Eric Lee / smarc-fsl-linux-kernel

Download as
Browse Code ยป

Commit 9eefe520c814f6f62c5d36a2ddcd3fb99dfdb30e

Authored by Nadia Derbey
Committed by Linus Torvalds
1 parent f1a43f93f0
Exists in master and in 40 other branches 8mp-imx_5.4.70_2.3.0, 8qm-imx_5.4.70_2.3.0, emb_imx_lf-5.15.y, emb_lf-6.1.y, emb_lf-6.6.y, imx_3.0.35_4.1.0, imx_3.10.17_1.0.1_ga, imx_3.10.53_1.1.0_ga, imx_3.14.28_1.0.0_ga, imx_4.1.15_1.0.0_ga, pitx_8mp_lf-5.10.y, rt-smarc-imx_4.1.15_1.0.0_ga, rt_linux_5.15.71, smarc-8m-android-11.0.0_2.0.0, smarc-imx6_4.14.98_2.0.0_ga, smarc-imx6_4.9.88_2.0.0_ga, smarc-imx7_4.14.98_2.0.0_ga, smarc-imx7_4.9.11_1.0.0_ga, smarc-imx7_4.9.88_2.0.0_ga, smarc-imx_3.10.53_1.1.0_ga, smarc-imx_3.14.28_1.0.0_ga, smarc-imx_4.1.15_1.0.0_ga, smarc-imx_4.9.11_1.0.0_ga, smarc-imx_4.9.51_imx8m_ga, smarc-imx_4.9.88_2.0.0_ga, smarc-m6.0.1_2.1.0-ga, smarc-n7.1.2_2.0.0-ga, smarc-rel_imx_4.1.15_1.2.0_ga, smarc_8m_00d0_imx_4.14.98_2.0.0_ga, smarc_8m_imx_4.14.78_1.0.0_ga, smarc_8m_imx_4.14.98_2.0.0_ga, smarc_8m_imx_4.19.35_1.1.0, smarc_8mm_imx_4.14.78_1.0.0_ga, smarc_8mm_imx_4.14.98_2.0.0_ga, smarc_8mm_imx_4.19.35_1.1.0, smarc_8mm_imx_5.4.24_2.1.0, smarc_8mp_lf-5.10.y, smarc_8mq_imx_5.4.24_2.1.0, smarc_8mq_lf-5.10.y, smarc_imx_lf-5.15.y

ipc: do not use a negative value to re-enable msgmni automatic recomputing 

This patch proposes an alternative to the "magical
positive-versus-negative number trick" Andrew complained about last week
in http://lkml.org/lkml/2008/6/24/418.

This had been introduced with the patches that scale msgmni to the amount
of lowmem.  With these patches, msgmni has a registered notification
routine that recomputes msgmni value upon memory add/remove or ipc
namespace creation/ removal.

When msgmni is changed from user space (i.e.  value written to the proc
file), that notification routine is unregistered, and the way to make it
registered back is to write a negative value into the proc file.  This is
the "magical positive-versus-negative number trick".

To fix this, a new proc file is introduced: /proc/sys/kernel/auto_msgmni.
This file acts as ON/OFF for msgmni automatic recomputing.

With this patch, the process is the following:
1) kernel boots in "automatic recomputing mode"
   /proc/sys/kernel/msgmni contains the value that has been computed (depends
                           on lowmem)
   /proc/sys/kernel/automatic_msgmni contains "1"

2) echo <val> > /proc/sys/kernel/msgmni
   . sets msg_ctlmni to <val>
   . de-activates automatic recomputing (i.e. if, say, some memory is added
     msgmni won't be recomputed anymore)
   . /proc/sys/kernel/automatic_msgmni now contains "0"

3) echo "0" > /proc/sys/kernel/automatic_msgmni
   . de-activates msgmni automatic recomputing
     this has the same effect as 2) except that msg_ctlmni's value stays
     blocked at its current value)

3) echo "1" > /proc/sys/kernel/automatic_msgmni
   . recomputes msgmni's value based on the current available memory size
     and number of ipc namespaces
   . re-activates automatic recomputing for msgmni.

Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
Cc: Solofo Ramangalahy <Solofo.Ramangalahy@bull.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Showing 3 changed files with 76 additions and 19 deletions Side-by-side Diff

include/linux/ipc_namespace.h
Diff comments   View file @ 9eefe52
... ... @@ -36,6 +36,7 @@
36 36 int msg_ctlmni;
37 37 atomic_t msg_bytes;
38 38 atomic_t msg_hdrs;
  39 + int auto_msgmni;
39 40  
40 41 size_t shm_ctlmax;
41 42 size_t shm_ctlall;
... ... @@ -53,7 +54,7 @@
53 54  
54 55 extern int register_ipcns_notifier(struct ipc_namespace *);
55 56 extern int cond_register_ipcns_notifier(struct ipc_namespace *);
56   -extern int unregister_ipcns_notifier(struct ipc_namespace *);
  57 +extern void unregister_ipcns_notifier(struct ipc_namespace *);
57 58 extern int ipcns_notify(unsigned long);
58 59  
59 60 #else /* CONFIG_SYSVIPC */
... ... @@ -27,15 +27,17 @@
27 27 }
28 28  
29 29 /*
30   - * Routine that is called when a tunable has successfully been changed by
31   - * hand and it has a callback routine registered on the ipc namespace notifier
32   - * chain: we don't want such tunables to be recomputed anymore upon memory
33   - * add/remove or ipc namespace creation/removal.
34   - * They can come back to a recomputable state by being set to a <0 value.
  30 + * Routine that is called when the file "auto_msgmni" has successfully been
  31 + * written.
  32 + * Two values are allowed:
  33 + * 0: unregister msgmni's callback routine from the ipc namespace notifier
  34 + * chain. This means that msgmni won't be recomputed anymore upon memory
  35 + * add/remove or ipc namespace creation/removal.
  36 + * 1: register back the callback routine.
35 37 */
36   -static void tunable_set_callback(int val)
  38 +static void ipc_auto_callback(int val)
37 39 {
38   - if (val >= 0)
  40 + if (!val)
39 41 unregister_ipcns_notifier(current->nsproxy->ipc_ns);
40 42 else {
41 43 /*
... ... @@ -71,7 +73,12 @@
71 73 rc = proc_dointvec(&ipc_table, write, filp, buffer, lenp, ppos);
72 74  
73 75 if (write && !rc && lenp_bef == *lenp)
74   - tunable_set_callback(*((int *)(ipc_table.data)));
  76 + /*
  77 + * Tunable has successfully been changed by hand. Disable its
  78 + * automatic adjustment. This simply requires unregistering
  79 + * the notifiers that trigger recalculation.
  80 + */
  81 + unregister_ipcns_notifier(current->nsproxy->ipc_ns);
75 82  
76 83 return rc;
77 84 }
78 85  
... ... @@ -87,10 +94,39 @@
87 94 lenp, ppos);
88 95 }
89 96  
  97 +static int proc_ipcauto_dointvec_minmax(ctl_table *table, int write,
  98 + struct file *filp, void __user *buffer, size_t *lenp, loff_t *ppos)
  99 +{
  100 + struct ctl_table ipc_table;
  101 + size_t lenp_bef = *lenp;
  102 + int oldval;
  103 + int rc;
  104 +
  105 + memcpy(&ipc_table, table, sizeof(ipc_table));
  106 + ipc_table.data = get_ipc(table);
  107 + oldval = *((int *)(ipc_table.data));
  108 +
  109 + rc = proc_dointvec_minmax(&ipc_table, write, filp, buffer, lenp, ppos);
  110 +
  111 + if (write && !rc && lenp_bef == *lenp) {
  112 + int newval = *((int *)(ipc_table.data));
  113 + /*
  114 + * The file "auto_msgmni" has correctly been set.
  115 + * React by (un)registering the corresponding tunable, if the
  116 + * value has changed.
  117 + */
  118 + if (newval != oldval)
  119 + ipc_auto_callback(newval);
  120 + }
  121 +
  122 + return rc;
  123 +}
  124 +
90 125 #else
91 126 #define proc_ipc_doulongvec_minmax NULL
92 127 #define proc_ipc_dointvec NULL
93 128 #define proc_ipc_callback_dointvec NULL
  129 +#define proc_ipcauto_dointvec_minmax NULL
94 130 #endif
95 131  
96 132 #ifdef CONFIG_SYSCTL_SYSCALL
97 133  
98 134  
... ... @@ -142,15 +178,12 @@
142 178 rc = sysctl_ipc_data(table, name, nlen, oldval, oldlenp, newval,
143 179 newlen);
144 180  
145   - if (newval && newlen && rc > 0) {
  181 + if (newval && newlen && rc > 0)
146 182 /*
147 183 * Tunable has successfully been changed from userland
148 184 */
149   - int *data = get_ipc(table);
  185 + unregister_ipcns_notifier(current->nsproxy->ipc_ns);
150 186  
151   - tunable_set_callback(*data);
152   - }
153   -
154 187 return rc;
155 188 }
156 189 #else
... ... @@ -158,6 +191,9 @@
158 191 #define sysctl_ipc_registered_data NULL
159 192 #endif
160 193  
  194 +static int zero;
  195 +static int one = 1;
  196 +
161 197 static struct ctl_table ipc_kern_table[] = {
162 198 {
163 199 .ctl_name = KERN_SHMMAX,
... ... @@ -221,6 +257,16 @@
221 257 .mode = 0644,
222 258 .proc_handler = proc_ipc_dointvec,
223 259 .strategy = sysctl_ipc_data,
  260 + },
  261 + {
  262 + .ctl_name = CTL_UNNUMBERED,
  263 + .procname = "auto_msgmni",
  264 + .data = &init_ipc_ns.auto_msgmni,
  265 + .maxlen = sizeof(int),
  266 + .mode = 0644,
  267 + .proc_handler = proc_ipcauto_dointvec_minmax,
  268 + .extra1 = &zero,
  269 + .extra2 = &one,
224 270 },
225 271 {}
226 272 };
ipc/ipcns_notifier.c
Diff comments   View file @ 9eefe52
... ... @@ -55,25 +55,35 @@
55 55  
56 56 int register_ipcns_notifier(struct ipc_namespace *ns)
57 57 {
  58 + int rc;
  59 +
58 60 memset(&ns->ipcns_nb, 0, sizeof(ns->ipcns_nb));
59 61 ns->ipcns_nb.notifier_call = ipcns_callback;
60 62 ns->ipcns_nb.priority = IPCNS_CALLBACK_PRI;
61   - return blocking_notifier_chain_register(&ipcns_chain, &ns->ipcns_nb);
  63 + rc = blocking_notifier_chain_register(&ipcns_chain, &ns->ipcns_nb);
  64 + if (!rc)
  65 + ns->auto_msgmni = 1;
  66 + return rc;
62 67 }
63 68  
64 69 int cond_register_ipcns_notifier(struct ipc_namespace *ns)
65 70 {
  71 + int rc;
  72 +
66 73 memset(&ns->ipcns_nb, 0, sizeof(ns->ipcns_nb));
67 74 ns->ipcns_nb.notifier_call = ipcns_callback;
68 75 ns->ipcns_nb.priority = IPCNS_CALLBACK_PRI;
69   - return blocking_notifier_chain_cond_register(&ipcns_chain,
  76 + rc = blocking_notifier_chain_cond_register(&ipcns_chain,
70 77 &ns->ipcns_nb);
  78 + if (!rc)
  79 + ns->auto_msgmni = 1;
  80 + return rc;
71 81 }
72 82  
73   -int unregister_ipcns_notifier(struct ipc_namespace *ns)
  83 +void unregister_ipcns_notifier(struct ipc_namespace *ns)
74 84 {
75   - return blocking_notifier_chain_unregister(&ipcns_chain,
76   - &ns->ipcns_nb);
  85 + blocking_notifier_chain_unregister(&ipcns_chain, &ns->ipcns_nb);
  86 + ns->auto_msgmni = 0;
77 87 }
78 88  
79 89 int ipcns_notify(unsigned long val)