg.data_len = msg->rsp_size - 4; - memcpy(recv_msg->msg_data, msg->rsp + 4, - msg->rsp_size - 4); - if (deliver_response(intf, recv_msg)) - ipmi_inc_stat(intf, unhandled_commands); - else - ipmi_inc_stat(intf, handled_commands); - } + /* + * We couldn't allocate memory for the message, so + * requeue it for handling later. + */ + rv = 1; } return rv; @@ -4141,7 +4114,7 @@ static int handle_lan_get_msg_cmd(struct ipmi_smi *intf, unsigned char chan; struct ipmi_user *user = NULL; struct ipmi_lan_addr *lan_addr; - struct ipmi_recv_msg *recv_msg; + struct ipmi_recv_msg *recv_msg = NULL; if (msg->rsp_size < 12) { /* Message not big enough, just ignore it. */ @@ -4162,9 +4135,8 @@ static int handle_lan_get_msg_cmd(struct ipmi_smi *intf, rcvr = find_cmd_rcvr(intf, netfn, cmd, chan); if (rcvr) { user = rcvr->user; - kref_get(&user->refcount); - } else - user = NULL; + recv_msg = ipmi_alloc_recv_msg(user); + } rcu_read_unlock(); if (user == NULL) { @@ -4195,49 +4167,44 @@ static int handle_lan_get_msg_cmd(struct ipmi_smi *intf, * causes it to not be freed or queued. */ rv = -1; + } else if (!IS_ERR(recv_msg)) { + /* Extract the source address from the data. */ + lan_addr = (struct ipmi_lan_addr *) &recv_msg->addr; + lan_addr->addr_type = IPMI_LAN_ADDR_TYPE; + lan_addr->session_handle = msg->rsp[4]; + lan_addr->remote_SWID = msg->rsp[8]; + lan_addr->local_SWID = msg->rsp[5]; + lan_addr->lun = msg->rsp[9] & 3; + lan_addr->channel = msg->rsp[3] & 0xf; + lan_addr->privilege = msg->rsp[3] >> 4; + + /* + * Extract the rest of the message information + * from the IPMB header. + */ + recv_msg->recv_type = IPMI_CMD_RECV_TYPE; + recv_msg->msgid = msg->rsp[9] >> 2; + recv_msg->msg.netfn = msg->rsp[6] >> 2; + recv_msg->msg.cmd = msg->rsp[10]; + recv_msg->msg.data = recv_msg->msg_data; + + /* + * We chop off 12, not 11 bytes because the checksum + * at the end also needs to be removed. + */ + recv_msg->msg.data_len = msg->rsp_size - 12; + memcpy(recv_msg->msg_data, &msg->rsp[11], + msg->rsp_size - 12); + if (deliver_response(intf, recv_msg)) + ipmi_inc_stat(intf, unhandled_commands); + else + ipmi_inc_stat(intf, handled_commands); } else { - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { - /* - * We couldn't allocate memory for the - * message, so requeue it for handling later. - */ - rv = 1; - kref_put(&user->refcount, free_ipmi_user); - } else { - /* Extract the source address from the data. */ - lan_addr = (struct ipmi_lan_addr *) &recv_msg->addr; - lan_addr->addr_type = IPMI_LAN_ADDR_TYPE; - lan_addr->session_handle = msg->rsp[4]; - lan_addr->remote_SWID = msg->rsp[8]; - lan_addr->local_SWID = msg->rsp[5]; - lan_addr->lun = msg->rsp[9] & 3; - lan_addr->channel = msg->rsp[3] & 0xf; - lan_addr->privilege = msg->rsp[3] >> 4; - - /* - * Extract the rest of the message information - * from the IPMB header. - */ - recv_msg->user = user; - recv_msg->recv_type = IPMI_CMD_RECV_TYPE; - recv_msg->msgid = msg->rsp[9] >> 2; - recv_msg->msg.netfn = msg->rsp[6] >> 2; - recv_msg->msg.cmd = msg->rsp[10]; - recv_msg->msg.data = recv_msg->msg_data; - - /* - * We chop off 12, not 11 bytes because the checksum - * at the end also needs to be removed. - */ - recv_msg->msg.data_len = msg->rsp_size - 12; - memcpy(recv_msg->msg_data, &msg->rsp[11], - msg->rsp_size - 12); - if (deliver_response(intf, recv_msg)) - ipmi_inc_stat(intf, unhandled_commands); - else - ipmi_inc_stat(intf, handled_commands); - } + /* + * We couldn't allocate memory for the message, so + * requeue it for handling later. + */ + rv = 1; } return rv; @@ -4259,7 +4226,7 @@ static int handle_oem_get_msg_cmd(struct ipmi_smi *intf, unsigned char chan; struct ipmi_user *user = NULL; struct ipmi_system_interface_addr *smi_addr; - struct ipmi_recv_msg *recv_msg; + struct ipmi_recv_msg *recv_msg = NULL; /* * We expect the OEM SW to perform error checking @@ -4288,9 +4255,8 @@ static int handle_oem_get_msg_cmd(struct ipmi_smi *intf, rcvr = find_cmd_rcvr(intf, netfn, cmd, chan); if (rcvr) { user = rcvr->user; - kref_get(&user->refcount); - } else - user = NULL; + recv_msg = ipmi_alloc_recv_msg(user); + } rcu_read_unlock(); if (user == NULL) { @@ -4303,48 +4269,42 @@ static int handle_oem_get_msg_cmd(struct ipmi_smi *intf, */ rv = 0; + } else if (!IS_ERR(recv_msg)) { + /* + * OEM Messages are expected to be delivered via + * the system interface to SMS software. We might + * need to visit this again depending on OEM + * requirements + */ + smi_addr = ((struct ipmi_system_interface_addr *) + &recv_msg->addr); + smi_addr->addr_type = IPMI_SYSTEM_INTERFACE_ADDR_TYPE; + smi_addr->channel = IPMI_BMC_CHANNEL; + smi_addr->lun = msg->rsp[0] & 3; + + recv_msg->user_msg_data = NULL; + recv_msg->recv_type = IPMI_OEM_RECV_TYPE; + recv_msg->msg.netfn = msg->rsp[0] >> 2; + recv_msg->msg.cmd = msg->rsp[1]; + recv_msg->msg.data = recv_msg->msg_data; + + /* + * The message starts at byte 4 which follows the + * Channel Byte in the "GET MESSAGE" command + */ + recv_msg->msg.data_len = msg->rsp_size - 4; + memcpy(recv_msg->msg_data, &msg->rsp[4], + msg->rsp_size - 4); + if (deliver_response(intf, recv_msg)) + ipmi_inc_stat(intf, unhandled_commands); + else + ipmi_inc_stat(intf, handled_commands); } else { - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { - /* - * We couldn't allocate memory for the - * message, so requeue it for handling - * later. - */ - rv = 1; - kref_put(&user->refcount, free_ipmi_user); - } else { - /* - * OEM Messages are expected to be delivered via - * the system interface to SMS software. We might - * need to visit this again depending on OEM - * requirements - */ - smi_addr = ((struct ipmi_system_interface_addr *) - &recv_msg->addr); - smi_addr->addr_type = IPMI_SYSTEM_INTERFACE_ADDR_TYPE; - smi_addr->channel = IPMI_BMC_CHANNEL; - smi_addr->lun = msg->rsp[0] & 3; - - recv_msg->user = user; - recv_msg->user_msg_data = NULL; - recv_msg->recv_type = IPMI_OEM_RECV_TYPE; - recv_msg->msg.netfn = msg->rsp[0] >> 2; - recv_msg->msg.cmd = msg->rsp[1]; - recv_msg->msg.data = recv_msg->msg_data; - - /* - * The message starts at byte 4 which follows the - * Channel Byte in the "GET MESSAGE" command - */ - recv_msg->msg.data_len = msg->rsp_size - 4; - memcpy(recv_msg->msg_data, &msg->rsp[4], - msg->rsp_size - 4); - if (deliver_response(intf, recv_msg)) - ipmi_inc_stat(intf, unhandled_commands); - else - ipmi_inc_stat(intf, handled_commands); - } + /* + * We couldn't allocate memory for the message, so + * requeue it for handling later. + */ + rv = 1; } return rv; @@ -4402,8 +4362,8 @@ static int handle_read_event_rsp(struct ipmi_smi *intf, if (!user->gets_events) continue; - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { + recv_msg = ipmi_alloc_recv_msg(user); + if (IS_ERR(recv_msg)) { mutex_unlock(&intf->users_mutex); list_for_each_entry_safe(recv_msg, recv_msg2, &msgs, link) { @@ -4424,8 +4384,6 @@ static int handle_read_event_rsp(struct ipmi_smi *intf, deliver_count++; copy_event_into_recv_msg(recv_msg, msg); - recv_msg->user = user; - kref_get(&user->refcount); list_add_tail(&recv_msg->link, &msgs); } mutex_unlock(&intf->users_mutex); @@ -4441,8 +4399,8 @@ static int handle_read_event_rsp(struct ipmi_smi *intf, * No one to receive the message, put it in queue if there's * not already too many things in the queue. */ - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { + recv_msg = ipmi_alloc_recv_msg(NULL); + if (IS_ERR(recv_msg)) { /* * We couldn't allocate memory for the * message, so requeue it for handling @@ -4858,12 +4816,10 @@ static void smi_work(struct work_struct *t) list_del(&msg->link); - if (refcount_read(&user->destroyed) == 0) { + if (refcount_read(&user->destroyed) == 0) ipmi_free_recv_msg(msg); - } else { - atomic_dec(&user->nr_msgs); + else user->handler->ipmi_recv_hndl(msg, user->handler_data); - } } mutex_unlock(&intf->user_msgs_mutex); @@ -5179,27 +5135,51 @@ static void free_recv_msg(struct ipmi_recv_msg *msg) kfree(msg); } -static struct ipmi_recv_msg *ipmi_alloc_recv_msg(void) +static struct ipmi_recv_msg *ipmi_alloc_recv_msg(struct ipmi_user *user) { struct ipmi_recv_msg *rv; - rv = kmalloc(sizeof(struct ipmi_recv_msg), GFP_ATOMIC); - if (rv) { - rv->user = NULL; - rv->done = free_recv_msg; - atomic_inc(&recv_msg_inuse_count); + if (user) { + if (atomic_add_return(1, &user->nr_msgs) > max_msgs_per_user) { + atomic_dec(&user->nr_msgs); + return ERR_PTR(-EBUSY); + } } + + rv = kmalloc(sizeof(struct ipmi_recv_msg), GFP_ATOMIC); + if (!rv) { + if (user) + atomic_dec(&user->nr_msgs); + return ERR_PTR(-ENOMEM); + } + + rv->user = user; + rv->done = free_recv_msg; + if (user) + kref_get(&user->refcount); + atomic_inc(&recv_msg_inuse_count); return rv; } void ipmi_free_recv_msg(struct ipmi_recv_msg *msg) { - if (msg->user && !oops_in_progress) + if (msg->user && !oops_in_progress) { + atomic_dec(&msg->user->nr_msgs); kref_put(&msg->user->refcount, free_ipmi_user); + } msg->done(msg); } EXPORT_SYMBOL(ipmi_free_recv_msg); +static void ipmi_set_recv_msg_user(struct ipmi_recv_msg *msg, + struct ipmi_user *user) +{ + WARN_ON_ONCE(msg->user); /* User should not be set. */ + msg->user = user; + atomic_inc(&user->nr_msgs); + kref_get(&user->refcount); +} + static atomic_t panic_done_count = ATOMIC_INIT(0); static void dummy_smi_done_handler(struct ipmi_smi_msg *msg)

2 months, 4 weeks

FAILED: patch "[PATCH] ipmi: Rework user message limit handling" failed to apply to 6.6-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.6-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.6.y git checkout FETCH_HEAD git cherry-pick -x b52da4054ee0bf9ecb44996f2c83236ff50b3812 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025101647-undated-train-2b88@gregkh' --subject-prefix 'PATCH 6.6.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From b52da4054ee0bf9ecb44996f2c83236ff50b3812 Mon Sep 17 00:00:00 2001 From: Corey Minyard <corey(a)minyard.net> Date: Fri, 5 Sep 2025 11:33:39 -0500 Subject: [PATCH] ipmi: Rework user message limit handling The limit on the number of user messages had a number of issues, improper counting in some cases and a use after free. Restructure how this is all done to handle more in the receive message allocation routine, so all refcouting and user message limit counts are done in that routine. It's a lot cleaner and safer. Reported-by: Gilles BULOZ <gilles.buloz(a)kontron.com> Closes: https://lore.kernel.org/lkml/aLsw6G0GyqfpKs2S@mail.minyard.net/ Fixes: 8e76741c3d8b ("ipmi: Add a limit on the number of users that may use IPMI") Cc: <stable(a)vger.kernel.org> # 4.19 Signed-off-by: Corey Minyard <corey(a)minyard.net> Tested-by: Gilles BULOZ <gilles.buloz(a)kontron.com> diff --git a/drivers/char/ipmi/ipmi_msghandler.c b/drivers/char/ipmi/ipmi_msghandler.c index b78cc359534d..d2fbf2203bd2 100644 --- a/drivers/char/ipmi/ipmi_msghandler.c +++ b/drivers/char/ipmi/ipmi_msghandler.c @@ -38,7 +38,9 @@ #define IPMI_DRIVER_VERSION "39.2" -static struct ipmi_recv_msg *ipmi_alloc_recv_msg(void); +static struct ipmi_recv_msg *ipmi_alloc_recv_msg(struct ipmi_user *user); +static void ipmi_set_recv_msg_user(struct ipmi_recv_msg *msg, + struct ipmi_user *user); static int ipmi_init_msghandler(void); static void smi_work(struct work_struct *t); static void handle_new_recv_msgs(struct ipmi_smi *intf); @@ -955,7 +957,6 @@ static int deliver_response(struct ipmi_smi *intf, struct ipmi_recv_msg *msg) * risk. At this moment, simply skip it in that case. */ ipmi_free_recv_msg(msg); - atomic_dec(&msg->user->nr_msgs); } else { /* * Deliver it in smi_work. The message will hold a @@ -1611,8 +1612,7 @@ int ipmi_set_gets_events(struct ipmi_user *user, bool val) } list_for_each_entry_safe(msg, msg2, &msgs, link) { - msg->user = user; - kref_get(&user->refcount); + ipmi_set_recv_msg_user(msg, user); deliver_local_response(intf, msg); } } @@ -2277,22 +2277,15 @@ static int i_ipmi_request(struct ipmi_user *user, int run_to_completion = READ_ONCE(intf->run_to_completion); int rv = 0; - if (user) { - if (atomic_add_return(1, &user->nr_msgs) > max_msgs_per_user) { - /* Decrement will happen at the end of the routine. */ - rv = -EBUSY; - goto out; - } - } - - if (supplied_recv) + if (supplied_recv) { recv_msg = supplied_recv; - else { - recv_msg = ipmi_alloc_recv_msg(); - if (recv_msg == NULL) { - rv = -ENOMEM; - goto out; - } + recv_msg->user = user; + if (user) + atomic_inc(&user->nr_msgs); + } else { + recv_msg = ipmi_alloc_recv_msg(user); + if (IS_ERR(recv_msg)) + return PTR_ERR(recv_msg); } recv_msg->user_msg_data = user_msg_data; @@ -2303,8 +2296,7 @@ static int i_ipmi_request(struct ipmi_user *user, if (smi_msg == NULL) { if (!supplied_recv) ipmi_free_recv_msg(recv_msg); - rv = -ENOMEM; - goto out; + return -ENOMEM; } } @@ -2315,10 +2307,6 @@ static int i_ipmi_request(struct ipmi_user *user, goto out_err; } - recv_msg->user = user; - if (user) - /* The put happens when the message is freed. */ - kref_get(&user->refcount); recv_msg->msgid = msgid; /* * Store the message to send in the receive message so timeout @@ -2347,8 +2335,10 @@ static int i_ipmi_request(struct ipmi_user *user, if (rv) { out_err: - ipmi_free_smi_msg(smi_msg); - ipmi_free_recv_msg(recv_msg); + if (!supplied_smi) + ipmi_free_smi_msg(smi_msg); + if (!supplied_recv) + ipmi_free_recv_msg(recv_msg); } else { dev_dbg(intf->si_dev, "Send: %*ph\n", smi_msg->data_size, smi_msg->data); @@ -2358,9 +2348,6 @@ static int i_ipmi_request(struct ipmi_user *user, if (!run_to_completion) mutex_unlock(&intf->users_mutex); -out: - if (rv && user) - atomic_dec(&user->nr_msgs); return rv; } @@ -3851,7 +3838,7 @@ static int handle_ipmb_get_msg_cmd(struct ipmi_smi *intf, unsigned char chan; struct ipmi_user *user = NULL; struct ipmi_ipmb_addr *ipmb_addr; - struct ipmi_recv_msg *recv_msg; + struct ipmi_recv_msg *recv_msg = NULL; if (msg->rsp_size < 10) { /* Message not big enough, just ignore it. */ @@ -3872,9 +3859,8 @@ static int handle_ipmb_get_msg_cmd(struct ipmi_smi *intf, rcvr = find_cmd_rcvr(intf, netfn, cmd, chan); if (rcvr) { user = rcvr->user; - kref_get(&user->refcount); - } else - user = NULL; + recv_msg = ipmi_alloc_recv_msg(user); + } rcu_read_unlock(); if (user == NULL) { @@ -3904,47 +3890,41 @@ static int handle_ipmb_get_msg_cmd(struct ipmi_smi *intf, * causes it to not be freed or queued. */ rv = -1; + } else if (!IS_ERR(recv_msg)) { + /* Extract the source address from the data. */ + ipmb_addr = (struct ipmi_ipmb_addr *) &recv_msg->addr; + ipmb_addr->addr_type = IPMI_IPMB_ADDR_TYPE; + ipmb_addr->slave_addr = msg->rsp[6]; + ipmb_addr->lun = msg->rsp[7] & 3; + ipmb_addr->channel = msg->rsp[3] & 0xf; + + /* + * Extract the rest of the message information + * from the IPMB header. + */ + recv_msg->recv_type = IPMI_CMD_RECV_TYPE; + recv_msg->msgid = msg->rsp[7] >> 2; + recv_msg->msg.netfn = msg->rsp[4] >> 2; + recv_msg->msg.cmd = msg->rsp[8]; + recv_msg->msg.data = recv_msg->msg_data; + + /* + * We chop off 10, not 9 bytes because the checksum + * at the end also needs to be removed. + */ + recv_msg->msg.data_len = msg->rsp_size - 10; + memcpy(recv_msg->msg_data, &msg->rsp[9], + msg->rsp_size - 10); + if (deliver_response(intf, recv_msg)) + ipmi_inc_stat(intf, unhandled_commands); + else + ipmi_inc_stat(intf, handled_commands); } else { - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { - /* - * We couldn't allocate memory for the - * message, so requeue it for handling - * later. - */ - rv = 1; - kref_put(&user->refcount, free_ipmi_user); - } else { - /* Extract the source address from the data. */ - ipmb_addr = (struct ipmi_ipmb_addr *) &recv_msg->addr; - ipmb_addr->addr_type = IPMI_IPMB_ADDR_TYPE; - ipmb_addr->slave_addr = msg->rsp[6]; - ipmb_addr->lun = msg->rsp[7] & 3; - ipmb_addr->channel = msg->rsp[3] & 0xf; - - /* - * Extract the rest of the message information - * from the IPMB header. - */ - recv_msg->user = user; - recv_msg->recv_type = IPMI_CMD_RECV_TYPE; - recv_msg->msgid = msg->rsp[7] >> 2; - recv_msg->msg.netfn = msg->rsp[4] >> 2; - recv_msg->msg.cmd = msg->rsp[8]; - recv_msg->msg.data = recv_msg->msg_data; - - /* - * We chop off 10, not 9 bytes because the checksum - * at the end also needs to be removed. - */ - recv_msg->msg.data_len = msg->rsp_size - 10; - memcpy(recv_msg->msg_data, &msg->rsp[9], - msg->rsp_size - 10); - if (deliver_response(intf, recv_msg)) - ipmi_inc_stat(intf, unhandled_commands); - else - ipmi_inc_stat(intf, handled_commands); - } + /* + * We couldn't allocate memory for the message, so + * requeue it for handling later. + */ + rv = 1; } return rv; @@ -3957,7 +3937,7 @@ static int handle_ipmb_direct_rcv_cmd(struct ipmi_smi *intf, int rv = 0; struct ipmi_user *user = NULL; struct ipmi_ipmb_direct_addr *daddr; - struct ipmi_recv_msg *recv_msg; + struct ipmi_recv_msg *recv_msg = NULL; unsigned char netfn = msg->rsp[0] >> 2; unsigned char cmd = msg->rsp[3]; @@ -3966,9 +3946,8 @@ static int handle_ipmb_direct_rcv_cmd(struct ipmi_smi *intf, rcvr = find_cmd_rcvr(intf, netfn, cmd, 0); if (rcvr) { user = rcvr->user; - kref_get(&user->refcount); - } else - user = NULL; + recv_msg = ipmi_alloc_recv_msg(user); + } rcu_read_unlock(); if (user == NULL) { @@ -3990,44 +3969,38 @@ static int handle_ipmb_direct_rcv_cmd(struct ipmi_smi *intf, * causes it to not be freed or queued. */ rv = -1; + } else if (!IS_ERR(recv_msg)) { + /* Extract the source address from the data. */ + daddr = (struct ipmi_ipmb_direct_addr *)&recv_msg->addr; + daddr->addr_type = IPMI_IPMB_DIRECT_ADDR_TYPE; + daddr->channel = 0; + daddr->slave_addr = msg->rsp[1]; + daddr->rs_lun = msg->rsp[0] & 3; + daddr->rq_lun = msg->rsp[2] & 3; + + /* + * Extract the rest of the message information + * from the IPMB header. + */ + recv_msg->recv_type = IPMI_CMD_RECV_TYPE; + recv_msg->msgid = (msg->rsp[2] >> 2); + recv_msg->msg.netfn = msg->rsp[0] >> 2; + recv_msg->msg.cmd = msg->rsp[3]; + recv_msg->msg.data = recv_msg->msg_data; + + recv_msg->msg.data_len = msg->rsp_size - 4; + memcpy(recv_msg->msg_data, msg->rsp + 4, + msg->rsp_size - 4); + if (deliver_response(intf, recv_msg)) + ipmi_inc_stat(intf, unhandled_commands); + else + ipmi_inc_stat(intf, handled_commands); } else { - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { - /* - * We couldn't allocate memory for the - * message, so requeue it for handling - * later. - */ - rv = 1; - kref_put(&user->refcount, free_ipmi_user); - } else { - /* Extract the source address from the data. */ - daddr = (struct ipmi_ipmb_direct_addr *)&recv_msg->addr; - daddr->addr_type = IPMI_IPMB_DIRECT_ADDR_TYPE; - daddr->channel = 0; - daddr->slave_addr = msg->rsp[1]; - daddr->rs_lun = msg->rsp[0] & 3; - daddr->rq_lun = msg->rsp[2] & 3; - - /* - * Extract the rest of the message information - * from the IPMB header. - */ - recv_msg->user = user; - recv_msg->recv_type = IPMI_CMD_RECV_TYPE; - recv_msg->msgid = (msg->rsp[2] >> 2); - recv_msg->msg.netfn = msg->rsp[0] >> 2; - recv_msg->msg.cmd = msg->rsp[3]; - recv_msg->msg.data = recv_msg->msg_data; - - recv_msg->msg.data_len = msg->rsp_size - 4; - memcpy(recv_msg->msg_data, msg->rsp + 4, - msg->rsp_size - 4); - if (deliver_response(intf, recv_msg)) - ipmi_inc_stat(intf, unhandled_commands); - else - ipmi_inc_stat(intf, handled_commands); - } + /* + * We couldn't allocate memory for the message, so + * requeue it for handling later. + */ + rv = 1; } return rv; @@ -4141,7 +4114,7 @@ static int handle_lan_get_msg_cmd(struct ipmi_smi *intf, unsigned char chan; struct ipmi_user *user = NULL; struct ipmi_lan_addr *lan_addr; - struct ipmi_recv_msg *recv_msg; + struct ipmi_recv_msg *recv_msg = NULL; if (msg->rsp_size < 12) { /* Message not big enough, just ignore it. */ @@ -4162,9 +4135,8 @@ static int handle_lan_get_msg_cmd(struct ipmi_smi *intf, rcvr = find_cmd_rcvr(intf, netfn, cmd, chan); if (rcvr) { user = rcvr->user; - kref_get(&user->refcount); - } else - user = NULL; + recv_msg = ipmi_alloc_recv_msg(user); + } rcu_read_unlock(); if (user == NULL) { @@ -4195,49 +4167,44 @@ static int handle_lan_get_msg_cmd(struct ipmi_smi *intf, * causes it to not be freed or queued. */ rv = -1; + } else if (!IS_ERR(recv_msg)) { + /* Extract the source address from the data. */ + lan_addr = (struct ipmi_lan_addr *) &recv_msg->addr; + lan_addr->addr_type = IPMI_LAN_ADDR_TYPE; + lan_addr->session_handle = msg->rsp[4]; + lan_addr->remote_SWID = msg->rsp[8]; + lan_addr->local_SWID = msg->rsp[5]; + lan_addr->lun = msg->rsp[9] & 3; + lan_addr->channel = msg->rsp[3] & 0xf; + lan_addr->privilege = msg->rsp[3] >> 4; + + /* + * Extract the rest of the message information + * from the IPMB header. + */ + recv_msg->recv_type = IPMI_CMD_RECV_TYPE; + recv_msg->msgid = msg->rsp[9] >> 2; + recv_msg->msg.netfn = msg->rsp[6] >> 2; + recv_msg->msg.cmd = msg->rsp[10]; + recv_msg->msg.data = recv_msg->msg_data; + + /* + * We chop off 12, not 11 bytes because the checksum + * at the end also needs to be removed. + */ + recv_msg->msg.data_len = msg->rsp_size - 12; + memcpy(recv_msg->msg_data, &msg->rsp[11], + msg->rsp_size - 12); + if (deliver_response(intf, recv_msg)) + ipmi_inc_stat(intf, unhandled_commands); + else + ipmi_inc_stat(intf, handled_commands); } else { - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { - /* - * We couldn't allocate memory for the - * message, so requeue it for handling later. - */ - rv = 1; - kref_put(&user->refcount, free_ipmi_user); - } else { - /* Extract the source address from the data. */ - lan_addr = (struct ipmi_lan_addr *) &recv_msg->addr; - lan_addr->addr_type = IPMI_LAN_ADDR_TYPE; - lan_addr->session_handle = msg->rsp[4]; - lan_addr->remote_SWID = msg->rsp[8]; - lan_addr->local_SWID = msg->rsp[5]; - lan_addr->lun = msg->rsp[9] & 3; - lan_addr->channel = msg->rsp[3] & 0xf; - lan_addr->privilege = msg->rsp[3] >> 4; - - /* - * Extract the rest of the message information - * from the IPMB header. - */ - recv_msg->user = user; - recv_msg->recv_type = IPMI_CMD_RECV_TYPE; - recv_msg->msgid = msg->rsp[9] >> 2; - recv_msg->msg.netfn = msg->rsp[6] >> 2; - recv_msg->msg.cmd = msg->rsp[10]; - recv_msg->msg.data = recv_msg->msg_data; - - /* - * We chop off 12, not 11 bytes because the checksum - * at the end also needs to be removed. - */ - recv_msg->msg.data_len = msg->rsp_size - 12; - memcpy(recv_msg->msg_data, &msg->rsp[11], - msg->rsp_size - 12); - if (deliver_response(intf, recv_msg)) - ipmi_inc_stat(intf, unhandled_commands); - else - ipmi_inc_stat(intf, handled_commands); - } + /* + * We couldn't allocate memory for the message, so + * requeue it for handling later. + */ + rv = 1; } return rv; @@ -4259,7 +4226,7 @@ static int handle_oem_get_msg_cmd(struct ipmi_smi *intf, unsigned char chan; struct ipmi_user *user = NULL; struct ipmi_system_interface_addr *smi_addr; - struct ipmi_recv_msg *recv_msg; + struct ipmi_recv_msg *recv_msg = NULL; /* * We expect the OEM SW to perform error checking @@ -4288,9 +4255,8 @@ static int handle_oem_get_msg_cmd(struct ipmi_smi *intf, rcvr = find_cmd_rcvr(intf, netfn, cmd, chan); if (rcvr) { user = rcvr->user; - kref_get(&user->refcount); - } else - user = NULL; + recv_msg = ipmi_alloc_recv_msg(user); + } rcu_read_unlock(); if (user == NULL) { @@ -4303,48 +4269,42 @@ static int handle_oem_get_msg_cmd(struct ipmi_smi *intf, */ rv = 0; + } else if (!IS_ERR(recv_msg)) { + /* + * OEM Messages are expected to be delivered via + * the system interface to SMS software. We might + * need to visit this again depending on OEM + * requirements + */ + smi_addr = ((struct ipmi_system_interface_addr *) + &recv_msg->addr); + smi_addr->addr_type = IPMI_SYSTEM_INTERFACE_ADDR_TYPE; + smi_addr->channel = IPMI_BMC_CHANNEL; + smi_addr->lun = msg->rsp[0] & 3; + + recv_msg->user_msg_data = NULL; + recv_msg->recv_type = IPMI_OEM_RECV_TYPE; + recv_msg->msg.netfn = msg->rsp[0] >> 2; + recv_msg->msg.cmd = msg->rsp[1]; + recv_msg->msg.data = recv_msg->msg_data; + + /* + * The message starts at byte 4 which follows the + * Channel Byte in the "GET MESSAGE" command + */ + recv_msg->msg.data_len = msg->rsp_size - 4; + memcpy(recv_msg->msg_data, &msg->rsp[4], + msg->rsp_size - 4); + if (deliver_response(intf, recv_msg)) + ipmi_inc_stat(intf, unhandled_commands); + else + ipmi_inc_stat(intf, handled_commands); } else { - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { - /* - * We couldn't allocate memory for the - * message, so requeue it for handling - * later. - */ - rv = 1; - kref_put(&user->refcount, free_ipmi_user); - } else { - /* - * OEM Messages are expected to be delivered via - * the system interface to SMS software. We might - * need to visit this again depending on OEM - * requirements - */ - smi_addr = ((struct ipmi_system_interface_addr *) - &recv_msg->addr); - smi_addr->addr_type = IPMI_SYSTEM_INTERFACE_ADDR_TYPE; - smi_addr->channel = IPMI_BMC_CHANNEL; - smi_addr->lun = msg->rsp[0] & 3; - - recv_msg->user = user; - recv_msg->user_msg_data = NULL; - recv_msg->recv_type = IPMI_OEM_RECV_TYPE; - recv_msg->msg.netfn = msg->rsp[0] >> 2; - recv_msg->msg.cmd = msg->rsp[1]; - recv_msg->msg.data = recv_msg->msg_data; - - /* - * The message starts at byte 4 which follows the - * Channel Byte in the "GET MESSAGE" command - */ - recv_msg->msg.data_len = msg->rsp_size - 4; - memcpy(recv_msg->msg_data, &msg->rsp[4], - msg->rsp_size - 4); - if (deliver_response(intf, recv_msg)) - ipmi_inc_stat(intf, unhandled_commands); - else - ipmi_inc_stat(intf, handled_commands); - } + /* + * We couldn't allocate memory for the message, so + * requeue it for handling later. + */ + rv = 1; } return rv; @@ -4402,8 +4362,8 @@ static int handle_read_event_rsp(struct ipmi_smi *intf, if (!user->gets_events) continue; - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { + recv_msg = ipmi_alloc_recv_msg(user); + if (IS_ERR(recv_msg)) { mutex_unlock(&intf->users_mutex); list_for_each_entry_safe(recv_msg, recv_msg2, &msgs, link) { @@ -4424,8 +4384,6 @@ static int handle_read_event_rsp(struct ipmi_smi *intf, deliver_count++; copy_event_into_recv_msg(recv_msg, msg); - recv_msg->user = user; - kref_get(&user->refcount); list_add_tail(&recv_msg->link, &msgs); } mutex_unlock(&intf->users_mutex); @@ -4441,8 +4399,8 @@ static int handle_read_event_rsp(struct ipmi_smi *intf, * No one to receive the message, put it in queue if there's * not already too many things in the queue. */ - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { + recv_msg = ipmi_alloc_recv_msg(NULL); + if (IS_ERR(recv_msg)) { /* * We couldn't allocate memory for the * message, so requeue it for handling @@ -4858,12 +4816,10 @@ static void smi_work(struct work_struct *t) list_del(&msg->link); - if (refcount_read(&user->destroyed) == 0) { + if (refcount_read(&user->destroyed) == 0) ipmi_free_recv_msg(msg); - } else { - atomic_dec(&user->nr_msgs); + else user->handler->ipmi_recv_hndl(msg, user->handler_data); - } } mutex_unlock(&intf->user_msgs_mutex); @@ -5179,27 +5135,51 @@ static void free_recv_msg(struct ipmi_recv_msg *msg) kfree(msg); } -static struct ipmi_recv_msg *ipmi_alloc_recv_msg(void) +static struct ipmi_recv_msg *ipmi_alloc_recv_msg(struct ipmi_user *user) { struct ipmi_recv_msg *rv; - rv = kmalloc(sizeof(struct ipmi_recv_msg), GFP_ATOMIC); - if (rv) { - rv->user = NULL; - rv->done = free_recv_msg; - atomic_inc(&recv_msg_inuse_count); + if (user) { + if (atomic_add_return(1, &user->nr_msgs) > max_msgs_per_user) { + atomic_dec(&user->nr_msgs); + return ERR_PTR(-EBUSY); + } } + + rv = kmalloc(sizeof(struct ipmi_recv_msg), GFP_ATOMIC); + if (!rv) { + if (user) + atomic_dec(&user->nr_msgs); + return ERR_PTR(-ENOMEM); + } + + rv->user = user; + rv->done = free_recv_msg; + if (user) + kref_get(&user->refcount); + atomic_inc(&recv_msg_inuse_count); return rv; } void ipmi_free_recv_msg(struct ipmi_recv_msg *msg) { - if (msg->user && !oops_in_progress) + if (msg->user && !oops_in_progress) { + atomic_dec(&msg->user->nr_msgs); kref_put(&msg->user->refcount, free_ipmi_user); + } msg->done(msg); } EXPORT_SYMBOL(ipmi_free_recv_msg); +static void ipmi_set_recv_msg_user(struct ipmi_recv_msg *msg, + struct ipmi_user *user) +{ + WARN_ON_ONCE(msg->user); /* User should not be set. */ + msg->user = user; + atomic_inc(&user->nr_msgs); + kref_get(&user->refcount); +} + static atomic_t panic_done_count = ATOMIC_INIT(0); static void dummy_smi_done_handler(struct ipmi_smi_msg *msg)

2 months, 4 weeks

FAILED: patch "[PATCH] ipmi: Rework user message limit handling" failed to apply to 6.12-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.12-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.12.y git checkout FETCH_HEAD git cherry-pick -x b52da4054ee0bf9ecb44996f2c83236ff50b3812 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025101646-stretch-gyration-6815@gregkh' --subject-prefix 'PATCH 6.12.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From b52da4054ee0bf9ecb44996f2c83236ff50b3812 Mon Sep 17 00:00:00 2001 From: Corey Minyard <corey(a)minyard.net> Date: Fri, 5 Sep 2025 11:33:39 -0500 Subject: [PATCH] ipmi: Rework user message limit handling The limit on the number of user messages had a number of issues, improper counting in some cases and a use after free. Restructure how this is all done to handle more in the receive message allocation routine, so all refcouting and user message limit counts are done in that routine. It's a lot cleaner and safer. Reported-by: Gilles BULOZ <gilles.buloz(a)kontron.com> Closes: https://lore.kernel.org/lkml/aLsw6G0GyqfpKs2S@mail.minyard.net/ Fixes: 8e76741c3d8b ("ipmi: Add a limit on the number of users that may use IPMI") Cc: <stable(a)vger.kernel.org> # 4.19 Signed-off-by: Corey Minyard <corey(a)minyard.net> Tested-by: Gilles BULOZ <gilles.buloz(a)kontron.com> diff --git a/drivers/char/ipmi/ipmi_msghandler.c b/drivers/char/ipmi/ipmi_msghandler.c index b78cc359534d..d2fbf2203bd2 100644 --- a/drivers/char/ipmi/ipmi_msghandler.c +++ b/drivers/char/ipmi/ipmi_msghandler.c @@ -38,7 +38,9 @@ #define IPMI_DRIVER_VERSION "39.2" -static struct ipmi_recv_msg *ipmi_alloc_recv_msg(void); +static struct ipmi_recv_msg *ipmi_alloc_recv_msg(struct ipmi_user *user); +static void ipmi_set_recv_msg_user(struct ipmi_recv_msg *msg, + struct ipmi_user *user); static int ipmi_init_msghandler(void); static void smi_work(struct work_struct *t); static void handle_new_recv_msgs(struct ipmi_smi *intf); @@ -955,7 +957,6 @@ static int deliver_response(struct ipmi_smi *intf, struct ipmi_recv_msg *msg) * risk. At this moment, simply skip it in that case. */ ipmi_free_recv_msg(msg); - atomic_dec(&msg->user->nr_msgs); } else { /* * Deliver it in smi_work. The message will hold a @@ -1611,8 +1612,7 @@ int ipmi_set_gets_events(struct ipmi_user *user, bool val) } list_for_each_entry_safe(msg, msg2, &msgs, link) { - msg->user = user; - kref_get(&user->refcount); + ipmi_set_recv_msg_user(msg, user); deliver_local_response(intf, msg); } } @@ -2277,22 +2277,15 @@ static int i_ipmi_request(struct ipmi_user *user, int run_to_completion = READ_ONCE(intf->run_to_completion); int rv = 0; - if (user) { - if (atomic_add_return(1, &user->nr_msgs) > max_msgs_per_user) { - /* Decrement will happen at the end of the routine. */ - rv = -EBUSY; - goto out; - } - } - - if (supplied_recv) + if (supplied_recv) { recv_msg = supplied_recv; - else { - recv_msg = ipmi_alloc_recv_msg(); - if (recv_msg == NULL) { - rv = -ENOMEM; - goto out; - } + recv_msg->user = user; + if (user) + atomic_inc(&user->nr_msgs); + } else { + recv_msg = ipmi_alloc_recv_msg(user); + if (IS_ERR(recv_msg)) + return PTR_ERR(recv_msg); } recv_msg->user_msg_data = user_msg_data; @@ -2303,8 +2296,7 @@ static int i_ipmi_request(struct ipmi_user *user, if (smi_msg == NULL) { if (!supplied_recv) ipmi_free_recv_msg(recv_msg); - rv = -ENOMEM; - goto out; + return -ENOMEM; } } @@ -2315,10 +2307,6 @@ static int i_ipmi_request(struct ipmi_user *user, goto out_err; } - recv_msg->user = user; - if (user) - /* The put happens when the message is freed. */ - kref_get(&user->refcount); recv_msg->msgid = msgid; /* * Store the message to send in the receive message so timeout @@ -2347,8 +2335,10 @@ static int i_ipmi_request(struct ipmi_user *user, if (rv) { out_err: - ipmi_free_smi_msg(smi_msg); - ipmi_free_recv_msg(recv_msg); + if (!supplied_smi) + ipmi_free_smi_msg(smi_msg); + if (!supplied_recv) + ipmi_free_recv_msg(recv_msg); } else { dev_dbg(intf->si_dev, "Send: %*ph\n", smi_msg->data_size, smi_msg->data); @@ -2358,9 +2348,6 @@ static int i_ipmi_request(struct ipmi_user *user, if (!run_to_completion) mutex_unlock(&intf->users_mutex); -out: - if (rv && user) - atomic_dec(&user->nr_msgs); return rv; } @@ -3851,7 +3838,7 @@ static int handle_ipmb_get_msg_cmd(struct ipmi_smi *intf, unsigned char chan; struct ipmi_user *user = NULL; struct ipmi_ipmb_addr *ipmb_addr; - struct ipmi_recv_msg *recv_msg; + struct ipmi_recv_msg *recv_msg = NULL; if (msg->rsp_size < 10) { /* Message not big enough, just ignore it. */ @@ -3872,9 +3859,8 @@ static int handle_ipmb_get_msg_cmd(struct ipmi_smi *intf, rcvr = find_cmd_rcvr(intf, netfn, cmd, chan); if (rcvr) { user = rcvr->user; - kref_get(&user->refcount); - } else - user = NULL; + recv_msg = ipmi_alloc_recv_msg(user); + } rcu_read_unlock(); if (user == NULL) { @@ -3904,47 +3890,41 @@ static int handle_ipmb_get_msg_cmd(struct ipmi_smi *intf, * causes it to not be freed or queued. */ rv = -1; + } else if (!IS_ERR(recv_msg)) { + /* Extract the source address from the data. */ + ipmb_addr = (struct ipmi_ipmb_addr *) &recv_msg->addr; + ipmb_addr->addr_type = IPMI_IPMB_ADDR_TYPE; + ipmb_addr->slave_addr = msg->rsp[6]; + ipmb_addr->lun = msg->rsp[7] & 3; + ipmb_addr->channel = msg->rsp[3] & 0xf; + + /* + * Extract the rest of the message information + * from the IPMB header. + */ + recv_msg->recv_type = IPMI_CMD_RECV_TYPE; + recv_msg->msgid = msg->rsp[7] >> 2; + recv_msg->msg.netfn = msg->rsp[4] >> 2; + recv_msg->msg.cmd = msg->rsp[8]; + recv_msg->msg.data = recv_msg->msg_data; + + /* + * We chop off 10, not 9 bytes because the checksum + * at the end also needs to be removed. + */ + recv_msg->msg.data_len = msg->rsp_size - 10; + memcpy(recv_msg->msg_data, &msg->rsp[9], + msg->rsp_size - 10); + if (deliver_response(intf, recv_msg)) + ipmi_inc_stat(intf, unhandled_commands); + else + ipmi_inc_stat(intf, handled_commands); } else { - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { - /* - * We couldn't allocate memory for the - * message, so requeue it for handling - * later. - */ - rv = 1; - kref_put(&user->refcount, free_ipmi_user); - } else { - /* Extract the source address from the data. */ - ipmb_addr = (struct ipmi_ipmb_addr *) &recv_msg->addr; - ipmb_addr->addr_type = IPMI_IPMB_ADDR_TYPE; - ipmb_addr->slave_addr = msg->rsp[6]; - ipmb_addr->lun = msg->rsp[7] & 3; - ipmb_addr->channel = msg->rsp[3] & 0xf; - - /* - * Extract the rest of the message information - * from the IPMB header. - */ - recv_msg->user = user; - recv_msg->recv_type = IPMI_CMD_RECV_TYPE; - recv_msg->msgid = msg->rsp[7] >> 2; - recv_msg->msg.netfn = msg->rsp[4] >> 2; - recv_msg->msg.cmd = msg->rsp[8]; - recv_msg->msg.data = recv_msg->msg_data; - - /* - * We chop off 10, not 9 bytes because the checksum - * at the end also needs to be removed. - */ - recv_msg->msg.data_len = msg->rsp_size - 10; - memcpy(recv_msg->msg_data, &msg->rsp[9], - msg->rsp_size - 10); - if (deliver_response(intf, recv_msg)) - ipmi_inc_stat(intf, unhandled_commands); - else - ipmi_inc_stat(intf, handled_commands); - } + /* + * We couldn't allocate memory for the message, so + * requeue it for handling later. + */ + rv = 1; } return rv; @@ -3957,7 +3937,7 @@ static int handle_ipmb_direct_rcv_cmd(struct ipmi_smi *intf, int rv = 0; struct ipmi_user *user = NULL; struct ipmi_ipmb_direct_addr *daddr; - struct ipmi_recv_msg *recv_msg; + struct ipmi_recv_msg *recv_msg = NULL; unsigned char netfn = msg->rsp[0] >> 2; unsigned char cmd = msg->rsp[3]; @@ -3966,9 +3946,8 @@ static int handle_ipmb_direct_rcv_cmd(struct ipmi_smi *intf, rcvr = find_cmd_rcvr(intf, netfn, cmd, 0); if (rcvr) { user = rcvr->user; - kref_get(&user->refcount); - } else - user = NULL; + recv_msg = ipmi_alloc_recv_msg(user); + } rcu_read_unlock(); if (user == NULL) { @@ -3990,44 +3969,38 @@ static int handle_ipmb_direct_rcv_cmd(struct ipmi_smi *intf, * causes it to not be freed or queued. */ rv = -1; + } else if (!IS_ERR(recv_msg)) { + /* Extract the source address from the data. */ + daddr = (struct ipmi_ipmb_direct_addr *)&recv_msg->addr; + daddr->addr_type = IPMI_IPMB_DIRECT_ADDR_TYPE; + daddr->channel = 0; + daddr->slave_addr = msg->rsp[1]; + daddr->rs_lun = msg->rsp[0] & 3; + daddr->rq_lun = msg->rsp[2] & 3; + + /* + * Extract the rest of the message information + * from the IPMB header. + */ + recv_msg->recv_type = IPMI_CMD_RECV_TYPE; + recv_msg->msgid = (msg->rsp[2] >> 2); + recv_msg->msg.netfn = msg->rsp[0] >> 2; + recv_msg->msg.cmd = msg->rsp[3]; + recv_msg->msg.data = recv_msg->msg_data; + + recv_msg->msg.data_len = msg->rsp_size - 4; + memcpy(recv_msg->msg_data, msg->rsp + 4, + msg->rsp_size - 4); + if (deliver_response(intf, recv_msg)) + ipmi_inc_stat(intf, unhandled_commands); + else + ipmi_inc_stat(intf, handled_commands); } else { - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { - /* - * We couldn't allocate memory for the - * message, so requeue it for handling - * later. - */ - rv = 1; - kref_put(&user->refcount, free_ipmi_user); - } else { - /* Extract the source address from the data. */ - daddr = (struct ipmi_ipmb_direct_addr *)&recv_msg->addr; - daddr->addr_type = IPMI_IPMB_DIRECT_ADDR_TYPE; - daddr->channel = 0; - daddr->slave_addr = msg->rsp[1]; - daddr->rs_lun = msg->rsp[0] & 3; - daddr->rq_lun = msg->rsp[2] & 3; - - /* - * Extract the rest of the message information - * from the IPMB header. - */ - recv_msg->user = user; - recv_msg->recv_type = IPMI_CMD_RECV_TYPE; - recv_msg->msgid = (msg->rsp[2] >> 2); - recv_msg->msg.netfn = msg->rsp[0] >> 2; - recv_msg->msg.cmd = msg->rsp[3]; - recv_msg->msg.data = recv_msg->msg_data; - - recv_msg->msg.data_len = msg->rsp_size - 4; - memcpy(recv_msg->msg_data, msg->rsp + 4, - msg->rsp_size - 4); - if (deliver_response(intf, recv_msg)) - ipmi_inc_stat(intf, unhandled_commands); - else - ipmi_inc_stat(intf, handled_commands); - } + /* + * We couldn't allocate memory for the message, so + * requeue it for handling later. + */ + rv = 1; } return rv; @@ -4141,7 +4114,7 @@ static int handle_lan_get_msg_cmd(struct ipmi_smi *intf, unsigned char chan; struct ipmi_user *user = NULL; struct ipmi_lan_addr *lan_addr; - struct ipmi_recv_msg *recv_msg; + struct ipmi_recv_msg *recv_msg = NULL; if (msg->rsp_size < 12) { /* Message not big enough, just ignore it. */ @@ -4162,9 +4135,8 @@ static int handle_lan_get_msg_cmd(struct ipmi_smi *intf, rcvr = find_cmd_rcvr(intf, netfn, cmd, chan); if (rcvr) { user = rcvr->user; - kref_get(&user->refcount); - } else - user = NULL; + recv_msg = ipmi_alloc_recv_msg(user); + } rcu_read_unlock(); if (user == NULL) { @@ -4195,49 +4167,44 @@ static int handle_lan_get_msg_cmd(struct ipmi_smi *intf, * causes it to not be freed or queued. */ rv = -1; + } else if (!IS_ERR(recv_msg)) { + /* Extract the source address from the data. */ + lan_addr = (struct ipmi_lan_addr *) &recv_msg->addr; + lan_addr->addr_type = IPMI_LAN_ADDR_TYPE; + lan_addr->session_handle = msg->rsp[4]; + lan_addr->remote_SWID = msg->rsp[8]; + lan_addr->local_SWID = msg->rsp[5]; + lan_addr->lun = msg->rsp[9] & 3; + lan_addr->channel = msg->rsp[3] & 0xf; + lan_addr->privilege = msg->rsp[3] >> 4; + + /* + * Extract the rest of the message information + * from the IPMB header. + */ + recv_msg->recv_type = IPMI_CMD_RECV_TYPE; + recv_msg->msgid = msg->rsp[9] >> 2; + recv_msg->msg.netfn = msg->rsp[6] >> 2; + recv_msg->msg.cmd = msg->rsp[10]; + recv_msg->msg.data = recv_msg->msg_data; + + /* + * We chop off 12, not 11 bytes because the checksum + * at the end also needs to be removed. + */ + recv_msg->msg.data_len = msg->rsp_size - 12; + memcpy(recv_msg->msg_data, &msg->rsp[11], + msg->rsp_size - 12); + if (deliver_response(intf, recv_msg)) + ipmi_inc_stat(intf, unhandled_commands); + else + ipmi_inc_stat(intf, handled_commands); } else { - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { - /* - * We couldn't allocate memory for the - * message, so requeue it for handling later. - */ - rv = 1; - kref_put(&user->refcount, free_ipmi_user); - } else { - /* Extract the source address from the data. */ - lan_addr = (struct ipmi_lan_addr *) &recv_msg->addr; - lan_addr->addr_type = IPMI_LAN_ADDR_TYPE; - lan_addr->session_handle = msg->rsp[4]; - lan_addr->remote_SWID = msg->rsp[8]; - lan_addr->local_SWID = msg->rsp[5]; - lan_addr->lun = msg->rsp[9] & 3; - lan_addr->channel = msg->rsp[3] & 0xf; - lan_addr->privilege = msg->rsp[3] >> 4; - - /* - * Extract the rest of the message information - * from the IPMB header. - */ - recv_msg->user = user; - recv_msg->recv_type = IPMI_CMD_RECV_TYPE; - recv_msg->msgid = msg->rsp[9] >> 2; - recv_msg->msg.netfn = msg->rsp[6] >> 2; - recv_msg->msg.cmd = msg->rsp[10]; - recv_msg->msg.data = recv_msg->msg_data; - - /* - * We chop off 12, not 11 bytes because the checksum - * at the end also needs to be removed. - */ - recv_msg->msg.data_len = msg->rsp_size - 12; - memcpy(recv_msg->msg_data, &msg->rsp[11], - msg->rsp_size - 12); - if (deliver_response(intf, recv_msg)) - ipmi_inc_stat(intf, unhandled_commands); - else - ipmi_inc_stat(intf, handled_commands); - } + /* + * We couldn't allocate memory for the message, so + * requeue it for handling later. + */ + rv = 1; } return rv; @@ -4259,7 +4226,7 @@ static int handle_oem_get_msg_cmd(struct ipmi_smi *intf, unsigned char chan; struct ipmi_user *user = NULL; struct ipmi_system_interface_addr *smi_addr; - struct ipmi_recv_msg *recv_msg; + struct ipmi_recv_msg *recv_msg = NULL; /* * We expect the OEM SW to perform error checking @@ -4288,9 +4255,8 @@ static int handle_oem_get_msg_cmd(struct ipmi_smi *intf, rcvr = find_cmd_rcvr(intf, netfn, cmd, chan); if (rcvr) { user = rcvr->user; - kref_get(&user->refcount); - } else - user = NULL; + recv_msg = ipmi_alloc_recv_msg(user); + } rcu_read_unlock(); if (user == NULL) { @@ -4303,48 +4269,42 @@ static int handle_oem_get_msg_cmd(struct ipmi_smi *intf, */ rv = 0; + } else if (!IS_ERR(recv_msg)) { + /* + * OEM Messages are expected to be delivered via + * the system interface to SMS software. We might + * need to visit this again depending on OEM + * requirements + */ + smi_addr = ((struct ipmi_system_interface_addr *) + &recv_msg->addr); + smi_addr->addr_type = IPMI_SYSTEM_INTERFACE_ADDR_TYPE; + smi_addr->channel = IPMI_BMC_CHANNEL; + smi_addr->lun = msg->rsp[0] & 3; + + recv_msg->user_msg_data = NULL; + recv_msg->recv_type = IPMI_OEM_RECV_TYPE; + recv_msg->msg.netfn = msg->rsp[0] >> 2; + recv_msg->msg.cmd = msg->rsp[1]; + recv_msg->msg.data = recv_msg->msg_data; + + /* + * The message starts at byte 4 which follows the + * Channel Byte in the "GET MESSAGE" command + */ + recv_msg->msg.data_len = msg->rsp_size - 4; + memcpy(recv_msg->msg_data, &msg->rsp[4], + msg->rsp_size - 4); + if (deliver_response(intf, recv_msg)) + ipmi_inc_stat(intf, unhandled_commands); + else + ipmi_inc_stat(intf, handled_commands); } else { - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { - /* - * We couldn't allocate memory for the - * message, so requeue it for handling - * later. - */ - rv = 1; - kref_put(&user->refcount, free_ipmi_user); - } else { - /* - * OEM Messages are expected to be delivered via - * the system interface to SMS software. We might - * need to visit this again depending on OEM - * requirements - */ - smi_addr = ((struct ipmi_system_interface_addr *) - &recv_msg->addr); - smi_addr->addr_type = IPMI_SYSTEM_INTERFACE_ADDR_TYPE; - smi_addr->channel = IPMI_BMC_CHANNEL; - smi_addr->lun = msg->rsp[0] & 3; - - recv_msg->user = user; - recv_msg->user_msg_data = NULL; - recv_msg->recv_type = IPMI_OEM_RECV_TYPE; - recv_msg->msg.netfn = msg->rsp[0] >> 2; - recv_msg->msg.cmd = msg->rsp[1]; - recv_msg->msg.data = recv_msg->msg_data; - - /* - * The message starts at byte 4 which follows the - * Channel Byte in the "GET MESSAGE" command - */ - recv_msg->msg.data_len = msg->rsp_size - 4; - memcpy(recv_msg->msg_data, &msg->rsp[4], - msg->rsp_size - 4); - if (deliver_response(intf, recv_msg)) - ipmi_inc_stat(intf, unhandled_commands); - else - ipmi_inc_stat(intf, handled_commands); - } + /* + * We couldn't allocate memory for the message, so + * requeue it for handling later. + */ + rv = 1; } return rv; @@ -4402,8 +4362,8 @@ static int handle_read_event_rsp(struct ipmi_smi *intf, if (!user->gets_events) continue; - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { + recv_msg = ipmi_alloc_recv_msg(user); + if (IS_ERR(recv_msg)) { mutex_unlock(&intf->users_mutex); list_for_each_entry_safe(recv_msg, recv_msg2, &msgs, link) { @@ -4424,8 +4384,6 @@ static int handle_read_event_rsp(struct ipmi_smi *intf, deliver_count++; copy_event_into_recv_msg(recv_msg, msg); - recv_msg->user = user; - kref_get(&user->refcount); list_add_tail(&recv_msg->link, &msgs); } mutex_unlock(&intf->users_mutex); @@ -4441,8 +4399,8 @@ static int handle_read_event_rsp(struct ipmi_smi *intf, * No one to receive the message, put it in queue if there's * not already too many things in the queue. */ - recv_msg = ipmi_alloc_recv_msg(); - if (!recv_msg) { + recv_msg = ipmi_alloc_recv_msg(NULL); + if (IS_ERR(recv_msg)) { /* * We couldn't allocate memory for the * message, so requeue it for handling @@ -4858,12 +4816,10 @@ static void smi_work(struct work_struct *t) list_del(&msg->link); - if (refcount_read(&user->destroyed) == 0) { + if (refcount_read(&user->destroyed) == 0) ipmi_free_recv_msg(msg); - } else { - atomic_dec(&user->nr_msgs); + else user->handler->ipmi_recv_hndl(msg, user->handler_data); - } } mutex_unlock(&intf->user_msgs_mutex); @@ -5179,27 +5135,51 @@ static void free_recv_msg(struct ipmi_recv_msg *msg) kfree(msg); } -static struct ipmi_recv_msg *ipmi_alloc_recv_msg(void) +static struct ipmi_recv_msg *ipmi_alloc_recv_msg(struct ipmi_user *user) { struct ipmi_recv_msg *rv; - rv = kmalloc(sizeof(struct ipmi_recv_msg), GFP_ATOMIC); - if (rv) { - rv->user = NULL; - rv->done = free_recv_msg; - atomic_inc(&recv_msg_inuse_count); + if (user) { + if (atomic_add_return(1, &user->nr_msgs) > max_msgs_per_user) { + atomic_dec(&user->nr_msgs); + return ERR_PTR(-EBUSY); + } } + + rv = kmalloc(sizeof(struct ipmi_recv_msg), GFP_ATOMIC); + if (!rv) { + if (user) + atomic_dec(&user->nr_msgs); + return ERR_PTR(-ENOMEM); + } + + rv->user = user; + rv->done = free_recv_msg; + if (user) + kref_get(&user->refcount); + atomic_inc(&recv_msg_inuse_count); return rv; } void ipmi_free_recv_msg(struct ipmi_recv_msg *msg) { - if (msg->user && !oops_in_progress) + if (msg->user && !oops_in_progress) { + atomic_dec(&msg->user->nr_msgs); kref_put(&msg->user->refcount, free_ipmi_user); + } msg->done(msg); } EXPORT_SYMBOL(ipmi_free_recv_msg); +static void ipmi_set_recv_msg_user(struct ipmi_recv_msg *msg, + struct ipmi_user *user) +{ + WARN_ON_ONCE(msg->user); /* User should not be set. */ + msg->user = user; + atomic_inc(&user->nr_msgs); + kref_get(&user->refcount); +} + static atomic_t panic_done_count = ATOMIC_INIT(0); static void dummy_smi_done_handler(struct ipmi_smi_msg *msg)

2 months, 4 weeks

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror